2011 The Palm Beach County Family Study (Full Report)

FifthMarciaGouvêa2011

SupportingLow-IncomeParentsofYoungChildrenThePalm BeachCountyFamilyStudyFifthAnnualReportJulieSpielbergerLaurenRichCarolynWinjeMolyScannelMarciaGouvêaSpielberger,J.,Rich,L.,Winje,C.,Scannel,M.,&Gouvêa,M.(2011).

AcknowledgmentsThis fifth report of the Palm Beach County Longitudinal Study would not have been possible without theparticipation of a large number of people in Palm Beach County. We acknowledge Tana Ebole, NateNichols, Grace Watson, Jeff Goodman, Laura Fleischman, Kim Lu, Anna Reardon, Tanya Palmer, SarahGosney, Scott Davey, and Carol Scott of the Children’s Services Council; Denise Pagan of the FloridaDepartment of Health; Donghai Xie at the School District of Palm Beach County; and Fran Smith andMark Holsapfel of the Florida Department of Children and Families (DCF). All of these individualshelped in various ways to clarify the goals and activities of the service system in Palm Beach County, toreview our methodology and data collection instruments, and to obtain administrative data. In addition toassisting us with our sample development and shepherding the interview process through some difficultperiods, the staff of Westat—Crystal MacAllum, Gail Thomas, and their field staff—were an invaluablesource of information and guidance every step of the way as we developed our interview and consentprotocols for the study. They also contributed information that assisted us in our analysis andinterpretation of findings.We also wish to acknowledge the contributions of other Chapin Hall staff who assisted with datacollection and analysis, including LaShaun Brooks, Lisa Michels, Erna Dinata, Alana Gunn, and KristinBerg who assisted with qualitative analysis, and Allen Harden for assistance with administrative dataanalysis. We also appreciate Anne Clary’s help with editing and Rosemary Gill’s help in preparing thisreport. In addition, we appreciate the support and guidance we have received from the following ChapinHall staff and former staff over the 6 years of this study: Deborah Daro, Robert Goerge, Harold Richman,John Schuerman, Ada Skyles, and Matthew Stagner.It goes without saying that we are greatly indebted to the families who are participating in this study.Although, for reasons of confidentiality, we cannot thank them by name, they have opened their doors tous, given us a glimpse into their lives, and shared their goals and hopes for their children and themselves.We have learned and continue to learn an enormous amount from them.Last, we are indebted to the Children’s Services Council of Palm Beach County for its vision of how toimprove the lives of children and families and its commitment to learning from research about itscommunity and how to serve it better, as well as for its generous support of this study.Chapin Hall at the University of Chicagoii

Use of Formal Services............................................................................................................................... 84The Maternal Child Health System ........................................................................................................ 84Other Formal Services............................................................................................................................ 92Service Needs and Help-Seeking in Year 5............................................................................................ 97Use of the Maternal Child Health System and Other Formal Services................................................ 101Summary............................................................................................................................................... 102Service Use, Maternal Functioning, and Child Outcomes........................................................................ 105Maternal Depression............................................................................................................................. 106Parental Stress....................................................................................................................................... 107Parenting Practices ............................................................................................................................... 109Children’s Development....................................................................................................................... 111Kindergarten Readiness........................................................................................................................ 119Summary............................................................................................................................................... 130Summary and Conclusions........................................................................................................................ 131Family Characteristics .......................................................................................................................... 131Health, Healthcare, and Child Development........................................................................................ 132Parenting Practices ............................................................................................................................... 134Childcare Arrangements ....................................................................................................................... 135Social Support....................................................................................................................................... 136Service Use ........................................................................................................................................... 136Service Use, Maternal Functioning, and Child Development.............................................................. 138Study Limitations ................................................................................................................................. 139Conclusions and Implications............................................................................................................... 140Bibliography.............................................................................................................................................. 146Appendix A: Longitudinal Study Sample and Methods ........................................................................... 153Research Questions, Sample, and Methods.......................................................................................... 153Administrative Records ........................................................................................................................ 155Household Surveys............................................................................................................................... 157Qualitative Interviews........................................................................................................................... 162Appendix B: Additional Data on PBC and TGA Birth Cohorts, 2004–2009 ........................................... 165Appendix C: Additional Data on Year 5 Study Sample ........................................................................... 189Chapin Hall at the University of Chicagoiv

Chapin Hall at the University of Chicagov

List of FiguresFigure 1. Conceptual Framework for Longitudinal Study............................................................................ 4Figure 2. Demographic Characteristics of PBC Mothers by Nativity, 2004–2009 a,b ................................... 8Figure 3. Primary Childcare Arrangements for Focal Children over Time a ............................................... 59Figure 4. Perceptions of Access to Support from All Sources over Time a ................................................. 81Figure 5. Access to Support by TGA over Time a ....................................................................................... 82Figure B- 1. Characteristics of Mothers with Newborns in Palm Beach County a .................................... 168Figure B- 2. Characteristics of Mothers with Newborns in Non-TGAs a .................................................. 169Figure B- 3. Characteristics of Mothers with Newborns in TGAs a .......................................................... 170Figure B- 4. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 a .......................... 172Figure B- 5. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 a .......................... 173Figure B- 6. Characteristics of Mothers in TGA Birth Cohort a ................................................................ 176Figure B- 7. Characteristics of Mothers in TGA Birth Cohort with MCHP Service Treatment a ............. 177Figure C- 1. Mean Days of MCHP Services for Year 5 Sample by Age of Child Through 2008 a .......... 202Chapin Hall at the University of Chicagovi

List of TablesTable 1. Components of CSC System of Care.............................................................................................. 2Table 2. Characteristics of Mothers of Newborns in Palm Beach County, 2004–2005 a .............................. 7Table 3. Baseline Characteristics of Mothers in Year 5 Sample a ............................................................... 10Table 4. Household Characteristics of Year 5 Sample over Time a ............................................................. 12Table 5. Language and Literacy Skills of Foreign-Born Mothers over Time a ........................................... 13Table 6. Family Income and Economic Support of Year 5 Sample over Time a ......................................... 15Table 7. Selected Characteristics of Year 5 Sample Mothers by TGA a ..................................................... 16Table 8. Selected Characteristics of Year 5 Sample Mothers by Nativity a ................................................. 18Table 9. Family Living Conditions over Time a .......................................................................................... 20Table 10. Reasons for Transportation Difficulties over Time a ................................................................... 21Table 11. Mothers’ Health and Maternal Functioning over Time a ............................................................. 24Table 12. Healthcare of Mothers over Time a .............................................................................................. 26Table 13. Children’s Health over Time a ..................................................................................................... 28Table 14. Healthcare of Children over Time a ............................................................................................. 30Table 15. Reasons Children Not Covered by Health Insurance over Time a ............................................... 31Table 16. Mothers’ and Children’s Health Status by TGA, Race/Ethnicity, and Nativity at Year 5 a ........ 34Table 17. Social and Emotional Behaviors of Focal Children at Year 5 a ................................................... 37Table 18. Communication Skills of Focal Children at Year 5 a ................................................................... 37Table 19. Language Skills of Focal Children over Time a ........................................................................... 38Table 20. Emerging Literacy Skills of Focal Children at Year 5 a .............................................................. 38Table 21. Emerging Pre-academic Skills of Focal Children at Year 5 a ...................................................... 39Table 22. Positive Parenting Practices during Previous 3 Months at Year 5 a ............................................ 43Table 23. Access to Reading Materials at Home at Year 5 a ....................................................................... 44Table 24. Mothers’ Positive Parenting Practices over Time a ...................................................................... 45Table 25. Fathers’ Positive Parenting Practices over Time a ....................................................................... 46Table 26. Year 5 Negative Parenting Practices during Previous 3 Months a ............................................... 47Table 27. Mothers’ Negative Parenting Practices over Time a .................................................................... 48Table 28. Fathers’ Negative Parenting Activities over Time a .................................................................... 49Table 29. Parent Involvement Activities during Previous 3 Months at Year 5 a ......................................... 50Table 30. Parent Involvement Activities of Mothers over Time a ............................................................... 51Chapin Hall at the University of Chicagovii

Table 31. Parent Involvement Activities of Fathers over Time a ................................................................. 51Table 32. Mothers’ Positive Parenting Activities during Previous 3 Months by Selected MaternalCharacteristics at Year 5 a ............................................................................................................................ 53Table 33. Mothers’ Negative Parenting Practices during Previous 3 Months by Selected MaternalCharacteristics at Year 5 a ............................................................................................................................ 54Table 34. Mothers’ Parent Involvement Activities during Previous 3 Months for Children in Childcare orSchool by Selected Maternal Characteristics at Year 5 a ............................................................................. 55Table 35. Minimum Number of Transitions in Focal Children’s Care Arrangements Year 1 to Year 5 a .. 61Table 36. Use of Nonparental Preschool Childcare by Mothers’ Characteristics at Year 5 a ...................... 62Table 37. Logistic Regression Predicting Use of Nonparental Preschool Childcare at Year 5 a ................. 63Table 38. Nonparental Childcare Arrangements for Focal Children at Year 5 by Race/Ethnicity andNativity a ...................................................................................................................................................... 64Table 39. Nonparental Childcare Arrangements for Focal Children at Year 5 by Maternal Education,Employment, and Income a .......................................................................................................................... 65Table 40. Characteristics of Mothers Receiving and Not Receiving Subsidy for Childcare at Year 5 a ..... 67Table 41. Percentage of Preschool Children in Study Families Receiving Care from QIS Providers at Year5 a .................................................................................................................................................................. 69Table 42. Percentage of Preschool Childcare Providers for Study Families Participating in the QIS at Year5 a .................................................................................................................................................................. 69Table 43. Types of Afterschool Activities and Programs for 77 Children at Year 5 a ................................ 70Table 44. Types of Afterschool Activities and Programs for 77 Children at Year 5 a ................................ 71Table 45. Types of Husband/Partner Support over Time a .......................................................................... 75Table 46. Frequency of Husband/Partner Support over Time a ................................................................... 76Table 47. Frequency of Husband/Partner Support over Time a ................................................................... 77Table 48. Family and Friends Support over Time a ..................................................................................... 79Table 49. Type of Family/Friends Support Received at least Weekly over Time a,b ................................... 80Table 50. Frequency of Community Support at Year 5 a ............................................................................. 81Table 51. Help Received for Basic Family Needs over Time a ................................................................... 94Table 52. Help Received for Concerns about Children’s Health and Development over Time a ................ 96Table 53. Use of Selected Services by Mothers in Glades and Non-Glades TGA at Year 5 a .................... 97Table 54. Service Needs: Concerns of Mothers Who Did Not Receive Services over Time a .................... 98Table 55. Mothers Who Sought Help for Concerns over Time a ................................................................. 99Table 56. Mothers Seeking Help Who Did Not Receive Services at Year 5 a ........................................... 101Table 57. Logistic Regression of Mothers’ Depression Symptoms at Year 5 a ......................................... 107Chapin Hall at the University of Chicagoviii

Table 58. Logistic Regression of Mothers’ Parenting Stress at Year 5 a ................................................... 109Table 59. Logistic Regression of Mothers’ Parenting Scores at Year 5 a .................................................. 110Table 60. Logistic Regression of Child’s Likelihood of Speaking in Long Sentences at Year 5 a ........... 114Table 61. Logistic Regression of Child’s Communication Skills at Year 5 a ............................................ 115Table 62. Logistic Regression of Child’s Social-Emotional Skills at Year 5 a .......................................... 116Table 63. Logistic Regression of Child’s Use of Books at Year 5 a .......................................................... 117Table 64. Logistic Regression of Child’s Emergent Literacy/Pre-Academic Skills at Year 5 a ................ 118Table 65. Logistic Regression of Child’s Drawing/Writing Skills at Year 5 a .......................................... 119Table 66. Selected Maternal and Child Characteristics by ECHOS Scores a ............................................ 123Table 67. Logistic Regression Predicting ECHOS Scores ....................................................................... 124Table 68. Selected Maternal and Child Characteristics by FAIR Scores ................................................. 126Table 69. First Logistic Regression Predicting FAIR Scores ................................................................... 127Table 70. Second Logistic Regression Predicting FAIR Scores............................................................... 129Table A- 1. FOCiS Codes for MCHP Services Defined as Treatment ..................................................... 156Table A- 2. Interview Study Sampling Plan ............................................................................................. 158Table A- 3. Household Survey Topics over Time .................................................................................... 160Table A- 4. Sample and Response Rates for Longitudinal Study over Time ........................................... 161Table A- 5. Reasons for Sample Attrition over Time ............................................................................... 162Table B- 1. Risk Screen Scores of Mothers with Newborns, 2004–2009 a ............................................... 165Table B- 2. Characteritics of TGA Mothers with Maternal Child Health Partnership Risk Screens, 2004–2009 a .......................................................................................................................................................... 166Table B- 3. MCHP and Chapin Hall Mean Risk Index for TGA Birth Cohorts a ...................................... 167Table B- 4. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 a ............................ 171Table B- 5. Logistic Regression Predicting Likelihood of Receiving Treatment a .................................... 174Table B- 6. Characteristics of Mothers in the TGA Birth Cohorts a .......................................................... 175Table B- 7. Characteristics of TGA Birth Cohort by Type of Services in MCHP System a ..................... 178Table B- 8. Logistic Regression Predicting Likelihood of Receiving Intensive Care Coordination a ....... 179Table B- 9. Logistic Regression Predicting Likelihood of Receiving Care Coordination Only a ............. 180Table B- 10. Use of Enhanced Services by Mothers in TGA Birth Cohorts a ........................................... 181Chapin Hall at the University of Chicagoix

Table B- 11. MCHP Prenatal and Postnatal Services and Referrals, 2004–2009 a .................................... 183Table B- 12. Logistic Regression Predicting Likelihood of Receiving Enhanced Services a .................... 184Table B- 13. Outside Service Referrals for 2004–2009 TGA Cohorts Who Received Treatment Services a................................................................................................................................................................... 185Table B- 14. Logistic Regression Predicting Likelihood of Receiving Family Support Planning a .......... 187Table B- 15. OLS Logistic Regression Predicting Days of Treatment Services a ..................................... 188Table C- 1. Selected Characteristics of Mothers by Nativity a .................................................................. 190Table C- 2. Selected Health Characteristics of Mothers by Nativity a,b ..................................................... 191Table C- 3. Services, Economic Supports, and Social Support by Nativity a,b .......................................... 191Table C- 4. Social and Emotional Behaviors of Focal Children at Year 4 and Year 5 a ........................... 192Table C- 5. Social and Emotional Behaviors of Focal Children by Nativity at Year 5 a ........................... 193Table C- 6. Communication Skills of Focal Children at Year 4 and Year 5 a ........................................... 193Table C- 7. Communication Skills of Focal Children by Nativity at Year 5 a ........................................... 194Table C- 8. Language Skills over Time by Sex of Focal Child a ............................................................... 194Table C- 9.Focal Children’s Language Skills by Nativity at Year 5 a ....................................................... 194Table C- 10. Preliteracy and Pre-academic Skills of Focal Children at Year 4 and Year 5 a .................... 195Table C- 11. Emerging Literacy and Pre-academic Skills of Focal Children by Nativity at Year 5 a ....... 196Table C- 12. Access to Reading Materials at Home by Nativity at Year 5 a ............................................. 197Table C- 13. Mothers’ Positive Parenting Practices during Previous 3 Months by Nativity at Year 5 a ... 197Table C- 14. Mothers’ Negative Parenting Practices during Previous 3 Months by Nativity at Year 5 a . 198Table C- 15. Parent Involvement Activities during Previous 3 Months by Nativity at Year 5 a ............... 198Table 16. Social Support by Nativity a,b ..................................................................................................... 198Table C- 17. Number and Most Frequent Types of Focal Children’s Care Arrangements between Birthand 5 Years a .............................................................................................................................................. 199Table C- 18. Days of MCHP Services for Year 5 Sample by Age of Child through 2008 a ..................... 203Table C- 19. Characteristics of Mothers by Timing of Receipt of MCHP Treatment Services a,b ............ 204Table C- 20. Characteristics of Mothers Who Received MCHP Services by Number of TreatmentServices Days a ........................................................................................................................................... 205Table C- 21. Number of Service Areas Used by Mothers in Glades and Non-Glades TGA over Time a . 206Table C- 22. Patterns of Service Use by Nativity in 12 Months Prior to Surveys over Time a ................. 207Table C- 23. Number of Services Used by Nativity over Time a ............................................................... 208Chapin Hall at the University of Chicagox

Table C- 24. Mothers’ Baseline and Year 4 Characteristics by Number of Services Reported at Year 5 a................................................................................................................................................................... 209Table C- 25. Linear Regression Analysis of Number of Services Used in Year 5 a .................................. 210Table C- 26. Baseline and Year 4 Characteristics of Mothers Who Changed/Did Not Change Service Usebetween Year 2 and Year 5 a ...................................................................................................................... 211Table C- 27. Baseline Characteristics of Mothers Who Changed/Did Not Change Service Use betweenYear 1 and Year 5 a .................................................................................................................................... 212Table C- 28. Changes in Circumstances between Year 3 and Year 4 and Number of Services Used at Year5 a ................................................................................................................................................................ 213Chapin Hall at the University of Chicagoxi

Executive SummaryThis is the fifth and final report of a longitudinal study examining the use of a comprehensive system ofprevention and early intervention services in Palm Beach County, and how its use relates to the outcomesof children and families living in four targeted geographic areas (TGAs) with high rates of poverty, teenpregnancy, crime, and child abuse and neglect. Parenting in any circumstance is challenging, but lowmothersface additional hurdles—including fewer economic and social resources, limited education,unstable living arrangements, and transportation difficulties, etc.—that place them and their children atrisk of poor outcomes. For more than a decade, the Children’s Service’s Council (CSC) of Palm BeachCounty has been engaged in an effort to strengthen the services offered to families in the TGAcommunities, in order to increase the number of healthy births, reduce the incidence of child abuse andneglect, and enhance parents’ abilities to raise socially, emotionally, and physically healthy children whoare eager to learn and ready for school.A fundamental goal of the emerging service system in Palm Beach County is to enhance the availabilityand coordination of services and supports to at-risk families and children. A fundamental assumption isthat improving service quality and increasing families’ access to services will have positive effects onmaternal functioning and parenting practices, which, in turn, will positively affect children’s behavior anddevelopment. At the same time, the underlying ecological model (e.g., Bronfenbrenner 1977, 1986) forthis study recognizes that a number of factors influence service use, including child, family, program, andcommunity characteristics, and including not only demographic characteristics but also individual beliefsand attitudes about services, perceptions of need and the costs of service use, the availability of socialsupport, and willingness to seek help.The primary goal of this study was to better understand families’ patterns of service use over time andhow service use was related to child and family outcomes. Consistent with the ecological model, wegathered a wide variety of information about family characteristics, health and healthcare, parentingChapin Hall at the University of Chicagoxii

• In addition, we observed a potential positive impact of center-based care and other formal childcarearrangements—as opposed to parental and other informal care—in the year prior to kindergartenentry on mothers’ reports of child development.• Based on mothers’ reports, we also found that the children of foreign-born Hispanic mothers werelagging in their cognitive and social-emotional development, when compared to the children of blackmothers, both foreign- and U.S.-born.• Teachers’ assessments of children’s development on the FLKRS soon after they entered kindergartenwere mixed. There were no significant race/ethnicity or nativity differences in teachers’ ratings on theECHOS, a comprehensive child development screen, after controlling for family and childcharacteristics. However, children of foreign-born Hispanic mothers were not assessed as highly onthe FAIR, a screen of literacy skills.• In addition, children in the study sample did not perform as well as other children enteringkindergarten in the school district. This was not surprising given their overall higher riskcharacteristics.Conclusions and RecommendationsBy concentrating services in the four areas of Palm Beach County with the highest rates of poverty, teenpregnancy, crime, and child abuse and neglect, CSC is developing a system of care to assist familieswhose children are most vulnerable to starting school behind their peers. To inform the development ofthe system, the purpose of the longitudinal study was to understand how these services were used by asample of families who, because of their risk characteristics, were likely to use services from the CSCsystem of care. Results indicate that the MCHP system initially identified mothers with many importantrisk factors, and that mothers with greater needs were more likely to receive services and more days ofservices than other mothers. Of mothers screened “at risk” who did not receive services, however, almosta third could not be located, and 15 percent declined services.Service use was affected by many other factors, including individual family characteristics, service type,and program characteristics. Receipt of MCHP services in year 1 was also associated with other serviceuse in years 2, 3, and 5. One reason, according to mothers’ reports, was that MCHP providers served anessential bridging or “brokering” role between parents and basic services, including Medicaid, foodassistance, and childcare subsidies. This support was particularly important for mothers with poorerlanguage and literacy skills, fewer personal resources, and low social support. Service use was alsoimpacted by the accuracy of information about service eligibility and cost and provider responsiveness.One of the clearest findings was the disadvantaged status of children born to foreign-born mothersrelative to those born to U.S.-born mothers. At the end of the study, as the focal children were gettingready for kindergarten, children of foreign-born mothers were more likely to be living at or below thepoverty level, even though their caregivers were more likely to be married or living together. In addition,Chapin Hall at the University of Chicagoxv

some groups of foreign-born mothers reported significantly higher levels of depression and parentingstress, and others reported significantly lower levels of parenting skills. Most importantly, based onmothers’ and teachers’ assessments, the children of foreign-born Hispanic mothers appeared to be laggingin at least some areas of their development when compared to the children of other mothers.In light of their relative disadvantaged status, it is encouraging to note that foreign-born mothers, andparticularly foreign-born Hispanic mothers, were more likely to receive treatment services from theMCHP system. On the other hand, they used fewer services outside of the MCHP system compared toU.S.-born mothers. Given that we did not find evidence that overall service use or parent educationservices in the first year after a child’s birth was associated with improved maternal or child outcomeslater on, it would be useful to follow up this finding with a more in-depth analysis of outcomes relative tolevels of participation in particular types of services. A more in-depth analysis might also suggest waysservices offered by the system could be made more effective for all mothers, and especially for foreignbornmothers, and their children.In conclusion, study findings make clear that the emerging system of care in Palm Beach County issuccessfully engaging many at-risk families in needed services through the MCHP around the birth of achild. It appears, moreover, unlike other formal services used by the study families, the MCHP has moreflexibility to adapt to the diverse circumstances and daily routines of the families they serve. At the sametime, there were challenges in keeping some of the mothers engaged, identifying and addressing newservice needs, and monitoring service use over time during their children’s early years. These findingssuggest there are opportunities to improve service access and use in the TGAs in general and strengthenthe CSC system of prevention and early intervention services in particular. Thus, we end with thefollowing set of recommendations:• Improve the quality and effectiveness of parenting supports and education. A consistent findingfrom the first year was the role of parenting practices in children’s development. This suggests a needto better understand early parenting practices and increase the availability and quality of interventionsdesigned to improve them. Mothers with lower educational backgrounds or literacy skills, especiallythose for whom English is a second language, might need additional support to strengthen their abilityto prepare their children for school during the preschool years.• Improve access to and quality of early care and education. Another important factor in children’sschool readiness outcomes was the type of childcare arrangement in the year before kindergarten, afinding that is supported by other research. Children who were at home with their parents were lesslikely than children who attended a Quality Counts center-based program to be screened as “ready”for kindergarten on the state standardized readiness screen. At the same time, there was a strongrelationship between having a childcare subsidy and enrollment in a center-based program. Manymothers in the qualitative study, regardless of race/ethnicity or nativity, expressed interest in someChapin Hall at the University of Chicagoxvi

kind of out-of-home educational experience for their children, but could not afford to enroll theirchild without assistance.In addition to increasing access, it also will be important to ensure that early care and educationprograms—both family childcare as well as center programs—address the particular needs of childrenfrom low-income but especially language-minority backgrounds.• Increase efforts to help families stay involved in or become re-connected to needed services overtime. In this study, families’ service use varied over time for multiple reasons, including parents’perceptions of need, access to other resources, difficulties with re-application processes, or actualimprovements in their circumstances. On one hand, it was not surprising that use of formal servicesfor parenting information declined after the first year, and likely reflects the lack of connections toother kinds of available services and supports once mothers leave the MCHP. (Because of datalimitations, it was not possible to easily track individual families’ engagement in different types ofservices over time in this study, but it would be useful to examine this hypothesis further.) It alsomight reflect, as several mothers in the qualitative study told us, an increasing confidence as theirchildren approached their first birthday in their parenting and a desire to be independent of family orcommunity supports. On the other hand, as children grow, new developmental stages are likely tobring new challenges for parents. As children approached 2 years of age, and began developinglanguage and more autonomy, parents expressed new questions and concerns about whether theirchildren’s behavior and development were on track. At the age of 3, as children’s language and selfcareskills improved, parents became more open to considering preschool or home-based educationalprograms but were not aware of what might be available to them or how to access them. Theseobservations suggest that there might be “touch points” when parents are more receptive to servicesbut there are not enough formal structures to help them get engaged or re-engaged in services.• Enhance training of service providers. Another strategy for keeping families engaged in services isto improve the knowledge and responsiveness of service providers by enhancing training in culturallyappropriate and family-strengths-based approaches, as well as special needs of families. CSC mightnot be able to directly impact service delivery in public and other agencies not funded by CSC, butmight help to raise the public’s awareness of the literacy and educational needs of families, inaddition to their service needs, in the targeted communities. Staff who are trained to help families in arespectful way could reduce future duplication of paperwork and client and staff frustration, as wellas make families feel more positive about seeking and accepting help earlier.• Making location and timing of services convenient for families. Of the many factors that constrainservice use, the locations of program offices, their hours, and excessive waiting times pose significantbarriers for families, especially if they have transportation or childcare problems. Strategies that CSCfundedprograms use, such as home visits and traveling service vans, are good alternatives to officevisits, especially if they are available during evening and weekend hours. Basing services at schools,Beacon Centers, or childcare centers is another option for reaching families who have childrenenrolled in school or formal childcare. As it may be difficult to persuade employers to allow familiestime off for appointments with teachers, doctors, or service agencies without jeopardizing theirwages, it may be more feasible to persuade healthcare providers, schools, and service agencies toexpand locations and hours of services to make them more convenient for families.Chapin Hall at the University of Chicagoxvii

• Improve channels of information and communication about services. During the time of thisstudy, CSC has expanded use of other vehicles (such as radio, television, faith-based organizations,and public libraries) to disseminate information that will reach families with limited education orliteracy skills, families who do not receive information through family or friends, and families whoare not already using other services like childcare. Healthcare providers might be engaged moreeffectively in providing information to families. The local offices of federal benefit programs are alsochannels for disseminating information about CSC-funded programs; for example, one of the studymothers was referred by a nurse in the WIC office to a provider in the MCHP system.• Strengthen relationships between the CSC system of care and other community supports andservices. Improving the quality of childcare and providing referrals through childcare programs is away to reach families who use these services. However, this approach will not reach many motherswho are not working, who are either not eligible or on a waiting list for a childcare subsidy, or whoprefer other childcare settings. Other strategies are needed to reach these families, for example,through WIC, public health clinics, and community outreach. Our finding of an increase in reportedlevels of community support, especially by medical and childcare providers in the third year, alsosuggests the importance of improving the knowledge of these professionals about parenting andparenting information and supports in the community as well as their ability to assess service needs.• Improve data systems and other sources of information on service availability, use (duration,intensity), and need. Although our findings are strengthened by the use of multiple data sources andmixed methods, there also were limitations in our ability to interpret changes in service use over timethrough the quantitative data. The database for the MCHP was an important source of information onthe types of services families received in the system and the kinds of referrals made to providersoutside the system for this study. However, if mothers received referrals to other services, we couldnot ascertain with current data if they actually followed up and got connected to these services. Norcould we determine, if a connection was made, for what duration or intensity they received theservices. This indicates a need for additional longitudinal data and data system capabilities tounderstand how families enter and leave the system over time as their families grow. Additionalsources of information on the location of services, community needs for services, and the outcomes ofreferrals would assist funders and providers of services in planning and monitoring the systems’ability to ensure families follow up and get connected to the services they need.Chapin Hall at the University of Chicagoxviii

IntroductionOver the last three decades, considerable progress has been made in understanding the ecological andcultural context for children’s development and, in particular, the harmful effects of poverty and itscorrelates on family functioning and child development (e.g., Bronfenbrenner, 1979, 1986; Brooks-Gunn,2003; Gomby, 2005; National Research Council and Institute of Medicine, 2000; Olds, Kitzman, Hankset al., 2007; Weisner, 2002). At the same time, a variety of early intervention strategies have beendesigned to ameliorate the effects of poverty on children’s development and readiness for school.Increasingly, comprehensive, integrated systems of health, educational, and social services are viewed asa promising strategy for supporting healthy family functioning and child development in low-income, atriskfamilies (Brooks-Gunn 2003; Brooks-Gunn, Duncan & Aber, 1997; Gomby 2005; Olds et al,. 2007;Reynolds, Ou, & Topitzes 2004).This growing body of evidence prompted the Children’s Services Council (CSC) of Palm Beach County(FL) to undertake a long-term initiative to build an integrated system of care to promote and support thehealthy development of children, with a focus on the first 5 years of life. The primary goals for the PalmBeach County system of care are to increase the number of healthy births, to reduce the incidence of childabuse and neglect, and to increase school readiness, as indicated by the number of children who enterkindergarten ready to learn. 1 These goals are based on the assumption that strengthening the system ofcommunity supports and services available to families in the targeted geographical areas (TGAs) willenhance families’ abilities to raise their children in healthy ways and, in turn, improve children’sdevelopment and well-being. With improved family functioning and improved child health anddevelopment, it is further expected that children will be better prepared for school and families will bebetter able to support them in school. Moreover, it is believed that by strengthening the system of1 “Palm Beach County’s Pathway to Early Childhood Development,” CSC draft planning document, August 2007.Chapin Hall at the University of Chicago 1

informal community supports and prevention and early intervention services, families are less likely toneed more intensive mental health, child welfare, and juvenile justice services.To pursue these aims, CSC and other stakeholders have developed a set of prevention and earlyintervention programs and systems serving low-income families and their young children in four targetedgeographic areas (TGAs). 2 The primary programs and systems designed to support children at differentstages of their development are presented below.Table 1. Components of CSC System of CareProgram/System NameMaternal Child HealthPartnership (MCHP)Early Care and EducationSchool Behavioral HealthProgramsAfterschool ProgramsProgram DescriptionA network of health and social services for high-risk pregnant women andmothers, which includes universal risk screening before and after birth; targetedassessment and home visitation; and coordinated services for familiesexperiencing medical, psychological, social, and environmental risks thatnegatively impact pregnancy and birth outcomesSeveral initiatives intended to identify and provide services for children withdevelopmental delays and to improve children’s school readiness, and QualityCounts, a quality improvement system for childcare programsDesigned to improve children’s adjustment to school and enhance their schoolsuccess by identifying social-emotional and other developmental problems andproviding referrals and interventions to respond to these problemsA network of afterschool programs for elementary and middle school youthsupported by Prime Time, an intermediary working to improve the quality ofschool-based and community programsThe Palm Beach County Longitudinal StudyA central concern for CSC and other stakeholders in the county is the effectiveness of this emergingsystem. Is the service system functioning and being used by families as expected? Is it achieving itsintended outcomes? Separate evaluations have been conducted on several individual programs andnetworks that are part of the system (e.g., Spielberger, Haywood, Schuerman, Richman, & Michels, 2005;Lyons, Karlstrom, & Haywood, 2007). Yet, these evaluations alone cannot provide information on howfamilies use the system of services or the relationship between use of multiple services and children’swell-being and development.Thus, CSC funded Chapin Hall at the University of Chicago to conduct a 6-year longitudinal study toexamine the use and potential effectiveness of an array of services in the county in promoting school2 About the time this study was conceived, 75 to 93 percent of children in these targeted communities received free or reduced-costlunch; the rate of child abuse and neglect was between 4.1 and 6.6 times the county average; and crime rates in the TGAs rangedfrom 14 to 93 percent above the county rate (Children’s Services Council report, State of the Child in Palm Beach County, 2003).Chapin Hall at the University of Chicago 2

eadiness and school success and in improving family functioning among children and families most inneed of support. The goal of the study is to describe the characteristics and needs of families the servicesystem is intended to serve, how they use the services that make up the service system in Palm BeachCounty, and how service use relates to child and family outcomes. It began in 2004 and addressesquestions in the following areas:What services and supports are available and how do families of young children in the TGAs usethem? Are there patterns of service use?What are the correlates of service use, including demographic and other family characteristics,indicators of risk and service need, geographic location, nativity, and prior service use?How does service use relate to child and family outcomes, including children’s school readiness;school success; and physical, social-emotional, and behavioral health; and to family functioning, ratesof abuse and neglect, and parent involvement in schools? 3Conceptual Framework for this StudyWe use an ecological model to guide this study. As described above, the Palm Beach County system ofcare is being built on the premise that families in the TGAs will function better and be better equipped toraise their children to be healthy and ready for school—cognitively, socially, emotionally, andphysically—with the support of a system of prevention and early intervention programs. The conceptualframework for the longitudinal study, shown in Figure 1, reflects this premise.This model assumes that families’ access to and use of the system of care will affect proximal outcomes(e.g., parenting practices, maternal functioning, and reports of child abuse and neglect), which, in turn,will affect children’s long-term outcomes. The model also suggests that a number of factors influenceservice use, including child, family, and community characteristics, and use of other services. Forexample, preliminary findings indicated that service use differs as a function of demographiccharacteristics such as age, education, and employment, as well as individual beliefs and attitudes aboutservices, perceptions of need and the costs of service use, and willingness to seek help (Spielberger,Lyons, Gouvea, Haywood, & Winje, 2007). In addition, mothers may be more likely to use services iftheir prior experiences with services were positive, or if the services used are a source of informationabout other services. Moreover, the conceptual framework suggests that the availability of social support3 It is important to note that, because families voluntarily chose whether or not to use services, and were not randomly assigned toreceive services, we cannot draw any conclusions about causal relationships between families’ service use and child and familyoutcomes. We are, however, able to examine the strength and direction of these relationships, after controlling for a wide varietyof child and family characteristics associated with both service use and child/family outcomes.Chapin Hall at the University of Chicago 3

may modify the relationship between child and family characteristics and service use.Figure 1. Conceptual Framework for Longitudinal StudyStudy Design and MethodsTo examine the use of the service system in Palm Beach County, and its relationship to children’s earlydevelopment and school readiness, we felt it important to track families during the early years of theirchildren’s lives, when they are most likely to come into contact with the service system. Thus, weselected as our primary study group families with newborns living in the TGAs, with the intent offollowing them for at least 8 years into the children’s early school years. The study uses a mixed-methodsapproach to examine the relations among the service systems in Palm Beach County, indicators of childwell-being and family functioning, and child and family outcomes. Methods include the following:An analysis of administrative data on service use and key outcomes for all children born in the TGAsand in the county during 2004 and 2005 and who remained in the county at various data collectionpoints during a 6-year period. Sources of administrative data are the Department of Health (DOH)Vital Statistics database; FOCiS, the database for the MCHP system; and the Department of Childrenand Families (DCF) database on reports of child abuse and neglect.A 6-year longitudinal survey of the service use experiences of a sample of families with youngchildren in the TGAs. Methods include annual in-person interviews with a baseline sample of 531mothers of newborn children and brief phone interviews with the same parents about 6 months aftereach interview for a period of 5 years. A total of 310 mothers were interviewed in the fifth year.A 3-year embedded qualitative study involving in-depth interviews and observations of a sample of40 families to enhance what is learned through analysis of the structured interviews andadministrative data. The qualitative study has focused in particular on service use, motivations to useservices, and how services fit into families’ lives. Six waves of interviews were conducted withmothers. In addition, in 2010, we interviewed a sample of 16 male caregivers.Chapin Hall at the University of Chicago 4

The study’s comprehensive, longitudinal design has allowed us to examine in depth the relationshipsamong child and family characteristics, use of the service system, and child and family outcomes. It hasalso allowed us to explore explanations for any relationships that are found and to document changes thatoccur within families over time that might be attributable to use of particular services. Appendix A, aswell as previous reports (Spielberger et al., 2006, 2007, 2009, 2010), provide other information about thestudy design and methodology.Organization of This ReportThis report presents findings from the fifth year of the in-person household survey, conducted when thefocal children were between 48 and 54 months of age, and makes comparisons with findings from theprevious years of the study. In all analyses, we use 310 as the sample population, which is the number ofmothers who completed all five in-person surveys. This report also draws on administrative data fromFOCiS, Vital Statistics, the Enhanced Field System (EFS) childcare subsidy data, and DCF data on childabuse and neglect. 4Chapter 2 begins with an overview of the demographic characteristics of the 2004–2005 birth cohort, aswell as a description of the 310 mothers who participated in all 5 years of in-person interviews. We thenpresent, in chapter 3, a description of the mothers’ health and healthcare practices, as well as theirchildren’s health and development. In chapter 4, we describe mothers’ reports of their parenting activitiesand those of their husbands/partners. In chapter 5, we discuss findings on families’ experiences withchildcare and the factors that affect their use of childcare. In chapter 6, findings on informal andcommunity supports are presented. In chapter 7, we report on families’ use of formal services by drawingon administrative data and mothers’ self-reports. We go on, in chapter 8, to examine the patterns andcorrelates of service use in year 5 and present findings on the relationships between patterns of serviceuse and maternal functioning and child development. In the final chapter, we summarize the findings andconsider their implications for the Palm Beach County service system.4 Other study reports will include results of analyses of qualitative interview data with a subsample of mothers and malecaregivers.Chapin Hall at the University of Chicago 5

Family CharacteristicsIn this chapter, we begin with a review of the demographic characteristics of the 2004–2005 birth cohortin Palm Beach County, based on an analysis of Vital Statistics data, and compare them with thecharacteristics of the cohort of mothers who gave birth in 2006, 2007, 2008, and 2009. We then describein detail the characteristics of the year 5 sample, composed of 310 mothers who participated in all five ofthe in-person structured interviews, including their demographics, living conditions, and healthcharacteristics. The descriptive data for the sample were weighted in the analyses to account for theoversampling of mothers from the Glades and mothers who were identified as being “at risk” on thehospital screen or home assessment.The 2004–2005 Birth Cohort in Palm Beach CountyTable 2 presents characteristics of the cohort of mothers who gave birth in Palm Beach County in 2004and 2005, from which the study interview sample was drawn. As shown in Table 2, mothers in the TGAswere more likely than mothers outside the TGAs to be unmarried, to be teen mothers, to have less than ahigh school education, to be black or Hispanic, to be foreign-born, and to have used WIC while pregnant. 5All of these differences were statistically significant at the .001 level or lower. Mothers in the TGAs werealso significantly more likely to have low birth-weight babies, although the difference was notsubstantive. Finally, they also had higher scores on a risk index we created for the study. 65 Within the TGAs, the Glades had the highest proportion of unmarried and teen mothers. Whereas large proportions of mothersin Lake Worth and West Palm Beach were foreign-born (62% and 50%, respectively) and Hispanic (46% and 34%, respectively),mothers in the Glades and Riviera Beach were predominantly U.S.-born and black (Spielberger et al., 2007).6 Because not all mothers in the county were screened at birth, we calculated a risk index based on a count of 11 demographicand health characteristics recorded in Vital Statistics: (1) no or late prenatal care, (2) mother does not have high school diplomaor GED, (3) mother is not married, (4) mother age 19 or less at birth, (5) mother not born in U.S., (6) mother received WIC whilepregnant, (7) mother smoked, (8) mother had medical complications (other than previous C-section), (9) mother had deliverycomplications, (10) baby’s weight less than 2,500 grams, and (11) baby’s gestational age 36 weeks or less.Chapin Hall at the University of Chicago 6

Table 2. Characteristics of Mothers of Newborns in Palm Beach County, 2004–2005 aCharacteristic b Palm Beach County Non-TGAs TGAs(N=29,620) (n=17,185) (n=12,435)Not married (%) 40 27 57Teen mother (%) 9 6 14< HS education (%) 22 13 35Black (%) c 26 18 36Hispanic (%) d 29 22 39Foreign-born (%) 41 36 47Late or no prenatal care 27 23 32Used WIC while pregnant (%) 35 23 52Low-birth-weight baby (%) 8 8 9Study risk index mean (SD) d 2.4 (1.7) 1.9 (1.6) 3.0 (1.7)a Source: Vital Statistics. Sample numbers exclude mothers who gave birth in Palm Beach County but were residents of othercounties. Multiple births were counted as one birth event.b Does not include mothers who responded “I don’t know” to any of the Vital Statistics questionsc To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.d Because not all mothers in the county were screened at birth, we calculated a risk index based on their number of 11demographic and health characteristics recorded in Vital Statistics. These were the following: (1) no or late prenatal care, (2)mother does not have high school diploma or GED, (3) mother is not married, (4) mother age 19 or lower at birth, (5) mother notborn in U.S., (6) mother received WIC while pregnant, (7) mother smoked, (8) mother had medical complications (other thanprevious C-section), (9) mother had delivery complications, (10) baby’s weight less than 2,500 grams, and (11) baby’sgestational age 36 weeks or less. Mothers who responded “I don’t know” to any of these items were not coded “at risk” for thatitem.For comparison purposes, we also analyzed administrative data on the cohort of mothers who gave birthduring the next 4 years, 2006 through 2009. These results, which are presented in Table B- 4, Figure B- 1,Figure B- 2, and Figure B- 3 in Appendix B, indicate little change with respect to most of thecharacteristics available in Vital Statistics of mothers who gave birth in the county over the 5-year period.The most significant change was a decline in the percentage of mothers who received late or no prenatalcare. In the county, this was 27 percent for the 2004–2005 birth cohort, and only 14 percent for the 2009cohort; in the TGAs, the percentage declined from 32 percent in 2004–2005 to 18 percent in 2009. Oneindicator of increasing risk in the population is the use of WIC during pregnancy, which increased from35 percent in the county in 2004–2005 to 44 percent in 2009; the corresponding change in the TGAs wasan increase from 52 percent in 2004–2005 to 61 percent in 2009. In addition, there was a modest increasein the percentages of new mothers who were unmarried and who had less than a high school educationfrom 2004–2005 to 2009 across the county.Chapin Hall at the University of Chicago 7

Results in Table 2 and Table 3 also indicate a small increase—from 41 to 44 percent—in the percentageof foreign-born mothers giving birth in Palm Beach County over the 5-year period. Most of this growthoccurred between 2004 and 2007. 7In addition to looking at changes in demographics over time, we also looked at demographic differencesby race/ethnicity and nativity (see Figure 2). We found that U.S.-born blacks were more likely to beunmarried, to be teen mothers, to have a low-birth-weight baby, and to live in the TGAs. Similarly,foreign-born Hispanics were also more likely to be unmarried and to live in the TGAs; in addition, theywere also more likely than the other groups to have less than a high school education. Foreign-bornmothers who were not black or Hispanic were the least likely to be unmarried, to be teen mothers, to haveless than a high school education, and to live in the TGAs.Figure 2. Demographic Characteristics of PBC Mothers by Nativity, 2004–2009 a,ba Source: Vital Statistics and FOCiS (2004–2009). The birth of twins, triplets, and other multiples are counted as one birth event.b To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.7 U.S. Census data indicate that the percentage of the population in Palm Beach County who say they are foreign-born grew from17 percent in the 2000 Census to 22 percent in the 2006–2008 American Community Survey estimate(http://factfinder.census.gov). The percentage of those who describe themselves as Hispanic or Latino grew from 12 to 17 percentfor the same period.Chapin Hall at the University of Chicago 8

The Year 5 Study SampleTable 4 shows changes in selected family characteristics over time. The most noteworthy changecontinues to be an increase in the proportion of the mothers who are working. At baseline, just 13 percentof the sample mothers were working part time or full time, whereas nearly half (46%) were working at thetime of the fifth interview. There were additional, modest changes in educational levels, community ofresidence, and marital status over time. These changes are described below.Household CharacteristicsAt the beginning of the study, 13 percent of the sample families lived in the Glades TGA and 87 percentlived in the other three TGAs. 8 Over the 5 years of the study, there has been a modest increase in thepercentage of study families living outside the TGAs (13% in year 5). Most of this movement has comefrom families in the non-Glades TGAs; consequently, in the fifth year, 75 percent live in the non-GladesTGAs, compared with 87 percent in the first year. 9 The proportion of families who live in the Glades hasbeen fairly stable.The proportion of married mothers has also remained fairly stable over time. However, the percentage ofmothers who are single and living with a partner declined from 39 percent to 31 percent from the first tothe fifth interview, while the percentage of mothers who are single and not in a relationship increasedfrom 20 percent to 26 percent. Overall, 64 percent of mothers were either married or living with a partnerin the fifth year.In terms of household composition, at the time of the year 5 interview, 25 percent of the sample had onechild, 32 percent had two, and 44 percent had three or more. Over 40 percent of mothers reported thatthey had had another child since the birth of their focal child. 10 Seventeen mothers (or 6% of the sample)were pregnant at the time of the year 5 interview; for five mothers, this was at least the second pregnancysince the birth of their focal child. Household sizes at the time of the interview ranged from 1 to 14members in the fifth year, with an average of 4.9 members per household. Ten percent of the mothersreported they had other children under the age of 18 who were not living in their households. 118 For our descriptive analyses, we weighted the survey data to adjust for the oversampling of mothers from the Glades TGA andmothers who were screened through the MCHP system to be “at risk.”9 At year 5, a majority of the mothers living outside the TGAs had moved to one of three areas (zip codes 33411, 33414, and 33463).10 In 2002 the interval between first and second births was less than 12 months for 5 percent and between 13 and 24 months for 23percent of low-income (0%–149% of poverty level) mothers ages 20 to 44 (CDC, U.S. Department of Health & Human Services, 2005).11 Four mothers reported that their children were living with another relative at year 5, although two expected their children toreturn home in the next 6 months. Among the four, one reported that DCF removed the children, and another said that the courtshad awarded custody to the father. We exclude mothers who permanently lose custody of the focal child, but keep those whoreport temporary separations from the child, even if they are unsure about when the child will return.Chapin Hall at the University of Chicago 9

Table 3. Baseline Characteristics of Mothers in Year 5 Sample aCharacteristicTGA (%)Year 1 Sample atBaseline(N=531)Year 5 Sample atBaseline(N=310)Glades 13 16Non-Glades 87 84FOCiS initial risk screen (%)At risk/high need screen score 31 30Study risk index (Vital Statistics)Mean (SD) 3.7 (1.6) 3.5 (1.5)Range 0–10 0–8Age of motherMean age (SD) 24.9 (5.9) 24.9 (5.7)Age range 15–46 15–43Teen mother at child’s birth (%)Age 15–19 18 15Mother’s marital status (%)Married 26 29Mother’s race (%)Hispanic 55 52Black, not Hispanic 36 42White, not Hispanic/other 9 7Main language spoken in home (%)English 45 47Spanish 47 45Haitian Creole, Kanjobal, other 9 8Mother’s nativity (%)United States 42 45Mexico 19 17Guatemala 13 13Haiti 7 7Other Caribbean, Central or South Americancountry19 19Mother’s education (%)Less than high school diploma 58 53High school graduate 42 47Mother’s employment (%)Employed full or part time 14 13Mother’s yearly income (%)Below $20,000 66 66Mother’s below specific income requirements (%)At or below income requirements 69 69aData were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chapin Hall at the University of Chicago 10

Household Structure, Employment, and EducationThe most marked change in maternal employment occurred from the first to the second year, when it rosefrom 13 percent to 45 percent; from the second year on it has remained fairly stable. 12 Also, 8 percent ofthe year 5 sample reported being employed at each of the five interviews; however, 28 percent have neverreported being employed at any point during the study. Sixteen percent were not employed at the time ofthe first interview, but have been employed at each point since then. Finally, 49 percent have reportedsporadic employment.Of those mothers who were employed in the fifth year, half reported working 36 hours or more per week,with an overall average of 32.6 hours per week. As in previous years, most mothers work one job; onlyeight reported having two or more jobs. More than three-fourths of mothers who work described their jobas a regular daytime shift, and another 9 percent reported working a regular evening shift. Eight percentof the mothers reported rotating shifts; 3 percent reported regular night shifts; and 2 percent reportedworking either a split shift or some other schedule. The largest group of mothers (55%) described theirwork as a “service” occupation, and 12 percent reported being in an administrative position. Smallerpercentages (9% or less) described their jobs as marketing/sales, production, nursing, or health technicianpositions.Almost half (47%) of the mothers in the TGAs who gave birth in 2004 and 2005 were foreign-born. Thelargest percentages of foreign-born mothers in the TGAs came from Mexico (9%), Haiti (8%), andGuatemala (7%); 16 percent came from other Caribbean, Central American, and South Americancountries, and 6 percent came from other countries.Paternal employment has remained fairly stable over the 5 years. Overall, when looking at both familystructure and employment over time, we see an increase in the percentage of two-parent households inwhich both parents are working (from 5% in the first year to 23% in the fifth year). We also see anincrease in the percentage of single-parent households with a working parent (from 7% in the first year to20% in the fifth year).12 The unemployment rate in January 2008 was reported to be Florida’s highest since October 2004. During the first 3 years ofthe study, Florida boasted lower unemployment rates than the national average but by January 2008, the two rates wereconverging, with both expected to grow (Hundley, 2008).Chapin Hall at the University of Chicago 11

Table 4. Household Characteristics of Year 5 Sample over Time a % Mothers (N=310)CharacteristicYear 1 Year 2 Year 3 Year 4 Year 5Marital status (%)Married, living with husband 28 31 32 31 32Married, not living with husband 1 0 0 0 1Single b , living with partner 39 35 31 34 31Single, in a relationship but not livingw/partner13 10 11 9 11Single, not in a relationship 20 23 26 26 26Number of children (%)One c 42 40 30 26 25Two 31 34 36 35 32Three or more 27 26 34 39 44Number of childrenMean (SD) 2.0 (1.1) 2.0 (1.1) 2.2 (1.2) 2.3 (1.2) 2.4 (1.3)Range 1–9 1–9 0–10 0–10 0–10Mean (SD) age of focal child in months 1.8 (1.1) 14.7 (1.4) 26.2 (1.5) 39.1 (1.4) 50.2 (1.7)Mother’s education level (%)Less than high school diploma 55 52 52 49 51High school graduate 27 24 22 33 28Post–high school education 19 23 26 17 21Mother employed full or part time 13 45 49 45 46Mother attending school 8 8 10 11 16Husband/partnerMean age (SD) 28.8 (6.8) 29.6 (6.7) 30.6 (6.5) 31.2 (6.5) 32.2 (6.7)High school graduate (%) 21 19 19 22 17Currently employed (%) 87 91 89 87 84Household size (children and adults)Mean (SD) 5.0 (1.7) 4.9 (1.6) 4.8 (1.7) 4.9 (1.8) 4.9 (1.8)Range 2–11 2–10 1–12 1–12 1–14Family structure and employment (%)Two-parent household; one works 57 43 38 37 36Two-parent household; both work 5 22 23 24 23Two-parent household; neither works 5 1 2 4 4Single-parent household; parent works 7 20 25 18 20Single-parent household; parent does not work 26 13 13 17 17a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b “Single” includes mothers who identified themselves as single, never married, divorced, separated, or widowed.c Changes in number of children occur primarily because of subsequent births but also because of changes in other children in the home.*Paired sample t-tests indicate these year-to-year differences are statistically significant at p < .05 or less: marital1 vs. marital3; marital1vs. marital4; martial1 vs. marital5; children1 vs. children4; children1 vs. children5; children2 vs. children3; children2 vs. children4;children2 vs. children5; children3 vs. children4; children3 vs. children5; children4 vs. children5; partner educ1 vs. partner educ3; partnereduc1 vs. partner educ4; partner educ1 vs. partner educ5; partner works2 vs. partner works4; partner works2 vs. partner works5; motherworks1 vs. mother works2; mother works1 vs. mother works3; mother works1 vs. mother works4; mother works1 vs. mother works5;mother in school1 vs. mother in school5; mother in school2 vs. mother in school5; mother in school3 vs. mother in school5; mother inschool4 vs. mother in school5; mother educ1 vs. mother educ2; mother educ1 vs. mother educ3; mother educ1 vs. mother educ5; mothereduc2 vs. mother educ5; mother educ3 vs. mother educ4; mother educ3 vs. mother educ5; mother educ4 vs. mother educ5.Chapin Hall at the University of Chicago 12

Although mothers continued to report relatively low levels of education in the fifth year of the study,there was a slight increase in the percentage of mothers reporting some additional education beyond highschool. On the other hand, there was a decrease in the percentage of husbands/partners whose highestlevel of education was high school. 13In addition, 16 percent of the mothers also said they are currently in a full-time or part-time educationalprogram, an increase from the 8 percent who reported being in such a program in the first year. Over athird of those in an educational program are enrolled full time. One-third of those currently enrolled in aprogram reported being in a vocational or technical program, while just under one-third reported being ina 2-year or associate’s degree program. Eighteen percent reported being in a 4-year or bachelor’s degreeprogram, and 16 percent of the sample reported being in an English-as-a-second-language program. Theremaining 3 percent of mothers reported being in a GED program.Language and Literacy AbilitiesBeginning in the third year of the study, mothers whose primary language was not English—a little morethan half of the sample each year—were asked how well they spoke English. Across the 3 years, about 80percent of the mothers whose primary language was not English reported speaking English either “alittle” or “not at all” versus “well” or “very well” (see Table 5).Table 5. Language and Literacy Skills of Foreign-Born Mothers over Time aAbility Level% Mothers % Mothers % MothersYear 3 Year 4 Year 5How well you speak English (n=145) (n=145) (n=143)Very well 6 4 4Well 13 14 17A little 43 43 50Not at all 37 40 29How well you read English (n=167) (n=167) (n=167)Very well 16 15 13Well 15 13 17A little 34 35 35Not at all 35 38 35How well you write English (n=167) (n=167) (n=167)Very well 14 14 14Well 11 11 12A little 28 34 34Not at all 46 40 41a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: speak3vs. speak5; speak4 vs. speak5.13 In some families, the person the mother identified as her partner changed, so some of this decrease may reflect the educationallevel of new partners. In year 1, all but three husbands/partners living in the household were the fathers of the focal child; in year2, all but nine; and in year 3, all but eight. In year 4, there was a dramatic increase, with 21 husbands/partners living in thehousehold who were not the fathers of the focal child; and in year 5, 25 husbands/partners living in the household were not thefathers of the focal child.Chapin Hall at the University of Chicago 13

There was, however, a significant change in English ability between years 4 and 5, when fewer mothers(40% in year 4 vs. 29% in year 5) reported that they did not speak any English “at all” (p < .01). Motherswhose primary language was not English assessed their literacy skills similarly; that is, across the 3 years,about 75 percent described their abilities to read and write English as only “a little” or “not at all.”Family Income and Economic SupportCorresponding in part to the increase in maternal employment, the percentage of mothers who reportedfamily incomes less than $20,000 declined from 66 percent in the first year to 46 percent in the fifth year(see Table 6). At the same time, the use of some income support programs has declined over time. Inparticular, 84 percent of the mothers in the year 5 sample reported using WIC in year 1, whereas only 52percent participated in the program in year 5. In addition, the relatively small percentage of familiesreceiving Temporary Assistance for Needy Families (TANF) in year 1 (7%) declined further to 2 percentin year 5. These changes might reflect higher income levels from employment, which may render somestudy families ineligible for support; but they may also reflect other barriers to service use. Also,consistent with the increase in maternal employment, there was a significant increase in the proportion offamilies who received the Earned Income Tax Credit (EITC), as well as unemployment insurance (UI).Finally, it is of interest to note a significant increase in the proportion of families receiving rent vouchersbetween the first and fifth years of the study.The survey also asked the mothers about home ownership. A large majority (71%) continued to live inrented homes in the fifth year of the study, although this was less than the 80 percent in the first year. 14 Inaddition, we calculated the percentage of families living in poverty using the household size, the numberof children under age 18, and the federal poverty thresholds. In year 1, 69 percent of our sample reportedincomes at or below the federal poverty threshold for the previous year. In subsequent years, thepercentage of poor families remained fairly constant, and the percentage of those at or below thethreshold stood at 52 percent in year 5. 1514 The survey asked mothers to estimate their “total household income from all sources” for the previous year in broad categories(i.e., $10,000–$19,999 and $20,000–$39,999); thus, we categorized families according to whether or not they were above orbelow $20,000. It should be noted that estimating income can be difficult when income is irregular and unstable (e.g., Edin &Lein, 1997). This was the case in some of the study families, in which fathers who worked in construction or landscaping couldnot work when it rained, lost wages if they had to take time off to take a mother or child to the clinic, and experienced frequentchanges in the days and hours of their work.15 Because mothers were asked to estimate the income for the previous year, we calculated the income-to-need ratio based on thepoverty levels for the preceding year that corresponded to the year the majority of the interviews were conducted. (For example,three-quarters of the sample were first interviewed in 2005, so we used the 2004 income threshold under the Federal PovertyGuidelines to calculate the ratio for year 1.)Chapin Hall at the University of Chicago 14

Table 6. Family Income and Economic Support of Year 5 Sample over Time aCharacteristicAnnual income previous year (%)% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Less than $20,000 66 49 50 54 46Income-to-need ratio (%)Living at or below poverty threshold 69 54 52 58 52Use of income support programs (%)Women, Infants, and Children (WIC) 84 79 56 50 52Food Stamps 37 46 35 38 47Earned Income Tax Credit (EITC) 18 23 19 22 35Social Security Insurance (SSI) 11 8 9 7 11Temporary Assistance for Needy Families(TANF)7 8 3 2 2Rent voucher 4 4 5 6 11Unemployment Insurance (UI) 3 2 1 3 9Home ownership (%)Home owned by mother or other family20 26 26 27 29membera Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Paired sample t-tests indicate that these year-to-year differences are statistically significant: income:need1 vs. income:need2;income:need1 vs. income:need3; income:need1 vs. income:need4; income:need1 vs. income:need5; FoodStamps1 vs. FoodStamps2;FoodStamps1 vs. FoodStamps5; FoodStamps2 vs. FoodStamps3; FoodStamps2 vs. FoodStamps4; FoodStamps3 vs. FoodStamps5;FoodStamps4 vs. FoodStamps5; wic1 vs. wic2; wic1 vs. wic3; wic1 vs. wic4; wic1 vs. wic5; wic2 vs. wic3; wic2 vs. wic4; wic2 vs.wic5; wic3 vs. wic4; tanf1 vs. tanf2; tanf1 vs. tanf4; tanf1 vs. tanf5; tanf2 vs. tanf3; tanf2 vs. tanf4; tanf2 vs. tanf5; rentvoucher1 vs.rentvoucher5; rentvoucher2 vs. rentvoucher5; rentvoucher3 vs. rentvoucher5; rentvoucher4 vs. rentvoucher5; ssi4 vs. ssi5; ui1 vs. ui3;ui1 vs. ui5; ui2 vs. ui5; ui3 vs. ui5; ui4 vs. ui5; eitc1 vs. eitc5; eitc2 vs. eitc5; eitc3 vs. eitc5; eitc4 vs. eitc5; home1 vs. home2; home1 vs.home3; home1 vs. home4; home1 vs. home5; home2 vs. home4; home2 vs. home5; home3 vs. home4; home3 vs. home5.Sample Characteristics by TGAThe service system in Palm Beach County targeted four low-income communities. One question ofinterest at the start of this study was whether service use differs in an area in the western part of thecounty called the Glades from that in three other, more urban areas along the east coast of the county.Because it is a larger, more rural, and more remote area than those in eastern Palm Beach County, fundersand service providers want to ensure that the service system serves families in this area well. Thus, tounderstand service use and the effects of service, it is important to understand the characteristics offamilies living in different parts of Palm Beach County. Also, subsequent analyses of the relationshipbetween geographic location and service use and the effects of service use by location, for example, musttake into consideration differences in the characteristics of the families who live in the Glades and ofthose in other parts of the county.Table 7 presents selected characteristics of families living in the Glades, the non-Glades TGAs, andoutside the TGAs elsewhere in Palm Beach County at year 5. Statistically significant differences wereobserved between these groups of mothers in terms of their race/ethnicity, nativity, and education.Chapin Hall at the University of Chicago 15

Table 7. Selected Characteristics of Year 5 Sample Mothers by TGA aCharacteristicTotal Glades Non-Glades Outside TGA(N=310) (n=38) (n=232) (n=40)FOCiS initial risk screen (%)At-risk/high-need screen score 31 29 30 38Study risk index (Vital Statistics)Mean (SD) 3.5 (1.5) 3.6 (1.6) 3.5 (1.5) 3.6 (1.4)Range 0–8 1–8 0–8 0–6Age of motherMean age (SD) 29.0 (5.7) 28.1 (5.8) 29.2 (5.7) 28.8 (5.7)Age range 19–47 19–44 19–47 20–41Teen mother at child’s birth (%) 15 24 13 20Mother’s race/ethnicity (%)**Hispanic 51 29 55 49Black, not Hispanic 41 68 38 34Main language spoken in home (%)English 50 74 46 50Spanish 43 21 47 43Haitian Creole 6 5 5 8Mother’s nativity (%)**United States 44 78 39 37Mexico 17 16 17 17Guatemala 13 0 16 7Haiti 7 3 8 7Honduras 3 0 4 2Other country 16 3 16 29Mother’s education (%)**High school graduate/GED (or more) 50 62 44 73Mother employed (%) 46 42 46 46Marital status (%)Married, living with husband 31 24 32 38Single c , living with partner 31 16 35 23Single, in relationship, but not living with partner 11 16 10 10Single, not in a relationship 26 45 22 28Number of children (%)One 25 21 25 33Two 32 33 32 25Three or more 44 46 43 43Mean (SD) household size (all children and adults) 4.9 (1.8) 5.0 (2.0) 5.0 (1.8) 4.4 (1.4)Annual income previous year less than $20,000 (%) 46 51 45 46Income-to-need ratio at/below poverty threshold (%) 52 50 55 42Home ownership (%)Mother owns home 20 19 20 22Someone else in household owns home 8 19 6 12a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bZip codes of areas outside the TGAs were: 33408, 33410, 33411, 33413, 33414, 33418, 33432, 33435, 33436, 33437, 33458,33463, 33467, 33469, and 33470c “Single” includes mothers who identified themselves as either single, never married, divorced, separated, or widowed.Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 16

Specifically, mothers living outside of the Glades were more likely than those residing in the Glades to beHispanic, and those living in the Glades were more likely to be black. In addition, those living outside theTGAs were more likely than mothers in the other two groups to be white or another race. Thesedifferences were reflected in differences with respect to nativity. Thus, mothers living in the Glades weremore likely to have been born in the United States than those living outside the Glades. Finally, mothersliving in the Glades were more likely to have a high school diploma or more than those living outside theGlades.Sample Characteristics by NativityAnother important factor to consider in determining the effect of service use on family functioning andchildren’s development is nativity. Research indicates that children growing up in these families facenumerous risks, including poverty, low parental education, linguistic isolation, and poor neighborhoodand school quality (e.g., Chase-Lansdale, Valdovinos D'Angelo, & Palacios, 2007; Hernandez, Denton, &Macartney, 2007). The 1996 Personal Responsibility and Work Opportunity Reconciliation Act prohibitsmost foreign-born individuals who have lived in the United States for less than 5 years from receivingpublic benefits such as Medicaid and Food Stamps (e.g., King 2007). Findings presented in previousreports of this study, as well as a growing body of literature, indicate that even among eligible groups ofpeople, foreign-born individuals are less likely to take up these and other services than U.S.-bornindividuals (e.g., Dinan, 2005a, 2005b). Given the growing population of foreign-born people in PalmBeach County, it is particularly important to examine service use in relation to the characteristics,experiences, and outcomes of foreign-born families with young children (many of whom are U.S.citizens) in the county. The vast majority of these children are U.S. citizens, and their successfulintegration into the educational system will increase the likelihood that they grow up to be productiveadults.Table 8 presents selected characteristics of the sample families as a function of nativity. 16 Along withrace/ethnic and TGA differences between the two groups, we found statistically significant differences inmother’s age, the likelihood of being a teen mother at the birth of the focal child, education, maritalstatus, employment, use of income supports, and home ownership. Foreign-born mothers wereapproximately 4 years older than U.S.-born mothers, more likely to be married or single and living with apartner than U.S.-born mothers (20% and 24%, respectively), and less likely to have been a teen mother16 In previous reports, we compared characteristics of mothers who had been in the United States for less than 5 years versus 5 yearsor more, but we did not do so here because only a few mothers still reported being in the country for less than 5 years at the year 5survey.Chapin Hall at the University of Chicago 17

Table 8. Selected Characteristics of Year 5 Sample Mothers by Nativity aCharacteristicTGA (%)***All Mothers U.S.-born Foreign-born(N=310) (n=143) (n=167)Glades 12 21 5Non-Glades 75 68 80FOCiS initial risk screen (%)At-risk/high-need screen score 31 33 29Study risk index***Mean (SD) 3.5 (1.5) 3.1 (1.6) 3.9 (1.4)Mother’s race (%)***Hispanic 51 20 77Black, not Hispanic 41 65 20Age of mother***Mean age (SD) 29.0 (5.7) 26.8 (4.5) 30.9 (6.0)Teen mother at focal child’s birth (%)*** 15 24 7Mother’s education (%)***High school graduate/GED (or more) 49 68 34Marital status (%)***Married, living with husband 32 20 41Single, living with a partner 31 24 38Single, in a relationship, but not living with partner 11 21 2Single, not in a relationship 26 34 19Number of children (%)One 25 22 27Two 32 35 29Three or more 44 43 44Employment (%)Mother currently working 46 47 45Full-time*** b 29 35 23Part-time 17 12 22Husband/partner working*** 85 73 93Main language spoken in home (%)***English 50 92 15Spanish 43 7 73Haitian Creole 6 1 10Income-to-need ratio (%)Living at or below poverty threshold 52 48 56Income support (%)WIC*** 52 36 65Food Stamps** 47 57 38Home ownership (%)***Mother owns home 20 13 27Someone else in household owns home 9 16 2a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b “ Full-time” was defined as 35 or more hours/week. Foreign-born mothers worked an average of 29.4 (11.4) hours and U.S.-bornmoms, an average of 36.3 (9.6) hours. A one-way ANOVA indicated this difference was statistically significant at p < .001.Chi-square tests indicated differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 18

at the birth of the focal child than U.S.-born mothers (7%). Additionally, foreign-born mothers were morelikely to own their own home (27%) than U.S.-born mothers (13%). On the other hand, U.S.-bornmothers were more likely to have a high school education or above and to be employed full time (versuspart time). Foreign-born mothers also are more likely to use WIC than U.S.-born mothers (65% vs. 36%),whereas U.S.-born mothers are more likely than foreign-born mothers to use Food Stamps (57% vs. 38%)Mobility and Living CircumstancesAs noted above, 13 percent of the study families had moved from the TGAs (primarily the non-GladesTGAs) to other areas of the county. In addition, a number of families moved within the TGAs. About athird of the families moved at least once in each of the study years; the exception was year 4, when 42percent of the families moved. Also, most mothers who moved did so just once in the previous year.When we compared the stability of mothers’ residential housing over the course of the 5 years, we foundthat 25 percent did not move at all during this time period, and 3 percent moved just once. The remaining72 percent of mothers reported other mobility patterns. A small percentage (2% to 8% each year) hadmoved more than once each year.Over time, the percentage of mothers reporting one or more negative housing conditions declined,although this percentage increased significantly between years 4 (30%) and 5 (40%). With regard tosafety, the percentage of mothers who felt their neighborhood was unsafe remained stable at 14–16percent. Also, when mothers were asked if a concern about safety ever kept them from leaving orreturning to their homes, only 6 percent of the sample responded that they had missed doctors’appointments, limited grocery or other shopping, and/or stayed with other family members or friends forthis reason.The percentage of mothers reporting that it is easy to get places increased significantly, from 58 percent inyear 1 to 78 percent in year 5. As in previous years, less than a quarter of the sample mothers reporteddifficulty with transportation in the fifth year. However, some of the explanations given for theirdifficulties changed. As shown in Table 10, mothers continued to report the lack of a working car as theprimary reason. However, they were more likely to report in year 5 than they were in year 4 that publictransportation is not available (35% in year 5 vs. 20% in year 4). In addition, they were less likely toreport that their children made it hard for them to get places in year 4 and year 5 than in year 2 or year3—perhaps because their children are getting older and more self-sufficient and, therefore, are easier totake places. Finally, the expense associated with transportation was also cited fairly frequently, with 21percent mentioning this in year 5 compared to 13 percent in year 4. This increase could be the result offewer mothers currently being employed, and an increase in the number of study families currently livingChapin Hall at the University of Chicago 19

at or below the poverty threshold in the fifth year of the study.Table 9. Family Living Conditions over Time aConditionTGA (%)% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Glades 13 13 12 12 12Non-Glades 87 81 79 76 75Outside TGAs — 6 9 12 13Residential mobility (%)Did not move during past year — b 64 58 67 67Moved once during past year 29 37 31 30Moved more than once during past year 8 5 2 3HousingOne or more negative housing conditions (%) 47 48 38 30 40Heat or air conditioning did not work 21 19 13 13 14Electricity did not work 18 9 5 7 10Plumbing did not work 18 17 11 7 7Cooking appliances did not work 17 13 10 4 8Broken windows or doors 13 20 12 7 10Peeling paint 13 11 8 7 10Not enough basics for cooking, eating, sleeping 9 6 4 5 5Overcrowded; not enough space 8 6 4 6 9Bare electric wires 4 3 1 2 2Mean number of negative housing conditions (SD) 1.2 (1.7) 1.0 (1.4) 0.7 (1.1) 0.6 (1.1) 0.8 (1.2)Hurricane-related loss of $50 or more (%) — 78 21 1 1Neighborhood (%)Unsafe neighborhood because of illegal activities 14 15 16 16 16Transportation (%)Find it easy to get places 58 70 70 75 78Know how to drive 70 75 — — —Have a driver’s license 69 70 — — —Have regular use of a car 79 81 — — —a Data were weighted to adjust for the oversampling of Glades mothers and mothers screened “at risk.”b In the baseline survey, mothers were asked how many different places they had lived in the last 2 years; 67 percent of the sample hadmoved at least once in the previous 2 years.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: TGA1 vs. TGA2;TGA2 vs. TGA3; TGA2 vs. TGA4; TGA2 vs. TGA5; TGA3 vs. TGA5; mobility2 vs. mobility4; mobility3 vs. mobility4; mobility3 vs.mobility5; electricity1 vs. electricity2; electricity1 vs. electricity3; electricity1 vs. electricity4; electricity1 vs. electricity5; electricity2 vs.electricity3; electricity3 vs. electricity5; plumbing1 vs. plumbing3; plumbing1 vs. plumbing4; plumbing1 vs. plumbing5; plumbing2 vs.plumbing3; plumbing2 vs. plumbing4; plumbing2 vs. plumbing5; plumbing3 vs. plumbing4; plumbing3 vs. plumbing5; cooking1 vs.cooking3; cooking1 vs. cooking4; cooking1 vs. cooking5; cooking2 vs. cooking4; cooking2 vs. cooking5; cooking3 vs. cooking4;cooking4 vs. cooking5; windows1 vs. windows2; windows1 vs. windows4; windows2 vs. windows3; windows2 vs. windows4; windows2vs. windows5; windows3 vs. windows4; paint1 vs. paint3; paint1 vs. paint4; paint2 vs. paint4; heat1 vs. heat3; heat1 vs. heat4; heat1 vs.heat5; heat2 vs. heat3; heat2 vs. heat4; overcrowded1 vs. overcrowded3; overcrowded3 vs. overcrowded5; basics1 vs. basics3;#conditions1 vs. #conditions3; #conditions1 vs. #conditions4; #conditions1 vs. #conditions5; #conditions2 vs. #conditions3; #conditions2vs. #conditions4; #conditions2 vs. #conditions5; #conditions4 vs. #conditions5; condscore1 vs. condscore3; condscore1 vs. condscore4;condscore1 vs. condscore5; condscore2 vs. condscore3; condscore2 vs. condscore4; condscore2 vs. condscore5; condscore3 vs.condscore4; condscore4 vs. condscore5; hurricane2 vs. hurricane3; hurricane2 vs. hurricane4; hurricane2 vs. hurricane5; hurricane3 vs.hurricane4; hurricane3 vs. hurricane5; easy1 vs. easy2; easy3 vs. easy5; drive1 vs. drive2.Chapin Hall at the University of Chicago 20

Table 10. Reasons for Transportation Difficulties over Time a% % % % %Mothers Mothers Mothers Mothers MothersReasons b at Year 1 at Year 2 at Year 3 at Year 4 at Year 5(n=130) (n=94) (n=92) (n=78) (n=69)Do not own or have access to a car (or car does notwork)71 75 77 68 73Do not drive, no license, or no one to providetransportation45 44 32 44 34Public transportation not accessible or schedule isinconvenient40 21 23 20 35Children make it difficult to get places — c 16 4 1 1Too expensive (e.g., do not have bus fare or gasmoney)17 7 9 13 21Other (e.g., physical limitations, don’t know howto take bus)17 7 3 1 10Afraid to go out — — 0 0 1a Data were weighted to adjust for the oversampling of Glades mothers and mothers screened “at risk.”b Just 10 of the 70 mothers in the year 5 sample reported experiencing transportation difficulties each year they were interviewed.Multiple responses allowed.c These items were not included in the year 1 or year 2 surveys.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:easy1 vs. easy2; easy1 vs. easy3; easy1 vs. easy4; easy1 vs. easy5; easy2 vs. easy5; easy3 vs. easy5; nocar1 vs. nocar3; nocar1vs. nocar4; nocar1 vs. nocar5; nocar2 vs. nocar4; don’tdrive1 vs. don’tdrive3; don’tdrive1 vs. don’tdrive4; don’tdrive1 vs.don’tdrive5; don’tdrive2 vs. don’tdrive3; don’tdrive2 vs. don’tdrive4; don’tdrive2 vs. don’tdrive5; publictransportation1 vs.publictransportation3; publictransportation1 vs. publictransportation4; publictransportation1 vs. publictransportation5;publictransportation2 vs. publictransportation3; publictransportation2 vs. publictransportation4; publictransportation2 vs.publictransportation5; publictransportation3 vs. publictransportation5; expense1 vs. expense4; expense3 vs. expense5; expense4vs. expense5.SummaryIn the fifth year of the study, most of the study families still lived in one of the TGAs; 12 percent lived inthe Glades TGA, and 75 percent lived in the non-Glades TGAs. Thirteen percent of the sample lived inother areas of Palm Beach County, which was fairly consistent with the 12 percent who lived outside theTGAs in the previous year. Although this suggests some stability in the mothers who remained in thesample at year 5, we should also note that 33 percent of the families reported moving at least once duringthe year; this is also consistent with the percentage of families who reported moving at least once in year4 (33%).The percentage of mothers who reported living in a home owned by a family member changed minimallyfrom year 4 to year 5, with 27 percent reporting home ownership in year 4 and 29 percent in year 5. Norwas there a change in the percentage of mothers who reported living in an unsafe neighborhood (16%).On the other hand, one indication of decline in living conditions was reflected in an increase in theChapin Hall at the University of Chicago 21

percentage of mothers who reported one or more negative housing conditions, such as electrical orplumbing problems, from year 4 (30%) to year 5 (40%).Overall, almost half of the mothers in the year 5 sample have worked at some point since the beginning ofthe study. Similar to year 4, nearly half (46%) of the mothers were working at the time of the year 5interview. The mothers’ estimates of their families’ income for the preceding year were similar to those inyear 4, with a little less than half (46%) reporting household incomes of less than $20,000 for theprevious year, as compared to 54 percent in year 4. Calculation of an income-to-need ratio based onhousehold size, the number of children under age 18, and the federal poverty thresholds indicates that 52percent of the families in the year 5 sample were living at or below the federal poverty threshold theprevious year.Household sizes remained fairly constant during the first 5 years of the study. The percentage of motherswho reported they were married in the fifth year (32%) is almost the same as in the fourth year, althoughthe percentage of unmarried mothers living with a partner (31%) continued to decline from the first year,when it was 39 percent. At the same time, the percentage of mothers with two or more children increased,with three-quarters of the sample having two or more children at the time of the fifth interview. More than4 in 10 of the mothers had had another child since the birth of their focal child, and 6 percent werepregnant at the time of the year 5 interview.Chapin Hall at the University of Chicago 22

Health, Healthcare, and ChildDevelopmentMothers’ Health and HealthcareWhen asked about the state of their health in year 5, a large majority (83%) of the mothers described it infavorable terms as “good,” “very good,” or “excellent.” This percentage is similar to the percentage ofmothers who described their health in these terms the previous year (see Table 11). Consistent with thefindings of the previous 4 years, the percentage of mothers who reported physical or mental healthproblems that kept them from working or attending school, or limited the kind of work they could do, wassmall (between 5 and 7% in each year).Mental Health and FunctioningWe again used two instruments to assess maternal functioning in the fifth year: the 20-item Center forEpidemiologic Studies Depression Scale (CES-D; Radloff, 1977) and the Parenting Stress Index ShortForm (PSI/SF; Abidin, 1995). Scores on the CES-D can range from 0 to 60, with higher scores indicatingthe presence of more depressive symptoms. CES-D scores for the sample in year 5 ranged from 0 to 43,with an average of 9.7 (see Table 11). Twenty-one percent of the mothers had scores of 16 or higher,suggesting some level of depression. This percentage was fairly stable between years 2 and 4, andrepresents a significant decrease compared to year 1. Five percent of the mothers had scores of 30 orhigher, indicating the possibility of severe depression. Compared with other mothers in the sample, thesemothers were more likely to be single and not currently in a relationship. They also had slightly lowerlevels of education and demonstrated somewhat higher levels of parenting stress on the PSI/SF than didthe rest of the sample.Chapin Hall at the University of Chicago 23

Scores on the PSI/SF can range from 0 to 180, with higher scores indicating greater stress; a score at orabove the 85 th percentile, defined as a raw score of 86 or higher, is considered to indicate clinicallysignificant levels of stress. Total stress scores for mothers in the year 5 sample ranged from a low of 36 toa high of 147, with a mean score of 66.5. Seventeen percent of the sample scored at or above the 85 thpercentile, which is similar to the level in the previous year, but is significantly higher than year 3. Also,consistent with earlier findings, there was a significant correlation between the CES-D and PSI/SF scoresat year 5 (r 2 = 0.413, p < .001).Table 11. Mothers’ Health and Maternal Functioning over Time aCharacteristic% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Health “good/verygood/excellent”83 78 85 85 83Depression (CES-D) bCES-D score ≥16 34 23 18 23 21Mean (SD) depression score 13.0 (10.1) 10.1 (8.8) 8.8 (9.1) 10.2 (9.8) 9.7 (8.8)Parental Stress (PSI/SF) cPSI/SF score ≥ 86 — 17 12 15 17Mean (SD) PSI/SF score — 65.0 (22.8) 62.7 (19.5) 63.7 (21.5) 66.5 (20.9)Use of alcohol (any) 7 8 10 11 18Use of tobacco d — — 9 6 11DCF report of abuse or neglect eInvestigated 10 12 9 7 8Indicated 6 8 5 4 2aData were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b A raw score of 16 or higher is above the normal range for the CES-D assessment.cIf a mother skipped two or more questions in any domain on the PSI/SF, her score was not included in the sample mean; theseresults are based on responses of 289 year 2 mothers, 295 year 3 mothers, 308 year 4 mothers, and 276 year 5 mothers. ThePSI/SF was not administered in year 1.d Mothers were not asked about smoking in the first 2 years. According to Vital Statistics, 5 percent of the sample smoked or quitsmoking during pregnancy.eSource: DCF data for focal child only. Year 1 refers to the first year after the focal child’s birth (0–1 year), and year 2, thesecond year (1–2 years). We have not yet requested new data for year 3 (2–3 years) or year 4 (3–4 years) because of transitions inagency data systems, but will provide these in a subsequent report.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05or lower:health2 vs. health3; health2 vs. health4; health2 vs. health5; depressed1 vs. depressed2; depressed1 vs. depressed3; depressed1vs. depressed4; depressed1 vs. depressed5; depressionscore1 vs. depressionscore2; depressionscore1 vs. depressionscore3;depressionscore1 vs. depressionscore4; depressionscore1 vs. depressionscore5; depressionscore2 vs. depressionscore3;depressionscore3 vs. depressionscore4; stressscore2 vs. stressscore3; stressscore3 vs. stressscore5; stressscore4 vs. stressscore5;alcohol1 vs. alcohol4; alcohol1 vs. alcohol5; alcohol2 vs. alcohol5; alcohol3 vs. alcohol5; alcohol4 vs. alcohol5; smoke3 vs.smoke 4; smoke4 vs. smoke5; investigated1 vs. investigated2; investigated1 vs. investigated3; investigated1 vs. investigated4;investigated2 vs. investigated3; investigated2 vs. investigated4; investigated2 vs. investigated5; investigated3 vs. investigated4;investigated3 vs. investigated5; investigated4 vs. investigated5; indicated1 vs. indicated2; indicated1 vs. indicated3; indicated2vs. indicated4; indicated2 vs. indicated5.As expected from other data on child abuse and neglect rates, the proportion of families in the PBC studysample who were either investigated or indicated for child abuse and/or neglect was highest in the firstChapin Hall at the University of Chicago 24

two years of the study. In year 1 10 percent of the families were investigated for child abuse and/orneglect, and 6 percent were indicated; year 2 showed the largest percentage of families investigated, with12 percent of the sample families being investigated and 8 percent being indicated for child abuse and/orneglect. (This bump from 6 to 8 percent from year 1 to year 2 was statistically significant.) As reported inthe third study report (Spielberger, Rich, Gouvea, et al. 2009), when compared with the indicated reportsfor the county and combined TGA populations, our study sample had an indicated rate comparable to thatof the county but lower than the rate for the TGA population in the first year. In the second year,however, the rate for the sample was higher than the rates for both the county and the TGA population.These rates improved over time, with a significant decline in the percent of families who wereinvestigated and indicated for child abuse and neglect in years 3, 4, and 5. With regard to indicatedreports of child abuse and/or neglect, just 5 percent of families were indicated of child abuse and/orneglect in year 3 and this decline continued in year 4 (4%) and year 5 (2%).Alcohol and Tobacco UseEighteen percent of the mothers in the year 5 sample reported they currently drank alcoholic beverages.Although the percentage was modest, it was a significant increase over the 7 percent reported by the samegroup of mothers in the year 1 interview. The average amount of alcohol consumed by these mothers alsoincreased somewhat from the previous year, although only 10 mothers reported having four or moredrinks per week. Thirty-one percent of the mothers who drank alcohol said they had less than one drinkper week, which was consistent with what they reported in the previous year; another 52 percent reporteddrinking between one and three alcoholic drinks per week. With regard to smoking, 11 percent of themothers in the year 5 sample said they smoked, which was a significant increase from year 4 where 6percent reported smoking. 17Mothers’ HealthcareThe percentages of sample mothers who were covered by health insurance and who received regularmedical care were stable in years 2 to 5 (see Table 12). Forty-five percent of the mothers were covered by17 Mothers were not asked about smoking in either the year 1 or year 2 interviews, although Vital Statistics data indicated thatonly 3 percent smoked during pregnancy. These percentages may underreport actual use of alcohol and tobacco by studymothers. Percentage of sample mothers who reported drinking alcohol is much lower than the 2005 average of 51 percentreported for Florida or 50 percent nationally for all women 18 through 44 years of age (U.S. Department of Health and HumanServices Centers for Disease Control and Prevention, 2005). In addition, the percentage obtained from Vital Statistics for mothersin the sample who smoked during pregnancy is much lower than the results of the 2006 Florida Adult Tobacco Survey indicatingthat 14 percent of women 18 years and older smoke.Chapin Hall at the University of Chicago 25

insurance in year 5. A little less than one-quarter of the sample did not receive regular medical care. 18U.S.-born mothers were somewhat more likely to receive regular care (83%) than foreign-born mothers(72%), and the difference was statistically significant at the 5 percent level. At the same time, only 19percent of foreign-born mothers had health insurance in the fifth year compared with 75 percent of U.S.-born mothers (χ 2 = 96.318, p < .001).The most common locations for routine medical care were public health clinics and doctors’ offices.However, there were differences between U.S.-born mothers and foreign-born mothers in the primarylocation of their medical care, which seem to reflect the aforementioned differences in health insurancecoverage. Among foreign-born mothers, 43 percent used a public health clinic, 15 percent received care ata doctor’s office, and 14 percent received care in another facility. In contrast, 72 percent of U.S.-bornmothers received regular care at a doctor’s office, and 6 percent at a public health clinic.It is possible that, unless they are pregnant, some women may not feel the need to see a physician on aregular basis—or, if they have a medical need, they may decide to postpone a visit to the doctor if theylack health insurance or have inadequate coverage. Also, data from the qualitative study indicates thatmothers are often more concerned about their children’s healthcare than their own. For example, aforeign-born mother from Mexico with one child reported that when her husband lost his job they did nothave insurance for themselves. However, she did not worry because they had Medicaid for their child.Table 12. Healthcare of Mothers over Time aCharacteristic% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Health insuranceMother covered 58 43 42 41 45Mother gets regular medical care 85 73 74 73 77Location of mother’s routine medical careDoctor’s office 46 40 41 43 41Public health department clinic 34 30 26 26 26Other clinic, health center, or emergencyroom5 3 6 3 10Mother does not get regular medical care 15 27 26 27 23a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .001:insured1 vs. insured2; insured1 vs. insured3; insured1 vs. insured4; insured1 vs. insured5; location1 vs. location2; location1 vs.location3; location1 vs. location4: location1 vs. location5.18 These results are despite the fact that, according to the Vital Statistics, slightly over half (51%) had had a child subsequent tothe focal child. Most of these mothers presumably would have been eligible to receive coverage through Mom Care, a Medicaidfundedprogram authorized by the Sixth Omnibus Budget Reconciliation Act (SOBRA); however, this coverage is limited to thefirst 60 days of the postpartum period.Chapin Hall at the University of Chicago 26

Subsequent Pregnancies and Prenatal CareAt the time of the fifth-year interview, 140 (45%) of the 310 sample mothers had given birth to at leastone child after the birth of their focal child. 19 Of these 140 mothers, 85 percent gave birth once after thefocal child, 12 percent gave birth twice, and 3 percent gave birth 3 times. Just six mothers (4%) were 19years of age or younger at the birth of the first subsequent child, and one mother was still a teen at thebirth of the second subsequent child. 20For the 140 mothers who had at least one subsequent child, the mean (SD) number of days in betweenbirths was 816 (311.5) days, or approximately 27 months, with a range of 277 to 1,494 days. The majority(53%) of these 140 mothers spaced the births of their focal child and the first subsequent child at least 2years apart. Just over a quarter (28%) of these mothers spaced these two children a year and a half to 2years apart, 13 percent spaced these two children a year to a year and a half apart, and 6 percent spacedthese two children less than a year apart.The spacing between births was smaller for mothers who gave birth to more than one child after theirfocal child. For the 21 mothers who gave birth to two children after their focal child, the mean (SD)number of days between the first subsequent child and the second subsequent child was 617 (173.3) days,or approximately 21 months, with a range of 306 to 945 days. Of these 21 mothers, a majority (56%) hadtheir second subsequent child a year and a half to 2 years after the birth of the first subsequent child.Twenty-one percent had their second subsequent child at least 2 years after the first subsequent child; 21percent gave birth to their second subsequent child a year to a year and a half after the first subsequentchild; and 3 percent gave birth to their second subsequent child a year or less after the birth of the firstsubsequent child.Finally, for the four mothers who gave birth three times after their focal child, the mean (SD) number ofdays between the birth of the second subsequent child and the third subsequent child was 371 (53.6) daysor approximately 12 months, with a range of 338 to 457 days.Children’s Health and HealthcareAs in previous years, mothers’ assessments of the health of the focal children were overwhelminglypositive, with 94 percent describing the child’s health as “good,” “very good,” or “excellent.” At the same19 Vital Statistics data through 2009, however, indicate a total of 187 subsequent births for these 310 mothers around the time ofthe year 5 interview; the reason for this discrepancy is not clear.20 These fertility statistics are higher than a report from the Department of Health and Human Services Centers for DiseaseControl and Prevention (2005), which indicates that in 2002 the interval between first and second births was less than 12 monthsfor 5 percent of low-income mothers and between 13 and 24 months for 23 percent of low-income mothers ages 20 to 44.Chapin Hall at the University of Chicago 27

time, however, almost a fifth (19%) of the sample reported that their focal child had special medicalneeds; a similar proportion reported that other children in the family had special medical needs. Inaddition, 16 (7%) of the year 5 mothers reported that their focal child, as well as at least one of their otherchildren, had special medical needs.As shown in Table 13, the percentage of mothers reporting that a medical professional told them that theirfocal child has special medical needs increased significantly from the first year to the second year, butremained stable thereafter. This is perhaps not surprising, given that some conditions in the focal childrenwould not have been apparent or diagnosed at the time of the baseline interview. There also appeared tobe some variability/instability in reports of special medical needs over time; for example, over half (57%)of the mothers who reported that their focal child had special needs at the fifth interview had not reportedthem the previous year.When asked to describe their children’s specific needs, a handful of mothers reported specific permanentdisabilities or conditions in their children, for example, sickle cell anemia or spina bifida. The mostfrequently reported medical condition for the focal children was asthma, mentioned by 44 percent of themothers reporting that their focal child had special medical needs. The next most commonly cited medicalissue was speech problems, with 11 percent of mothers reporting this problem. The frequency of differentconditions varied some by race/ethnicity. Within the group of mothers who reported that their focalchildren had medical problems, mothers who were black or Hispanic reported a higher incidence ofasthma and other respiratory conditions (50% and 39%, respectively) than mothers who were white orother (20%). 21Table 13. Children’s Health over Time aCharacteristicFocal child’s health% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Child’s health good/very good/excellent 91 88 94 93 94Child has special medical needs 9 21 19 17 1921 19 18 15 22Other children have special medical needs(n=185) (n=184) (n=214) (n=226) (n=230)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:health2 vs. health3; health2 vs. health4; health2 vs. health5; medical1 vs. medical2; medical1vs. medical3; medical1 vs.medical4; medical1 vs. medical5; other2 vs. other5; other3 vs. other5; other4 vs. other5.21 This last finding is consistent with other published research. According to the National Center for Health Statistics (Akinbami,2006), in 2004 and 2005, Puerto Rican (19%) and non-Hispanic black children (13%) had higher rates of asthma than did non-Hispanic white children (8%).Chapin Hall at the University of Chicago 28

Children’s HealthcareMost (80%) of the mothers reported taking their children to the doctor for at least one well-child check-upin the 6 months prior to the fifth-year interview (see Table 14). On average, mothers said their childrenhad been to the doctor approximately once in the past 6 months for routine care. Overall, by the time ofthe year 5 interview, mothers reported taking their focal child to the doctor an average of 19 times forroutine care since birth; the number of times ranging from a low of 9 times to a high of 81 times. 22A majority (87%) of mothers reported taking their children to a doctor’s office for routine medical care,whereas 9 percent took their children to a public health clinic and 3 percent to another clinic or healthcenter. Relative to the first year, these results represent an increase in the use of a medical doctor and adecrease in the use of a public health clinic or another health center for children’s care. As with their owncare, the frequency with which foreign-born and U.S.-born mothers used a doctor’s office versus a publichealth clinic for their children’s healthcare varied. Although a majority of both groups were more likely touse a doctor’s office rather than a public health clinic, the percentage of U.S.-born mothers who did so(94%) was significantly higher than the percentage of foreign-born mothers who did so (81%).Correspondingly, 14 percent of foreign-born mothers used public health clinics for their children’s care,whereas only 4 percent of U.S.-born mothers did so (χ 2 = 14.354, p < .01).Analysis of qualitative data suggests that a primary reason for the decline in the use of public healthclinics over time is the high cost of services for uninsured children and adults. These data suggest thatmothers may start prenatal care at a public health clinic when they become pregnant and then continue touse the clinic for a while after the birth of their focal child, especially if they are also using the WICProgram, which is usually located in the same facility. However, because of the expense, mothers use theclinic only sporadically for care for themselves, for children not eligible for public insurance, and forMedicaid-eligible children during lapses in Medicaid coverage. Among these mothers, the clinic is usedmost often for family planning services, school physicals, and illnesses that cannot be treated at home.Almost all of the mothers (97%) reported that their focal child had received all required immunizations.Ten mothers (3%) said their children had not received all of their shots. When asked why, eight motherssaid they did not have insurance to pay for the shots; one said she “had missed the appointment” andreported that “the child was sick and could not receive the scheduled shot.”22 Mothers were asked every 6 months, both at the in-person annual interview and in the brief telephone interview, how manytimes they had taken their focal child to the doctor for “well-child” visits during the previous 6 months, or, in the case of thebaseline interview, since birth.Chapin Hall at the University of Chicago 29

Table 14. Healthcare of Children over Time aCharacteristicHealth insurance% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Focal child has insurance 65 83 80 80 85All children in family covered 59 74 71 74 76Some children in family covered 20 11 11 10 11No children in family covered 22 15 18 17 13Types of health insurance for children bMedicaid 67 75 82 78 77Private plan or HMO 9 10 14 18 16SCHIP, KidCare, MediKds, Healthy Kids 4 2 5 6 8Location of children’s routine medical careDoctor’s office 61 79 84 87 87Public health department clinic 32 18 13 7 9Other clinic or health center 5 3 3 6 3No place most often 1 0 0 0 0Children do not get regular care 2 0 1 0 1Focal child had check-up in past 6 months c — 98 90 80 80Focal child got recommended immunizations — — 93 97 97a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Percentages are based on the total sample of 310. Multiple responses were allowed. Between 1 and 3percent of mothers listed otherpublicly funded programs.c In the baseline interview, we asked mothers how many times they had taken their babies to the doctor since birth rather than the past6 months, so year 1 responses are not included. All mothers took their newborns to the doctor at least once during the first year.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: insured1vs.insured2,insured3,insured4, and.insured5; insuracecoverage1 vs. insurancecoverage, insurancecoverage3, insurancecoverage4, andinsuracecoverage5;private1 vs. private3; private1 vs. private4; private1 vs. private5; private2 vs. private3; private2 vs. private4;private2 vs. private5;medicaid1 vs. medicaid3; medicaid1 vs. medicaid4; medicaid1 vs. medicaid5; medicaid2 vs. medicaid3;medicaid2 vs. medicaid4; medicaid2 vs. medicaid5; schip1 vs.schip3; schip1 vs.schip4; schip1 vs.schip5; schip2 vs.schip3; schip2vs.schip4; schip2 vs.schip5;hcd4 vs. hcd5; location1 vs. location3; location1 vs. location4; location1 vs. location5; location2 vs.location3; location2 vs. location4; location2 vs. location5;checkup2 vs. checkup3; checkup2 vs. checkup4; checkup2 vs. checkup5;checkup3 vs. checkup4; checkup3 vs. checkup5;immunizations3 vs. immunizations4; immunizations3 vs. immunizations5.Three-quarters of the mothers in the year 5 sample had health insurance for all of their children, andanother 11 percent had coverage for some of their children (see Table 14). These percentages were aboutthe same as they were in the previous 3 years. 23 Medicaid coverage for children was reported by 77percent of the year 5 sample. Smaller percentages had coverage through a private plan or HMO (16%) orfrom the State Children’s Health Insurance Program (SCHIP) or Florida KidCare (8%). 24 Those with23 Various reports estimate that between 16 and 24 percent of Florida’s children are uninsured and one-fourth of uninsuredchildren do not receive any medical care during the year (Henry J. Kaiser Family Foundation, 2008).24 Participation in Florida’s KidCare program declined statewide and in Palm Beach County between 2004 and 2006 (CSC, February2007). One-fourth of uninsured children in Florida had no medical care for the entire year (Covering Kids and Families 2005).Chapin Hall at the University of Chicago 30

private plans or HMOs increased significantly from the first to the fifth years.A quarter of mothers did not have insurance for some or all of their children, which is consistent with theprevious 3 years (but represented a decrease compared to the first year). Given that more foreign-bornthan U.S.-born mothers were without health insurance for some or all of their children (81% vs. 25%), itis not surprising that the reason cited most often for lack of insurance for children was that the childrenwere ineligible because they were not born in the United States. 25 As shown in Table 15, the next mostcommonly cited reasons were that the paperwork was in process and that Medicaid coverage was lost.Table 15. Reasons Children Not Covered by Health Insurance over Time aReasonProgram factors% MothersYear 1 Year 2 Year 3 Year 4 Year 5(n=128) (n=82) (n=89) (n=82) (n=74)Paperwork is in progress 58 31 14 19 22Child/family is not eligibleNot eligible because children not born in U.S. — b 35 29 27 32Not eligible/became ineligible (other reason) 15 20 18 11 10Lost Medicaid 9 22 16 17 21Job change 1 6 16 15 21Insurance company refused coverage 1 6 13 16 1Tried but did not qualify 1 4 7 2 5Child on waiting list 4 4 2 0 0Change in marital status 0 0 1 0 1Individual factorsCost too high 13 2 13 9 16Coverage not offered by employer 13 0 4 12 4Too much trouble to apply 0 2 6 3 7Did not know how to apply 12 0 0 3 0Did not know was eligible 4 2 0 2 2Other c 21 7 9 8 10a Data were weighted to adjust for the oversampling of Glades mothers and mothers screened “at risk.” Mothers in the year 5sample who reported some or none of their children were covered each year they were interviewed. Multiple responses wereallowed.b This response option was not included in the year 1 survey.c “Other” includes “[Insurance] expired and has not renewed it,” “[It] has to be done by computer and I don’t have a computer,”“respondent told to provide additional information,” “respondent will apply tomorrow,” “they keep sending letters, but not thecards,”“DCFS took children; father has custody,” “respondent has not looked for insurance for older children,” “no transportation.”25 To be eligible for Florida KidCare, children generally must be U.S. citizens. However, noncitizen children classified by the as“qualified aliens” are eligible, including children who have been legal permanent residents for at least 5 years and Cuban andHaitian immigrants (http://www.floridadoh.com, accessed April 22, 2008).Chapin Hall at the University of Chicago 31

While some of the children without health insurance were older children in the family who were born inanother country, 15 percent of the focal children were without health insurance at the time of the year 5interview. This is noteworthy since all of these children were born in the United States and thereforeeligible for Medicaid if the household fell below a certain income threshold. These weredisproportionately children of foreign-born mothers (69%) and families whose income-to-need ration wasabove the poverty level (63%).Although 40 percent of focal children have had sporadic insurance coverage since birth and 1 percenthave never been covered by insurance, another 40 percent have had consistent insurance coverage sincebirth. An additional 16 percent did not have coverage at the time of the first interview, but have beencovered since the second interview. The qualitative data provide more detail on some of the reasonsmothers do not have insurance for their children. Newborn children in families who meet certain incomerequirements are eligible for Medicaid or the SCHIP program for 1 year after birth, and, according tomothers in the study, enrollment occurs at the hospital. However, there may be a waiting period beforecoverage is confirmed. In addition, to continue to receive health coverage for their children after their firstbirthday, mothers must recertify every year. Thus, there can be gaps in insurance while paperwork isbeing processed or if mothers do not reapply early enough before it expires. For foreign-born parents, inparticular, the process of renewing their children’s insurance can be time consuming and difficultdepending on their ability to use a computer and the availability of service providers, family members, orfriends to assist with translation.Thus, as we have reported in previous years, mothers’ lack of knowledge of public health insuranceprograms, their eligibility requirements, or the application process could be other reasons that childrenwere uninsured in the fifth year. Among mothers who did not have coverage for some or all of theirchildren, almost all (97%) said they had heard of KidCare, Florida’s public health insurance program forchildren; in contrast, in year 4, only 66 percent of these mothers had heard of KidCare. Almost half (47%)were aware of the Healthy Kids program, a decrease from 56 percent in year 4; and 62 percent of motherswho had uninsured children knew about MediKids, an increase from year 4 (47%). Smaller percentagesof mothers were aware of Children’s Medical Services (25%) and SCHIP (6%).Health and Healthcare by Selected Family CharacteristicsTable 16 summarizes the status of the health and healthcare of mothers and their children in the fifth yearby TGA residence, race/ethnicity, and nativity. With respect to TGA residence, mothers living in theGlades were significantly more likely than mothers in the other two areas to report being covered byinsurance. They were also more likely than mothers living in the non-Glades TGAs and those livingChapin Hall at the University of Chicago 32

outside the TGAs to report that all or some of their children have insurance coverage. Finally, mothers inthe Glades were significantly more likely than mothers in the non-Glades TGAs to report receivingregular medical care.There were also differences in self-reported health information and insurance coverage according tomothers’ race/ethnicity. Black mothers were significantly more likely than Hispanic mothers to report thattheir health and their focal child’s health was “good” to “excellent.” At the same time, however, blackmothers were more likely than Hispanic mothers to report having a physical or mental health problem(11% vs. 1%), while white/other mothers reported more health problems (29%) than both black andHispanic mothers. In addition, black and white/other mothers were more likely than Hispanic mothers toreport that they had insurance coverage for themselves, while black mothers were more likely than eitherHispanic or white/other mothers to report that all or some of their children were covered by healthinsurance. Finally, black mothers were more likely than Hispanic or white/other mothers to reportreceiving regular medical care themselves.According to Table 16, differences by nativity reflect the race/ethnicity results. That is, U.S.-born motherswere more positive about their children’s health and their own health than foreign-born mothers, but morelikely to report having a physical or mental health issue themselves. The reasons for these differences arenot clear. One reason might be differences in interpretations of the survey question about general health;another might be different views of what it means to have special medical needs. Also, U.S.-born motherswere more likely to report that they have insurance, and that at least some of their children haveinsurance. Perhaps because they were more likely to have insurance, U.S.-born mothers were also morelikely to get regular medical care for themselves.Chapin Hall at the University of Chicago 33

Table 16. Mothers’ and Children’s Health Status by TGA, Race/Ethnicity, and Nativity at Year 5 aIndicatorHealthTGA Race/Ethnicity Nativity% AllMothersNon- OutsideWhite/ ForeignbornbornU.S.-GladesBlack HispanicGlades TGAOther(N=310) (n=38) (n=232) (n=40) (n=126) (n=159) (n=25) (n=167) (n=143)Mother’s health “good” to “excellent”* 83 92 81 83 91 75 92 77 90Mother has physical or mental health problem* 7 3 7 13 11 1 29 1 14Focal child’s health “good” to “excellent”* 95 95 94 98 98 91 100 92 98Focal child has special medical needs 19 29 17 20 21 16 21 16 22Other children have medical problems 22 20 23 18 26 19 20 21 23Health insuranceMother covered by insurance* 45 82 39 39 68 23 64 19 75All/some children covered by insurance* 87 95 88 76 97 80 84 83 93Focal child has insurance* 85 95 85 76 94 78 80* 81 90HealthcareMother receives regular medical care* 77 92 75 78 88 69 68* 72 83Child received well-child check-up in last 680 87 78 90 84 79 72 80 81monthsa Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: [By TGA] motherinsured(1) vs. motherinsured(2);motherinsured(1) vs. motherinsured(3); some/allchildreninsured(1) vs. some/allchildreninsured(3); [By Race/Ethnicity] mothershealth(1) vs. mothershealth(2);motherhealthproblem(1) vs. motherhealthproblem(2); focalchildhealth(1) vs. focalchildhealth(2); motherinsured(1) vs. motherinsured(2); motherinsured(2) vs. motherinsured(3);some/allchildreninsured(1) vs. some/allchildreninsured(2); focalchildinsured(1) vs. focalchildinsured(2); mothermedicalcare(1) vs. mothermedicalcare(2); mothermedicalcare(1)vs. mothermedicalcare(3); [By Nativity] mothershealth(1) vs. mothershealth(2); motherhealthproblem(1) vs. motherhealthproblem(2); focalchildhealth(1) vs. focalchildhealth(2);motherinsured(1) vs. motherinsured(2); some/allchildreninsured(1) vs. some/allchildreninsured(2); focalchildinsured(1) vs. focalchildinsured(2);mothermedicalcare(1) vs.mothermedicalcare(2).Chapin Hall at the University of Chicago 34

Children’s Behavior and DevelopmentAt the time of the fifth in-person interview, the focal children ranged in age from 47 to 56 months, withan average age of 50 months. Fifty-four percent of the children are boys. 26 In the fifth year, mothers wereasked questions drawn from the Early Childhood Longitudinal Study Birth Cohort (ECLS-B) preschoolparent interview (National Center for Education Statistics, 2003; Andreassen & Fletcher, 2007) regardingthe behavior and development of the focal children. 27 These questions assessed children’s language andcommunication skills, social-emotional competencies, and emergent literacy/pre-academic behaviors (useof books, drawing, writing, letter-recognition, counting, etc.).Development is a function of both individual characteristics and the social environment. Thus, althoughthe behaviors and abilities mentioned above serve as useful indicators about how a child is developing,the actual age at which a normally developing child exhibits them can vary considerably (Andreassen &Fletcher, 2007). Also, prior to the ECLS-B, few national norms were available to link the age at whichthese milestones are achieved with future development, except for evidence suggesting that delays inreaching milestones are linked to poorer outcomes later.The tables in this section present the percentages of sample children whose mothers reported that theywere demonstrating selected behaviors that are developmentally characteristic of 4- and 5-year-oldchildren at the time of year 5 survey. We also describe differences in their reported behaviors by age andnativity. In a subsequent report, we will compare the percentages of children in our sample with those inthe national ECLS-B study sample reaching milestones used in the year 5 survey.Social and Emotional DevelopmentStudy mothers described their children’s social and emotional competencies in very positive terms. Asshown in Table 17, 83 percent or more of the sample characterized their focal child as happy, cooperative,and well liked by other children, as well as a child who enjoys playing with other children and who iseager to learn and try new things “most of the time.” Although a sizable percentage (42%) reported thattheir child also has difficulty sitting still, this is not atypical behavior for 4-year-olds. The fact that at leasta quarter of the mothers reported that their child gets “upset easily” or “worries a lot” is a concern,although the reason for this response is not clear. Overall, mothers’ reports on their children’s social-26 Five of the 310 mothers in the year 5 sample had multiple births, so in total, the year 5 sample represents 316 children, 114girls and 172 boys. When responding to the child development questions, mothers of twins and triplets were asked to refer to theoldest child. There were no differences in the mean ages of boys and girls; as of July 2010, both boys and girls had a mean age of5.5 years (with a standard deviation of 0.34 and 0.35 respectively).27 Another potential source of developmental information on these children are scores on the Ages and Stages Questionnaire(ASQ), which is filled out by parents who receive some maternal health services or whose children are enrolled in childcare.However, completion rates tend to be low; ASQ data were available for just 15 percent of the sample children in the fifth year.Chapin Hall at the University of Chicago 35

emotional development are in line with their reports in the previous year (see Table C- 4 in Appendix C),but they also show modest change in a positive direction for most items at year 5. The most importantdifferences noted are declines in the percentages of children who were reported to do “things withoutthinking” and who were described as “aggressive.”We also looked at whether mothers’ reports on their children’s social-emotional development differed bymother’s nativity (Table C- 5). We did not find any differences between foreign- and U.S.-born mothersin terms of reports of positive characteristics. However, we did find that foreign-born mothers weresignificantly more likely than U.S.-born mothers to report that their child is “overly active,” “worriesabout things,” and “does things without thinking” at least some of the time. These responses may reflect,in part, cultural differences in expectations for the behavior of preschool children.Communication and Language DevelopmentChildren develop language skills largely through social interactions—playing, listening, talking, andreading with other people—and children learn the style and rules of communication that are characteristicof their family and culture. As with other areas of development, children learn language at very differentrates depending on both individual characteristics and the social context. However, most have said theirfirst words by their first birthday, and most have started combining words by their second birthday.Children’s early language and communication skills are important in their own right, but they are alsoclosely linked to their cognitive and social development and, particularly, their future language andliteracy development. Thus, we believe it is especially important in the PBC family study to attempt toassess children’s language development during the preschool years.In terms of general communication skills, almost two-thirds or more of the mothers reported that theirfocal child listens well, speaks clearly, uses “I” to refer to him- or herself, uses appropriate socialgreetings and can say his or her first and last name “almost all of the time” (see Table 18). “Waiting for aturn to talk” did not come as easily, however, for these 4-year-olds. About half of the mothers reportedthat their children could “sometimes” wait their turn, while 20 percent responded “almost all of the time.”Also, as shown in Table C- 7, there were some small but statistically significant differences in reports byforeign- and U.S.-born mothers. On the one hand, foreign-born mothers were more likely to report thattheir focal child is a good listener and refers to him- or herself as “I.” On the other hand, U.S.-bornmothers were more likely to report that their focal child speaks clearly, uses appropriate social greetings,and can say his or her first and last name. 28 However, the latter two differences may reflect differences in28 However, these differences are only significant at the 10 percent level.Chapin Hall at the University of Chicago 36

cultural norms regarding what sorts of greetings are appropriate for children and when they are taught tosay their first and last names.Table 17. Social and Emotional Behaviors of Focal Children at Year 5 a % Mothers (N=310) bSocial-Emotional Indicator“Almost All the Time” “Sometimes”Positive behaviorsFocal child seems happy 95 5Focal child is eager to learn new things 92 7Focal child likes playing with other children close to his/her age 87 11Focal child is accepted and liked by other children 86 14Focal child likes to try new things 83 14Focal child helps or cooperates with adults 70 27Focal child adjusts easily to new situations 65 28Focal child pays attention well 61 33Focal child finishes what he/she is asked to do 45 45Negative behaviorsFocal child is overly active and unable to sit still 42 30Focal child gets upset easily 29 45Focal child worries about things 24 33Focal child does things without thinking 10 33Focal child is aggressive 8 25a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mothers used a 3-point scale to rate their child’s behaviors: “almost never,” “sometimes,” and “almost all the time.”Table 18. Communication Skills of Focal Children at Year 5 aCommunication IndicatorFocal child…% Mothers b (N=310)“Almost All the Time”“Sometimes”If asked, can say his or her first and last name 81 8Refers to him or herself as “I” 73 18Uses appropriate social greetings 73 15Speaks clearly so strangers understand 71 22Is a good listener 64 29Waits his or her turn to talk 20 49aData were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bMothers used a 3-point scale to rate their children’s behaviors: “almost never,” “sometimes,” and “almost all the time.”When asked how the focal children communicate, 70 percent of mothers reported that their children weretalking in long, complicated sentences (see Table 19). About a quarter said that their children speak inshort sentences, and very few (5%) reported that their children typically talk in two- or three-wordphrases. Of potential concern is the fact that 2 percent of the children reportedly are still using one-wordphrases to communicate, although this is lower than the 8 percent reported the previous year. As shown inChapin Hall at the University of Chicago 37

Appendix C, girls in the study families were more likely than boys to be combining words frequently andtalking in sentences (see Table C- 8). This is consistent with most research on early languagedevelopment. Also, as shown in Table C- 9, foreign-born mothers were less likely than U.S.-born mothersto report that their focal child speaks in either long, complicated sentences or short sentences.Table 19. Language Skills of Focal Children over Time a% MothersHow Child CommunicatesYear 3 Year 4 Year 5(n=305) (n=306) (n=306)Does not talk yet 1 0 0Mostly talks in 1-word sentences 23 8 2Talks in 2- to 3-word phrases 50 20 5Talks in fairly complete, short sentences 23 37 23Talks in long complicated sentences 2 35 70a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Four to fivemothers did not respond to these questions.Chi-square tests at year 3 and at year 5 indicate that girls are significantly more likely to use complete sentences than boys(p < .01).Emerging Literacy and Pre-Academic SkillsAt year 5, most of the focal children appeared to be engaging in a variety of preliteracy activities,including drawing, writing, and looking at books, that are typical of their age (see Table 20).As shown in Table 21, a large majority of mothers reported that their children could count, recognize orname colors, and say letters of the alphabet. All but a few of the children were drawing, and at least twothirdswere drawing or attempting to draw recognizable shapes, numbers, or letters and pictures of people.Almost all of the mothers said their children have started counting, although just 48 percent reported thatthe child could count to 10. Table C- 10 in the Appendix indicates that all of these behaviors and skillsincreased between the year 4 and year 5 surveys. In a subsequent report, we will compare these resultswith data from the national ECLS-B study.Table 20. Emerging Literacy Skills of Focal Children at Year 5 a% Mothers (N=310)Emergent Literacy/Pre-Academic Behaviors“Almost All“Sometimes”the Time”Child looks at picture books on his/her own 91 8Child points to pictures while looking at picture books 90 8Child pretends to read the words in a book 75 16Child can tell what is in each picture in a picture book 82 16Child reads the written words in a book 4 24a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chapin Hall at the University of Chicago 38

Table 21. Emerging Pre-academic Skills of Focal Children at Year 5 aPre-Academic Behaviors% Mothers(N=310)Child scribbles or draws on paper b 99Child says the names of most of the colors (red, yellow, blue, and green) d 96Child points to colors (red, yellow, blue, and green) when you say the names 90Child tells you in words what he/she has drawn b 89Child tries to draw shapes, numbers, or letters b 88Child can draw one or more shapes that you can recognize b 82Child points to letters of the alphabet when you say the names most of the time d 79Child draws pictures of people or objects b 72Child can draw one or more letters that you can recognize b 67Child can say the names of the letters most of the time d 60Child can count up to 10 c 48aData were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bPercentages indicate mothers who responded “yes” (versus “no”) to item.cMothers were asked how high child can count in five categories: “not at all,” “up to 5,” “up to 10,” “up to 20,” and “up to 50 ormore.” This indicates the percentage of mothers who responded “up to 10” or a higher number.dMothers were asked whether their children can point or say “all,” “most,” “some,” or “none” of the names, colors, and letters.Similar to the results for other developmental outcomes, we found some differences between the reportsof foreign- and U.S.-born mothers (see Table C- 11 in the Appendix). Specifically, U.S.-born motherswere significantly more likely to report that their focal child: (1) says the names of letters; (2) draws oneor more recognizable shapes; (3) points to letters of the alphabet when the mother says the names; (4)draws one or more recognizable letters; and (5) read the written words in a book. These results supportfindings in the literature. For example, De Feyter and Winsler (2009) report differences in the cognitiveand language development of 4-year-old children of low-income foreign- and native-born parents inMiami, Florida. Specifically, assessments based on the Learning Accomplishment Profile-Diagnostic(LAP-D) show that children of U.S.-born parents display higher cognitive skills than the children offoreign-born parents. Further, children of U.S.-born parents reach the national average of the 50 thpercentile by the end of the year in which they are observed, whereas the children of foreign-born parentsstart and end the year below national averages.In addition, children of U.S.-born parents also demonstrate stronger language skills, although some of thedifferences disappear when the language of administration of the assessment is taken into account. Overthe past 2 years of interviews, we have found some significant changes in the focal children’s preliteracyand pre-academic skills. At year 5, focal children are significantly more likely to either read or pretend toread books, talk about the pictures they see in books, recognize and name colors and letters, count, draw,draw recognizable shapes and letters, and tell in words what they have drawn.Chapin Hall at the University of Chicago 39

It should be noted that it is not clear whether mothers’ responses primarily reflect children’s behaviorsand development or opportunities to draw, write, and look at books, which has been known to vary infamilies (e.g., Dickinson, 1994; Beals, DeTemple & Dickinson, 1994; Hart & Risley, 1995; Raikes et al.,2006; Snow, Barnes, Chandler, Goodman & Hemphill, 1991). As discussed in the next chapter, weobserved some variations in the reported availability of print materials in the home between U.S.-bornand foreign-born families.We also asked mothers to indicate the level of education they expected the focal children to achieve. Themajority (76%) of the mothers said they expected their children to graduate from college, and another 11percent expected their children would attend (but not necessarily graduate from) college. Thirteen percentexpected their focal child to graduate from high school, but not go any further. These figures differ fromwhat we found in year 4, when 87 percent of mothers expected their focal child to graduate from college;5 percent expected them to attend college; and 7 percent expected them to graduate from high schoolonly. There were no differences in these results as a function of maternal characteristics; however, therewere differences as a function of children’s reported development. Specifically, mothers who reportedthat they expected their children to attend or graduate from college were more likely to report that theirchildren demonstrate a high level of emerging literacy and pre-academic skills than were mothers who didnot expect their children to attend college.SummaryIndicators of maternal functioning examined in this chapter include a measure of depression, mothers’self-reported health, a measure of parental stress, and administrative data on reports of child abuse andneglect. The percentage of mothers reporting depressive symptoms was fairly stable from years 2 to 5 andsignificantly lower relative to the first year. The percentage of mothers with elevated parental stressscores was also stable from years 2 through 5 (but was not measured in year 1). Finally, similarproportions of mothers described themselves as in “good” to “excellent” health in all 5 years.Access to healthcare is another indicator related to maternal health. A majority (77%) of the mothersreported receiving regular medical care for themselves at the time of the year 5 interview, but this stillindicates that, as in year 4, almost a quarter of the sample mothers are going without routine healthcare.This group of mothers, therefore, is less likely to have access to services that will keep them healthybetween pregnancies, should they become pregnant again. Use of prenatal care by mothers who had givenbirth to children subsequent to their focal child followed the same pattern observed with the focalchildren. That is, among the 14 percent of the mothers who had had a subsequent pregnancy at the time ofChapin Hall at the University of Chicago 40

the year 5 interview, over three-quarters (77%) said they had initiated care in the first trimester, while 16percent reported doing so in the second trimester.More than three-fourths (76%) of the focal children were covered by health insurance in year 5. However,the fact that 13 percent of all children in the study (focal as well as other children in the study families)were not covered is a concern. An additional concern is that only 45 percent of the sample mothersreported having health insurance for themselves; this percentage is fairly stable from years 2 through 5and is significantly lower than the percentage in year one. There were marked differences betweenforeign-born and U.S.-born mothers with respect to health insurance coverage; 75 percent of U.S.-bornmothers had health insurance, compared to only 19 percent of foreign-born mothers. In addition, the focalchildren of foreign-born mothers were less likely than the children of U.S.-born children to have healthinsurance, although the difference is much smaller (81% vs. 90%).The results for children’s health and development are generally positive. Most mothers reported theirfocal child to be in “good” to “excellent” physical health, although 19 percent of the mothers reported thattheir focal child had “special needs,” with asthma and other respiratory problems being the dominantneed. In terms of development, a majority of the children were demonstrating a range of early literacy,social-emotional, and communicative behaviors typical for their ages. More than two-thirds of thechildren were reported to be talking in sentences, and 90 percent or more were demonstrating appropriatelistening and speaking skills when conversing with other people. However, we did find that foreign-bornmothers were more likely than U.S.-born mothers to report some social-emotional problem behaviors(such as the inability to sit still) of their focal child, which may be attributed, in part, to culturaldifferences in their expectations for the behavior of preschool children. In addition, foreign-born motherswere less likely than U.S.-born mothers to report that their child was demonstrating certain preliteracybehaviors. Such differences might be the result of differences in opportunities to develop and use theseskills.Chapin Hall at the University of Chicago 41

Parenting PracticesA variety of factors shape child development. In the early years, home environments and interactions withparents and other family members particularly influence children. Thus, in the effort to understand theindependent effects of the Palm Beach County service system on children’s development, it is importantto gather information about these other influences, including home and neighborhood characteristics,children’s daily activities, and the kinds of parenting practices they experience. In this section, wedescribe the kinds of parenting activities that study mothers reported for themselves and theirhusbands/partners in the fifth-year interview for any of their children. We clustered these activities intotwo broad categories, “positive” and “negative,” to reflect their potential for either beneficial or harmfuleffects on children’s development.Positive Parenting PracticesThe vast majority of mothers (90% or more) reported that they engaged in the following activities withtheir children in the 3 months prior to the survey: praised them, took them on errands, took them outsidefor a walk or to play, read books to them, encouraged them to read, sang songs with them, and told storiesto them (see Table 22). Somewhat smaller percentages, though still a majority, did household chores withtheir children, played a game, did a puzzle, or played with a building toy with children, and talked to theirchildren about television programs. However, a much lower percentage of the mothers (43%) took theirchildren to the library during the previous 3 months.In families in which husbands/partners had had contact with their children over the previous 3 months, thedistribution of their parenting activities was similar to that of mothers, although the percentage of fatherswho engaged in each activity tended to be lower than the percentage of mothers who did. Mothersreported that their husbands/partners most often engaged in the following activities: praised children, tookchildren outside for walk or play, took children on errands, encouraged children to read, read books tochildren, told stories to children, and sang songs with children. When asked about the frequency ofChapin Hall at the University of Chicago 42

various activities, mothers reported that they or their husbands/partners tended to engage in all of theparenting activities, on average, about once per week, with the exception of taking their children to thelibrary, which was done much less frequently, and praising children, which was done with far greaterfrequency.Table 22. Positive Parenting Practices during Previous 3 Months at Year 5 aActivity%Mothers(N=310)%Husbands/Partners(n=225)MeanFrequency b (SD)Praised children 99 97 1.1 (.36)Took children on errands (e.g., post office or store) 98 89 1.5 (.57)Took children outside for walk or play 96 90 1.6 (.58)Encouraged children to read a book 95 87 1.4 (.56)Read books to children 95 83 1.5 (.61)Sang songs with children 92 75 1.3 (.51)Told stories to children 90 78 1.6 (.66)Did household chores with children 79 51 1.5 (.62)Played with a game, puzzle, or building toy 77 63 1.9 (.69)Talked to children about a television program 75 65 1.5 (.63)Played with clay, drew pictures, or other arts/crafts 75 51 1.8 (.68)Took children to library 43 30 2.5 (.59)Mean (SD) Positive Parenting Score (range: 0–1) c 0.92 (.13) 0.81 (.21)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Lower number indicates higher frequency. Average frequency reported by only those mothers who responded affirmatively thateither they or their husbands/partners had done each activity during the previous 3 months, using a 3-point scale: 3=“once or twice amonth,” 2=“at least once a week,” and 1=“daily or most days.”c The mean parenting score is based on only the items included in all surveys so that comparisons could be made over time. Thefollowing items were excluded in calculation of the mean: “played with a game, puzzle, or building toy,” “did household chores,”“talked about a television program,” and “took child to library.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: books(mother) vs. books(partner); stories(mother) vs. stories(partner); songs(mother) vs. songs(father); errands(mother) vs. errands(partner); outside(mother) vs. outside(partner); arts(mother) vs. arts(partner); game(mother) vs. game(partner); library(mother) vs.library(partner); chores(mother) vs. chores(partner); praised(mother) vs. praised(partner); encouragedreading(mother) vs.encouragedreading(partner); television(mother) vs. television(partner).As shown in Table C- 13, there were some differences in the proportions of foreign- and U.S.-bornmothers who reported engaging in certain activities with their children. U.S.-born mothers were morelikely to report: (1) encouraging their children to read; (2) doing household chores with their children; (2)playing games, working puzzles, or playing with building toys with their children; (3) talking to childrenabout television programs; (4) doing arts and crafts with children; and (5) taking their children to thelibrary. Most of these differences were not only statistically significant but substantive as well, with somedifferences equivalent to twenty percentage points or more. In addition, many of the activities for whichthere were differences are those associated with children’s cognitive and language development.Chapin Hall at the University of Chicago 43

significant increase in the percentage of mothers who reported taking their children to the library. Therewas also a small but statistically significant increase in the percentage of mothers who reported tellingstories to their children. These increases are, again, consistent with the increasing age of the children.Similar to the trend for mothers, the percentage of fathers or partners engaging in most of the positiveparenting activities increased significantly between years 1 and 2, and then remained fairly stablethereafter (see Table 25). Between years 4 and 5, however, there were significant increases in thepercentages of fathers/partners who told stories to their children, took the children to the library,encouraged the children to read a book, did household chores with children, and who talked to theirchildren about television programs.Table 25. Fathers’ Positive Parenting Practices over Time aActivity% FathersYear 1 Year 2 Year 3 Year 4 Year 5(n=250) (n=238) (n=225) (n=225) (n=225)Read books to children 52 77 81 77 83Told stories to children 53 67 64 61 78Sang songs with children 70 80 80 75 75Took children on errands (e.g., post office,store)76 90 87 89 89Took children outside for walk or play 69 90 89 86 90Played with clay, drew pictures, or otherarts/crafts28 53 52 42 51Played with a game, puzzle, or building toy(Y2)— b 65 58 53 63Took children to library (Y2) — 15 13 17 30Did household chores with children (Y2) — 41 38 38 51Praised children 92 95 97 96 97Encouraged children to read a book 84 81 73 73 87Talked to children about a television program— — 55 52 65(Y3)Mean (SD) Positive Parenting Score(range: 0–1) c 0.63 (.29) 0.79 (.22) 0.78 (.22) 0.75 (.23) 0.81 (.21)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” “Father” includesbiological fathers and other men mothers identified as partnersb Asked only of mothers with children 1 year or older (n=137 in year 1 and n=236 in year 2).c Mean parenting scores are based on parenting items included in all three surveys. (Y2) and (Y3) indicate items added in year 2or year 3 that were excluded in the calculation of mean scores.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: books1 vs.books2, books3, books4, and books5; books2 vs. books5; stories1 vs. stories2; stories1 vs. stories5; stories2 vs. stories5; stories3 vs.stories5; stories4 vs. stories5; songs1 vs. songs2; songs1 vs. songs3; errands1 vs. errands2; errands1 vs. errands3; errands1 vs.errands4; errands1 vs. errands5; outside1 vs. outside2; outside1 vs. outside3; outside1 vs. outside4; outside1 vs. outside5; arts1 vs.arts2; arts1 vs. arts3; arts1 vs. arts4; arts1 vs. arts5; arts2 vs. arts4; games2 vs. games4; library2 vs. library5; library3 vs. library5;library4 vs. library5; praised1 vs. praised3; praised1 vs. praised5; praised2 vs. praised5; encouraged1 vs. encouraged3; encouraged1vs. encouraged4; encouraged2 vs. encouraged3; encouraged3 vs. encouraged5; encouraged4 vs. encouraged5; chores2 vs. chores5;chores3 vs. chores5; chores4 vs. chores5; tv3 vs. tv5; tv4 vs. tv5; score1 vs. score2; score1 vs. score3; score1 vs. score4; score1 vs.score5; score3 vs. score5; score4 vs. score5.Chapin Hall at the University of Chicago 46

Negative Parenting PracticesThere is a considerable amount of evidence suggesting that maternal depression and harsh parenting canhave harmful effects on children’s social and emotional development (e.g., Bradley & Corwyn, 2007;Chang et al., 2004). 29 Thus, we also asked mothers about their so-called negative parenting practicesduring the previous 3 months. Just over half (60%) of all mothers reported that they had lost their temperwith their children; as shown in Table C- 14, foreign-born mothers were significantly more likely thanU.S.-born mothers to report that they’d lost their temper. A quarter of mothers said they had found hittingor spanking is a good way to get their children to listen; and another quarter (26%) said they got angrierwith their children than they had intended during the previous 3 months. Smaller percentages said that inthe past 3 months they had punished their children for not finishing the food on their plate (9%) or blamedtheir children for something that was not their fault (18%). U.S.-born mothers were significantly morelikely than foreign-born mothers to report the latter practice (see Table C- 14).Table 26. Year 5 Negative Parenting Practices during Previous 3 Months a% Husbands/% MothersMeanActivityPartnersFrequency b (SD)(N=310) (n=225)Lost temper with children 60 44 2.3 (.71)Found hitting/spanking a good way to get children to listen 25 22 2.5 (.64)Got more angry than meant to with children 26 14 2.4 (.68)Punished children for not finishing food on plate 9 7 2.1 (.79)Blamed children for something not their fault 18 12 2.7 (.58)Mean (SD) Negative Parenting Score (range: 0–1) c 0.19 (.23) 0.14 (.21)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bLower number indicates higher frequency. Mean frequency calculated only for mothers who said either they or theirhusbands/partners had done each activity during the previous 3 months, using a 3-point scale: 3=“once or twice a month,” 2=“atleast once a week,” and 1=“daily or most days.”cThe mean score uses only items included in all surveys; the item “lost temper with child(ren)” was excluded in calculating themean.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:temper(mother) vs. temper(partner); hitting(mother) vs. hitting(partner); angry(mother) vs. angry(partner); blamed(mother) vs.blamed(partner).Interestingly, mothers whose husbands/partners had had contact with their children in the previous 3months reported lower percentages of some negative parenting practices for their husband or partner thanthey reported for themselves (see Table 26). Specifically, mothers reported significantly lowerpercentages of husbands/partners losing their temper with the children and getting angrier than intended29 It is also important to note that the literature on parenting suggests that the relationships between positive and negativeparenting and children’s outcomes are complex and mediated by a number of factors, including children’s own personalities andtemperament and the other contexts, e.g., school, in which they develop.Chapin Hall at the University of Chicago 47

with the children. When asked about the frequency with which they or their husbands/partners engaged inany of the negative parenting practices, mothers reported that each practice occurred, on average, betweenonce or twice a month and once a week. However, blaming a child for something that was not the child’sfault occurred less frequently, on average less than once a month.Negative Parenting Practices over TimeBetween years 1 and 5, there was a significant decline in the percentage of mothers who said theypunished their children for not finishing the food on their plates (see Table 27); in addition, mothersreported a similar pattern with respect to husbands/partners who had had contact with the children duringthe past 3 months. Otherwise, the percentages of mothers reporting use of negative parenting practicesthemselves was relatively stable during the study period, although there were significant fluctuationsacross time for some practices. Specifically, there was an increase between years 1 and 2 in thepercentage of mothers who found hitting or spanking to be a good way to get their children to listen, butthe percentage remained stable in the years thereafter. Also, there was a significant decline between years1 and 4 in the percentage of mothers who said they had become angrier with their children than intended,but this percentage increased significantly in year 5. A similar pattern was found for mothers who saidthey blamed their children for something that was not their fault. There was more fluctuation in mother’sreports of the negative parenting practices of husbands/partners. The most significant changes to note aredecreases in the percentages of husbands/partners who were reported to have lost their temper with thechildren or to have found hitting or spanking to be a good way to get the children to listen.Table 27. Mothers’ Negative Parenting Practices over Time a % Mothers (N=310)ActivityYear 1 Year 2 Year 3 Year 4 Year 5Lost temper with children (Y2) — 54 53 57 60Found hitting/spanking a good way to getchildren to listen22 29 30 29 25Got more angry than meant to with children 28 22 23 19 26Punished children for not finishing food onplate b 17 6 6 8 9Blamed children for something not theirfault b 20 13 11 10 18Mean (SD) Negative Parenting Score(range:0–1) c 0.19 (.26) 0.18 (.24) 0.17 (.22) 0.16 (.22) 0.19 (.23)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Asked only of mothers with children 1 year or older (n=163 in year 1 and n=306 in year 2).c Mean parenting scores are based on parenting items included in all five surveys. (Y2) and (Y3) indicate items added in year 2 oryear 3 that were excluded in the calculation of mean scores.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: hitting1vs. hitting2; hitting1 vs. hitting3; hitting1 vs. hitting4; hitting3 vs. hitting5; angry1 vs. angry2; angry1 vs. angry4; angry4 vs.angry5; food1 vs. food2; food1 vs. food3; food1 vs. food4; food1 vs. food5; blame3 vs. blame5; blame4 vs. blame5; temper2 vs.temper5; temper3 vs. temper5; score4 vs. score5.Chapin Hall at the University of Chicago 48

Table 28. Fathers’ Negative Parenting Activities over Time aActivity% Father/Partner% Father/Partner% Father/Partner% Father/Partner% Father/PartnerYear 1 Year 2 Year 3 Year 4 Year 5(n=250) (n=238) (n=225) (n=225) (n=225)Lost temper with children (Y2) — 40 48 54 44Found hitting/spanking a good way toget children to listen17 22 24 31 22Got more angry than meant to withchildren20 11 10 16 14Punished children for not finishingfood on plate b 15 4 7 8 7Blamed children for something nottheir fault b 17 10 8 7 12Mean (SD) Negative ParentingScore (range:0–1) c 0.15 (.25) 0.12 (.20) 0.12 (.19) 0.16 (.21) 0.14 (.21)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Asked only of mothers with children 1 year or older (n=137 in year 1 and n=236 in year 2).c Mean parenting scores are based on parenting items included in all five surveys. (Y2) and (Y3) indicate items added in year 2 oryear 3 that were excluded in the calculation of mean scores.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: hitting1vs. hitting3; hitting1 vs. hitting4; hitting2 vs. hitting4; hitting4 vs. hitting5; angry1 .v angry2; angry1 vs. angry3; angry1 vs.angry5; angry3 vs. angry4; food1 vs. food2; food1 vs. food3; blame3 vs. blame5; blame4 vs. blame5; temper2 vs. temper3;temper2 vs. temper4; temper3 vs. temper4; temper4 vs. temper5; score1 vs. score3; score2 vs. score4; score3 vs. score4.Parent Involvement ActivitiesAnother important area of parenting is involvement in children’s out-of-home care and education.Research has shown the value of parental involvement in both early childhood education and laterschooling with respect to children’s academic outcomes (e.g., Epstein, 2001; Hill & Craft, 2003; Jeynes,2003; Meidel & Reynolds, 1999). At the time of the year 5 interview, over three-fourths of the mothershad children who were in childcare or school and, thus, were asked additional questions about theirinvolvement in their children’s out-of-home care and education.As shown in Table 29, about three-fourths of these mothers reported talking with their children’s teacherand helping their children with homework during the previous 3 months. Over half of the mothers hadattended a parent-teacher conference (62%); as shown in Table C- 15, foreign-born mothers were morelikely than U.S.-born mothers to report attending such a conference. Almost half (44%) of all mothersreported they had participated in a field trip or school event for families, with U.S.-born mothers morelikely than foreign-born mothers to report doing so. A third reported attending a PTA meeting during theChapin Hall at the University of Chicago 49

past 3 months. The percentage of husbands or partners who engaged in any of these activities wassignificantly smaller than the percentage of mothers. 30Table 29. Parent Involvement Activities during Previous 3 Months at Year 5 a% Mothers % Husbands/PartnersActivity b (n=255) (n=182)Talked to teacher about child’s progress (at times other thanparent-teacher conference)75 38Helped children with homework 72 62Attended parent-teacher conference 62 38Participated in field trip or family event at school 44 27Attended PTA or other parent meetings 33 19Mean (SD) Parent Involvement Score (range: 0–1) c 0.57 (.30) 0.37 (.30)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Only mothers who had children in childcare or school were asked about these activities.c Mean parent involvement score based on all items asked at each time point.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:homework(mother) vs. homework(partner); conference(mother) vs. conference(partner); progress(mother) vs. progress(partner);fieldtrip(mother) vs. fieldtrip(partner); meetings(mother) vs. meetings(partner).We asked mothers in all 5 years of the study about whether they helped children with homework orattended a parent-teacher conference. The percentage of mothers who answered yes to these questionsdeclined significantly between years 1 and 2. This decline may reflect the mother’s needs to focus theirenergies on the newborn focal child, as the percentages answering yes return to the year 1 levels by year 5(see Table 30). The percentages of mothers participating in the other activities are relatively stable overtime, with the exception of significant fluctuations in the percentage of mothers who said they talked to ateacher about their children’s progress. Finally, the reported patterns for husbands/partners are similar tothose of the mothers (see Table 31).30 It should be mentioned that the types of activities in Table 29 to Table 31, except for helping with homework, differ from otherparenting activities in that they are unlikely to occur as regularly.Chapin Hall at the University of Chicago 50

Table 30. Parent Involvement Activities of Mothers over Time aActivity% Mothers % Mothers % Mothers % Mothers % Mothersat Year 1 at Year 2 at Year 3 at Year 4 at Year 5(n=110) (n=205) (n = 233) (n=231) (n=255)Helped child with homework 77 50 51 61 72Attended parent-teacherconference65 49 52 57 62Talked to teacher about child’sprogress (at times other than— — 70 83 75parent-teacher conference)Participated in field trip or familyevent— 35 39 42 44Attended parent or PTA meetings — 38 35 27 33Mean (SD) Parent InvolvementScore (range: 0–1) b 0.71 (.40) 0.43 (.41) 0.49 (.37) 0.54 (.32) 0.57 (.30)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mean parent involvement score based on all items asked at each time point.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:homework2 vs. homework4; homework2 vs. homework5; homework3 vs. homework5; homework4 vs. homework5; conference3vs.conference5; fieldtrip2 vs. fieldtrip5; pta2 vs. pta4; pta3 vs. pta4; pta4 vs. pta5; progress3 vs. progress4; progress4 vs.progress5; score1 vs. score2; score2 vs. score3; score2 vs. score4; score2 vs. score5; score3 vs. score5.Table 31. Parent Involvement Activities of Fathers over Time aActivity% Father/Partnerat Year 1% Father/Partnerat Year 2% Father/Partnerat Year 3% Father/Partnerat Year 4% Father/Partnerat Year 5(n=86) (n=145) (n=159) (n = 160) (n=182)Helped child with homework 66 47 50 52 62Attended parent-teacher conference 36 22 33 33 38Talked to teacher about child’sprogress (at times other than parentteacher— — 36 44 38conference)Participated in field trip or familyevent— 19 23 25 27Attended parent or PTA meetings — 24 19 20 19Mean (SD) Parent InvolvementbScore (range: 0–1)0.51 (.40) 0.28 (.34) 0.33 (.33) 0.35 (.34) 0.37 (.30)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mean parent involvement score based on all items asked at each time point.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:homework3 vs. homework5; conference2 vs. conference3; conference2 vs. conference5; fieldtrip2 vs. fieldtrip4; pta2 vs. pta3;score1 vs. score2; score1 vs. score5.Correlates of Parenting ActivitiesWe have already examined how positive and negative parenting activities differ according to nativity. Asshown in Table 32 and Table 33, the frequency of particular positive and negative parenting activities alsodiffers significantly by educational background, race/ethnicity, and relationship status, all of which areChapin Hall at the University of Chicago 51

highly correlated with nativity. It should be emphasized that, because of these relationships, the results inthese tables do not imply causation. However, we present them here as suggestive of the kinds of factorsthat may impact parenting practices; complex interactions and relationships among the variables will beexplored later.As shown in Table 32, mothers who had not graduated from high school reported significantly lower ratesof all positive parenting activities, with the exception of taking children on errands, taking childrenoutside for a walk or to play, and praising children. Perhaps not surprisingly, since Hispanic and foreignmothers in the sample possess relatively lower levels of education, we found that these mothers reportedlower rates of many of the positive parenting activities in comparison to black and white/other and U.S.-born mothers.By the same token, we found evidence that being single or not living with a partner is associated with ahigher likelihood of some positive parenting practices; this result is most likely related to the fact thatHispanic and foreign-born mothers are more likely to be married than mothers from other groups.We found fewer associations between maternal characteristics and negative parenting practices. We didfind that mothers with less education were more likely to report losing their tempers with their children.This was also true of Hispanic (as well as white/other) mothers, foreign-born mothers, and those in arelationship, in comparison to black, U.S.-born, and single mothers, respectively. Similarly, we also foundthat Hispanic (as well as white/other) mothers were more likely to report getting angrier than intendedwith their children.Once more, we caution readers not to draw causal inferences about the relationships between parentingpractices and the sociodemographic characteristics presented here, because many of these characteristicsare related to one another. We provide more analysis of demographic characteristics and parentingpractices in relation to service use in a later section of this report. In addition,. a number of other factorsshape parenting practices, as suggested by data from the embedded qualitative study (Spielberger et al.,2009) reported previously. These include family’s beliefs about who should care for their children and thepriority of children’s needs over those of other family members, work and financial circumstances,neighborhood context (including safety), and future goals for their children.Chapin Hall at the University of Chicago 52

Table 32. Mothers’ Positive Parenting Activities during Previous 3 Months by Selected Maternal Characteristics at Year 5 aParenting ActivitiesAllMothersEducation (%) Race/Ethnicity (%) Nativity (%) Relationship Status (%)Single/HSWhite/ ForeignbornbornU.S.-Black HispanicMarried live w/graduateotherpartnerNot HSgraduateSingle ornot livingw/partner(N=310) (n=157) (n=153) (n=126) (n=159) (n=25) (n=167) (n=143) (n=102) (n=95) (n=113)Read books to children 95 91 98* 98 93 92 94 96 96 93 96Told stories to children 90 86 93* 91 89 88 91 88 86 88 93Sang songs with children 92 88 95* 95 88 96 90 94 93 90 92Took children on errands 98 98 99 98 99 96 98 98 99 98 97Took children outside for walk orto play96 95 97 95 96 100 96 97 94 99 95Played with clay, drew pictures,or did other arts and crafts with 75 62 87* 81 66 96* 63 87* 73 65 84*childrenPlayed games, puzzles withchildren (Y2)77 68 85* 78 73 88 71 83* 73 77 80Took children to library (Y2) 43 32 54* 52 35 52* 30 58* 44 37 46Did household chores withchildren (Y2)79 72 86* 81 74 96* 68 91* 78 68 87*Praised children 99 99 100 100 99 100 99 100 100 98 100Encouraged children to read abook96 92 99* 98 93 100* 93 98* 94 94 97Talked to children about TVprogram (Y3)75 63 87* 82 66 96* 70 81* 72 73 79Mean (SD) Positive Parenting 0.92 0.89 0.96 0.94 0.90 0.96 0.90 0.95 0.92 0.91 0.94Score (range: 0–1) b (.13) (.15) (.09)*** (.10) (.15) (.10)** (.14) (.11)** (.13) (.13) (.13)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mean parenting scores are based only on parenting items included in all five surveys. (Y2) and (Y3) indicate items added in year 2 or year 3.One-way ANOVA tests indicate means are statistically significant at *p < .05, ** p < .01, or *** p < .001.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: [Education] lookedatbooks(2) vs. lookedatbooks(1);toldstories(2) vs. toldstories(1); sangsongs(2) vs. sangsongs(1); artscrafts(2) vs. artscrafts(1); playedgames(2) vs. playedgames(1); library(2) vs. library(1); chores(2) vs. chores(1);encouragedreading(2) vs. encouragedreading(1); talkedtelevision(2) vs. talkedtelevision(1); [Race/Ethnicity] artscrafts(1) vs. artscrafts(2), artscrafts(3) vs. artscrafts(2); library(1) vs.library(2); chores(3) vs. chores(2); encouragedreading(1) vs. encouragedreading(2); talkedtelevision(1) vs. talkedtelevision(2); talkedtelevision(3) vs. talkedtelevision(2); [Nativity]artscrafts(2) vs. artscrafts(1); playedgames(2) vs. playedgames(1); library(2) vs. library(1); chores(2) vs. chores(1); encouragedreading(2) vs. encouragedreading(1); talkedtelevision(2)vs. talkedtelevision(1); [Relationship status] artscrafts(3) vs. artscrafts(2); chores(3) vs. chores(2).Chapin Hall at the University of Chicago 53

Table 33. Mothers’ Negative Parenting Practices during Previous 3 Months by Selected Maternal Characteristics at Year 5 aParentingActivitiesAllMothersEducation (%) Race/Ethnicity (%) Nativity (%) Marital Status (%)Single/HSWhite/ ForeignbornbornU.S.-Black HispanicMarried live w/graduateotherpartnerNot HSgraduateSingle ornot livingw/partner(N=310) (n=157) (n=153) (n=126) (n=159) (n=25) (n=167) (n=143) (n=101) (n=97) (n=112)Lost temper withchildren (Y2)60 68 51* 38 77 64* 69 49* 60 69 51*Found hitting orspanking a goodway to get children25 22 28 26 23 36 23 27 31 26 20to listenGot more angry thanmeant to with27 29 24 17 33 33* 29 22 33 25 22childrenPunished children fornot finishing food 8 10 7 5 11 12 10 8 9 10 9on plateBlamed children forsomething not their 18 15 21 16 17 32 14 22* 23 16 16faultMean (SD) Negative0.19 0.19 0.20 0.16 0.21 0.27 0.19 0.20 0.23 0.19 0.16Parenting Score(range: 0–1) b (.23) (.23) (.23) (.21) (.25) (.22)* (.23) (.23) (.24) (.23) (.21)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mean parenting scores are based only on parenting items included in all five surveys. (Y2) and (Y3) indicate items added in year 2 or year 3.One-way ANOVA tests indicate means are statistically significant at *p < .05, ** p < .01, or *** p < .001.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: [Education] losttemper(1) vs. losttemper(2);[Race/Ethnicity] losttemper(2) vs. losttemper(1); losttemper(3) vs. losttemper(1); angry(2) vs. angry(1); [Nativity] lostempter(1) vs. losttemper(2); blamed(2) vs. blamed(1);[Relationship status] losttemper(2) vs. losttemper(3).Chapin Hall at the University of Chicago 54

Table 34. Mothers’ Parent Involvement Activities during Previous 3 Months for Children in Childcare or School by Selected MaternalCharacteristics at Year 5 aParentInvolvementActivitiesHelped childwithhomeworkAttended aparent-teacherconferenceTalked to teacherabout child’sprogress atother timesAttended fieldtrip or familyevent at child’sschool/centerAttended PTAmeetings atchild’s schoolor centerMean (SD)ParentInvolvementScoreAllMothersBlackRace/Ethnicity (%) Nativity (%) Marital Status (%) Employment (%)HispanicWhite/otherForeignbornU.S.-bornMarriedSingle/live w/partnerSingle/Notlivingw/partnerNotworkingWorkingYear 1 TeenMother (%)(n=255) (n=109) (n=126) (n=20) (n=133) (n=122) (n=84) (n=75) (n=96) (n=121) (n=134) (n=31) (n=224)72 74 73 53 68 76 70 70 75 79 66* 45 75**62 54 71 50* 70 53* 66 67 54 73 52** 39 65**75 82 69 79 72 78 73 75 76 76 74 77 7544 57 36 25** 33 56*** 42 38 50 42 46 29 4633 44 25 25** 33 32 33 26 39 36 30 19 350.57 0.62 0.55 0.45 0.55 0.59 0.57 0.55 0.59 0.61 0.53 0.42(.30) (.30) (.29) (.28)* (.29) (.30) (.27) (.30) (.32) (.26) (.32)* (.30)(range: 0-1) ba Sample includes only mothers who have children in childcare or school. Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Mean parenting scores are based only on parenting items included in all five surveys. (Y2) and (Y3) indicate items added in year 2 or year 3.One-way ANOVA tests indicate means are statistically significant at *p < .05, ** p < .01, or *** p < .001.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: [Race/Ethnicity] parentteacherconferences(2) vs.parentteacherconferences(1); fieldtrip(1) vs. fieldtrip(2); fieldtrip(1) vs. fieldtrip(3); ptameeting(1) vs. ptameeting(2); [Nativity] parentteacherconferences(1) vs.parentteacherconferences(2); fieldtrip(2) vs. fieldtrip(1); [Relationship status] none; [Employment] homework(1) vs. homework(2); parentteacherconferences(1) vs.parentteacherconferences(2); [Teen parent] homework(2) vs. homework(1); parentteacherconferences(2) vs. parentteacherconferences(1); ptameeting(2) vs. ptameeting(1).Teen(15–19)Nonteen(≥ 20)0.59(.29)**Chapin Hall at the University of Chicago 55

SummaryWhen surveyed in the fifth year, mothers were asked to report on three general kinds of parentingpractices that they and their husbands/partners used during the previous 3 months: positive activities withtheir children in their home and neighborhood; negative parenting practices, usually for disciplinarypurposes; and activities that involved them with their children’s school or childcare. More than threequartersof the mothers reported that they engaged in a variety of positive parenting activities. Theseincluded praising their children; taking children on errands and outside to play; reading books andencouraging their children to read; singing songs with children and telling stories; doing household choreswith children; playing with a game, puzzle, or building toy with children; and talking to their childrenabout television programs. For families in which husbands/partners had contact with their children,mothers reported that at least half of husbands/partners engaged in most positive parenting activities.Smaller percentages of mothers reported that they or their husbands/partners used negative parentingpractices. A little over half (60%) of the mothers reported that they had lost their temper with theirchildren; a quarter (25%) said they had found hitting or spanking was a good way to get their children tolisten; and a quarter (26%) said they got angrier with their children than they had intended during theprevious 3 months. No more than 18 percent said that in the past 3 months they had blamed their childrenfor something that was not their fault or punished their children for not finishing the food on their plate.Mothers whose husbands/partners had had contact with their children reported somewhat lowerpercentages of negative parenting practices for their husbands/partners than they reported for themselves.Over time, positive parenting activities increased for both mothers and husbands/partners, but most of thechanges occurred between the first and second years. The only noteworthy differences between years 4and 5 was an increase in the percentage of mothers who reported telling stories to their children andtaking their children to the library, which is consistent with the increasing age of their children. Similarly,there was an increase from year 1 to year 5 in the percentage of husbands/partners engaging in most of thepositive parenting practices. Also, between years 4 and 5, there was an increase in the percentage ofhusbands/partners who told stories to their children, took the children to the library, encouraged thechildren to read a book, did household chores with children, and talked to their children about televisionprograms.In terms of negative parenting practices, there was a significant increase in mean negative parentingscores for mothers, but not husbands/partners between years 4 and 5; the former was due to an increase inthe percentage of mothers who said they had become angrier than intended with their children or hadblamed their children for something that was not their fault.Chapin Hall at the University of Chicago 56

The most frequent parent involvement activity, among mothers with children in school or childcare, wastalking with their children’s teachers, which was reported by three-quarters (75%) of the mothers. Almostthree-quarters of the mothers said they had helped their children with homework, and over half hadattended a parent-teacher conference during the previous 3 months; less than half reported that they hadparticipated in a field trip; and one-third reported attending a PTA or other parent meeting. Thepercentage of husbands/partners who engaged in any of these activities is smaller than the percentage ofmothers.We also found associations between parenting practices and a number of maternal/family characteristics,including educational background, race/ethnicity, relationship status and nativity, although it is importantto keep in mind that these characteristics are highly correlated.Chapin Hall at the University of Chicago 57

Childcare ArrangementsChildcare for preschool and school-age children is an important support for the study families, especiallyamong working parents or those enrolled in school. In the fifth year, a total of 191 mothers, representing260 preschool-age children, used some form of childcare, defined as care on a regular basis from someoneother than a parent. This amounts to almost two-thirds (62%) of the mothers in the year 5 sample andrepresents a significant increase relative to the 54 percent reported in the fourth-year interview (p < .05).This increase may correspond to a slight increase in the percentage of working mothers at the time of theyear 5 interview, the percentage of mothers currently enrolled in school, and the percentage of familiesreceiving a childcare subsidy. More than three-quarters of mothers using childcare reported that theirchildren are in their childcare arrangements 5 days a week. The number of hours spent in care per weekranges from a low of 2 to a high of 75, although just 4 percent spend more than 50 hours in childcare eachweek. Typically, each child spends about 34 hours a week in childcare.Preschool Childcare ArrangementsWhen asked about nonparental childcare arrangements for their preschool-age focal child, mothersreported using a variety of arrangements at year 5. Informal care by relatives, friends, or neighbors wasused less than half as often as formal arrangements, which include center-based programs and familychildcare (see Figure 3). The most frequently mentioned type of care for the focal children at year 5 wascenter care (34%), followed by relative care (11%), and care by a friend or neighbor (6%). Perhaps notsurprisingly, the percentage of parents using center care for their focal child has increased over time, from1 percent in year 1 to 34 percent in year 5. 3131 When we asked mothers to report on childcare arrangements for all of their preschool-age children, the results were verysimilar to those presented in Figure 3.Chapin Hall at the University of Chicago 58

Figure 3. Primary Childcare Arrangements for Focal Children over Time aa Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chapin Hall at the University of Chicago 59

These changes in the types of arrangements over time likely reflect the older age of the children andmothers’ growing knowledge about and comfort with out-of-home care. Another reason, as reported inSpielberger et al., (2009), is the greater availability of center-based programs, such as Head Start, forolder children. The qualitative data also suggest that an important source of information about childcare ismothers’ social networks. Although foreign-born mothers seemed less familiar and less comfortable withthe concept of out-of-home childcare when their children were young, over time they learned more aboutit from relatives and friends. Some foreign-born mothers also began to see it as a way for their children tolearn English and prepare for school. The age of the child surfaced as another important factor mothersconsidered in making childcare decisions. Many mothers talked about waiting until their children wereabout 3 years old before enrolling them in center care, because at this age children can communicate andcontrol their behavior enough to protect themselves against poor-quality care. In the words of one mother,“Some teachers don’t treat them good. So, I’m going to put her in school when she can go to thebathroom on her own and is like talking reasonable.”Stability of Childcare ArrangementsConsistency in caregiving is generally considered to be an important factor in children’s development,especially in the first few years of life (e.g., Bornstein, 1991; Landry, Smith, Swank, Assel & Vellet,2001; Landry, Smith & Swank, 2006). We attempted to assess the consistency of childcare arrangementsby looking at both the number of different arrangements and the number of transitions from onearrangement to another experienced by the focal children during the first 5 years of their lives. These werebased on maternal reports at roughly 6-month intervals, which were obtained in the five annual in-personsurveys and the telephone surveys that occurred between each of the in-person surveys. We conductedthis analysis only for the 284 mothers who had participated in all nine of the surveys. 32It should be noted that the number of childcare arrangements and the number of transitions in care are twodifferent measures. A child may be in only two types of care arrangement, for example, parental care athome and a childcare center, but experience four transitions if placed in a different arrangement every 6months. 33 For example, one of the study mother’s two young children were being cared for by theirparents at the baseline interview, and then by a relative (the children’s grandmother) 6 months later. Sixmonths after that, the children were in the care of their parents again. Subsequently, because of a DCFSreport, the mother received a childcare subsidy that allowed her to place her children in a childcare center32 Most of the analyses for this report are based on a sample of 310 mothers, the number interviewed at each of the five in-personsurveys; however, 26 of these mothers missed at least one of the four follow-up telephone interviews.33 Other instabilities in childcare arrangements, such as changes in the parent who cares for the child or changes in the particularchildcare center used, may be obscured in the survey data.Chapin Hall at the University of Chicago 60

for 6 months, but when the subsidy ran out, the children were back home with their parents andgrandmother again.Regarding the number of different types of childcare, 4 in 10 of the children were in fairly stable carearrangements, meaning that they were either in the same setting (15%) or in no more than two or threedifferent settings across the nine time points (see Table 35). Most children who were in the samearrangement were at home with a parent. Children who were in two or three different arrangements mostoften experienced a mix of parental care and care by a relative. Nonetheless, although a majority of focalchildren experienced three or fewer different childcare arrangements during the first 5 years of their lives,over half (55%) of them had three or more transitions in care; the remainder experienced four or moretransitions in their childcare arrangements during this time. Table C- 15 provides additional detail on therange of childcare arrangements for the focal children over time.Table 35. Minimum Number of Transitions in Focal Children’s Care Arrangements Year 1 to Year 5 aNumber of Transitions across Nine Time Points bFrequency Percent of Total(n=284)TransitionsNone 43 15One 30 11Two 41 14Three 43 15Four 49 17Five 45 16Six 18 6Seven 9 3Eight 6 2a Consistent with the analysis of types of childcare arrangements over time, we did not weight the data to account for theoversampling of mothers in the Glades and mothers assessed “at risk.” Twenty-six mothers who were not interviewed at one ormore of the 6-month telephone follow-up points, although interviewed at all five yearly points, are not included in this analysis aswe do not know what type of childcare arrangements their focal child was in at all nine time points.b The nine time points include year 1 (1–6 mos.), 1.5 (6–12 mos.), 2 (12–18 mos.), 2.5 (18–24 mos.), 3 (24–30 mos.), 3.5 (30–36mos.), 4 (36–42 mos.) 4.5 (42–48 mos.), and 5 (48–54 mos.).Correlates and Predictors of Childcare UseBivariate associations presented in Table 36 indicate family characteristics associated with use ofchildcare, including the work status of mothers, whether or not they are attending school or a job trainingprogram, and a variety of sociodemographic characteristics. Consistent with the previous year, we foundthat U.S.-born mothers were more likely than foreign-born mothers to use childcare; similarly, black andwhite/other mothers were more likely than Hispanic mothers to use childcare. In addition, mothers withmore education and those living above the poverty line were more likely to use childcare. On the otherhand, mothers without a husband/partner were also more likely to use childcare.Chapin Hall at the University of Chicago 61

Table 36. Use of Nonparental Preschool Childcare by Mothers’ Characteristics at Year 5 aCharacteristicUsing ChildcareTGA^% Glades TGA (n=38) 79% Non-Glades TGA (n=232) 60% Outside TGAs (n=41) 56Employment***% Working (n=142) 90% Not working (n=169) 37School***% Currently in school (n=49) 90% Not in school (n=261) 56Job training% Participating in job training (n=14) 71% Not participating in job training (n=296) 61Nativity**% U.S.-born (n=143) 70% Foreign-born (n=168) 54Race/ethnicity*% Black, not Hispanic (n=126) 71% Hispanic (n=160) 54% White/other (n=25) 64Education***% Less than high school diploma (n=157) 48% High school graduate (n=88) 72% Post–high school education (n=65) 79Partner status*% Has husband/partner (n=231) 58% Does not have husband/partner (n=79) 72Number of children% One child (n=77) 64% Two children (n=97) 67% Three or more children (n=134) 57Income-to-poverty ratio**% Living at or below poverty threshold (n=153) 52% Living above poverty threshold (n=141) 71a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chi-square analysis indicates that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or *** p< .001.Because all of these variables are likely to be correlated, we performed a regression analysis to determinethe most important characteristics associated with mothers’ use of childcare for any of their preschoolchildren. The race/ethnicity and nativity variables were combined in the analysis to determine whetherthere are significant differences in use of childcare among U.S.-born blacks and Hispanics, foreign-bornblacks and Hispanics, and “other” mothers. The indicator for residing in the Glades TGA was includedChapin Hall at the University of Chicago 62

ecause it is of particular interest in this study. This variable was found to be a significant predictor ofchildcare use in year 2 and year 4 but not in year 3.As presented in Table 37, the factors associated with increased odds of using childcare included themothers’ employment status, enrollment status, residence in the Glades TGA, and race/nativity. As inprevious years, the strongest predictors of childcare use were maternal employment and enrollment inschool. The odds that mothers who were currently working (and not in school) were using childcare were23 times those of mothers who were neither working nor in school. Similar results were found formothers currently in school or combining work and school. We also found that the odds that mothersresiding in the Glades used childcare were nearly 3 times those of mothers outside of the Glades.Finally, we found that black mothers—both U.S.- and foreign-born—were more likely than foreign-bornHispanic mothers to use childcare. Together, all of the regression variables explained 34 percent of thevariation in childcare use, which suggests that other factors also influence childcare arrangements. 34Table 37. Logistic Regression Predicting Use of Nonparental Preschool Childcare at Year 5 aPredictor Variable Log Odds Ratio Sig.Mother currently working (not in school) 3.1 23.1 ***Mother currently in school or working and in school 3.2 23.5 ***Mother neither working nor in school (excluded category) — — —Mother high school graduate or GED 0.5 1.7 NSMother has post–high school education 0.2 1.3 NSHousehold is at or below the poverty level 0.3 1.4 NSIn Glades TGA 1.0 2.8 **Currently has a husband/partner 0.3 1.3 NSNumber of children 0.1 1.1 NSAll other racial/ethnic groups 0.8 2.2 NSBlack and foreign-born 0.9 2.5 *Black and U.S.-born 0.9 2.5 ^Hispanic and U.S.-born 0.2 1.2 NSHispanic and foreign-born (excluded category) — — —Frequency of support from friends/family 0.1 1.1 NSχ 2 (13, N=335) 150.98Pseudo R 2 .34a Weights were not applied to this analysis.^p < .10, *p < .05, **p < .01, or ***p < .001.34 We also ran a regression to examine the predictors of using formal childcare (i.e., a center-based program or family childcare).We found that mothers’ employment/enrollment status and residence in the Glades are still significant. In addition, black foreignbornmothers are significantly more likely than Hispanic foreign-born mothers to use formal care, although black U.S.-bornmothers are not. Finally, we found that mothers with a high school education and those with more children are also more likely touse formal care.Chapin Hall at the University of Chicago 63

Additional bivariate analyses suggest that different types of childcare arrangements are associated with anumber of maternal characteristics, including race/ethnicity, nativity, education, employment, and income(see Table 38 and Table 39). As shown in Table 39, black mothers were significantly more likely thanHispanic mothers to use center care (65% vs. 46%). At the same time, Hispanic mothers weresignificantly more likely than black mothers to rely on friends or neighbors for childcare (20% vs. 1%).In the fifth year, differences in the use of center care by nativity became far less apparent. A little morethan half of both U.S.- and foreign-born mothers used center-based care. In addition, about the samepercentages of foreign- and U.S.-born mothers had their children enrolled in prekindergarten (see Table38). This result is a change from the previous year when U.S.-born mothers were more likely to use centercare than were foreign-born mothers. On the other hand, foreign-born mothers were significantly morelikely to use neighbors or friends for childcare.Table 38. Nonparental Childcare Arrangements for Focal Children at Year 5 by Race/Ethnicityand Nativity a Race/Ethnicity (%) Nativity (%)ChildcareArrangementAllMothersBlack Hispanic White/other Foreign-born U.S.-born(n=190) (n=89) (n=85) (n=15) (n=89) (n=101)Relative 17 14 18 33 14 21Friend/neighbor 10 1 20 0* 16 4*Childcare center/HeadStart55 65 46 47* 55 55Prekindergarten (schoolbased)8 6 9 13 9 7Family childcare 3 5 0 7 0 5Other/someone else 1 0 4 0 3 0Multiple arrangements 6 10 4 0 3 9a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower:[Race/ethnicity] friendneighbor(2) vs. friendneighbor(1) and childcareHeadStart(1) vs. childcareHeadStart(2). [Nativity]friendneighbor(1) vs. friendneighbor(2).Table 39 shows additional variations in childcare arrangements associated with mothers’ education levels,employment status, and income-to-needs ratio. Mothers who had graduated from high school weresignificantly more likely to use childcare centers (61% vs. 45%) and less likely to use care by friends orneighbors (4% vs. 19%) than mothers who had not graduated from high school. In addition, workingmothers were more likely than mothers who were not working to use care by relatives (21% vs. 8%) andfriends or neighbors (13% vs. 2%) and less likely to use a prekindergarten program (3% vs. 18%).However, the difference between employed and unemployed mothers in the use of center care or HeadStart programs is not significant. Also, there are no statistically significant differences in use of differentChapin Hall at the University of Chicago 64

types of care between mothers whose family incomes are at or below the poverty level versus those withhigher family incomes.The qualitative data suggest a variety of reasons for these differences in types of childcare arrangements,which include parents’ perceptions of the quality of care, ease of access, and cost. As we report in thenext section, the survey data suggest that use of center care, in particular, is related to having a childcaresubsidy through Family Central. Of those mothers who received a childcare subsidy, well over threequarters(82%) had their focal child enrolled in a childcare center, whereas less than half (42%) ofmothers who did not have a childcare subsidy had their child in center care. Among mothers whoparticipated in qualitative interviews, those who were working—both foreign-born and U.S.-born—frequently cited cost as the main barrier to using center care. 35Table 39. Nonparental Childcare Arrangements for Focal Children at Year 5 by MaternalEducation, Employment, and Income aChildcareArrangementMothersIncome-to-NeedEducation (%) Employment (%)Ratio (%)Not HS HS NotAt/below Abovegraduate graduate Working Working poverty poverty(n=190) (n=75) (n=114) (n=61) (n=127) (n=82) (n=98)Relative 17 20 16 8 21* 15 21Friend/neighbor 10 19 4* 2 13* 6 12Childcarecenter/Head Start55 45 61* 62 52 59 50Prekindergarten(school-based)8 8 8 18 3 7 9Family childcare 3 3 3 3 2 4 2Other/someone else 1 4 0 5 0 4 0Multiple6 1 10* 2 9* 6 5arrangementsa Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: [Mother’sEducation] friendneighbor(1) vs. friendneighbor(2); childcareHeadStart(2) vs. childcareHeadStart(1); mulitplearrangements(2)vs. mulitplearrangements(1). [Employment] relatives(2) vs. relatives(1); friendneighbor(2) vs. friendneighbor(1);prekindergarten(1) vs. friendneighbor(2); mulitplearrangements(2) vs. mulitplearrangements(1).35 A body of research indicates that foreign-born mothers tend to prefer informal care arrangements to formal, center-based care(Hernandez, Denton, & Macartne, 2007; Lippman et al., 2008; Fuller, Holloway, & Liang, 1996; Hirshberg, Huang, & Fuller,2005; Kinukawa, Guzman, & Lippman, 2004; Lawrence & Kreader, 2005; Liang, Fuller, & Singer, 2000). At the same time,some research suggests that such differences may be explained by other sociodemographic characteristics, the age of the child,and the availability, cost, and quality of care in different regions of the country. For example, according to a 2005 Child Trendsreport (http://www.childtrends.org), nationally, 53 percent of children ages 0 to 4 living in a low-income family were innonparental care in 2003, but the use of childcare by low-income families varied from state to state.Chapin Hall at the University of Chicago 65

Costs of Preschool ChildcareWhen asked if they pay for their childcare, half of the mothers using childcare reported that they pay allor some of their children’s childcare costs. Among the other 50 percent of mothers who reported notpaying for childcare, about two-thirds do not pay because they receive a childcare subsidy through FamilyCentral or financial assistance from a social service agency, relative, or other individual. Most of theremaining mothers who do not pay for childcare reported placing their children in the care of relatives.The qualitative data suggest these mothers may pay small amounts to reimburse relatives for childcare;for example, mothers talked about paying their mother or a sister $100 on an occasional basis or givingher money for gas, bus fare, or food in exchange for caring for their children.The amounts mothers paid for childcare on a regular basis ranged from as little as $5 per week to as muchas $275 per week, amounts that are similar to those reported in the fourth year. 36 The average weekly costof childcare reported by mothers was $70, which is slightly higher than what was reported in year 4 ($67).Based on these figures, we estimate that families who pay for some or all of their childcare spend from aslittle as $20 to as much as $1,100 per month for all children in care. Although these monthly costestimates were not adjusted to the numbers of hours per week children are in care, most children are incare on a full-time basis. The monthly average of childcare costs reported by mothers in the fifth year wasapproximately $280, slightly down from $297 in the previous year. To obtain a rough estimate of theproportion of family income these mothers spent on childcare, we calculated the percentage of themidpoint of the family’s reported income range for the previous year spent on childcare. Based on thiscalculation, the study families spent approximately 14 percent of their yearly income on childcare. 37More than half of the mothers using childcare in the fifth year reported receiving some financialassistance to help pay for that care. One-third of the mothers using childcare reported having a childcaresubsidy, which is significantly more than the 29 percent who reported receiving the subsidy in year 4. Anadditional 22 percent of mothers using childcare in year 5 obtained financial help from a social serviceagency. A small percentage (5%) said they received help from an individual, such as a friend or relative;only four mothers (2%) reported receiving assistance with childcare expenses from an employer. Again,this financial assistance was linked to the type of childcare arrangement mothers provided for their focalchild. As shown in Table 40, mothers who received a subsidy were much more likely to have their focal36 Interviewers reported that mothers sometimes had difficulty figuring the amount they pay for childcare, and some of thefigures reported for childcare seemed high in relation to household incomes. At the same time, although estimates vary, studiesindicate that low-income families spend a greater proportion—20 percent or more—of their income on childcare than do familieswith higher incomes (e.g., Chase et al., 2005; Henly & Lyons, 2000; Koppelman, 2002).37 One mother reported spending 274 percent of her yearly income on childcare and was excluded from this analysis.Chapin Hall at the University of Chicago 66

Table 40. Characteristics of Mothers Receiving and Not Receiving Subsidy for Childcare at Year 5 aCharacteristicAll Mothers UsingChildcareMothers withSubsidyMothers withoutSubsidy b(n=190) (n=60) (n=130)Age of motherMean age (SD) 29.4 (6.3) 29.3 (5.9) 29.4 (6.4)Number of children (%)One 25 15 30Two 35 35 35Three or more 40 50 35Mean number of children (SD)* 2.4 (1.2) 2.7 (1.2) 2.2 (1.2)Mother’s race (%)Black, not Hispanic 47 55 43Hispanic 45 38 48White/other c 9 7 9Mother’s nativity (%)U.S.-born 52 52 52Mother’s education (%)Less than high school diploma 40 35 42High school graduate 33 42 29Post–high school education 27 23 29Mother’s employment (%)**Employed full or part time 67 52 74Mother’s school status (%)In school full or part time 23 25 23Income-to-need ratio (%)*Living at or below poverty threshold 44 56 39Income supports received (%)Women Infants and Children (WIC) 46 52 44Food Stamps 42 52 38Social Security Insurance (SSI) 11 15 9TANF 3 5 2Rent voucher program 9 15 6Earned Income Tax Credit 41 37 43Childcare for focal child (%) (n=189)***Relative 18 2 25Friend/neighbor 10 0 14Childcare center/Head Start 55 82 42Prekindergarten (school-based) 8 3 10Family childcare 3 5 2Other/multiple arrangements 7 8 6a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Some (30%) of the mothers in this group reported receiving help with childcare costs from social service or welfare agencies(n=41); relatives (n=3); employer (n=4); and another source (n=9).c “White/other” includes mothers who self-identified as being white, Asian, American Indian, multiracial, and other.Chi-square tests indicate differences between mothers with and without childcare subsidies are statistically significant at *p < .05,**p < .01, or ***p < .001.Chapin Hall at the University of Chicago 67

child in a childcare center or Head Start (82%) than mothers who did not receive a subsidy (42%). Almost4 in 10 mothers not using a subsidy use friends or relatives to care for their children.Table 40 also shows socio-economic and demographic differences between the group of study motherswho received a subsidy to assist with childcare costs during the previous year and those who did not. Forexample, as might be expected, mothers using subsidies were significantly more likely to have an incometo-needratio at or below the poverty level (56% vs. 39%). They were also less likely to be employed atthe time of the fifth interview than mothers who did not receive subsidies (52% vs. 74%). In addition,mothers who reported receiving a subsidy have more children than mothers who did not.Although the differences are not significant, mothers who received a subsidy for childcare were morelikely to use other income supports than were mothers who did not receive a subsidy. Other studies reporthigher use of childcare subsidies among mothers using services such as TANF and Food Stamps as wellas differences by maternal education, race/ethnicity, and age and number of children (e.g., Herbst, 2008;Lee et al., 2004; Lowe & Weisner, 2004; Shlay et al., 2004).Childcare Quality and Satisfaction with ChildcareThere is considerable evidence that the quality of care makes a difference in children’s development andreadiness for school (Burchinal et al., 2002; Howes, 1997; Howes & Smith, 1995; NICHD EarlyChildcare Research Network, 2000, 2007; Pianta et al., 2002). This is particularly true for young childrenfrom families with limited economic resources or other challenges (Karp, 2006). Thus, in this study, itwas important to examine the impact of childcare experiences and the type and quality of childcare onchildren’s outcomes. The primary source of information on childcare quality, other than a measure of themothers’ general satisfaction with their childcare arrangements, is assessment information for childcarecenters and family childcare participating in the Palm Beach County Early Childhood QualityImprovement System (QIS). Mothers who use formal childcare arrangements were asked to name theirchildcare center or individual caregiver in an effort to determine how many children were receiving carefrom providers participating in the QIS. Mothers named 106 different centers and homes as childcareproviders for 204 children; 41 (39%) were names of providers participating in the QIS as of March 18,2010 (seeChapin Hall at the University of Chicago 68

Table 41). Of the 204 children, 91 (45%) were served by these 41 QIS sites (see Table 43). This is anincrease over the previous year, when just 29 percent of the centers and homes named by mothers as thelocations of childcare were names of providers in the QIS.Chapin Hall at the University of Chicago 69

Table 41. Percentage of Preschool Children in Study Families Receiving Care from QIS Providersat Year 5 aChildren in QIS Children in Non-QISAll ChildrenProviderPrograms bProgramsn % N % n %Childcare center/HeadStart/Prekindergarten176 86 87 49 89 51Family childcare/other childcare 28 14 4 14 24 86Any formal childcare arrangement 204 91 45 113 55a The percentages in the “All Children” column are based on a denominator of 204; the total number of children in formalchildcare arrangements. All other percentages are based on a denominator of the total n for the row, the total number of childrenin that specific childcare arrangement. Data were not weighted in the analysis.b Refers to programs participating in the QIS as of March 18, 2010.Table 43. Percentage of Preschool Childcare Providers for Study Families Participating in the QISat Year 5 aProviderChildcare center/HeadStart/PrekindergartenProviders in QIS Providers in Non-All ProvidersPrograms b QIS Programsn % N % n %89 84 39 44 50 56Family childcare/other childcare 17 16 2 12 15 88Any formal childcare arrangement 106 41 39 65 61a The percentages in the “All Providers” column are based on a denominator of 106, the total number of formal childcareproviders. All other percentages are based on a denominator of the total n for the row; the total number of formal childcareproviders providing that specific type of childcare. Data were not weighted in the analysis.b Refers to programs participating in the QIS as of March 18, 2010.In addition, the survey also asked mothers using childcare how satisfied they were with their childcarearrangements. Over three-fourths of the sample rated their childcare arrangements as either “very good”or “excellent”. There were no difference between satisfaction levels of parents using childcarearrangements in the Quality Counts (QC) system versus those not in the QC system. Fifteen percentreported that their childcare options were “good,” and just 4 percent said they were “fair.” A largerproportion of mothers relying on childcare centers or Head Start programs (61%) reported that theirchildcare options were “excellent” compared with mothers using other forms of care (3% to 22%).In general, mothers who participated in the qualitative interviews spoke favorably about the care providedby relatives—and some mothers were more likely to entrust the care of their children to a relative than tosomeone unknown even as their children grew older. On the other hand, a few mothers talked about thedifficulty of patching together childcare from different members of their families. Some mothers alsoChapin Hall at the University of Chicago 70

commented on their children’s needs for socialization and learning in a structured environment with otherchildren as they got older.School-Age Childcare and ActivitiesA total of 162 mothers, or 52 percent of the study families, had school-age children in year 5. Fifty-fivemothers, or one-third of this group, reported that their school-age children were involved in activities orchildcare after school. Mothers reported that between 1 and 5 of their school-age children were enrolledin afterschool activities; the mean age of these children was 9.6 years and their ages ranged from 6 to 17years. The amount of time these children spent in afterschool activities ranged from 3 to 46 hours perweek, with an average of 14 hours per week. A majority (70%) of the children engaged in afterschoolprograms and activities participated 5 days per week. Almost half of the group of mothers with schoolagechildren also reported that their older children participated in organized care or programs during thesummer. In well over half of the cases, school-age children who were in structured care or programsduring the school year also participated in summer activities. 38As with mothers’ use of preschool childcare arrangements, bivariate analyses indicates use of afterschoolactivities differs by a variety of sociodemographic characteristics and family circumstances. For example,64 percent of mothers living in the Glades involved their school-age children in formal afterschoolactivities, compared to 33 percent of mothers living in the non-Glades TGAs and 6 percent of those livingoutside the TGAs (χ 2 = 14.102, p < .01). Foreign-born mothers were less likely to involve children inafterschool activities than were U.S.-born mothers (24% vs. 47%, χ 2 = 9.444, p < .01). In addition,working mothers were more likely than nonworking mothers to involve their children in afterschoolactivities (54% vs. 19%, χ 2 = 21.588, p < .001).Table 44. Types of Afterschool Activities and Programs for 77 Children at Year 5 aFrequency of Mention b % Activities % ChildrenType of Program/Activity(n=174)(n=77)Education 42 24 55Recreation 35 20 45Sports 32 18 42Childcare 32 18 42Arts or crafts 29 17 38Other (not specified) 4 2 5Total 17438 National estimates of participation in afterschool programs indicate variations as a function of age of child, race/ethnicity,family income, and maternal employment and education. Overall, in 2005, 43 percent% of elementary school children inkindergarten through eighth grade participated in at least one afterschool activity, with higher participation among students fromnonpoor families (56%) than from poor families (22%) (NCES, 2007).Chapin Hall at the University of Chicago 71

a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Frequency of mention refers to number of different types of afterschool activities reported by 55 mothers for 77 children.Multiple responses were allowed for each child.Because afterschool programs are typically characterized by more than one type of activity, mothers wereallowed to give multiple responses when describing their children’s afterschool activities. Over half(57%) of mothers with school-age children in afterschool programs mentioned two or more differenttypes when describing their activities. As shown in Table 45, the most frequent type of activity mentionedby mothers was educational, but with recreational, childcare, and sports programs being mentioned almostas frequently.All mothers with school-age children enrolled in afterschool programs were asked to rate the quality oftheir afterschool and summer care arrangements. When asked how they felt about afterschoolarrangements for their school-age children, a majority (77%) reported that their afterschool arrangementswere either “very good” or “excellent.” However, 60 percent reported that they did not have a lot ofchoice in the selection of afterschool activities. Overall, mothers were a little less happy with summeractivities for their school-age children than they were with afterschool activities. Almost two-thirds saidsummer arrangements were “very good” or “excellent.” At the same time, most (59%) of the sample alsosaid that they did not have a lot of opportunities for summer activities.Table 45. Types of Afterschool Activities and Programs for 77 Children at Year 5 aFrequency of Mention b % Activities % ChildrenType of Program/Activity(n=174)(n=77)Education 42 24 55Recreation 35 20 45Sports 32 18 42Childcare 32 18 42Arts or crafts 29 17 38Other (not specified) 4 2 5Total 174a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Frequency of mention refers to number of different types of afterschool activities reported by 55 mothers for 77 children.Multiple responses were allowed for each child.SummaryMore than half (61%) of the mothers in the fifth year of the study were using nonparental carearrangements for their focal child, motivated largely by their needs as working mothers and/or as mothersenrolled in school; indeed, the small increase in the use of childcare between years 4 and 5 may be due inpart to the increase in the percentage of mothers enrolled in school. The most frequently reported type ofChapin Hall at the University of Chicago 72

nonparental childcare arrangement for the focal children was center care, followed by relative care, andcare by a friend or neighbor.Maternal employment and/or school enrollment continues to be the strongest predictor of childcare use.The only other significant factors predicting childcare use are residence in the Glades TGA andrace/nativity. In particular, both U.S.-born and foreign-born black mothers were significantly more likelythan foreign-born Hispanic mothers to use childcare.Additional bivariate analyses suggest that different types of childcare arrangements are associated with anumber of maternal characteristics, including race/ethnicity, nativity, education, employment, andincome. Among black mothers, over half (65%) used center care, and 14 percent had a relative who tookcare of their children. Among Hispanic mothers, almost half (46%) used center care, and one-fifth (20%)relied on a friend or neighbor to care for their children. Furthermore, foreign-born mothers were morelikely than U.S.-born mothers to use friends or neighbors to care for their children, while U.S.-bornmothers were more likely to rely on relatives. However, equal proportions of U.S.-born and foreign-bornmothers used center care (55%). This is a change from the previous year when U.S.-born mothers weremore likely to use center care.In general, we attribute differences in the use of different types of childcare arrangements to a variety offactors, but one important factor is cost. The survey data indicate that use of center care, in particular, isrelated to having a childcare subsidy through Family Central. Mothers who received a subsidy were muchmore likely to have their child in a childcare center, Head Start, or prekindergarten program (82%) thanwere mothers who did not receive a subsidy (42%). Over one-third (39%) of mothers not using a subsidyrelied on friends or relatives to care for their children.As reported in the third study report (Spielberger et al., 2009), the qualitative data suggest several otherfactors, in addition to maternal employment, that shape mothers’ decisions about the use of childcare. Oneis the greater availability of formal childcare options for children 3 years and older than for infants andtoddlers and, correspondingly, the mothers’ growing knowledge about the options in their communities.In addition, as children become more communicative, more self-sufficient, and more in control of theirbehavior with age, mothers become more comfortable with the idea of nonparental childcare and begin torecognize the importance of experiences with other children and adults for their cognitive and socialdevelopment.Finally, an analysis of the stability of childcare over time indicates that a majority of the focal childrenexperienced fairly stable childcare arrangements during the first 4 years of their lives, meaning there wereonly a small number of transitions in childcare. However, about 44 percent of the children experiencedChapin Hall at the University of Chicago 73

four or more transitions in childcare arrangements during the first 4 years of their lives. The difficulty insecuring a stable childcare arrangement reflects, in part, the instability of a family’s work circumstancesand income; at the same time, the lack of stability in childcare also contributes to instability in their work.Chapin Hall at the University of Chicago 74

Informal and Community SupportAn underlying assumption of the service system in Palm Beach County is that if families have stronginformal and community supports and access to prevention and early intervention services, they will beless likely to need more intensive intervention services. Indeed, an extensive body of research documentsthe benefits of informal and community supports (e.g., Balaji et al., 2007; Martinez-Schallmoser,MacMullen, & Telleen, 2003; Uno, Florsheim, & Uchino, 1998). Healthier prenatal practices, fewer birthcomplications, and lower incidence of postpartum depression—all associated with enhanced maternalfunctioning and child development outcomes—characterize mothers with strong social support networks.Thus, an important topic of the interview study is the availability and use of various informal andcommunity supports by families in the TGAs. In this section, we focus on the range and quality of socialsupports—including emotional, instrumental, and informational types of support—from family, friends,neighbors, and other community members over the first 5 years of the study.Informal Support from Family and FriendsHusband/Partner SupportMothers who were either married or in a relationship with someone they considered a partner—74 percentof the mothers in the year 5 sample—were asked about the types of support they receive from theirhusbands/partners. As shown in Table 46, virtually all of the mothers reported talking with their partnersabout problems or personal issues; receiving help with money, food, or clothing; and receiving help withdisciplining children in year 5. About 9 in 10 or more of mothers also reported receiving help withshopping, with repairs around the house, and with advice regarding children or household problems.However, only about a quarter of mothers reported receiving help in “other” areas, such as help withchildcare and transportation to school and doctor’s appointments.Chapin Hall at the University of Chicago 75

Over time, the percentages of mothers reporting help in a number of areas, including help with shoppingand child discipline, significantly increased. In addition, there was a significant increase in the percentageof mothers reporting that their “partner gives advice on children or household problems.”Table 46. Types of Husband/Partner Support over Time aType of Support% Year 1 b % Year 2 b % Year 3 b % Year 4 b % Year 5 b(n=249) (n=239) (n=228) (n=228) (n=229)Any support from husband/partner 99 100 99 99 99Talk to husband/partner about problems orpersonal things96 95 99 99 99Husband/partner helps with money, food, orclothing94 95 98 99 99Husband/partner shops for food orhousehold items78 80 85 95 92Husband/partner helps around house withcleaning or repairs82 85 81 90 94Husband/partner helps with child discipline 77 92 92 97 97Husband/partner gives advice on children orhousehold problems73 74 70 90 89Husband/partner provides other help (e.g.,transportation, childcare)33 25 10 15 25Husband/partner support score c (mean(SD))5.3 (1.5) 5.5 (1.2) 5.3 (1.2) 5.8 (1.0) 5.9 (0.9)a Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”b Responses to these questions were mothers who reported that they were married or in a relationship with a partner.c The husband/partner support score is based on the frequency of all individual items (i.e., calculated by adding all “yes”responses to the items in the table). This differs from the calculation of the support score in previous reports.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: talk1 vs.talk4; talk1 vs. talk5; talk2 vs. talk3; talk2 vs. talk4; talk2 vs. talk5; money2 vs. money4; money2 vs. money5; shop1 vs. shop4;shop1 vs. shop5; shop2 vs. shop4; shop2 vs. shop5; shop3 vs. shop4; clean1 vs. clean4; clean1 vs. clean5; clean2 vs. clean5;clean3 vs. clean4; clean3 vs. clean5; discipline1 vs. discipline2; discipline1 vs. discipline3; discipline1 vs. discipline4; discipline1vs. discipline5; discipline2 vs. discipline4; discipline2 vs. discipline5; discipline3 vs. discipline4; advice1 vs. advice4; advice1vs. advice5; advice2 vs. advice4; advice2 vs. advice5; advice3 vs. advice4; advice3 vs. advice5; other1 vs. other3; other1 vs.other4; other2 vs. other3; other2 vs. other4; other3 vs. other5; other4 vs. other5; score1 vs. score4; score1 vs. score5; score2 vs.score3; score2 vs. score4; score2 vs. score5; score3 vs. score4; score3 vs. score5.In year 5, as in previous years, a majority of mothers reported talking to their husbands/partners aboutproblems on a daily or weekly basis and getting help daily with money, food, or clothing (see Table 48).Overall, there was a slight increase in the overall amount of support over time. In addition, there werenotable increases in the amount of “daily” support for help around the house with cleaning or repairs and,especially, child discipline, which increased from 60 percent in year 1 to 79 percent in year 5. Consistentacross the 5 years, husbands/partners were least likely to perform household tasks and most likely toprovide child discipline on a “daily” basis.Chapin Hall at the University of Chicago 76

Table 46. Frequency of Husband/Partner Support over Time aType ofSupport% Year 1 % Year 2 % Year 3 % Year 4 % Year 5(n=249) (n=239) (n=228) (n=228) (n=229)Daily Weekly Daily Weekly Daily Weekly Daily Weekly Daily WeeklyTalk aboutproblems53 32 51 32 49 41 53 37 54 33Help withmoney,food, or51 31 53 33 52 33 48 38 61 29clothingHelp withchild 60 14 71 15 70 19 82 8 79 14disciplineShop for foodorhousehold10 55 6 60 9 63 13 60 10 60itemsHelp aroundthe house(cleaning20 41 20 36 23 27 25 29 33 38or repairs)Advice onchildren or 35 26 35 26 34 23 44 25 46 27householda Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: talk1 vs.talk4; talk1 vs. talk5; talk2 vs. talk4; talk2 vs. talk5; talk3 vs. talk4; talk3 vs. talk5; money1 vs. money4; money1 vs. money5;money2 vs. money4; money2 vs. money5; money3 vs. money4; money3 vs. money5; shop1 vs. shop4; shop1 vs. shop5; shop2vs. shop4; shop2 vs. shop5; shop3 vs. shop4; shop3 vs. shop5; clean1 vs. clean4; clean2 vs. clean4; clean3 vs. clean4; clean4 vs.clean5; discipline1 vs. discipline2; discipline1 vs. discipline3; discipline1 vs. discipline4; discipline1 vs. discipline5; discipline2vs. discipline4; discipline2 vs. discipline5; discipline3 vs. discipline4; discipline3 vs. discipline5; advice1 vs. advice4; advice1vs. advice5; advice2 vs. advice4; advice2 vs. advice5; advice3 vs. advice4; advice3 vs. advice5; other1 vs. other3; other1 vs.other4; other1 vs. other5; other2 vs. other3; other2 vs. other4; other2 vs. other5; other4 vs. other5.Chapin Hall at the University of Chicago 77

Table 48. Frequency of Husband/Partner Support over Time aType ofSupportTalk aboutproblemsHelp withmoney,food, orclothingHelp withchilddisciplineShop for foodorhouseholditemsHelp aroundthe house(cleaningor repairs)Advice onchildren orhousehold% Year 1 % Year 2 % Year 3 % Year 4 % Year 5(n=249) (n=239) (n=228) (n=228) (n=229)Daily Weekly Daily Weekly Daily Weekly Daily Weekly Daily Weekly53 32 51 32 49 41 53 37 54 3351 31 53 33 52 33 48 38 61 2960 14 71 15 70 19 82 8 79 1410 55 6 60 9 63 13 60 10 6020 41 20 36 23 27 25 29 33 3835 26 35 26 34 23 44 25 46 27a Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: talk1 vs.talk4; talk1 vs. talk5; talk2 vs. talk4; talk2 vs. talk5; talk3 vs. talk4; talk3 vs. talk5; money1 vs. money4; money1 vs. money5;money2 vs. money4; money2 vs. money5; money3 vs. money4; money3 vs. money5; shop1 vs. shop4; shop1 vs. shop5; shop2vs. shop4; shop2 vs. shop5; shop3 vs. shop4; shop3 vs. shop5; clean1 vs. clean4; clean2 vs. clean4; clean3 vs. clean4; clean4 vs.clean5; discipline1 vs. discipline2; discipline1 vs. discipline3; discipline1 vs. discipline4; discipline1 vs. discipline5; discipline2vs. discipline4; discipline2 vs. discipline5; discipline3 vs. discipline4; discipline3 vs. discipline5; advice1 vs. advice4; advice1vs. advice5; advice2 vs. advice4; advice2 vs. advice5; advice3 vs. advice4; advice3 vs. advice5; other1 vs. other3; other1 vs.other4; other1 vs. other5; other2 vs. other3; other2 vs. other4; other2 vs. other5; other4 vs. other5.The qualitative data offer additional evidence for changes in the frequency of household support fromhusbands and partners and their roles in the family over time. For example, mothers described usingalternate sources of advice (e.g., pediatricians, books, and other printed materials). They also expressedincreased confidence in their parenting skills as their children became older. In addition, although mothersappeared to appreciate the support and contributions of their husbands/partners, some also expressed adesire for more independence. At the same time, as children became older, husbands/partners took onadditional responsibilities for disciplining them.Mothers also continued to report satisfaction with the support they received from their husbands/partnersin the fifth year. Almost three-fourths of the year 5 sample described themselves as “very satisfied,” 24percent were “somewhat satisfied,” and just 4 percent were “somewhat or very dissatisfied.”Chapin Hall at the University of Chicago 78

We also asked the mothers about the quality of the relationship with their husbands/partners. When askedhow often their husbands/partners show affection toward them, 79 percent of the sample responded“often,” and 19 percent said “sometimes.” When asked about the frequency of arguments with theirhusbands/partners, 72 percent reported that they argue “sometimes,” 24 percent said “never,” and 4percent said “often.” Qualitative data indicate that when there are conflicts, they center around money andparenting. As one mother explained, “We just don’t see eye to eye on how we’re supposed to raise thekids and how life is supposed to be for us. I refuse to compromise a little his way, he refuses tocompromise my way.”Family and Friends SupportIn addition to husbands and partners, the vast majority of mothers (92%) reported receiving assistancefrom family members or friends in year 5, and this percentage is comparable to the 89 percent of motherswho reported assistance in year 1. Siblings and mothers/stepmothers were the two most frequentlymentioned sources of support (see Table 48). More than half of the mothers reported receiving help fromsiblings, and half of them reported receiving help from mothers or stepmothers in year 5, which isconsistent with the first 4 years of the study. Reported support from all family members in general,including siblings, mothers/stepmothers, aunts, uncles, and cousins, increased between year 4 and year 5.It should be noted that the proportion of the mothers who were either married or partnered decreased fromyear 1 (80%) to year 5 (74%). Thus, the rise in support received from “former” husbands/partners mayreflect the increase in marital or partner separations within the sample families.The frequency of support from family and friends also declined over time (see Table 48). In year 1, 40percent of the mothers reported receiving “daily” advice on child rearing or household problems; by year5 it declined to 25 percent. Mothers were also less likely to talk to friends/family about personal matterson a “daily” basis. In year 1, 33 percent of the mothers reported “daily” support, but by year 5, only 21percent of the sample reported “daily” support for personal problems.Finally, although not shown in the tables below, the number of mothers who reported receiving noassistance or assistance just once a year with food, money, and clothing jumped from 40 percent in year 1to 53 percent of the cohort in year 5. Again, this general decline in high-frequency support from friendsand family might be related, in part, to participants’ growing skills and parenting competencies as well asthe growing independence of the focal children. It also might reflect, for some mothers, a reluctance toask for help or a lack of confidence in the quality of care friends or relatives might give their children.One mother told us, “I rarely like to ask for favors.” Another mother advised, “Let me tell you—savemoney so you can stay home because nobody is going to give your child better care than you.”Chapin Hall at the University of Chicago 79

Table 48. Family and Friends Support over Time aPerson Providing Support% Mothers (N=310) bYear 1 Year 2 Year 3 Year 4 Year 5Family memberMother/stepmother 52 51 47 49 50Father/stepfather 24 26 25 23 27Grandparent 23 21 18 16 18Sister/brother 55 55 58 55 58Aunt/uncle 31 28 22 21 29Cousin 27 24 18 19 27Grown children and otherrelatives25 21 23 20 20Mother-in-law/father-in-law 30 30 27 22 29Nonfamily memberFriend 51 44 49 52 56Neighbor 20 17 14 13 24Former husband/partner 7 9 9 10 11Coworker 6 11 7 9 11Others 2 4 3 3 4No family or friend support 11 9 7 9 8Family/friend support score b 4.1 (3.4) 4.1 (3.3) 3.7 (2.6) 3.5 (2.6) 4.3 (3.2)a Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”b The family/friend support score combines items that ask about the frequency with which mothers receive any assistance fromany of the individuals listed in this table (i.e., calculated by adding all “yes” responses to the items in the table). This differs fromthe calculation of the support score in previous reports.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: aunt1 vs.aunt3; aunt1 vs. aunt4; aunt2 vs. aunt3; aunt2 vs. aunt4; aunt3 vs. aunt5; aunt4 vs. aunt5; cousin1 vs. cousin3; cousin1 vs.cousin4; cousin2 vs. cousin3; cousin3 vs. cousin5; cousin4 vs. cousin5; in-law1 vs. in-law4; in-law2 vs. in-law4; in-law3 vs. inlaw4;in-law4 vs. in-law5; grandparent1 vs. grandparent3; grandparent1 vs. grandparent4; grandparent1 vs. grandparent5;grandparent2 vs. grandparent4; former husband1 vs. former husband5; friend1 vs. friend2; friend2 vs. friend4; friend2 vs.friend5; neighbor1 vs. neighbor4; neighbor2 vs. neighbor5; neighbor3 vs. neighbor5; neighbor4 vs. neighbor5; coworker1 vs.coworker2; coworker1 vs. coworker5; no one1 vs. no one3; score1 vs. score3; score1 vs. score4; score2 vs. score3; score2 vs.score4; score3 vs. score5; score4 vs. score5.Chapin Hall at the University of Chicago 80

Table 49. Type of Family/Friends Support Received at least Weekly over Time a,b% MothersType of SupportYear 1 Year 2 Year 3 Year 4 Year 5(n=276) (n=283) (n=288) (n=282) (n=285)Talk about personalproblems59 54 50 58 48Advice on children orhousehold problems64 52 45 43 43Help withhousework/childcare57 52 46 45 50Help with money, food, orclothing28 18 9 15 16a Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”bPercentages indicate the total of mothers who responded “weekly” or “daily” on a 5-point scale.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: talk1 vs.talk3; talk1 vs. talk5; talk2 vs. talk5; talk3 vs. talk4; talk4 vs. talk5; money1 vs. money2; money1 vs. money3; money1 vs.money4; money1 vs. money5; money2 vs. money3; money3 vs. money4; money3 vs. money5; housework1 vs. housework3;housework1 vs. housework4; housework2 vs. housework3; advice1 vs. advice2; advice1 vs. advice3; advice1 vs. advice4;advice1 vs. advice5; advice2 vs. advice3; advice2 vs. advice4; advice2 vs. advice5.Community Support“Community” encompasses a variety of professionals with whom mothers may have contact—everyonefrom medical to school personnel, social service personnel to members of the clergy. Overall, the mothersreport higher levels of community support as time progresses. Forty percent of the mothers reportedsupport from someone in the community in year 1, 45 percent in year 2, 57 percent in year 3, and 58percent in year 4. However, in year 5, the number of mothers who reported receiving help from someonein the community decreased dramatically; in this year, just 22 percent of mothers reported receiving helpfrom community members. 39With respect to the kinds of community support families received in year 5, mothers tended to reportreceiving advice on children or household problems somewhat more frequently than concrete supportsuch as help with money, food, or clothing (see Table 50). However, reported help from communitymembers with money, food, or clothing increased from the previous year. Lesser levels of concretesupport were received more frequently in all 5 years. In the previous year, just 2 percent of the samplereported receiving this help “weekly,” and 8 percent “monthly.” Reported help in the form of advice onchildren or household problems in year 5 also increased from that in year 4.39 More detail on changes in social support can be found in previous study reports (e.g., Spielberger et al., 2010). Over the first 4years of the study, mothers reported receiving increased support from healthcare professionals and childcare providers, butvariable support from social workers. In the year 5 interview, we did not ask about specific individuals in the community whoprovided support.Chapin Hall at the University of Chicago 81

Type of SupportFrequency (%)(n=69)Daily Weekly Monthly Annually Never bAdvice on children or household problems 8 27 22 25 18Help with money, food, or clothing 2 4 19 25 50a Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”b The “Never” response refers to mothers who reported receiving some types of community support but not the type indicated.Access to SupportAs shown in Figure 4, perceptions of access to support fluctuate over time, with few notable patterns. Forexample, the percentage of mothers who said it was “somewhat easy” or “very easy” to talk to someoneabout things that are very personal or private was stable in years 1 through 3, increased somewhat in year4, but declined again in year 5. We observed a similar pattern for those who said it was “somewhat easy”or “very easy” to get advice on caring for children or dealing with household problems. On the otherhand, the percentage of mothers who found it either “somewhat easy” or “very easy” to get help withhousework or childcare generally increased, though modestly, over time.Figure 4. Perceptions of Access to Support from All Sources over Time aa Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: talk1 vs.talk4; talk2 vs. talk4; talk4 vs. talk5; childcare1 vs. childcare4; childcare1 vs. childcare5; childcare2 vs. childcare3; childcare2 vs.childcare4; childcare2 vs. childcare5; advice1 vs. advice2; advice2 vs. advice3; advice2 vs. advice4; advice4 vs. advice5.Chapin Hall at the University of Chicago 82

Access to Support by TGAFigure 5 shows perceptions of support over time by TGA. Every year, mothers in the Glades were morelikely to report finding access to support to be either “very easy” or “somewhat easy,” when comparedwith their counterparts outside of the Glades. Across all 5 years, more than three-fourths of mothers in theGlades were more likely to report “very easy” or “somewhat easy” access to getting help with houseworkor caring for children. In most cases, however, the percentage of mothers outside the Glades reporting thislevel of access was 10 to 20 percentage points lower. One factor in these geographic differences may bethat families in the Glades appear to be less mobile (see chapter 2); thus, even though this area may havemore “risk” characteristics than do other areas, the level of perceived social support is higher. Becausefamilies are not moving and have lived longer in the Glades, mothers may have more family and friendsaccessible to them as well as more knowledge of community supports and services.Figure 5. Access to Support by TGA over Time aa Data were weighted to account for the oversampling of mothers in the Glades and mothers assessed “at risk.”b Based on a 5-point scale: “Very Easy,” “Somewhat Easy,” “Between Hard and Easy,” “Somewhat Hard,” and “Very Hard.”This table combines the “Very” and “Somewhat Easy” responses.Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 83

SummaryMothers in the fifth year continued to voice high satisfaction with the level of informal support theyreceive. Consistent with previous years, if the mothers had husbands/partners, they most often receivedsupport from them, although the mean level of support from husbands and partners was higher in year 5than in year 4. Otherwise, the mothers relied primarily on their family, especially siblings and mothers orstepmothers. In addition to family members, the mothers depended on friends for support. The meansupport score based on the mothers’ reports of help received from family and friends in the fifth year wassignificantly higher in the fifth year compared to the fourth year; on average, the fifth-year score washigher than it was in the first year. In addition, we also saw a decline in the frequency with which somekinds of support were provided by husbands/partners, other family members, and friends.In year 5, more (18%) mothers reported that they never received any advice on children or householdproblems as compared to the previous 3 years of the study (6% in year 2, 4% in year 3, and 6% in year 4).With regard to money, food, or clothing, far fewer mothers (50%) in year 5 reported that they neverreceived any assistance in this area as compared to the first 4 years of the study (74% in year 1, 67% inyear 2, 72% in year 3, and 75% in year 4). The mothers’ perceptions of their access to social support,regardless of source, fluctuated over time, and we are not able to discern any significant patterns.As described in Spielberger et al. (2009), the qualitative data suggest a variety of reasons for trends (andlack thereof) over time. Sometimes relatives are no longer able to help with childcare or turn out to beunreliable caregivers, so mothers have to find other sources of help. Over time, if they can afford it,mothers living with relatives or friends increasingly try to set up households of their own—although theymay continue to share resources, such as childcare and Food Stamps. As their children grow, somemothers convey an increasing desire to be independent. They also express more confidence in theirparenting abilities, and, although respondents’ mothers remain an important source of information andsupport, sample mothers appear to be turning more to doctors, teachers, and other nonfamily forinformation and support than they did when their children were younger.Chapin Hall at the University of Chicago 84

Use of Formal ServicesA central question of the longitudinal study is what formal services are available and used by familieswith young children in the TGAs. Data on service use come from both administrative data records and themothers’ self-reports. This chapter begins with a summary of analyses of information from the FOCiSdatabase about the use of maternal and child health services by mothers in the Palm Beach County 2004–2009 birth cohorts and mothers in the baseline study sample. 40,41 More details of these analyses and datatables can be found in previous study reports (Spielberger et al., 2007, 2009) and in Appendices B and C.We then present findings from the fifth-year interview, in which we asked the mothers to identify a rangeof service areas in which they received help during the previous year.The Maternal Child Health SystemThe Maternal Child Health Partnership (MCHP) is a growing network of integrated and coordinatedprevention and early intervention health and social services for pregnant women who are at risk of poorbirth outcomes and for children from birth to age 5 who are not on target for developmental or healthoutcomes, and who, without the support of the program, may have an increased risk of abuse or neglect ora diminished chance of being ready for school. 42 Focused on improving birth outcomes, reducing childmaltreatment, and promoting early childhood development and readiness to learn, the MCHP system40 Because the year 5 study sample is similar in most respects to the baseline sample (see Table 3), we expect that statistics onuse of maternal and child health services will be very similar for the two samples.41 The FOCiS database replaced the Right Track database as the source of information on mothers served by the maternal childhealth system in November 2007.42 This system has evolved over time. It began as the Healthy Start Coalition in 1992 and later became a larger, more formalsystem called the Maternal Child Family Health Alliance. In 2004, shortly after the start of the longitudinal study, the system wasreconstituted as the Maternal and Child Health Partnership (MCHP). Although the system was renamed Healthy Beginnings in2006, we refer to the system here as “MCHP” because it was in place at the time mothers were recruited for the study. We alsowant to distinguish this system from a new Healthy Beginnings system that formally began in July 2009 and includes some newpolicies and procedures for screening and engaging families in services. Over the period of the study, agencies providing servicesin the system have also changed; some changed their names, whereas others ceased operation or left the system. We did notinclude FOCiS data for these agencies after these dates in counts of services in the system.Chapin Hall at the University of Chicago 85

includes universal risk screening, targeted home visitation programs, and referrals and linkages to a rangeof services within and outside the system. Services also include assistance connecting with a payer sourceand/or medical home, health education/health literacy, identification of and treatment for perinataldepression, nutrition counseling, childbirth and breastfeeding education, and family support servicesranging from telephone counseling to intensive home visiting services. Qualified providers also deliverdevelopmentally appropriate early childhood therapeutic and family supportive services to eligiblefamilies and children birth to age 5 through home visits or consultations at childcare or other clientconvenientlocations. These services include parenting support/education and infant mental health/socialemotionalwellness, including parent/child bonding, family therapy, and early literacy.Early identification is a key preventive service for all pregnant and postnatal women and children birth toage 5. According to a recent description of the MCHP, in the future, the system will include two entryagencies based on assessment information. 43 There were three common entry points to the system duringthe period of this study, however: the Healthy Mothers/Healthy Babies Coalition of Palm Beach County;Healthy Start/Healthy Families; and the Women’s Health Initiative (WHIN). Mothers could enter thesystem prenatally or postnatally based on the Healthy Start Prenatal and Infant Risk Screen or a homeassessment, but a majority of mothers entered through the Healthy Mothers/Healthy Babies Coalition,whose staff of hospital liaisons administered the 10-item Healthy Start Risk Screen to as many newlydelivered mothers as possible. On this instrument, a score of 4 or higher is considered an indicator ofpossible risk, and mothers who scored 4 or higher were encouraged to accept a home visit from a HealthyStart nurse. Mothers who received lower scores were not offered a home visit but could request one. 44Subsequently, in a home visit, mothers were reassessed and identified as having service needs of E, 1, 2,or 3. Mothers assigned levels 2 and 3 were thought to need more frequent or more varied services andthus were loosely referred to as “high risk,” whereas mothers with service levels of E or 1 weredesignated as “not high risk.” Mothers who were screened or assessed at a level 3 were automaticallyassigned to intensive care coordination services. 45In the following sections, we summarize what we have learned about mothers’ use of the MCHP systemin Palm Beach County.43 Personal communication with Carol Scott and Regina Battle, CSC September 2008.44 At the time we began recruiting, the Healthy Start program included a universal home visiting component for all newly deliveredmothers. In spring 2005, the program changed to target mothers who were, based on a risk screen score, identified as most in needof and more likely to use services. Mothers in a “special low risk” group were also offered a home visit if they were younger than19, new to the county, had delivered their first child with no or only late-term prenatal care, had no identified pediatrician, haddifficulty bonding with their baby, or seemed to lack social support (personal communication with Tanya Palmer, Children’sServices Council, 2005; personal communication with Christine Walsh, Healthy Mothers/Healthy Babies Coalition, 2005).45 Risk screen and assessment scores also are likely to change over time with subsequent contacts with healthcare and otherservice providers.Chapin Hall at the University of Chicago 86

Correlates of At-Risk StatusA large majority of mothers who gave birth in Palm Beach County in 2004–2005 received a risk screen.As shown in Table B- 1 of the Appendix, 73 percent of all mothers in the 2004–2005 county birth cohorthad a risk screen score in the FOCiS database, and 81 percent of the mothers residing in the TGAs had arisk screen score. These percentages increased to 84 and 88 percent, respectively, for the 2008 birthcohort. However, they declined to 75 and 79 percent, respectively, for the 2009 cohort.One-third (32%) of the mothers in the 2004–2005 birth cohort who had a risk screen score were identifiedas “at risk”; the corresponding percentage for mothers in the TGAs was 44 percent. These numbersremained fairly stable over time, although the figures for the 2009 birth cohort were somewhat lower thanthose for previous cohorts.As reported in Spielberger et al. (2010), we found that 94 percent of the mothers in the baseline studysample had a risk screen score recorded in the administrative data for the MCHP system, and 56 percentof these mothers were designated to be “at-risk” based on the risk screen.Further, a logistic regression analysis of the survey and Vital Statistics data for the baseline sample inyear 4 (Spielberger et al., 2010) reveals that, controlling for the effects of all other variables, a mother’seducation, marital status, age at birth of first child, and baby’s birth weight are significantly related to thelikelihood of her being designated “at-risk.” In particular, the odds of a mother with less than a highschool education being designated “at-risk” are over 4 times those of mothers with a high school diplomaor above. Also, the odds of an unmarried mother being designated “at-risk” are over 7 times those of amarried mother, while those of a mother with a low–birth weight baby are almost 3 times those of amother with a normal weight baby.Receipt of MCHP ServicesAs shown in Table B- 4 of the Appendix, 27 percent of the mothers in the 2004–2005 birth cohortreceived MCHP treatment services; among mothers residing in the TGAs, this figure was 40 percent. Thepercentage of mothers receiving treatment services in the 2006, 2007, and 2008 birth cohorts was similar,but the percentage among mothers in the 2009 birth cohort was somewhat lower. This could be becausewe may not yet have complete data on service receipt for this cohort and/or the result of a change inMCHP intake procedures.We also looked at differences in receipt of treatment services across race/nativity groups in the 2004–2009 birth cohorts. We found that, on average, 51 percent of foreign-born Hispanic mothers receivedtreatment services in the Maternal Child Health system, compared to about 35 percent of foreign- andnative-born black mothers, and 24 percent of native-born Hispanic mothers. Mothers of “other”Chapin Hall at the University of Chicago 87

ace/ethnicities, whether foreign or native born, were the least likely to receive treatment services; only 10percent of these mothers received such services.A logistic regression analysis of the likelihood of service receipt among mothers in all five birth cohortscombined (the “county sample”) reveals that the following variables, obtained from Vital Statistics data,increase the likelihood of mothers’ receipt of treatment services: use of WIC during pregnancy, use oftobacco during pregnancy, late or no prenatal care, less than a high school education, not married at timeof birth, a teen at time of birth, residing in the TGAs, health complications during pregnancy, and deliverycomplications (see Table B- 5 in the Appendix). 46 In particular, the odds of receiving treatment servicesincrease most for mothers who are not married, mothers with less than a high school education, mothersliving in a TGA, mothers who use WIC, and mothers who were teens at the time of birth. We also foundthat blacks (both U.S.- and foreign- born), native Hispanics, and other racial/ethnic groups (U.S. andforeign born) are less likely than foreign-born Hispanics to receive treatment services. In particular, theodds of native-born blacks and Hispanics receiving treatment services are about one-third of the odds forforeign-born Hispanics. Finally, controlling for all other factors, mothers in the 2005–2009 birth cohortswere less likely to receive treatment services than mothers in the 2004 birth cohort. The odds of treatmentreceipt for the 2005–2008 cohorts were about 90 percent of the odds for the 2004 cohort, while the oddsof treatment receipt for the 2009 cohort were about 60 percent of those for the 2004 cohort. Again, thelatter result is probably due in part to the fact that we don’t have complete data on service receipt for the2009 cohort.The mothers in our baseline study sample were about twice as likely as other mothers in the TGAs tohave received MCHP treatment services: 80 percent of these mothers had records of treatment activity inthe FOCiS database. This result is not surprising, given the maternal characteristics targeted by the MCHPsystem and the fact that we recruited mothers through two core programs in the system—the HealthyMothers/Healthy Babies Coalition and Healthy Start/Healthy Families—and oversampled mother whowere screened to be at “at risk.”Most sample mothers who received services in the system—either care coordination or intensive carecoordination—did so during the 3 months before and the 6 months after they gave birth (Spielberger etal., 2009). Very few mothers who received only care coordination services continued to receive these46 Health complications included in this variable are diabetes, hypertension, previous preterm birth, previous poor pregnancyoutcome, pregnancy the result of infertility treatments, and “other” health complications. Delivery complications includedventilator was required, newborn given antibiotics, significant birth injury, hyaline membrane disease, newborn intensive careunit admission, seizures or serious neurologic dysfunction, surfactant given to newborn, and “other” child complications.Chapin Hall at the University of Chicago 88

services after 6 months. Among the mothers who received intensive care coordination, a little more than aquarter received services between 6 and 9 months after giving birth.A logistic regression analysis of the survey and Vital Statistics data for the baseline sample indicates that,controlling for all other variables, “at-risk” status is not significantly related to the likelihood of receivingservices (Spielberger et al., 2010). In addition, similar to the results for the county sample, blacks (bothU.S.- and foreign-born), native Hispanics, and other racial/ethnic groups (U.S. and foreign born) were lesslikely than foreign-born Hispanics to receive treatment services. 47 Also, similar to the county results,mothers who were teens at the birth of their focal child were more likely to receive treatment services. Inaddition, two variables that were not available for the county cohort—income below poverty and having afocal child with special needs—were also positively and significantly related to the likelihood ofreceiving treatment services. On the other hand, in contrast to the county results, mothers in the samplewho smoked during pregnancy were significantly less likely to receive treatment services. Finally, theresults indicate that mothers who were employed and those who owned their own home, or lived withsomeone who owned the home, were less likely to receive services.To learn more about why some sample mothers did not receive treatment services after the initialscreening, we examined the termination codes entered into the administrative data system (seeSpielberger et al., 2010). We found that one-third of “at-risk” mothers who did not receive treatmentservices could not be located. Additionally, for about 20 percent of these mothers, the initial assessmentcould not be completed. Based on conversations with MCHP staff, this could have occurred for severalreasons, including cancellation of the assessment appointment or a decision by the mother to end theappointment before completion of the assessment. 48Finally, for most of the remaining mothers, MCHP staff determined that no further services were neededafter the completion of the initial assessment, while 15 percent of the mothers declined services. We alsofound that MCHP staff members were unable to locate a much higher proportion of mothers determinedto be “not at risk”; almost 7 in 10 fell into this category. Also, as with “at-risk” mothers, MCHP staffmembers were unable to complete the initial assessment with about 20 percent of those identified as “notat risk.” Finally, smaller percentages were determined not to need further services (8%) or they declinedservices (3%).47 We also ran a regression including an indicator of foreign-born status only, instead of the indicators of race and nativity. Theresults indicate that foreign-born mothers are significantly more likely to receive treatment services, when compared to U.S.-bornmothers.48 Personal communication with Joanne Newman, Department of Health, Healthy Start/Healthy Families Program, January 2009.Chapin Hall at the University of Chicago 89

Receipt of More Intensive Treatment ServicesOverall, 45 percent of the mothers in the 2004–2009 birth cohorts who received treatment servicesreceived basic care coordination, while the remainder received intensive care coordination. Thesepercentages are fairly consistent across the five birth cohorts, except that larger percentages of mothers inthe 2008 and 2009 birth cohorts received intensive care coordination (see Table B- 7 of the Appendix).However, as reported in chapter 2, the demographic characteristics of the birth cohorts did not changesubstantially over time (aside from a decline in the percentage of mothers with late or no prenatal care);thus, it is not clear what accounts for the increase in the percentage of mothers who received intensivecare coordination.A logistic regression analysis using the county sample data reveals that the following variables increasethe odds of a mother receiving intensive care coordination: use of WIC during pregnancy, use of tobaccoduring pregnancy, late or no prenatal care, mother had less than a high school education, mother was notmarried at time of birth, mother was a teen at time of birth, mother resided in the TGAs, mother hadhealth complications during pregnancy, baby was underweight and baby was premature (see Table B- 8 inthe Appendix). Furthermore, the odds increase most for mothers with less than a high school education,mothers who smoked, mothers who were not married, mothers who were teens, and mothers with late orno prenatal care. We also found that U.S.-born blacks were significantly more likely than foreign-bornHispanics to receive intensive care coordination; however, the remaining race/nativity groups hadsignificantly lower odds of receiving intensive care coordination, when compared to foreign-bornHispanics. Finally, we found that mothers in the 2008 and 2009 birth cohorts were significantly morelikely to receive intensive care coordination, as compared to mothers in the 2004 cohort.Of the mothers in the baseline study sample who received MCHP treatment services, approximately 39percent received care coordination, while 61 percent received intensive care coordination. Comparing thetwo groups of mothers reveals that those who received intensive care coordination were significantlymore likely to show depressive symptoms, to be poor, to have a child with special needs, to have aphysical or mental health problem, or to live with someone with a health problem than were mothers whoreceived care coordination (Spielberger et al., 2010). They were also more likely to have both indicatedand investigated DCF reports. We did not, however, find any significant differences with respect torace/nativity, other demographic characteristics, or other variables available from Vital Statistics (e.g. lateor no prenatal care).Chapin Hall at the University of Chicago 90

Enhanced ServicesEnhanced services are considered to be more specialized or rigorous services and exclude screening andgeneral case management. 49 In the county sample, 31 percent of mothers who received care coordinationalso received enhanced services, and 56 percent of mothers who received intensive care coordination alsoreceived enhanced services.As shown in Table B- 10 of the Appendix, translation services is the most common category of enhancedservices, followed by breastfeeding services. Other enhanced services include parenting support andeducation, counseling for depression, nutrition education, and childbirth education, smoking cessationservices, and male support services.A logistic regression analysis of the likelihood of a mother in the county receiving enhanced services,given that she received care coordination, shows that mothers who were teens at birth and those who usedtobacco were less likely to receive enhanced services. On the other hand, mothers who (1) had less than ahigh school education, (2) were not married at the time of the birth, (3) had health or deliverycomplications, or (4) had an underweight baby, were more likely to receive enhanced services. Withrespect to demographic characteristics, we also found that all race/nativity groups were less likely thanforeign-born Hispanic mothers to receive enhanced services, while mothers living in the TGAs were morelikely to receive enhanced services. Finally, mothers in the 2005–2008 birth cohorts were more likely toreceive enhanced services, as compared to mothers in the 2004 birth cohort.We also conducted a logistic regression analysis of the likelihood of a mother in the county receivingenhanced services, given that she received intensive care coordination, and found results similar to thoseabove. For example, mothers who were teens at the time of birth were less likely to receive enhancedservices. In addition, mothers who (1) had less than a high school education, (2) were not married at thetime of the birth, or (3) had health or delivery complications were more likely to receive enhancedservices. Similarly, with respect to demographic characteristics, we found that all race/nativity groupswere less likely than foreign-born Hispanic mothers to receive enhanced services. Finally, mothers in the2005–2009 birth cohorts were more likely to receive enhanced services, when compared to mothers in the2004 birth cohort.Among the sample mothers who received care coordination, 44 percent also received enhanced services.A logistic regression analysis of the likelihood of receiving enhanced services shows that, similar to theresults for the county, blacks (both U.S.- and foreign-born) were significantly less likely than foreign-49 For all categories, with the exception of interconceptional education and childbirth education, more mothers used theseservices postnatally than prenatally.Chapin Hall at the University of Chicago 91

orn Hispanics to receive enhanced services. However, there was no difference between U.S.- andforeign-born Hispanics, or between mothers of other races/ethnicities and foreign-born Hispanics. Also,mothers with depressive symptoms or physical/mental health problems were significantly less likely toreceive enhanced services. On the other hand, similar to the results for the county, mothers who were highschool dropouts remained more likely to receive enhanced services; the odds of these mothers receivingsuch services were over 2½ times the odds of mothers with a high school diploma or above doing so.Among the sample mothers who received intensive care coordination, about 53 percent also receivedenhanced services. A logistic regression analysis shows that, after controlling for other variables, U.S.-born blacks and residents of the Glades were significantly less likely to receive enhanced services, aswere mothers who owned their own homes or lived with a homeowner.Days of Treatment ServicesAs shown in Table B- 11 of the Appendix, mothers in the county who received treatment servicesexperienced 14 days of treatment on average (SD = 20.7). In addition, across the cohorts, there has been adecline in the days of treatment services received, with the majority of the decline appearing to arise froma decline in the number of days of intensive care coordination.An OLS regression analysis of the county data reveals that the following variables are associated with anincrease in the days of treatment services received by mothers: use of alcohol during pregnancy, use ofWIC during pregnancy, use of tobacco during pregnancy, late or no prenatal care, mother had less than ahigh school education, mother was not married at time of birth, mother was a teen at time of birth, motherresided in the TGAs, mother had health complications during pregnancy, delivery complications, babywas underweight and baby was premature (see Table B- 15 in the Appendix). The largest increases werefound for mothers with less than a high school education, mothers who were teens, mothers who residedin the TGAs, mothers who were not married, mothers who used alcohol, and mothers who used WIC. Wealso found that all race/nativity groups received 2 to 2½ fewer days of treatment on average, whencompared to foreign-born Hispanics. Finally, we found that mothers in the 2008 and 2009 birth cohortsreceived 1 to 2 fewer days of treatment on average, when compared to mothers in the 2004 cohort.In the baseline survey sample, mothers who received treatment services experienced 20 days of treatmenton average (SD = 26). As reported in Spielberger et al. (2010), we did not find many differences betweenmothers receiving more as opposed to fewer days of treatment services. A linear regression analysis ofdays of treatment services reveals that, controlling for all other variables, each additional child isassociated with about 4 more days of service on average. Also, having a child with special needs isassociated with an increase of almost 9 days, while having depressive symptoms or a physical or mentalhealth problem is associated with an increase of about 13 days.Chapin Hall at the University of Chicago 92

Termination of MCHP ServicesAs previously reported, to understand why mothers in the sample may have stopped receiving any type oftreatment services, we examined the termination codes entered into the MCHP administrative data system(Spielberger et al., 2010). We found that, for over half of the mothers, MCHP staff determined that nofurther services were needed. On the other hand, it appears that MCHP staff either lost contact with, orwere unable to continue providing services for other reasons, for about 20 percent of the mothers. 50Similarly, based on treatment patterns observed in the data, it appears that approximately 8 percent of themothers who received services for their focal child had a subsequent child, and MCHP staff memberswere then unable to make an initial assessment for that child or to engage the mother in further services. 51In addition, 5 percent of the mothers explicitly declined to receive any additional services, and 10 percentdid not have a termination code recorded in the system.Finally, we did not find many differences between mothers with different termination codes. The mainsignificant differences suggest that mothers determined not to need further services were less likely to beblack and more likely to be Hispanic or to be foreign born.Other Formal ServicesAdministrative records on participation in MCHP services are just one source of information on mothers’service use. In the annual in-person surveys, we also asked the mothers about other services they receivedand their needs for services in a wide range of areas. These included meeting their family’s basic needs,such as food, clothing, and housing; childcare; medical and mental healthcare; and addressing concernsabout their children’s health and development. We defined service use broadly as “help received from anyagency, program, or professional” to meet these needs. For example, mothers coded as receiving helpwith food included those who received assistance through any of these sources, including the WIC orFood Stamp programs, or a church food pantry. If the mothers did not report receiving help in a particularservice area, we asked whether or not they had had concerns in that area and sought help for thoseconcerns.It should be noted that the fourth and fifth in-person household surveys did not ask the mothers abouttheir level of satisfaction with the services they received because, in previous years, responses variedlittle. Overall, satisfaction levels with all areas of service were high in year 3, with more than three-50 According to documentation for the MCHP administrative data, the code “unable to locate” can be used for “those participantswho have not responded to three attempts to contact or for those participants who covertly decline.”51 Mothers with “unable to complete initial assessment/contact” codes typically had treatment activity over a period of months,followed by a significant lapse in services just prior to the termination code.Chapin Hall at the University of Chicago 93

fourths of the mothers using services describing themselves as “very satisfied” with them on a 4-pointscale. 52In year 5, all but 19 mothers, or 94 percent of the sample, reported receiving help from a program,agency, or professional in the past year for at least one area of concern regarding basic family needs. 53Mothers received help for an average of three basic family needs, slightly more than the average numberof reported services in previous years. In year 5, 22 percent of the mothers received help in five or moreservice areas, which is a significant increase from year 4, when just 12 percent (p < .01) reported usingfive or more services; year 3, when 11 percent (p < .001) reported using five or more services; and years 1and 2, when just 14 percent (p < .01) each year reported using five or more services.Services for Basic Family NeedsIn general, across all 5 years, mothers were more likely to have concerns about and receive help in areasdefined as basic family needs than to have specific concerns related to their children’s health anddevelopment. Consistent with the first 4 years of the study, in the fifth year, mothers reported receivinghelp most often with healthcare (86%) and food (72%). In addition, 59 percent of the mothers reportedgetting help with dental care; similar to year 4, this represents a significant increase relative to the first 3years when just 22 to 27 percent of mothers reported receiving help with dental care. This might reflectthe older age of the children, as well as mothers’ increasing awareness of the importance of dental care tooverall health. 54The percentage of mothers who reported help with food in the fifth year was about the same as in the thirdand fourth years, but not as high as it was in the first and second years, which is consistent with trends inthe use of WIC shown in Table 6. 55 Also consistent with the increasing age of the children, the percentageof mothers reporting help with childcare has increased over time. Less than a third of the sample receivedhelp with any other area of service use. Family planning, paying rent/bills, and housing were the nextmost frequent areas in which mothers reported receiving assistance; also, trends in the use of these52 With regard to the year 3 satisfaction items for the year 5 sample, the two lowest “very satisfied” ratings were for theemergency shelter item (72% “very satisfied”) and the school attendance item (43% “very satisfied”). Nine of the 29 items had100 percent of the mothers who were “very satisfied” and in another 11 areas between 90 and 99 percent of the mothers reportedbeing “very satisfied” with services received.53 This pertains to just those items asked about at all five yearly time points.54 According to the American Academy of Pediatric Dentistry (www.aapd.com), children should first visit a dentist when theyget their first tooth or no later than their first birthday. Pediatricians and childcare providers might also encourage mothers’ use ofdental care. Florida participates in a “Head Start Dental Home Initiative,” which might influence families whose children areenrolled in Head Start and Early Head Start.55 Additionally, Food Stamps were used by 47 percent of the year 5 mothers and by 35 percent and 38 percent of the year 3 andyear 4 mothers, but in year 1 just 37 percent and in year 2 46 percent of the mothers used them. The use of Food Stamps wasdefinitely higher than in year 3 and year 4, but usage was not less than in year 1 or in year 2.Chapin Hall at the University of Chicago 94

services were fairly stable, although there has been a small drop-off in the use of family planning or birthcontrol services.Table 51. Help Received for Basic Family Needs over Time a % Mothers (N=310)Service AreaYear 1 Year 2 Year 3 Year 4 Year 5Healthcare for mother or children 75 95 94 89 86Getting enough food (includes receipt offood stamps and WIC)89 87 68 70 72Dental care 22 27 26 55 59Childcare (includes subsidy use) 17 27 28 40 41Family planning or birth control — b 33 28 20 25Paying rent or bills (includes receipt of ahousing voucher)11 10 10 11 15Housing or emergency shelter c 5 10 10 11 11Mental health or substance abuse 2 2 3 8 11Legal issues 2 2 6 7 6Parenting information 73 23 11 5 8Transportation 17 16 7 5 7Employment 5 5 3 5 9Clothes for children 6 5 3 2 5Domestic violence — — — 2 3Reading or writing skills — 0 0 1 1Translating things into English d 11728—(n=147) (n=149) (n=152) (n=142)a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b Item not included in year 1 survey.c The year 1 survey had one item combining housing and emergency shelter, which was separated into two items in allsubsequent surveys; they are combined in this analysis.d Only those mothers who do not speak English as their primary language were asked this question (n=147, 149, 152, 142).*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower: health1vs. health2; health1 vs. health3; health1 vs. health4; health1 vs. health5; health2 vs. health4; health2 vs. health5; health3 vs.health4; health3 vs. health5; food1 vs. food3; food1 vs. food4; food1 vs. food5; food2 vs. food3; food2 vs. food4; food2 vs.food5; familyplanning2 vs. familyplanning4; familyplanning2 vs. familyplanning5; familyplanning3 vs. familyplanning4; dental1vs. dental4; dental1 vs. dental5; dental2 vs. dental4; dental2 vs. dental5; dental3 vs. dental4; dental3 vs. dental5; childcare1 vs.childcare2; childcare1 vs. childcare3; childcare1 vs. childcare4; childcare1 vs. childcare5; childcare2 vs. childcare4; childcare2vs. childcare5; childcare3 vs. childcare4; childcare3 vs. childcare5; childcare4 vs. childcare5; parenting1 vs. parenting2;parenting1 vs. parenting3; parenting1 vs. parenting4; parenting1 vs. parenting5; parenting2 vs. parenting3; parenting2 vs.parenting4; parenting2 vs. parenting5; parenting3 vs. parenting4; rentbills2 vs. rentbills5; rentbills3 vs. rentbills5; rentbills4 vs.rentbills5; housing1 vs. housing2; housing1 vs. housing3; housing1 vs. housing4; housing1 vs. housing5;transportation1 vs.transportation3; transportation1 vs. transportation4; transportation1 vs. transportation5; transportation2 vs. transportation3;transportation2 vs. transportation4; transportation2 vs. transportation5; legal1 vs. legal3; legal1 vs. legal4; legal1 vs. legal5;legal2 vs. legal3; legal2 vs. legal4; legal2 vs. legal5; employment1 vs. employment5; employment2 vs. employment5;employment3 vs. employment5; employment4 vs. employment5; clothing1 vs. clothing3; clothing1 vs. clothing4; clothing4 vs.clothing5; mental1 vs. mental4; mental1 vs. mental5; mental2 vs. mental4; mental2 vs. mental5; mental3 vs. mental4; mental3 vs.mental5; translation2 vs. translation4; translation3 vs. translation4; translation4 vs. translation5.Chapin Hall at the University of Chicago 95

With respect to other services, although only 11 percent of mothers reported use of mental health orsubstance abuse services, this is a significant increase relative to use in the first few years. (Becausemothers might have been reluctant to admit using mental health services, this might represent both anincrease in service use and an increase in mothers’ willingness to share this information with ourinterviewer.) Finally, less than 10 percent of the mothers recalled receiving help with parentinginformation, legal issues, transportation, finding employment, translation services, and clothing forchildren. Just 8 percent of the mothers with limited English proficiency reported receiving help translatingthings into English. The percentages of mothers reporting help in these areas have either been relativelystable over time, or show no clear pattern, except for some decline in the percentage reporting help withparenting information and transportation.Child Development ServicesIn the fifth year, 68 percent of the sample reported that they received help from a program, agency, orprofessional for some aspect of their children’s health or development; this represents an increasecompared to previous years. As shown in Table 52, more than half received help with a concern abouttheir children’s physical health or illness. Sixteen percent received help with a child’s speech or otherlanguage problem, and smaller percentages received help in the other areas, including problems withpaying attention, physical development, and learning new things. In general, these percentages werestable over time, or showed no clear patterns. Of the mothers with children older than 5 years, 28 percentsaid they received help for concerns related to academic progress, and 22 percent reported receiving helpfor concerns related to their children’s homework.Service Use by TGAAs shown in Table C- 22, a larger percentage of mothers living in the Glades TGA group than thoseliving in the non-Glades TGAs or outside the TGAs reported that they received services in five or moreareas, and this is consistent with reports in previous years. In terms of specific areas of service use,mothers living in the Glades were significantly more likely than those living outside the TGAs to receivehelp with getting enough food. They were also significantly more likely than the other two groups ofmothers to receive help with childcare. Also, mothers in the non-Glades TGAs were significantly lesslikely than mothers outside the TGAs to receive help with legal services and domestic violence.These differences in service use might reflect some of the characteristics that differentiate the threesamples, as well as variations in the availability of services in these communities. It may be more difficultfor the larger percentage of foreign-born mothers who do not speak English as their primary language inthe non-Glades TGAs to get information about available services or to find services in their ownChapin Hall at the University of Chicago 96

languages. The fact that more mothers in the non-Glades TGAs reported that they do not drive or thattransportation is not easy for them also suggests that they may have more difficulty accessing someservices.Table 52. Help Received for Concerns about Children’s Health and Development over Time aService AreaChildren of all ages% Mothers (N=310)Year 1 Year 2 Year 3 Year 4 Year 5Physical health or illness 20 47 39 46 58Language and communication 3 8 7 8 16Physical development 3 5 4 3 4Problems paying attention 3 2 4 3 7Problems learning new things 1 2 1 3 5Anger: getting upset or angry 1 2 1 2 4Getting along with peers, siblings, oradults1 1 1 2 3Eating problems 2 5 2 1 3Child’s sadness, depression, shyness,or withdrawal b 1 1 1 1 2Older children > 5 cAcademic progress — d — d 16 23 28Doing homework 12 13 21 13 22School attendance 5 3 2 4 3Older children > 10 cUse of drugs or alcohol 2 0 2 0 1a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b The year 1 survey had two separate items about the child’s sadness, depression, shyness, or withdrawal, which are combinedhere.c Sample sizes for items about older children varied year to year depending on the number and age of older children in thehouseholds.d Item not included in year 1 or year 2 surveys.*Paired sample t-tests indicate these year-to-year differences are statistically significant at p < .05 or lower: health1 vs. health2;health1 vs. health3; health1 vs. health4; health1 vs. health5; health2 vs. health3; health2 vs. health5; health3 vs. health4; health3vs. health5; health4 vs. health5; language1 vs. language2; language1 vs. language3; language1 vs. language4; language1 vs.language5; language2 vs. language5; language3 vs. language5; language4 vs. language5; social1 vs. social5; social2 vs. social5;social3 vs. social5; anger1 vs. anger5; anger2 vs. anger5; anger3 vs. anger5; attention1 vs. attention5; attention2 vs. attention5;attention3 vs. attention5; attention4 vs. attention5; learning1 vs. learning5; learning3 vs. learning5; eating2 vs. eating3; eating2vs. eating4; homework1 vs. homework3; homework2 vs. homework3; homework4 vs. homework5; academics3 vs. academics5.Chapin Hall at the University of Chicago 97

Table 53. Use of Selected Services by Mothers in Glades and Non-Glades TGA at Year 5 aService Area% AllMothers% Glades % Non-Glades % Outside TGAs(N=310) (n=38) (n=232) (n=40)Healthcare for mother or children 86 95 85 85Getting enough food* 72 84 71 60Dental care 59 76 57 55Childcare (including subsidy)* 41 71 37 39Family planning or birth control 25 13 27 25Paying rent or bills 15 18 16 12Housing or emergency shelter b 11 16 9 15Mental health or substance abuse 10 11 9 17Legal issues* 6 5 5 15Parenting information 8 11 7 8Transportation 7 5 8 5Employment 9 11 10 3Translating things into English(n=142)9 0 11 0Clothes for children 5 5 5 5Domestic violence* 3 3 2 8aData were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”bThe year 1 survey had one item combining housing and emergency shelter, which was separated into two items in all subsequentsurveys; they are combined in this analysis.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: food(1) vs.food(3); childcare(1) vs. childcare(2); childcare(1) vs. childcare(3); legal(3) vs. legal(2); domesticviolence(3) vs. domesticviolence(2).Service Needs and Help-Seeking in Year 5The survey also asked mothers who did not receive help in a particular service area during the previousyear if they had had a concern in that area and, if so, whether they had tried to get help for that concern.Their responses are presented in the next two tables. First, Table 54 shows the percentage of the group ofmothers who did not get help in a particular area but said they had a concern in that area. Healthcare wasthe area mentioned most frequently among mothers not receiving services. In addition, almost 4 in 10mothers cited dental care as a concern, while about a quarter of mothers indicated concerns with respectto childcare, paying rent or bills, getting enough food, translating things into English, and employment.Ten to fifteen percent of mothers indicated concerns in the areas of housing or emergency shelter, legalissues, transportation, and mental health or substance abuse, while less than 10 percent indicated concernsin the remaining areas. Only small percentages of the mothers reported not receiving help for needsrelated to their children’s health or development. The largest area of concern with respect to childdevelopment was help with anger issues, mentioned by 6 percent of mothers with school-age children.Chapin Hall at the University of Chicago 98

In interpreting these results, it should be noted that the denominator or the number of mothers who did notreceive help varied by item (see Table 51) because only those mothers who did not get help in a particularservice area were asked if they had concerns in that area.Table 54. Service Needs: Concerns of Mothers Who Did Not Receive Services over Time aService Area% Mothers bYear 1 Year 2 Year 3 Year 4 Year 5Healthcare for mother or children 42 75 87 64 45Dental care 36 40 28 24 39Childcare (including subsidy) 29 30 38 39 27Paying rent or bills 32 23 30 28 25Getting enough food 32 14 32 24 24Translating things into English — c 25 39 30 24Housing or emergency shelter 23 28 27 24 15Employment 15 11 14 19 19Transportation 19 15 15 12 11Legal issues 11 10 11 15 12Clothes for children 17 12 10 11 6Child’s homework issues (children > 5) 10 11 11 2 4Parenting information 10 4 3 3 7Child’s language and communication 3 7 6 6 4Family planning or birth control — 10 5 5 5Child’s anger problems 1 5 6 5 6Mother’s reading or writing skills — 7 5 5 5Child’s attention problems 3 5 5 3 5Child’s academic progress (children > 5) — — 10 5 6Mental health or substance abuse 1 1 1 6 10Child’s social problems 1 3 3 4 3Child’s learning problems 3 1 1 2 4Child’s eating problems 2 2 2 2 4Child’s sadness, depression, shyness, orwithdrawal2 2 4 1 3Child’s school attendance (children > 5) 5 1 3 1 2Child’s physical development 1 2 1 1 3Domestic violence — — — 3 2a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”bPercentage of mothers who did not receive help and who said area was a concern.cItems were not included in these surveys.Next, in Table 55, we present the percentages of mothers with concerns in each service area who soughthelp for their concerns. These results suggest that mothers seek help at different rates depending on thearea of service need. That is, mothers with concerns were more likely to seek help in the areas ofhealthcare (98%), childcare (80%), and getting enough food (65%) than in other areas. They were nextChapin Hall at the University of Chicago 99

most likely to seek help with concerns related to legal issues (49%) and housing (39%). The percentagesof mothers who had concerns about their children’s development and who sought help were somewhatsmaller than the percentages who sought help for basic needs. Sixty percent sought assistance forconcerns they had regarding their child’s physical development, over half (56%) sought help for concernsover their child’s language and communication abilities, and over a third (40%) sought assistance forconcerns they had regarding their child’s getting along with peers, siblings, and adults. For children 5years of age or older, over three-quarters (89%) sought help with their older children’s academic progress,and 86 percent sought help for concerns they had regarding their children’s homework.Table 55. Mothers Who Sought Help for Concerns over Time aService Area% Mothers bYear 1 Year 2 Year 3 Year 4 Year 5Healthcare for mother or children 89 100 82 73 98Childcare (including subsidy) 33 52 55 54 80Child’s physical development 22 50 63 41 60Getting enough food 51 20 58 38 65Child’s doing homework (children > 5) 0 49 36 59 86Child’s language and communication 27 32 48 50 56Child’s school attendance (children > 5) 31 0 100 0 60Housing or emergency shelter 41 40 31 34 39Paying rent or bills 32 30 24 23 36Family planning or birth control — c 44 11 56 26Employment 17 30 20 31 38Legal issues 25 20 7 25 49Dental care 11 28 11 37 32Child’s learning problems 27 32 0 39 6Child’s academic progress (children > 5) — — 14 0 89Child’s social problems 20 12 9 21 40Child’s attention problems paying attention 16 32 19 0 25Mother’s reading or writing skills — 11 23 19 37Child’s eating problems 13 22 30 17 8Mental health or substance abuse 23 0 30 14 21Parenting information 37 17 19 10 5Transportation 17 18 3 18 31Child’s anger problems 17 12 11 23 19Domestic violence — — — 10 29Clothes for children 8 20 1 15 7Child’s sadness, depression, shyness, orwithdrawal16 0 20 0 13Translating things into English — 8 3 3 3a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”bPercentage of mothers who sought help for a concern. Note that the denominator or the number of mothers who did not receivehelp varied by item (see Table 51).c Item not included in survey.Chapin Hall at the University of Chicago 100

Overall, mothers reported seeking help at higher rates in year 5 compared to prior years. Most of thetrends help-seeking for specific concerns, however, fluctuated up and down and showed no clear patternsover time. The exceptions were trends in help-seeking for childcare, employment, reading/writing skills,dental care, and children’s language and communication, which tended to increase over time. In contrast,trends in help-seeking for family planning and parenting information tended to decrease over time.Mothers Perceptions of Reasons Services Not ReceivedWhen mothers reported they did not receive help they sought, we asked them why. As shown in Table 56,many of the reasons were related to program or provider limitations. Most often, mothers said they weretold they were not eligible for help when they applied for services because they did not meet the incomethreshold, did not live in the service area, or for some other reason. In some cases, mothers were eligiblefor help at the time they applied, but the service was not available, and they were put on a waiting list.This explanation was often given to mothers who did not receive a childcare subsidy. In other cases,mothers were told that services were no longer available (for example, as was the case with Section 8housing vouchers), or that the agency would not provide a service for some other reason.Another type of reason why mothers did not receive services was that they had started or completed anenrollment process and were waiting for services and, thus, were reasonably assured of obtaining theservice in the future. These included mothers who had submitted their applications and were waiting for aresponse and mothers who had an appointment scheduled. Another small group of mothers said they hadcontacted a service provider but had not received a response.A third category of reason includes characteristics of individual mothers that interfered with theirparticipation in services for which they might have been eligible, including a lack of follow-through orpersistence, losing paperwork or not having the necessary paperwork, and having accessibility issues,such as a lack of childcare or transportation, or difficulty scheduling an appointment. Some of thesereasons, of course, overlapped with program and provider factors, such as eligibility requirements andoffice hours.As discussed in the third study report (Spielberger et al., 2009), analysis of three waves of qualitative datasuggest that a number of other factors affect service use. Mothers frequently reported reasons related toprovider characteristics that adversely affected their receipt of services. For example, a mother whosechild was covered under Medicaid his first year lost coverage when she reapplied because she was toldher income from a new job was too high. Medicaid referred her to KidCare, but KidCare staff referred herback to Medicaid. After several months of being bounced back and forth between the two programs, sheChapin Hall at the University of Chicago 101

grew tired of the situation and stopped applying. At the time of her last interview, no one in the familyhad insurance of any kind. They reported using a local clinic for any medical needs.Table 56. Mothers Seeking Help Who Did Not Receive Services at Year 5 aReason Frequency Percent bMother/children not eligible for services (income too high, do not live inservice area)74 31Mother made contact but has not heard back 36 15Mother put on waiting list 31 13Service not available anymore (e.g., no vouchers left, no waiting list) 16 7In process of getting service (paperwork in process) 14 6Mother lost paperwork or did not have necessary papers 13 5Mother received response from provider but still has a concern 10 4Mother sought help for area of concern, but agency did not cover 9 4Transportation problems 7 3Mother frustrated and gave up or no longer interested so did not follow up 7 3Told no jobs available 4 2Childcare problems 4 2Service too costly 3 1Mother has not received help yet but has appointment scheduled 3 1Mother unable to contact service 3 1Mother missed appointment 2 1Language barrier 2 1Accessibility—difficult to schedule appointments due to scheduleconflicts1 0Husband/partner will not commit to help 1 0Total number of reasons given 240a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”bPercentage of all reasons. Multiple responses were allowed.Use of the Maternal Child Health System and Other Formal ServicesOne function of the MCHP system is to link families to other needed services and supports available inthe community. Thus, to better understand the relationship between participation in the maternal childhealth system and the use of other formal services, we previously conducted a series of analyses tocompare the experiences of mothers who last received MCHP treatment services more than 12 monthsbefore each survey date in years 2 through 4 to mothers who never received any MCHP treatmentservices (see Spielberger et al., 2010). 56 Data for the analyses came from the second-through-fourth-yearChapin Hall household surveys and from the Vital Statistics and FOCiS data systems.56 We focused on treatment service use that occurred prior to the 12-month period preceding the survey to help ensure thattreatment service receipt did not overlap with receipt of other services.Chapin Hall at the University of Chicago 102

We found that in years 2 and 3, after controlling for other factors, receipt of treatment services wasassociated with a statistically significant increase in the number of areas in which mothers reportedreceiving assistance. However, we did not find a statistically significant relationship between receipt oftreatment services and use of other services in year 4. When we repeated the analysis for year 5, we againfound a significant relationship between receipt of treatment services and use of other services.We also found that, in year 2, mothers who had received MCHP treatment services were significantlymore likely to report receiving help with child development in the 12 months preceding the survey. Inyear 3, mothers who had received MCHP treatment services were significantly more likely to reportreceiving help with family planning, child development, and parenting information in the 12 monthspreceding the survey. Finally, in year 5, mothers who had received MCHP treatment services weresignificantly more likely to report receiving help with dental care, paying rent or bills, and mental healthservices; however, the latter two results are only significant at the 10 percent level.We cannot rule out the possibility that at least some of the increased assistance in years 2, 3, and 5associated with MCHP treatment services may be due to reporting error on the part of mothers. We arefairly certain, however, that the relationship between MCHP treatment and receipt of other services is notdue to mothers reporting MCHP treatment services for subsequent children.Finally, we found little evidence of a relationship between the number of days of treatment servicesreceived by a mother and the number of other service areas in which she reported receiving assistance.We also did not find any evidence of a relationship between receiving intensive care coordination (asopposed to basic care coordination) or one or more referrals, and the number of areas in which mothersreceived assistance.Overall, it appears likely that the MCHP system is successful in linking mothers, either directly orindirectly, with other services and supports in the community. This is an important role for the system toplay, since the majority of mothers did not receive MCHP treatment services after their children were 6months old. To the extent that receipt of MCHP treatment services connected mothers to other servicesand supports, they and their children may have indirectly benefited from the MCHP system over a longerperiod of time. In the next chapter, we look at whether families and children who received other servicesand supports are better off than those who did not.SummaryA large majority of mothers who gave birth in 2004–2005—73 percent of the county cohort and 81percent of the TGA cohort—were screened. About a third of the county cohort were designated to be “atChapin Hall at the University of Chicago 103

isk” compared to 44 percent of the TGA cohort. In the study sample, 94 percent of the mothers werescreened and 56 percent had screen scores that designated them to be “at risk.”At the same time, we previously found that 80 percent of the mothers in the baseline study samplereceived services from the MCHP system around the birth of their focal child. This suggests that riskscreen score was not the only criterion for receiving services. Most mothers received these services duringthe 3 months prior to and 6 months after their baby’s birth. A little more than a fourth of the mothers whoreceived intensive care coordination (61% of the sample) received services up to 9 months after theirbaby’s birth.We found some evidence that mothers with greater needs were more likely to receive services. Forexample, mothers who were teens at the birth of their focal child, mothers living in poverty, and thosewith a focal child with special needs were more likely to receive services. Although smoking duringpregnancy was associated with greater likelihood of receiving services in the county sample, it was notfor the study sample; mothers in the study sample who smoked during pregnancy were less likely toreceive services. Also, with respect to the study sample, we found that mothers with physical or mentalhealth problems or depressive symptoms were just as likely to receive services as not. Finally, we foundunexplained racial/ethnic and nativity differences related to the likelihood of receiving services in boththe county cohort and the study sample.Of the mothers who were “at risk” but did not receive services, a majority could not be located or MCHPstaff members were unable to complete the initial assessment. This suggests that making additional effortto locate these mothers and/or simplifying the assessment process might have resulted in more “at-risk”mothers receiving services.Finally, we found that, among mothers who received services, those who received more days of servicesmay have had greater needs. Specifically, mothers with more children, one or more child with specialneeds, and depressive symptoms or a physical/mental health problem received more days of services.With regard to using other formal services, most (94%) of the year 5 sample reported receiving help inmeeting basic family needs from at least one program, agency, or professional in the past year. In year 5,mothers received help for basic family needs from an average of three services, slightly more than allprevious year’s average numbers. Consistent with previous years of the study, mothers in the fifth yearreported receiving help most often with healthcare. Other areas of frequent assistance noted in the fifthyear were in the areas of dental care and childcare. Less than a third of the sample received help in anyother service area, with family planning, paying rent or bills, and housing as the next most frequent areasin which mothers reported receiving assistance.Chapin Hall at the University of Chicago 104

We also found that mothers tended to seek help at different rates depending on the area of service need.Mothers were more likely to seek help when they had concerns (and had not received any help in theprevious year) in the areas of healthcare (98%), childcare (80%), and getting enough food (65%). Theywere next most likely to seek help with concerns related to legal issues (49%) and housing (39%). Thepercentages of mothers who had concerns about their children’s development and sought help weregenerally smaller than the percentages who sought help for basic needs.Because one goal of the service system is to link mothers to other needed services, we also examined therelationship between participation in MCHP and use of other formal services. We found some evidencethat mothers who received MCHP treatment services also received more from other formal services inyears 2, 3, and 5 (though not in year 4).In the next chapter, we turn to findings related to the associations between service use—including use ofservices in the MCHP system around the birth of the focal child—and maternal functioning and childdevelopment.Chapin Hall at the University of Chicago 105

Service Use, Maternal Functioning,and Child OutcomesExperts generally agree that the stresses associated with living in poverty are likely to impair effectiveparenting and child well-being (Bradley & Corwyn 2002; Duncan & Brooks-Gunn 2000; McLoyd, 1998).It has also been suggested that society may be able to improve the outcomes of children living in povertyby strengthening the level of support available to them and their parents by providing public services andsupports (Attree 2005). This conclusion stems, in part, from evidence that social support, both informaland formal, is an important resource for poor parents (Attree 2005; Henly et al. 2005). Building on thisidea, comprehensive, integrated service systems have been viewed increasingly as a promising strategyfor supporting healthy family functioning and child development in low-income, at-risk families (Brooks-Gunn, 2003; Gomby 2005; Olds et al. 2007). However, although a sizable body of literature focuses onthe effects of specialized services and interventions for children and families, few studies to date haveattempted to assess the impact of families’ use of comprehensive services (Leventhal et al., 2000).The Palm Beach County longitudinal study provides a unique opportunity to understand the potentialconsequences of mothers’ use of services, in the broadest sense, for their families’ well-being.Accordingly, in the fourth study report (Spielberger et al., 2010), we considered how overall service useby mothers, as well as other variables, related to their own and their children’s outcomes. We did not findmuch evidence of a relationship between mothers’ overall or specific service use and maternal and childoutcomes; the evidence we did find suggests that mothers with greater needs (and thus poorer outcomes)tend to use more services over time. 57 However, we did find significant differences with respect to57 It should be kept in mind that, because mothers voluntarily decided whether or not to use services, and were not randomlyassigned to receive services, it is not possible for this study to fully control for all differences between mothers who did and didnot receive services. Thus, it is not surprising that in some cases we found a significant relationship between greater service useand poorer outcomes.Chapin Hall at the University of Chicago 106

variables that could potentially be impacted by mothers’ service use, and that might then have an impacton children’s development. In particular, we found that mothers’ depressive symptoms and parentingscores in year 1—both of which have the potential to be influenced by service use—were significantlyrelated to poorer child outcomes in year 4.In this final chapter, we extend the analyses in the fourth study report by examining the relationshipbetween mothers’ service use, mothers’ characteristics, and maternal and child outcomes in year 5. Wefocus in particular on the extent to which the year 5 results resemble or differ from the year 4 results.Also, similar to last year, the maternal outcomes that we consider include depression, parental stress, andpositive and negative parenting practices. For children, we examine outcomes drawn from two sources:(1) mothers’ reports of the focal children’s language, social-emotional competencies, and emergentliteracy/pre-academic skills at the time of the fifth interview and (2) children’s scores on a schoolreadiness screen administered soon after they entered kindergarten.Maternal DepressionAs described previously, we used the Center for Epidemiologic Studies Depression Scale (CES-D) toassess mothers’ depression symptoms. Scores on the CES-D can range from 0 to 60, with higher scoresindicating the presence of more depressive symptoms, and scores of 16 or higher indicating the presenceof some depression. In the fifth year, the mothers’ CES-D scores ranged from 0 to 43, with an average of9.7. Also, 21 percent of the mothers had scores of 16 or higher.As in year 4, we conducted logistic regression analyses to determine if service use and/or mothers’characteristics are significantly associated with the likelihood of depressive symptoms in year 5, aftertaking into account possible correlations between the explanatory variables. We estimated and report theresults from two different models. In the first model, we employed average service use during years 2through 5 as the primary independent variable. In the second, we employed receipt of help with mentalhealth services as the primary independent variable. 58The results of the logistic regression analyses are presented in Table 57. Similar to the result for year 4,after controlling for mothers’ baseline characteristics, there is no significant relationship between serviceuse and mothers’ depressive symptoms in the fifth year. Also similar to the year 4 results, aside fromdepressive symptoms in year 1, the only other significant predictor of depressive symptoms in year 5 is58 The study risk index was not included in the regressions since it is most likely highly correlated with many of the explanatoryvariables. Also, because the partner support score is closely related to whether or not a mother is married or cohabiting, these twovariables were entered in separate regressions. However, neither variable has a statistically significant coefficient. For simplicity,only the estimated coefficient for living with a husband/partner is presented in the results table.Chapin Hall at the University of Chicago 107

eing black and foreign born; the odds of depression for these mothers are almost 4 times those forHispanic foreign-born mothers (the “excluded,” or comparison, group).Table 57. Logistic Regression of Mothers’ Depression Symptoms at Year 5 aPredictor VariableModel 1 b Model 2Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 1.1 NS — —Ever received help with mental health, years 2–5 — — 1.8 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 1.8 NS 1.9 NSBlack—foreign-born 3.6 * 3.8 *Hispanic—U.S.-born 1.4 NS 1.5 NSHispanic—foreign-born (excluded category) — — — —White/other 1.1 NS 1.0 NSEducationLess than high school diploma 0.6 NS 0.6 NSHigh school graduate 1.2 NS 1.1 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 1.0 NS 1.0 NSGlades 0.7 NS 0.7 NSNumber of children 1.0 NS 1.0 NSLives with husband/partner 1.0 NS 1.0 NSMother employed 0.5 NS 0.5 NSIncome at or below poverty 1.8 NS 1.8 NSChild with special needs 0.7 NS 0.7 NSNumber of housing problems 1.1 NS 1.2 NSDepression: CES-D score > 16 4.6 *** 4.7 ***Family/friend support score 0.9 NS 0.9 NSDCF involvement in child’s first year 1.9 NS 2.1 NSχ 2 59.8*** 62.0***a Data were not weighted in this analysis.b We estimated (and report the results from) two different models. In the first model, we employed average service use duringyears 2 through 5 as the primary independent variable. In the second, we employed receipt of help with mental health services asthe primary independent variable.∧ p < .10, * p < .05, ** p < .01, *** p < .001.Parental StressThe Parenting Stress Index Short Form (PSI/SF) was used to assess mothers’ symptoms of parentingstress. Scores on the PSI/SF can range from 0 to 180, with higher scores indicating greater stress; a scoreat or above the 85 th percentile, defined as a raw score of 86 or higher, is considered indicative of clinicallyChapin Hall at the University of Chicago 108

significant levels of stress. Parenting stress scores for the mothers in the year 5 sample ranged from 36 to147 with a mean of 66.5. Also, 17 percent of the sample had a score at or above 86.Similar to the analysis of maternal depression, we conducted logistic regression analyses on parentingstress in year 5 to examine which variables remain statistically significant after taking into accountpossible correlations among the explanatory variables (see Table 58). Note that because parenting stresswas not measured in year 1, and because we think it is important to account for previous levels of thisvariable, we measured this variable and other characteristics of the mothers at year 2, instead of atbaseline. 59 Accordingly, we looked at the average number of services used and the receipt of help withparenting information at any time during years 3 through 5.In year 4, we found that the average number of services used in previous years was related significantly tothe odds of experiencing high parenting stress, with use of more services associated with higher odds ofincreased stress. We interpreted this result as potentially reflective of greater underlying needs amongmothers with high stress scores. However, in year 5, we do not find a significant relationship between theaverage number of services used and parenting stress (see Table 58). We also do not find a significantrelationship between receipt of help with parenting information and parenting stress, which is consistentwith the year 4 results.Similar to the year 4 results, the odds of a black, foreign-born mother having a clinically significant levelof parenting stress are over 5 times the odds for a Hispanic, foreign-born mother in both models. The factthat these mothers also have significantly higher odds of experiencing depression, as shown in theprevious section, is of concern. Interestingly, we find that U.S.-born Hispanic mothers have significantlylower odds of high stress, relative to foreign-born Hispanic mothers (although this result was not found inyear 4). Finally, as expected, mothers with high levels of parenting stress in year 2 had significantlyincreased odds of having higher levels of parenting stress in year 5.59 The exception is the DCF involvement variable. At the time of this writing, we only had information on DCF involvementthrough the first year of the focal child’s life.Chapin Hall at the University of Chicago 109

Table 58. Logistic Regression of Mothers’ Parenting Stress at Year 5 aPredictor VariableModel 1 b Model 2Odds Ratio Sig. Odds Ratio Sig.Average number of services, years 3–5 0.9 NS — —Received help obtaining parenting information services,years 3–5— — 1.5 NSYear 2 CharacteristicsRace/nativityBlack—U.S.-born 0.4 NS 0.4 NSBlack—foreign-born 5.5 ** 5.6 **Hispanic—U.S.-born 0.1 * 0.1 *Hispanic—foreign-born (excluded category) — — — —White/other 0.5 NS 0.4 NSEducationLess than high school diploma 1.0 NS 1.0 NSHigh school graduate 0.6 NS 0.7 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 2.3 NS 2.4 NSGlades 1.1 NS 1.2 NSNumber of children 1.2 NS 1.2 NSLives with husband/partner 0.9 NS 0.9 NSMother employed 1.0 NS 1.1 NSIncome at or below poverty 1.0 NS 1.0 NSChild with special needs 0.8 NS 0.7 NSNumber of housing problems 1.0 NS 1.0 NSDepression: CES-D score > 16 1.6 NS 1.6 NSParenting stress index ≥ 86 6.5 *** 6.1 ***Family/friend support score 1.0 NS 1.0 NSDCF involvement in child’s first year 1.3 NS 1.4 NSχ 2 47.9*** 48.2***a Data were not weighted in this analysis.b We report results from two different models. In the first, we used average service use during years 3 through 5 as the primaryindependent variable; in the second, we used receipt of help with parenting information services during years 3 through 5.∧ p < .10, * p < .05, ** p < .01, *** p < .001.Parenting PracticesFor this section, we combined the positive and negative parenting scales discussed in chapter 4. Thus,mothers’ parenting practices were measured by the average frequency with which they engaged in 12Chapin Hall at the University of Chicago 110

positive parenting items and 5 negative items; the total score could range from 0 to 3. 60 The (weighted)mean overall parenting score for the year 5 sample was 1.9, with a standard deviation of 0.4.We conducted logistic regression analyses of the likelihood of having a parenting score in the upper 25 thpercentile. As shown in Table 59, similar to the year 4 analysis, we did not find a significant relationshipTable 59. Logistic Regression of Mothers’ Parenting Scores at Year 5 aPredictor VariableModel 1 b Model 2Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 0.8 NS — —Received help obtaining parenting informationservices, years 2–5— — 0.6 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 6.5 *** 5.2 **Black—foreign-born 3.3 * 2.6 ^Hispanic—U.S.-born 1.8 NS 1.7 NSHispanic—foreign-born (excluded category) — — — —White/other 2.6 NS 2.1 NSEducationLess than high school diploma 1.9 NS 2.1 NSHigh school graduate 1.5 NS 1.9 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 0.5 NS 0.5 NSGlades 2.6 ** 2.8 **Number of children 0.7 * 0.7 **Lives with husband/partner 0.6 NS 0.6 NSMother employed 0.6 NS 0.8 NSIncome at or below poverty 0.5 * 0.4 *Child with special needs 1.0 NS 0.9 NSNumber of housing problems 1.2 * 1.2 *Depression: CES-D score > 16 0.6 NS 0.6 ^Overall parenting score 26.0 ** 44.0 **Family/friend support score 1.0 NS 1.0 NSDCF involvement in child’s first year 1.5 NS 1.3 NSχ 2 58.6*** 62.7***a Data were not weighted in this analysis.b We report results from two different models. In the first, we used average service use during years 2 through 5 as the primaryindependent variable; in the second, we used receipt of help with parenting information services during years 2 through 5.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.60 See Table 22for a list of the positive parenting items in the year 5 survey. See Table 26 for a list of the negative parentingitems in the year 5 survey.Chapin Hall at the University of Chicago 111

etween mothers’ parenting scores and their average use of services or receipt of help with parentinginformation in years 2 through 5. Also similar to year 4, we found that black mothers, both foreign- andU.S.-born, were significantly more likely than foreign-born Hispanic mothers to have high parentingscores. Other variables significantly related to parenting scores included residence in the Glades, numberof children, poverty status, number of negative housing conditions, and the mother’s depression score inyear 1. Residence in the Glades and number of housing conditions increased the odds of having a highparenting score, while having an income-to-needs ratio below the poverty level and more childrendecreased the odds. 61Children’s DevelopmentAs discussed above, we asked mothers in the fifth-year interview whether their children had demonstrateda variety of behaviors typical of 4- and 5-year-olds in the areas of language and communication skills,social-emotional competencies, and emergent literacy/pre-academic skills. In this section, we focusspecifically on whether the child is speaking in long, complicated sentences, as well as the child’scommunication skills, 62 social-emotional skills, 63 use of books, 64 emergent literacy/pre-academic skills, 65and drawing/writing skills. 66 Similar to the analyses of maternal outcomes, we conducted logisticregression analyses for each developmental area, employing the average number of services used in years2 through 5 and receipt of help with parenting information at any time during years 2 through 5 as theprimary independent variables. In general, we measured the other explanatory variables at baseline. Theexceptions were child’s age and childcare arrangement (i.e., parent or relative, QIS Head Start orprekindergarten program, non-QIS Head Start or prekindergarten program, family childcare, or61 The result for number of negative housing conditions is counterintuitive and may be a consequence of the fact that povertystatus and number of negative housing conditions are highly correlated.62 Assessment of these skills was based on asking mothers how often the child is a good listener, speaks clearly so strangersunderstand, refers to him- or herself as “I,” uses appropriate social greetings, waits his or her turn to talk, and can say his or herfirst and last name.63 Assessment of these skills was based on questioning how often the child pays attention well, adjusts easily to new situations,or gets upset easily. For a full list of the items, see Table 17 in chapter 3.64 The child’s “score” in this area was based on the extent to which she or he looks at picture books on his or her own, points topictures while looking at books, pretends to read the words in books, reads the words in books, and can tell what is in a givenpicture when looking at a book.65 The child’s “score” in this area was based on the extent to which she or he recognizes and can say the names of colors andletters of the alphabet, as well as how high she or he can count.66 The child’s drawing/writing skills were determined through questions ascertaining whether the child scribbles or draws onpaper; draws pictures of people or objects; can say in words what she or he has drawn; tries to draw shapes, numbers, or letters;or is able to draw recognizable shapes or letters.Chapin Hall at the University of Chicago 112

friend/neighbor/other), which were measured in year 5. 67 We also used mothers’ reports on children’sspecial needs at year 2 because not all were likely to have been identified at baseline.Table 60 through Table 65 present the results of the logistic regression analyses. Similar to the results ofprevious analyses, we found little evidence of a relationship between mothers’ service use and childoutcomes. The evidence we did find suggests that greater service use is associated with a lower likelihoodof the focal child speaking in long sentences (see Table 60) and that having ever received parenting helpis associated with a lower likelihood of the focal child being in the top 25 th percentile for communicationskills (see Table 61). As explained previously, these results are likely a reflection of the fact that motherswith greater needs were more likely to receive services and to use more services on average than motherswith fewer needs. Furthermore, as explained previously, we cannot fully control for the differencesbetween mothers who did and did not receive services. Thus, it is not surprising to find that greaterservice use is associated with worse outcomes.We also found that, for four of the six child outcomes examined, the children of U.S.-born black mothershad significantly better outcomes (based on mother’s self-report) than the children of foreign-bornHispanic mothers. Specifically, the children of U.S.-born black mothers were significantly more likely tobe speaking in long sentences and had significantly greater odds of falling into the top 25 th percentile forschool readiness, social-emotional, and communication skills. The differences were especially large forsocial-emotional and school readiness skills, for which the odds of the children of U.S.-born blackmothers being in the top 25 th percentile were 12 and 7 to 8 times those of the children of foreign-bornHispanic mothers, respectively. These results are similar to those obtained in year 4, when we found thatthe children of U.S.-born black mothers were significantly more likely to be speaking in either short orlong sentences and had significantly greater odds of falling into the top 25 th percentile for all otheroutcomes.Also similar to year 4, we found that the children of foreign-born black mothers had significantly betteroutcomes (based on mother’s self-report) in four of the six developmental areas, when compared to thechildren of foreign-born Hispanic mothers. Finally, we did not find much evidence of differences inoutcomes between the children of U.S.- and foreign-born Hispanic mothers; this result is also similar tothat obtained in year 4. The one area where the children of U.S.-born Hispanic mothers appear to have67 Because only 8 percent of family childcare centers participated in the QIS program (seeTable 41), we did not create a separate variable for these centers; rather, we looked at the combined effect of QIS and non-QISfamily childcare on child outcomes.Chapin Hall at the University of Chicago 113

etter outcomes than the children of foreign-born Hispanic mothers is in social-emotional skills, althoughthis difference is only significant at the 10 percent level.Also consistent with the year 4 results, we found that in several of the regressions, mothers’ parentingscores in year 1 were significantly and positively related to their children’s outcomes in year 5. Forexample, an approximately one-standard-deviation increase (~.40 on a scale of 0–3) in the overallparenting score was associated with a 5- to 7- fold increase in the child’s odds of being in the top 25 thpercentile for communication skills. In addition, a one-standard-deviation increase in the parenting scorewas associated with an approximately 4-fold increase in the odds of a child being in the top 25 th percentilefor school readiness skills. These results suggest that effective interventions targeted at improvingparenting skills around the time of the birth may be able to positively influence children’s development. 68Finally, we found some evidence that children attending QIS center programs had better outcomes insome areas when compared to children cared for by parents, other relatives, friends or neighbors.Specifically, the odds of a child being cared for by parents/relatives being in the top 25 th percentile forcommunication skills were found to be about half those of a child attending a QIS center. 69 Also, the oddsof a child being cared for by parents/relatives being in the top 25 th percentile for drawing/writing skillswere about 40 percent of those of a child attending a QIS center.Similarly, the odds of a child being cared for by a neighbor or friend being in the top 25 th percentile fordrawing/writing skills were about 20 percent of those of a child attending a QIS center. It should be noted,however, that in two instances, we found better outcomes associated with attending a non-QIS center. Inparticular, we found that the odds of a child attending a non-QIS center being in the top 25 th percentile foruse of books were almost 3 times that of a child attending a QIS center. In addition, the odds of a childattending a non-QIS center being in the top 25 th percentile for school readiness skills were twice that of achild attending a QIS center; however, this result is only significant at the 10 percent level.68 Unlike the results for year 4, however, we did not find any significant relationships between mothers’ depression scores in year1 and child outcomes in year 5.69 However, this result was only significant at the 10 percent level.Chapin Hall at the University of Chicago 114

Table 60. Logistic Regression of Child’s Likelihood of Speaking in Long Sentences at Year 5 aPredictor VariableModel 1 b Model 2Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 0.8 * — —Ever received help with parenting information, years 2–5 — — 0.8 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 3.1 * 2.2 ^Black—foreign-born 0.9 NS 1.1 NSHispanic—U.S.-born 1.8 NS 1.8 NSHispanic—foreign-born (excluded category) — — — —White/other 1.3 NS 1.1 NSEducationLess than high school diploma 0.7 NS 0.8 NSHigh school graduate 0.5 NS 0.5 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 1.7 NS 1.6 NSGlades 1.6 NS 1.5 NSNumber of children 1.2 NS 1.1 NSLives with partner 1.1 NS 1.1 NSMother employed 0.5 NS 0.7 NSAt or below poverty line 1.1 NS 0.9 NSFocal child with special needs 0.6 NS 0.6 NSNumber of housing problems 1.1 NS 1.1 NSDepression: CES-D score > 16 1.2 NS 1.1 NSFamily/friend support score 1.0 NS 1.0 NSOverall parenting score 1.9 NS 2.1 NSInvestigated or indicated DCF report 0.1 *** 0.1 ***Childcare arrangements (year 5)Non-QIS center 1.2 NS 1.3 NSFamily childcare 1.0 NS 1.3 NSFriend/neighbor/other 0.7 NS 1.1 NSParent/relatives 0.7 NS 0.9 NSQIS center (excluded category) — — — —Age of focal child at year 5 1.2 * 1.2 *χ 2 40.6* 35.1^a Data were not weighted in this analysis.b We report results from two different models. In the first, we used average service use during years 2 through 4 as the primaryindependent variable; in the second, we used receipt of help with parenting information services during years 2 through 4.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.Chapin Hall at the University of Chicago 115

Table 61. Logistic Regression of Child’s Communication Skills at Year 5 aPredictor Variable Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 0.9 NS — —Ever received help with parenting information, years 2–5 — — 0.5 *Baseline CharacteristicsRace/nativityBlack—U.S.-born 2.1 NS 2.4 ^Black—foreign-born 2.6 ^ 2.9 *Hispanic—U.S-born 0.8 NS 0.9 NSHispanic—foreign-born (excluded category) — — — —White/other 3.0 ^ 3.1 ^EducationLess than high school diploma 0.5 NS 0.5 NSHigh school graduate 0.4 * 0.3 **Post–high school education (excluded category) — — — —Teen mother (at birth of focal child) 0.9 NS 1.0 NSGlades 1.3 NS 1.2 NSNumber of children 0.8 NS 0.7 ^Lives with partner 1.0 NS 1.0 NSMother employed 1.3 NS 1.0 NSAt or below poverty line 1.0 NS 1.0 NSFocal child with special needs 1.2 NS 1.2 NSNumber of housing problems 0.9 NS 1.0 NSDepression: CES-D score > 16 1.0 NS 1.0 NSFamily/friend support score 1.0 NS 1.0 NSOverall parenting score 13.0 * 17.9 **Investigated or indicated DCF report 0.5 NS 0.4 NSChildcare arrangements (year 5)Non-QIS center 1.5 NS 1.5 NSFamily childcare 3.2 NS 3.6 NSFriend/neighbor/other 0.5 NS 0.6 NSParent/relatives 0.6 NS 0.5 ^QIS center (excluded category) — — — —Age of focal child at year 5 0.9 NS 0.9 NSχ 2 43.7** 49.3**a Data were not weighted in this analysis.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.Chapin Hall at the University of Chicago 116

Table 62. Logistic Regression of Child’s Social-Emotional Skills at Year 5 aPredictor Variable Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 0.9 NS — —Ever received help with parenting information, years 2–5 — — 0.8 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 12.0 *** 11.8 ***Black—foreign-born 7.8 ** 7.2 **Hispanic—U.S.-born 3.2 ^ 3.4 ^Hispanic—foreign-born (excluded category) — — — —White/other 11.2 *** 11.4 ***EducationLess than high school diploma 1.6 NS 1.4 NSHigh school graduate 0.9 NS 0.8 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 1.0 NS 1.0 NSGlades 0.8 NS 0.8 NSNumber of children 1.1 NS 1.1 NSLives with partner 1.3 NS 1.2 NSMother employed 0.8 NS 0.7 NSAt or below poverty line 0.5 NS 0.6 NSFocal child with special needs 0.9 NS 0.9 NSNumber of housing problems 0.7 * 0.7 *Depression: CES-D score > 16 0.6 NS 0.6 NSFamily/friend support score 1.0 NS 1.0 NSOverall parenting score 1.7 NS 1.5 NSInvestigated or indicated DCF report 0.5 NS 0.5 NSChildcare arrangements (year 5)Non-QIS center 1.7 NS 1.6 NSFamily childcare 0.7 NS 0.7 NSFriend/neighbor/other 0.7 NS 0.5 NSParent/relatives 0.6 NS 0.6 NS—QIS center (excluded category) — — — —Age of focal child at year 5 0.8 ^ 0.8 ^χ 2 60.0*** 59.3***a Data were not weighted in this analysis.∧p ≤ .10, * p < .05, ** p < .01, *** p < .001.Chapin Hall at the University of Chicago 117

Table 63. Logistic Regression of Child’s Use of Books at Year 5 aPredictor Variable Odds Ratio Sig. Odds Ratio Sig.Average number of services, years 2–5 1.0 NS — —Ever received help with parenting information, years 2–5 — — 1.1 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 1.9 NS 2.1 NSBlack—foreign-born 2.4 NS 3.1 *Hispanic—U.S.-born 2.1 NS 2.3 NSHispanic—foreign-born (excluded category) — — — —White/other 2.0 NS 2.2 NSEducationLess than high school diploma 1.0 NS 0.9 NSHigh school graduate 0.9 NS 0.8 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 0.9 NS 1.0 NSGlades 0.6 NS 0.6 NSNumber of children 0.9 NS 0.9 NSLives with partner 0.7 NS 0.7 NSMother employed 0.6 NS 0.5 NSAt or below poverty line 1.2 NS 1.1 NSFocal child with special needs 0.8 NS 0.8 NSNumber of housing problems 0.9 NS 0.9 NSDepression: CES-D score > 16 0.6 NS 0.6 ^Family/friend support score 1.1 NS 1.1 NSOverall parenting score 1.5 NS 1.5 NSInvestigated or indicated DCF report (years 1–3) 0.7 NS 0.7 NSChildcare arrangements (year 5)Non-QIS center 2.8 * 2.7 *Family childcare 3.5 NS 3.5 NSFriend/neighbor/other 0.4 NS 0.4 NSParent/relatives 1.0 NS 1.0 NSQIS center (excluded category) — — — —Age of focal child at year 5 1.0 NS 1.0 NSχ 2 36.6* 38.5*a Data were not weighted in this analysis.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.Chapin Hall at the University of Chicago 118

Table 64. Logistic Regression of Child’s Emergent Literacy/Pre-Academic Skills at Year 5 aPredictor Variable Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 1.1 NS — —Ever received help with parenting information, years 2–5 — — 0.9 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 7.1 *** 7.6 ***Black—foreign-born 8.8 *** 7.7 ***Hispanic—U.S.-born 2.3 NS 2.5 NSHispanic—foreign-born (excluded category) — — — —White/other 5.2 * 5.5 *EducationLess than high school diploma 0.6 NS 0.5 NSHigh school graduate 0.7 NS 0.7 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 0.5 NS 0.5 NSGlades 3.0 ** 3.1 **Number of children 0.9 NS 0.9 NSLives with partner 1.0 NS 1.0 NSMother employed 2.9 * 2.8 *At or below poverty line 0.3 ** 0.4 *Focal child with special needs 1.4 NS 1.3 NSNumber of housing problems 0.8 * 0.8 *Depression: CES-D score > 16 0.7 NS 0.7 NSFamily/friend support score 1.1 NS 1.0 NSOverall parenting score 9.5 ^ 8.7 ^Investigated or indicated DCF report 0.1 * 0.1 *Childcare arrangements (year 5)Non-QIS center 2.2 ^ 2.0 NSFamily childcare 1.2 NS 1.1 NSFriend/neighbor/other 0.3 NS 0.2 NSParent/Relatives 0.6 NS 0.6 NSQIS center (excluded category) — — — —Age of focal child at year 5 0.9 NS 0.9 NSχ 2 120.6*** 126.3***a Data were not weighted in this analysis.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.Chapin Hall at the University of Chicago 119

Table 65. Logistic Regression of Child’s Drawing/Writing Skills at Year 5 aPredictor Variable Odds Ratio Sig. Odds Ratio Sig.Average number of services used, years 2–5 1.0 NS — —Ever received help with parenting information, years 2–5 — — 1.3 NSBaseline CharacteristicsRace/nativityBlack—U.S.-born 1.6 NS 1.5 NSBlack—foreign-born 0.8 NS 0.8 NSHispanic—U.S.-born 0.9 NS 0.8 NSHispanic—foreign-born (excluded category) — — — —White/other 1.5 NS 1.5 NSEducationLess than high school diploma 0.9 NS 0.9 NSHigh school graduate 1.0 NS 1.0 NSPost–high school education (excluded category) — — — —Teen mother (at birth of focal child) 0.8 NS 0.8 NSGlades 1.0 NS 1.0 NSNumber of children 0.8 NS 0.8 NSLives with partner 0.9 NS 0.8 NSMother employed 1.1 NS 1.1 NSAt or below poverty line 0.5 * 0.5 *Focal child with special needs 1.0 NS 1.0 NSNumber of housing problems 1.0 NS 1.0 NSDepression: CES-D score > 16 1.1 NS 1.1 NSFamily/friend support score 1.1 NS 1.1 NSOverall parenting score 4.8 ^ 4.9 ^Investigated or indicated DCF report 0.6 NS 0.7 NSChildcare arrangements (year 5)Non-QIS center 0.8 NS 0.8 NSFamily childcare 0.6 NS 0.6 NSFriend/neighbor/other 0.2 * 0.2 *Parent/relatives 0.4 ** 0.4 **QIS center (excluded category) — — — —Age of focal child at year 5 1.0 NS 1.0 NSχ 2 44.8** 47.2**a Data were not weighted in this analysis.∧ p ≤ .10, * p < .05, ** p < .01, *** p < .001.Kindergarten ReadinessFollowing the fifth year of the study, we were able to obtain results of the Florida Kindergarten ReadinessScreen (FLKRS) for 324 focal children who entered kindergarten in the School District of Palm BeachChapin Hall at the University of Chicago 120

County in either fall 2009 or fall 2010. Forty-five percent of these children were indicated in the schooldistrict data to have a home language that was not English, and 13 percent were designated as having“exceptionality status.” Most of these were minorities, described as either black (47%) or Hispanic (45%),and 8 percent were white or other races. In contrast, in the cohort of all children who entered kindergartenin the school district in 2009, 28 percent were described as black and 31 percent as Hispanic (CSC 2010).In addition, almost all (96%) of the study children were eligible for the free- or reduced-cost lunchprogram according to the school district data, which is considerably higher than the 57 percent ofkindergartners who were eligible in the county in 2009 (CSC 2010).The FLKRS, the state’s school readiness screening instrument, is administered within the first 30 days ofkindergarten by teachers and covers seven developmental domains—language and literacy, mathematics,social and personal skills, science, social studies, physical health and fitness, and creative arts—that alignwith the state’s Voluntary Pre-Kindergarten Education Standards. The FLKRS includes two sets ofmeasures. The first is a subset of the Early Childhood Observation System (ECHOS), an observationalinstrument that is used to monitor the skills, knowledge, and behaviors a student demonstrates or needs todevelop in each of the seven developmental domains listed above. Children’s total scores on the ECHOSfall into one of three readiness categories—not yet demonstrating, emerging/progressing, or consistentlydemonstrating—with the latter two indicating readiness for school. In 2009, 87 percent of children whowere screened in Palm Beach County were scored as ready on the ECHOS (CSC 2010).The other measure is the Florida Assessments for Instruction in Reading (FAIR), which is also a teacheradministeredscreening of emergent literacy composed of two tasks, Letter Naming and PhonemicAwareness. The scores on these tasks are used to determine the child's Probability of Reading Successscore, which is expressed as a percentage, from 1 percent to 99 percent. This Probability of ReadingSuccess is a statement about the probability of the student scoring at or above grade-level, the 40thpercentile, on the word reading subtest of the Stanford Early School Achievement Test (SESAT) at theend of the year. There are three rankings on the FAIR. Children with a probability of reading successscore at or above 67 percent are considered ready for kindergarten. In 2009, 65 percent of childrenscreened on the FAIR in Palm Beach County were scored as ready (CSC 2010).To analyze the school readiness of the focal children, we used as our sample 320 children who remainedin the study sample in either year 4 or year 5. 70 Of the children who were still part of the study sample inyear 4 or 5, ECHOS data were available for 283 (88%) of the children, and FAIR data were available for293 (90%) children. In other words, not all children were screened on both tests. A total of 271 (85%)70 These were the children whose mothers remained in the study at year 4 or 5 and who we also still had permission to accesstheir administrative records for the study.Chapin Hall at the University of Chicago 121

focal children had scores on both the ECHOS and FAIR screens. When we measured the correlationsbetween scores on the ECHOS and the FAIR assessments for children who were screened on bothassessments, we found that they were significantly correlated (Spearman R 2 =.443, p < .001).As displayed in Figure 6, of the focal children who were screened on the ECHOS, three-fourths (76%)were considered “ready” for kindergarten as determined by a rating of either consistently demonstratingor emerging/progressing. 71 These results are below those reported for all public school kindergartners inPalm Beach County (87%), as well as slightly below those for the subgroup of children who participate inthe free or reduced school lunch program (81%) in 2009 (CSC 2010).Figure 6 also present separate results by the focal children’s sex and race/ethnicity. The differencesbetween girls and boys were very small. However, the results vary based on race and ethnicity. Notably,children identified as black and white/other were much more likely than children identified as Hispanic toscore ready for kindergarten on the ECHOS (81% and 84% vs. 69%). The direction of these differences isconsistent with that of the racial and ethnic disparities for the county as a whole (CSC 2010).Results for the FAIR portion of the FLKRS are presented in Figure 7. Almost 38 percent of the samplechildren were assessed with a probability of reading success score at or above 67 percent, which isconsidered ready for kindergarten. This is considerably lower than the 65 percent of children screened bythe school district in 2009; it is also lower than the 50 percent of children participating in the free orreduced lunch program who were screened (CSC 2010). The pattern of racial and ethnic differences wassimilar to that observed in the ECHOS results.Next, we conducted analyses to determine if service use and characteristics of mothers and children aresignificantly associated with the likelihood of children being ready for school on the ECHOS and on theFAIR. First we looked at the maternal and child characteristics that might be associated with childrenbeing screened as ready for school on the ECHOS. The descriptive data in Table 65 suggest that scoring“ready for school” on the ECHOS is associated with the following baseline characteristics: languagespoken at home, the overall parenting score, and whether or not the mother was living with her husbandor partner. In addition, there appears to be an association with childcare arrangement in year 5, the yearprior to the child’s entry into kindergarten, in that children who did not score “ready for school” weremore likely to have been cared for at home or with a relative or friend than children who did.71 Unlike other descriptive data in this report, these are unweighted results because we used children who remained in the sampleat year 4 and year 5 and it would have been difficult to calculate one weight to apply to all children. When we looked just atchildren in the year 5 sample, the differences between the weighted and unweighted data were very small.Chapin Hall at the University of Chicago 122

Figure 6. Percent of Focal Children Scoring Ready for School on the ECHOS Portion of the FloridaKindergarten Readiness Screen in Fall 2009 or Fall 2010 a60.050.051.7 51.5 51.9 51.554.84840.030.020.023.8 24.5 25.423.124.4 23.719.129.431.014.31636Not YetDemonstratingEmerging/ProgressingConsistentlyDemonstrating10.00.0AllGirlsBoysBlackHispanicWhite/Othera Children scoring either “consistently demonstrating” or “emerging/progressing” are considered ready for kindergarten.Figure 7. Percent of Focal Children Scoring Ready for School on the FAIR Portion of the FloridaKindergarten Readiness Screen in Fall 2009 or Fall 2010 a60.050.040.050.047.437.5 36.945.4 45.641.238.050.043.552.230.031.30-1516-6620.015.013.116.613.218.767+10.04.40.0a Children with a probability of reading success score at or above 67 percent are considered ready for kindergarten.Chapin Hall at the University of Chicago 123

Table 66. Selected Maternal and Child Characteristics by ECHOS Scores aVital Statistics DataRace/ethnicity and NativityTotalNot YetDemonstratingEmerging/Progressing orConsistentlyDemonstratingN=283 n=68 n=215Black, U.S.-born 36 31 37Black, Foreign-born 12 6 14Hispanic, U.S.-born 8 4 9Hispanic, Foreign-born** 40 56 35Other 5 3 5Teen mother (at birth of focal child) 21 25 20First trimester prenatal care 40 34 42Year 1 (Baseline) Survey DataLanguage spoken at homeEnglish 50 40 53Spanish* 41 54 37Other 9 6 10Lives with partner* 61 74 57Less than high school diploma 59 68 56At or below poverty threshold 71 77 69Mother currently employed 14 10 15Depression score ≥ 16 37 35 38Focal child’s age in months (mean, sd) 1.8 (1.1) 1.6 (0.9) 1.9 (1.2)Number of children (mean, sd) 2.0 (1.2) 1.9 (1.3) 2.1 (1.2)Negative housing conditions (mean, sd) 1.3 (1.7) 1.4 (1.8) 1.2 (1.7)Overall Parenting Score (mean, sd)*** 0.78 (.15) 0.71 (.18) 0.80 (.14)Year 1-3 Survey DataNumber of well-child medical visits (mean, sd) 6.4 (2.6) 6.5 (2.7) 6.3 (2.7)Year 1-5 Survey DataNumber of years focal child has health insurance (mean, sd) 4.0 (1.1) 3.8 (1.3) 4.1 (1.1)Year 2 Survey DataFocal child has special needs 21 25 20Year 2-5 Survey DataNumber of services used (mean, sd) 3.6 (1.5) 3.4 (1.3) 3.6 (1.6)Year 5 Survey DataChildcare optionsQIS Center, Head Start, Prekindergarten 21 14 23Non-QIS Center, Head Start, Prekindergarten** 21 9 24Family Childcare 1 0 2Friend, Neighbor or Other* 6 12 4At home with parent or relative* 51 65 47a Analyses were not weighted and were not selected.Notes: With regard to number of well child visits, it should be noted that in Year 1 this item was phrased as "Since focal child was born, howmany times has s/he going for well-baby check-ups?" In years 2 and 3, this item was phrased as "During the last 6 months, how many times hasfocal child gone for well-baby check-ups?"Chapin Hall at the University of Chicago 124

Next, we conducted logistic regression analyses to determine if service use and characteristics of mothersand children were significantly associated with the likelihood of children being ready for school on theECHOS, after controlling for possible correlations between the explanatory variables. Independentvariables again included demographics, maternal depression, and parenting practices at baseline, child’sspecial needs at year 2, average service use for years 2 through 5, and childcare arrangement in year 5.Table 67. Logistic Regression Predicting ECHOS ScoresLog Odds Ratio Sig.Vital Statistics DataRace/EthnicityBlack; U.S.-born -0.54 .58 NSBlack; Foreign-born 0.04 1.04 NSHispanic; U.S.-born -0.72 .49 NSOther 0.23 1.26 NSHispanic; Foreign –born (Excluded category) — — —Teen mother 0.31 1.37 NSReceived prenatal care in first trimester or earlier -0.36 .70 NSYear 1 Survey DataLives with partner -0.76 .47 NSLess than high school diploma -0.02 .98 NSAt or below poverty threshold -0.56 .57 NSMother currently employed 0.06 1.06 NSDepression score 0.27 1.30 NSNumber of children -0.12 .89 NSNegative housing conditions -0.03 .97 NSOverall Parenting Score 4.11 61.13 **Year 1-3 Survey DataNumber of well-child medical visits -0.10 .90 NSYear 1-5 Survey DataNumber of years of health insurance 0.06 1.07 NSYear 2 Survey DataFocal child has special needs -0.34 .71 NSYear 2-5 Survey DataNumber of services used Year 2-5 0.11 1.12 NSYear 5 Survey DataChildcare optionsQIS Center, Head Start, Prekindergarten (Excluded — — —category)Non-QIS Center, Head Start, Prekindergarten 0.32 1.38 NSFamily childcare or Friend, Neighbor or Other -1.01 .36 NSAt home with parent or relative -1.10 .33 *χ 2 (21 , N=233) 34.496R 2 .138*p < .05, **p < .01, or *** p< .001 R 2 is the Cox & Snell R-squareChapin Hall at the University of Chicago 125

As shown in Table 66, the results of this analysis show little relationship between the ECHOS assessmentof children’s school readiness and overall service use. It also shows that when we control for otherbackground characteristics, the racial/ethnic and nativity differences we saw in the descriptive data arenot significant. On the other hand, the results do show an effect of parenting practices in the first year andchildcare arrangement the year prior to kindergarten. Specifically, an increase of one standard deviation(0.15) in the parenting score at year 1 for a mother meant her child was 9 times as likely to score ready forschool on the ECHOS. In addition, children who were cared for at home by a parent in the year beforekindergarten were only a third as likely to be assessed as ready for kindergarten as children in QIS sites.With regard to the FAIR assessment of literacy, we first examined if service use and selected mother andchild characteristics were associated with children being screened with a greater than 67 percentprobability of reading success on the FAIR. The descriptive data in Table 67 suggest that scores on theFAIR assessment are associated with several characteristics measured at baseline, including use ofprenatal care in the first trimester, language spoken at home, mother’s education, employment, povertystatus, and the overall parenting score. In addition, there appears to be an association with childcarearrangement in the year prior to the child’s entry into kindergarten, in that children who were screenedwith less than 67 percent probability of reading success on the FAIR were more likely to have been caredfor at home than children who did.Next, we conducted logistic regression analyses to determine which of these characteristics aresignificantly associated with the likelihood of children’s reading success, after controlling for possiblecorrelations between the explanatory variables. As shown in Table 68, we again found little relationshipbetween the FAIR assessment of children’s probability of reading success and our measure of overallservice use. However, the evidence does suggest that an increase of one standard deviation in theparenting score at year 1 for a mother means her child is more than 4 times as likely to have a score ofgreater than 67 percent likelihood of reading success. Type of childcare arrangement the year prior tokindergarten did not appear to be a significant predictor of FAIR scores. However, mothers who wereemployed in year 1 and who were not living at or below poverty were more likely to have children whoscored in the highest category on the FAIR. In addition, we also observed that black foreign-born motherswere more than 5 times as likely to have children who scored in the highest category than Hispanicforeign-born mothers.A contributing factor to these racial/ethnic differences is likely to be language spoken at home. Thus, weconducted a second regression that included Spanish as the home language as a separate variable. Theseresults are shown in Table 69. The findings were very similar to the results of the first regression in thatChapin Hall at the University of Chicago 126

Table 68. Selected Maternal and Child Characteristics by FAIR ScoresVital Statistics DataRace/ethnicity and NativityTotal Below 67% 67% or higherN=289 n=181 n=108Black, U.S.-born 36 36 36Black, Foreign-born* 11 7 17Hispanic, U.S.-born 8 7 10Hispanic; Foreign-born* 40 45 32Other 5 5 6Teen mother 20 22 19First trimester prenatal care* 40 35 47Year 1 (Baseline) Survey DataLanguage spoken at homeEnglish 50 46 57Spanish* 41 46 32Other 9 7 11Lives with partner 62 64 58Less than high school diploma** 59 65 49At or below poverty threshold*** 71 79 58Mother currently employed** 13 8 20Depression score ≥ 16 36 38 33Focal child’s age in months (mean, sd) 1.8 (1.1) 1.7 (1.1) 1.9 (1.1)Number of children (mean, sd) 2.0 (1.2) 2.0 (1.1) 2.0 (1.3)Negative housing conditions (mean, sd) 1.3 (1.7) 1.4 (1.8) 1.1 (1.6)Overall Parenting Score (mean, sd)** 0.78 (.16) 0.75 (.16) 0.81 (.14)Year 1-3 Survey DataNumber of well-child medical visits (mean, sd) 6.4 (2.7) 6.4 (2.8) 6.3 (2.5)Year 1-5 Survey DataNumber of years of health insurance (mean, sd)* 4.0 (1.1) 3.9 (1.2) 4.2 (1.0)Year 2 Survey DataFocal child has special needs 20 20 21Year 2-5 Survey DataNumber of services used Years 2-5 (mean, sd) 3.6 (1.6) 3.7 (1.6) 3.5 (1.6)Year 5 Survey DataChildcare optionsQIS Center, Head Start, Prekindergarten 21 18 24Non-QIS Center, Head Start, Prekindergarten 20 18 23Family Childcare 2 2 2Friend, Neighbor or Other 6 6 7At home with parent or relative* 51 56 43a Analyses were not weighted and were not selected.Notes: With regard to number of well child visits, it should be noted that in Year 1 this item was phrased as "Since focal childwas born, how many times has s/he going for well-baby check-ups?" In years 2 and 3, this item was phrased as "During the last 6months, how many times has focal child gone for well-baby check-ups?"Chapin Hall at the University of Chicago 127

parenting practices at year 1 were a positive factor in scoring ready for school on the FAIR. The odds ofchildren whose home language is Spanish were lower compared to other children of scoring ready on theFAIR, but the results were not statistically significant. 72Table 69. First Logistic Regression Predicting FAIR ScoresLog Odds Ratio Sig.Vital Statistics DataRace/EthnicityBlack; U.S.-born -0.12 .89 NSBlack; Foreign-born 1.65 5.22 **Hispanic; U.S.-born -0.08 .92 NSOther 0.28 1.32 NSHispanic; Foreign-born (Excluded category) — — —Teen mother 0.31 1.37 NSReceived prenatal care in first trimester or earlier 0.02 1.02 NSYear 1 Survey DataLives with partner -0.19 .83 NSLess than high school diploma -0.08 .93 NSAt or below poverty threshold -1.13 .32 **Mother currently employed 1.11 3.04 *Depression score -0.57 .57 NSNumber of children -0.03 .97 NSNegative housing conditions -0.16 .85 NSOverall Parenting Score 3.00 20.04 **Year 1-3 Survey DataNumber of well-child medical visits -0.06 .94 NSYear 1-5 Survey DataNumber of years of health insurance 0.21 1.23 NSYear 2 Survey DataFocal child has special needs -0.04 .96 NSYear 2-5 Survey DataNumber of services used Year 2-5 0.18 1.20 NSYear 5 Survey DataChildcare optionsQIS Center, Head Start, Prekindergarten (Excluded category) — — —Non-QIS Center, Head Start, Prekindergarten 0.09 1.10 NSFamily childcare or Friend, Neighbor or Other 0.26 1.30 NSAt home with parent or relative -0.05 .95 NSχ 2 (21 , N=236) 46.78972 We also conducted a separate regression analysis of the sample of foreign-born mothers to examine the relationship betweenyears in the U.S. and child's score on the FAIR and found that it was at best only marginally significant with an odds ratio of 0.9.The Year 1 parenting score continued to be highly significant in this regression.Chapin Hall at the University of Chicago 128

R 2 .180*p < .05, **p < .01, or *** p< .001 R 2 is the Cox & Snell R-squareChapin Hall at the University of Chicago 129

Table 70. Second Logistic Regression Predicting FAIR ScoresLog Odds Ratio Sig.Vital Statistics DataHispanic, U.S.-born -0.12 .89 NSTeen mother 0.12 1.13 NSReceived prenatal care in first trimester or earlier 0.02 1.02 NSYear 1 Survey DataSpanish is primary language spoken at home -0.59 .55 NSLives with partner -0.00 1.00 NSLess than high school diploma 0.03 1.03 NSAt or below poverty threshold -1.00 .37 **Mother currently employed 1.10 3.01 *Depression score -0.45 .64 NSFocal child’s age 0.01 1.01 NSNumber of children -0.03 1.00 NSNegative housing conditions -0.15 .86 NSOverall Parenting Score 2.77 16.02 **Year 1-3 Survey DataNumber of well-child medical visits -0.05 .96 NSYear 1-5 Survey DataNumber of years of health insurance 0.11 1.12 NSYear 2 Survey DataFocal child has special needs -0.16 .85 NSYear 2-5 Survey DataNumber of services used Year 2-5 0.06 1.06 NSYear 5 Survey DataChildcare optionsQIS Center, Head Start, Prekindergarten— — —(excluded category)Non-QIS Center, Head Start, Prekindergarten 0.09 1.09 NSFamily childcare, Friend, Neighbor or Other 0.06 1.06 NSAt home with parent or relative -0.22 .80 NSχ 2 (20 , N=236) 40.588R 2 .158*p < .05, ** p < .01, or *** p < .001 R 2 is the Cox & Snell R-squareChapin Hall at the University of Chicago 130

SummarySimilar to previous analyses, we did not find much evidence of a relationship between the mothers’overall (or specific) service use and maternal and child outcomes. The evidence we did find suggests thatmothers with greater needs tend to use more services over time. This implies that those who may havegreater need for services are using them but that, once observable differences in circumstances betweenmothers with high and low need are accounted for, greater use of services overall—at least as measuredby this study—is not associated with improved outcomes.However, consistent with the literature on child development, we did find evidence of the important roleof parenting skills and practices in children’s development and school readiness. That is, mothers’parenting scores in year 1 are significantly and positively related to several of the child outcomes weexamined. These results suggest that effective interventions targeted at improving parenting skills aroundthe time of a child’s birth may be able to positively influence his or her development.A third finding concerns the potential positive impact of QIS center-based care, as opposed to parentaland other informal care, on child outcomes. When we examined possible differences between childrenenrolled in programs participating in Quality Counts (QC) and other formal childcare arrangements, wealso found some evidence that children in non-QC centers have better outcomes related to use of booksand other preliteracy outcomes—as reported by mothers—than those QC centers. However, we do nothave specific information on the characteristics of these childcare arrangements or the attendance andparticipation of the study children in them. These findings should be explored further. 73Finally, as in year 4, we also found that the children of foreign-born Hispanic mothers may be lagging intheir development when compared to the children of black mothers, both foreign- and U.S.-born. It is notclear whether these data from mothers’ self-reports point to real differences in development, differencesin interpretations of survey questions, or both. Teachers’ assessments of children’s development on thestandardized Florida Kindergarten Readiness Screen soon after they entered kindergarten were mixed inthis regard. There were no significant race/ethnicity or nativity differences in teachers’ ratings on theECHOS, a comprehensive child development screen. However, children of foreign-born Hispanicmothers were not assessed as highly on the FAIR, a screen of literacy skills. In addition, children in thestudy sample did not perform as well as other children entering kindergarten in the school district. Thiswas not surprising given their overall higher risk characteristics.73 A recently completed study of the school readiness rates of children attending programs in the Quality Counts system suggestspositive effects of the QC (Shen, Tackett, Ma 2009).Chapin Hall at the University of Chicago 131

Summary and ConclusionsThis is the fifth and final report of a longitudinal study examining the use and effects of a comprehensivesystem of prevention and early intervention services in Palm Beach County and its effects on children andfamilies living in four targeted geographic areas (TGAs) with high rates of poverty, teen pregnancy,crime, and child abuse and neglect. Parenting in any circumstance is challenging, but low- mothers faceadditional hurdles—fewer economic and social resources, limited education, unstable livingarrangements, transportation difficulties, etc.—that place them and their children at risk of pooroutcomes. By strengthening the services offered to families in the TGA communities, CSC seeks toincrease the number of healthy births, reduce the incidence of child abuse and neglect, and enhanceparents’ abilities to raise socially, emotionally, and physically healthy children who are eager to learn andready for school.The primary goal of this study was to better understand families’ patterns of service use over time andtheir impact on child and family outcomes. Although the main concern of the study is the system ofprevention and early intervention services funded by CSC, the study also collected data on the use ofother services and community supports. Data sources included administrative records on Palm BeachCounty families who gave birth between 2004 and 2009; data from annual in-person household surveyswith a baseline sample of 531 mothers drawn from the 2004–2005 TGA birth cohort; and qualitative datafrom interviews with a subgroup of 40 mothers randomly selected from this larger survey sample. Thisreport covers information from 5 years of the study and focuses on findings for 310 mothers whoparticipated in all five of the annual surveys. Below we summarize the key findings for this sample ofmothers.Family CharacteristicsNewly delivered mothers in the TGAs of Palm Beach County in 2004–2005, from which the study samplewas drawn, are more likely than mothers outside the TGAs to be unmarried, to be a teen mother, to haveChapin Hall at the University of Chicago 132

less than a high school education, to be black or Hispanic, to be foreign born, and to have used WIC whilepregnant. Just over half of the study sample identified themselves as Hispanic, and two-fifths, as black orAfrican-American. More than half (55%) are foreign born. At the beginning of the study, 13 percent ofthe sample families lived in the Glades TGA, and 87 percent lived in the other three TGAs. Over the 5years of the study, there was a modest increase in the percentage of study families living outside theTGAs (13% in year 5), with most of this movement coming from families in the non-Glades TGAs. Thus,at year 5, 12 percent lived in the Glades TGA, 75 percent lived in the non-Glades TGAs, and 13 percentlived in other, nearby areas of Palm Beach County. As in the previous year, one-third of the familiesreported moving at least once during the year.Although household compositions changed during the course of the study, the number of reported familymembers remained fairly stable. The percentage of mothers who reported they were married in the fifthyear (32%) is almost the same as in the fourth year, although the percentage of unmarried mothers livingwith a partner (31%) continued to decline from the first year, when it was 39 percent. At the same time,the percentage of mothers with two or more children increased, with three-quarters of the sample havingtwo or more children at the time of the fifth interview. More than 4 in 10 mothers had had another childsince the birth of their focal child, and 6 percent were pregnant at the time of the year 5 interview.Overall, almost half of the mothers in the year 5 sample had worked at some point since the beginning ofthe study, and, similar to year 4, nearly half (46%) were working at year 5. The mothers’ estimates oftheir family income were slightly better in year 5, with 54 percent reporting household incomes above$20,000 for the previous year compared to 46 percent in year 4. Calculation of an income-to-need ratiobased on household size, the number of children under age 18, and the federal poverty thresholdsindicates that 52 percent of the families in the year 5 sample were living at or below the federal povertythreshold the previous year.Home ownership changed only slightly from year 4 to year 5, with 27 percent of mothers reporting homeownership in year 4 and 29 percent in year 5. The same percentage (16%) of mothers who reported livingin an unsafe neighborhood in year 4 did so in year 5. However, some evidence of a decline in livingconditions is reflected in an increase in the percentage of mothers who reported one or more negativehousing conditions, such as electrical or plumbing problems, from year 4 (30%) to year 5 (40%).Health, Healthcare, and Child DevelopmentIndicators of maternal health and functioning examined in this study include a measure of depression,mothers’ self-reported health, a measure of parental stress, and administrative data on reports of childabuse and neglect. The percentage of mothers reporting depressive symptoms was fairly stable from yearChapin Hall at the University of Chicago 133

2 to year 5, when 21 percent scored above the normal range on the CES-D, and significantly lowerrelative to the first year. The percentage of mothers with elevated parental stress scores—17 percent inyear 5—was also stable from years 2 through 5 (but not measured in year 1). Finally, about four-fifths ofmothers described their general health as “good” to “excellent” in all 5 years.The results for children’s health and development are generally positive. Most mothers reported theirfocal child to be in “good” to “excellent” physical health, although 19 percent of the mothers reported thattheir focal child has “special needs”—most often asthma and other respiratory problems. In terms ofdevelopment, a majority of the children demonstrate a range of early literacy, social-emotional skills, andcommunicative behaviors typical for their ages. For example, more than two-thirds of the children werereported to be talking in sentences, and 90 percent or more demonstrating appropriate listening andspeaking skills when conversing with other people. However, foreign-born mothers were more likely thanU.S.-born mothers to report their focal child has some social-emotional problem behaviors (such as theinability to sit still), which may be attributed, in part, to cultural differences in their expectations for thebehavior of preschool children. In addition, foreign-born mothers were less likely than U.S.-born mothersto report that their child demonstrates certain preliteracy behaviors. Such differences might be the resultof differences in opportunities to develop and use these skills.Another important health indicator is access to healthcare. A majority (77%) of the mothers reportedreceiving regular medical care for themselves at year 5. However, this still indicates that, as in theprevious year, almost a quarter of the sample mothers were going without routine healthcare, makingthem less likely to receive services to maintain interconceptional health. With regard to care duringpregnancy, mothers’ use of prenatal care for children born after their focal child followed the samepattern observed with their focal child; that is, about three-quarters said they had initiated care in the firsttrimester, while 16 percent reported doing so in the second trimester.Although three-fourths (76%) of the focal children were covered by health insurance in year 5, it is aconcern that 24 percent of the focal children (and 13 percent of all children in the study families) were notcovered. An additional concern is that only 45 percent of the sample mothers reported having healthinsurance for themselves, a percentage that was fairly stable from years 2 through 5 and much lower thanthe percentage in year 1 when mothers were likely to have Medicaid coverage around the birth of theirfocal child. There also were marked differences between foreign-born and U.S.-born mothers with respectto health insurance coverage; 75 percent of U.S.-born mothers had health insurance, compared to only 19percent of foreign-born mothers. In addition, the focal children of foreign-born mothers were less likelythan the children of U.S.-born children to have health insurance, although the difference was muchsmaller (81% vs. 90%).Chapin Hall at the University of Chicago 134

Parenting PracticesWhen surveyed in the fifth year, mothers were asked to report on three general kinds of parentingpractices: positive activities with their children in their homes and neighborhoods; negative parentingpractices, usually for disciplinary purposes; and activities that involved them with their children’s schoolor childcare. More than three-quarters of the mothers reported that they engaged in a variety of positiveparenting activities. These included praising their children; taking children on errands and outside to play;reading books and encouraging their children to read; singing songs with children and telling stories;doing household chores with children; playing with a game, puzzle, or building toy with children; andtalking to their children about television programs. For families in which husbands/partners had contactwith their children, mothers reported that at least half of husbands/partners engaged in most positiveparenting activities.Smaller percentages of mothers reported that they or their husbands/partners used negative parentingpractices. A little over half (60%) of the mothers reported that they had lost their temper with theirchildren; a quarter (25%) said they had found hitting or spanking was a good way to get their children tolisten; and a quarter (26%) said they got angrier with their children than they had intended during theprevious 3 months. No more than 18 percent said that in the past 3 months they had blamed their childrenfor something that was not their fault or punished their children for not finishing the food on their plate.Mothers whose husbands/partners had had contact with their children reported somewhat lowerpercentages of negative parenting practices for their husbands/partners than they reported for themselves.Over time, positive parenting activities increased for both mothers and husbands/partners, but most of thechanges occurred between the first and second years. The only noteworthy differences between years 4and 5 was an increase in the reported percentage of mothers and husbands/partners who told their childrenstories and took them to the library, which is consistent with the increasing age of their children. In termsof negative parenting practices, there was a significant increase in mean negative parenting scores formothers, but not those reported for husbands/partners between years 4 and 5. More specifically, moremothers said they had become angrier than intended with their children or had blamed their children forsomething that was not their fault in year 5.The most frequent parent involvement activity, among mothers with a child in school or childcare, wastalking with their child’s teachers, which was reported by three-quarters (75%) of the mothers. Almostthree-quarters of the mothers said they had helped their child with homework, and over half had attendeda parent-teacher conference during the previous 3 months; less than half reported that they hadparticipated in a field trip, and one-third reported attending a PTA or other parent meeting. The reportedChapin Hall at the University of Chicago 135

percentage of husbands/partners who engaged in any of these activities was smaller than the percentage ofmothers.We also found associations between parenting practices and a number of maternal/family characteristics,including educational background, race/ethnicity, relationship status and nativity, although it is importantto keep in mind that these characteristics are highly correlated.Childcare ArrangementsMore than half (61%) of the mothers in the fifth year of the study were using nonparental carearrangements for their focal child, motivated largely by their needs to work or attend school. Thus,maternal employment and/or school enrollment continued to be the strongest predictor of childcare use;the only other significant factors predicting childcare use were residence in the Glades TGA andrace/nativity. In particular, both U.S.-born and foreign-born black mothers were significantly more likelythan foreign-born Hispanic mothers to use childcare.The most frequently reported type of nonparental childcare arrangement for the focal children was centercare, followed by relative care, and care by a friend or neighbor. Different types of childcarearrangements were associated with a number of maternal characteristics, including race/ethnicity, nativity,education, employment, and income. Among black mothers, about two-thirds (65%) used center care, and14 percent used relative care. Among Hispanic mothers, almost half (46%) used center care, and one-fifth(20%) relied on a friend or neighbor to care for their children. Furthermore, foreign-born mothers weremore likely than U.S.-born mothers to use friends or neighbors to care for their children, while U.S.-bornmothers were more likely to rely on relatives. However, equal proportions of U.S.-born and foreign-bornmothers used center care (55%), which is a change from previous years when U.S.-born mothers weremore likely than foreign-born mothers to use center care.Cost was an important factor in families’ choice of childcare arrangements. Mothers who received asubsidy were much more likely to have their child in a childcare center, Head Start, or prekindergartenprogram (82%) than were mothers who did not receive a subsidy (42%). Over one-third (39%) of mothersnot using a subsidy relied on friends or relatives to care for their children. Other important factors inchoice of childcare type, suggested by the qualitative data, appear to be the greater availability of formalchildcare options for children 3 years and older than for infants and toddlers and, correspondingly, themothers’ growing knowledge about the options in their communities. In addition, as children becomemore communicative, more self-sufficient, and more in control of their behavior with age, mothersbecome more comfortable with the idea of nonparental childcare and begin to recognize the importance ofexperiences with other children and adults for their children’s cognitive and social development.Chapin Hall at the University of Chicago 136

We also looked at the stability of childcare. A majority of the focal children experienced fairly stablechildcare arrangements during the first 4 years of their lives, experiencing only a few transitions in typesof childcare. However, about 44 percent of the children experienced four or more transitions in types ofcare arrangements during the first 4 years of their lives. Instability in stable childcare arrangement wasrelated to the instability of a family’s work circumstances and income; at the same time, it is not clearwhat impact this instability might have on children’s development.Social SupportThroughout the 5 years of the study, mothers reported satisfaction with the level of informal support theyreceive from family members and friends. Consistent with previous years, if the mothers hadhusbands/partners, they most often received support from them; moreover, they reported a higher level ofsupport from husbands/partners, on average, in year 5 than in year 4. Otherwise, the mothers reliedprimarily on their family, especially siblings and mothers or stepmothers, and secondly, on friends forsupport. The mean support score based on the mothers’ reports of help received from family and friendsin the fifth year is significantly higher compared to the fourth year, on average; and the fifth-year score ishigher than it had been in the first year.Mothers’ perceptions of their access to social support, regardless of source, fluctuated over time, with nosignificant patterns. The frequency with which husbands/partners, other family members, and friendsprovided different kinds of support also varied over time. For example, almost twice as many mothers(50%) in year 5 reported receiving help with money, food, or clothing than did in the previous years ofthe study. On the other hand, mothers were much less likely to receive support in the form of advice onchildren or household problems in year 5 than in the previous 3 years of the study. The qualitative datasuggest a variety of reasons for these trends (and lack thereof). Sometimes relatives were no longer ableto help with childcare or turned out to be unreliable caregivers, so mothers turned to other sources of help.As their children grew, some mothers also conveyed more confidence in their parenting abilities and anincreasing desire to be out on their own. Although relatives remained an important source of informationand support, sample mothers also appeared to be relying more on doctors, teachers, and other nonfamilyfor information and support than they did when their children were younger.Service UseAs previously reported, 80 percent of the mothers in the study sample received services from the MCHPsystem around the birth of their focal child. Most mothers received these services during the 3 monthsprior to and 6 months after their baby’s birth. A little more than a fourth of the mothers who receivedintensive care coordination (61% of the sample) received services up to 9 months after their baby’s birth.Chapin Hall at the University of Chicago 137

We found some evidence that mothers with greater needs—for example, mothers who were teens at thebirth of their focal child, mothers living in poverty, and those with a focal child with special needs—weremore likely to receive services. Although smoking during pregnancy was associated with greaterlikelihood of receiving services in the county sample, it was not in the study sample. Also, with respect tothe study sample, we found that mothers with physical or mental health problems or depressive symptomswere just as likely to receive services as not. Finally, we found unexplained racial/ethnic and nativitydifferences related to the likelihood of receiving services in both the county cohort and the study sample.In addition, among mothers who received services, those who received more days of services may havehad greater needs. Specifically, mothers with more children, one or more children with special needs, anddepressive symptoms or a physical/mental health problem received more days of services. At the sametime, some mothers identified as needing services on a postnatal risk screen did not receive them. Amajority of these mothers could not be located, or MCHP staff members were unable to complete theinitial assessment. This suggests that making additional effort to locate these mothers and/or simplifyingthe assessment process might have resulted in more “at-risk” mothers receiving services.With regard to use of other formal services, most (94%) of the year 5 sample reported receiving help inmeeting basic family needs. On average, mothers received help for basic family needs in three differentservice areas, slightly more than the average number used in previous years. As in previous years,mothers in the fifth year reported receiving help most often with healthcare. Other areas of frequentassistance were dental care and childcare. Less than a third of the sample received help in any otherservice area, with family planning, paying rent or bills, and housing the next most frequent areas in whichmothers reported receiving assistance.Mothers tended to seek help at different rates depending on the area of service need. Mothers were morelikely to seek help when they had concerns in the areas of healthcare, childcare, and getting enough food.They were next most likely to seek help with concerns related to legal issues and housing. Thepercentages of mothers who had concerns about their children’s development and sought help weregenerally smaller than the percentages who sought help for basic needs.Because one goal of the service system is to link mothers to other needed services, we also examined therelationship between participation in MCHP and use of other formal services. We found some evidencethat mothers who received MCHP treatment services also received more assistance from other formalservices in years 2, 3, and 5 (though not in year 4).Chapin Hall at the University of Chicago 138

Service Use, Maternal Functioning, and Child DevelopmentSimilar to analyses conducted in year 4, we found little evidence of a relationship between the mothers’overall service use and maternal and child outcomes, or between specific types of services (e.g., parenteducation) and child outcomes. Instead, the evidence suggests that mothers with greater needs tend to usemore services over time. This implies that those who may have greater need for services are using thembut that, once observable differences in circumstances between mothers with high and low needs areaccounted for, greater use of services overall—at least as measured by this study—is not associated withimproved outcomes.However, consistent with the literature on child development, there is evidence of the important role ofparenting practices in children’s development. That is, mothers’ parenting scores in year 1 weresignificantly and positively related to several of the child outcomes we examined. These results suggestthat effective interventions targeted at improving parenting skills around the time of a child’s birth mightpositively influence his or her development.A third finding concerns the potential positive impact of center-based care and other formal childcarearrangements—as opposed to parental and other informal care—in the year prior to kindergarten entry onmothers’ reports of child development. When we looked at possible differences between children enrolledin programs that were part of Quality Counts (QC) and other formal arrangements, we found evidence ofbetter outcomes, as reported by mothers, in some areas of preliteracy development among children in thenon-QC centers. However, we do not know enough about the characteristics of these childcarearrangements or about the attendance and participation of the study children in these childcare settings tointerpret these results. These findings should be explored further. 74Also consistent with the year 4 results, we found that the children of foreign-born Hispanic mothers maybe lagging in their development, when compared to the children of black mothers, both foreign- and U.S.-born. Although it is not clear whether these data, which come from mothers’ self-reports, reflect realdifferences or differences in interpretations of survey questions and children’s development, these resultssuggest that interventions tailored specifically to the needs of these mothers and their children may bewarranted.Controlling for maternal education, employment, poverty status, year 1 parenting score, and childcarearrangements, Hispanic foreign-born mothers were less likely than mothers of other race/ethnic/immigrant groups to report their children to be in the top 25 percent on measures of pre-academic74 A recently completed study of the school readiness rates of children attending programs in the Quality Counts system suggestspositive effects of the QC (Shen, Tackett, Ma 2009).Chapin Hall at the University of Chicago 139

and social-emotional development. Maternal education, employment, poverty, year 1 parenting score andchildcare arrangements were also statistically significant. It is not clear whether these data from mothers’self-reports point to real differences in development, differences in interpretations of survey questions, orboth. Teachers’ assessments of children’s development on the standardized Florida KindergartenReadiness Screen soon after they entered kindergarten were mixed in this regard. There were nosignificant race/ethnicity or nativity differences in teachers’ ratings on the ECHOS, a comprehensivechild development screen. However, children of foreign-born Hispanic mothers were not assessed ashighly on the FAIR, a screen of literacy skills. In addition, children in the study sample did not perform aswell as other children entering kindergarten in the school district. This was not surprising given theiroverall higher risk characteristics.Study LimitationsThis study was designed to look at the effects of services funded by CSC in a broader ecological contextthat includes the multiple service systems that low-income families are likely to have to navigate. Thus, itprovides a wide-ranging look at service use (and service nonuse) in a diverse group of low-incomefamilies, including publicly funded healthcare and income supplements as well as a network of smaller,voluntary prevention and early intervention services. It also provides extensive information about otherfactors that may affect family functioning and children’s development. The findings are strengthened bythe use of multiple data sources and mixed methods, involving the ability to link individual-leveladministrative data across service systems and the opportunity to look in depth at the service experiencesof a small sample of the study families.At the same time, like any research study, this one has limitations. The various data sources used in thisstudy, including mothers’ self-reports on service use and service records in administrative data systems,are not perfect. We attempted to limit inaccuracies in any of our sources by not relying on a single sourceof data and integrating data from different sources. Another limitation is that this study isnonexperimental (i.e., families were not randomly assigned to services); thus, we cannot definitivelyattribute differences in family and child outcomes to services. On the other hand, the comprehensive andlongitudinal design of the study allows us to describe and suggest explanations for relationships betweenfamily characteristics, service use, and outcomes. In addition, it is important to note that our findings maynot generalize to other low-income families outside the targeted communities that are the focus of thisstudy.Chapin Hall at the University of Chicago 140

Conclusions and ImplicationsMore than four-fifths of our sample received services from the MCHP around the birth of the focal child,and most stopped receiving these services within the first 6 to 9 months after giving birth. Most familiesalso received help with healthcare and food assistance throughout the study, although use of these andother services tended to decline over the first 3 years of the study before increasing slightly in the last 2years. In some cases, changes in family circumstances changed eligibility for public supports. In others,families perceived less need for services or struggled to complete reapplication processes to maintain theirservices. This was reflected in our finding that more intensive and longer participation in the MCHPsystem was associated with greater use of other services in the second year of the study, but not in thefollowing year. One reason, according to mothers’ reports, was that MCHP providers served an essentialbridging or “brokering” role between parents and basic services, including Medicaid, food assistance, andchildcare subsidies. This support was particularly important for mothers with poorer language and literacyskills, fewer personal resources, and low social support. Service use was also impacted by the accuracy ofinformation about service eligibility and cost and provider responsiveness.One of the clearest findings was the disadvantaged status of children born to foreign-born mothersrelative to those born to U.S.-born mothers. At the end of the study, as the focal children were gettingready for kindergarten, children of foreign-born mothers were more likely to be living at or below thepoverty level, even though their caregivers were more likely to be married or living together. In addition,some groups of foreign-born mothers reported significantly higher levels of depression and parentingstress, and others exhibited significantly lower levels of parenting skills. Most importantly, based onmothers’ and teachers’ assessments, the children of foreign-born Hispanic mothers appeared to be laggingin at least some areas of their development when compared to the children of other mothers.Given their relative disadvantaged status, it is encouraging to note that foreign-born mothers, and foreignbornHispanic mothers in particular, were more likely to receive treatment services from the MCHPsystem. On the other hand, we did not find evidence that receipt of services or services described asparent education in the first year after a child’s birth—to the extent they could be measured here—had asignificant impact on later maternal or child outcomes,. It would be useful to follow up this finding with amore in-depth analysis of outcomes relative to levels of participation in particular types of MCHPservices.In addition, it would be beneficial to examine how services offered by the system could be made moreeffective for all mothers, and especially for foreign-born mothers, and their children. For example, ourfinding of lower parenting scores among foreign-born Hispanic mothers suggests CSC might review theappropriateness of the parenting education services offered to this group of mothers and explore otherChapin Hall at the University of Chicago 141

ways to engage them in parenting education for longer periods of time. In addition, results showinggreater depressive symptoms and parenting stress among foreign-born black mothers suggest a need toimprove screening for depression and ensure mothers are connected with appropriate counseling andother services. Finally, it is also important to note that MCHP services could be impacting outcomes thatwere not measured by this study, and it would be useful to consider what those outcomes might be.Consistent with other research, this study also highlighted the general challenges of engaging families involuntary programs. Given that the demographic characteristics of families living in the TGAs are theones associated with children’s poor developmental and educational outcomes, CSC’s strategy ofidentifying families who need services and targeting services to families in high-poverty areas remains asound one for reaching children most at-risk of not succeeding in school. However, to benefit fromservices, families must use them. There was wide variability in service use among the low-incomefamilies in this study, and many families who were eligible for and might have benefited from CSCfundedand other services were either not receiving them or not using them enough to obtain intendedbenefits.Although foreign-born mothers were more likely to receive MCHP services, overall they used fewerservices outside of the MCHP system compared to U.S.-born mothers. Given that the MCHP system wassuccessful in engaging foreign-born mothers, there may be ways that CSC can positively impact howother publicly-funded services are provided to eligible foreign-born families who are harder to reach withservices. Raising public awareness of families’ literacy, educational, and social needs—as well as theirservice needs—could be helpful. CSC might also consider ways to share its knowledge of andexperiences training service providers in culturally appropriate approaches more widely.With regard to the steep decline in the use of formal assistance with parenting information after the firstyear, one reason might have been that mothers no longer perceived a need for these services—perhapsbecause they feel more confident in their parenting skills; because they have other sources, includingfamily members, pediatricians, and early care and education providers; or because other concerns aremore pressing. But given that each new child developmental stage brings with it its own challenges forparents, it also could reflect a lack of knowledge about and connections to home-based or center-basedservices for parents of older children once they leave the MCHP system.Our ability to interpret changes in service use over time through quantitative data is limited becausealthough mothers might have received referrals to additional services within or outside the system, wecould not ascertain with current data if they actually followed up and got connected to these services. Norcould we determine, if a connection was made, for what duration or intensity they received the services.Chapin Hall at the University of Chicago 142

Meanwhile, our findings indicate that other strategies might be needed to reach these families, forexample, through other services providers, such as WIC, and community-based organizations.In conclusion, this study has provided a wealth of information about the needs of a sample of low-incomefamilies in Palm Beach County, their sources of information about services, their service experiences, thepotential effects of services, and the other factors that affect family functioning and children’sdevelopment. Additionally, we learned about some of the reasons for service disparities among families inthe study. The findings suggest not only the need but also the challenges of designing effective andflexible services and service delivery to fit the diverse needs and circumstances of these families. Thevariations in service use over time also imply the importance of collecting and analyzing data on serviceuse on an ongoing basis to monitor and improve the service system.Study findings make clear that the emerging system of care in Palm Beach County is successfullyengaging many at-risk families in needed services through the MCHP around the birth of a child. Itappears, moreover, unlike other formal services used by the study families, the MCHP has moreflexibility to adapt to the diverse circumstances and daily routines of the families they serve. At the sametime, there were challenges in keeping some of the mothers engaged, identifying and addressing newservice needs, and monitoring service use over time during their children’s early years. These findingssuggest there are opportunities to improve service access and use in the TGAs in general and strengthenthe CSC system of prevention and early intervention services in particular. Thus, we end with thefollowing recommendations:• Improve the quality and effectiveness of parenting supports and education. One of the mostconsistent findings in this study from the first year was the role of parenting practices in children’sdevelopment and school readiness. This suggests a need to better understand early parenting practicesand increase the availability and quality of interventions designed to improve them. Mothers withlower educational backgrounds or literacy skills, especially those for whom English is a secondlanguage, might need additional support to strengthen their ability to prepare their children for schoolduring the preschool years.• Improve access to and quality of early care and education. As suggested by other research,another important factor in children’s school readiness outcomes was the type of childcarearrangement in the year before they entered kindergarten. Children who were at home with theirparents were less likely than children who attended a QIS center-based program to be screened as“ready” for kindergarten on the state standardized readiness screen. At the same time, there was astrong relationship between having a childcare subsidy and enrollment in a center-based program.Qualitative data indicate that many mothers, regardless of race/ethnicity or nativity, were interested insome kind of out-of-home educational experience for their children, but could not afford to enrolltheir child without assistance. In addition to increasing access, it also will be important to ensure thatChapin Hall at the University of Chicago 143

early care and education programs—both family childcare as well as center programs—address theparticular needs of children from low-income but especially language-minority backgrounds.• Help families stay involved in or become re-connected to services over time as needs change.Declines in service use occur for multiple reasons, including parents’ perceptions of need, access toother resources, or actual improvements in their circumstances. But given that each newdevelopmental stage brings with it its own challenges for parents, it also may reflect the lack ofconnections to other kinds of available services and supports once they leave the MCHP. Because ofdata limitations, it was not possible to easily track individual families’ engagement in different typesof services over time in this study, but it would be useful to examine this hypothesis further.Indeed, both survey and qualitative data in this study suggest there are certain times in children’searly development when it is easier or harder to engage families in services, for example, before andafter the birth of a child—especially if a first child and if informal supports are not available. On theother hand, as children approached their first birthday, mothers expressed more confidence in theirparenting and a desire to be independent of family or community supports. Then, as childrenapproached 2 and were expressing more autonomy and developing language, parents seemed to havenew questions about whether their children’s behavior and development were on track. At the age of3, as children’s language and self-care skills improve, parents became more open to consideringpreschool or home-based educational programs. These observations of “touch points” when parentsare more receptive to services suggests the need to maintain some kind of ongoing engagement withthem.• Enhance training of service providers. Another strategy for keeping families engaged in services isto improve the knowledge and responsiveness of service providers by enhancing training in culturallyappropriate and family-strengths-based approaches, as well as special needs of families. CSC mightnot be able to directly impact service delivery in public and other agencies not funded by CSC, butraising the public’s awareness of the literacy and educational needs of families, in addition to theirservice needs, in the targeted communities might help. Families can be intimidated by programconcepts and requirements, and staff who are trained to help families through the process in arespectful way could reduce future duplication of paperwork and client and staff frustration, as wellas make families feel more positive about seeking and accepting help earlier.• Making location and timing of services convenient for families. Of the many factors that constrainservice use, the locations of program offices, their hours, and excessive waiting times pose significantbarriers for families, especially if they have transportation or childcare problems. Strategies that CSCfundedprograms use, such as home visits and traveling service vans, are good alternatives to officevisits, especially if they are available during evening and weekend hours. Basing services at schools,Beacon Centers, or childcare centers is another option for reaching families who have childrenenrolled in school or formal childcare. As it may be difficult to persuade employers to allow familiestime off for appointments with teachers, doctors, or service agencies without jeopardizing theirwages, it may be more feasible to persuade healthcare providers, schools, and service agencies toexpand locations and hours of services to make them more convenient for families.• Improve channels of information and communication about services. During the time of thisstudy, CSC has expanded use of other vehicles (such as radio, television, faith-based organizations,and public libraries) to disseminate information that will reach families with limited education orChapin Hall at the University of Chicago 144

literacy skills, families who do not receive information through family or friends, and families whoare not already using other services like childcare. Healthcare providers might be engaged moreeffectively in providing information to families. The local offices of federal benefit programs are alsochannels for disseminating information about CSC-funded programs; for example, one of the studymothers was referred by a nurse in the WIC office to a provider in the MCHP system.• Strengthen relationships between the CSC system of care and other community supports andservices. Improving the quality of childcare and providing referrals through childcare programs is away to reach families who use these services. However, this approach will not reach many motherswho are not working, who are either not eligible or on a waiting list for a childcare subsidy, or whoprefer other childcare settings. Other strategies are needed to reach these families, for example,through WIC, public health clinics, and community outreach. Our finding of an increase in reportedlevels of community support, especially by medical and childcare providers in the third year, alsosuggests the importance of improving the knowledge of these professionals about parenting andparenting information and supports in the community as well as their ability to assess service needs.• Improve data systems and other sources of information on service availability, use (duration,intensity), and need. This study was designed to look at the effects of services funded by CSC in abroader ecological context that includes the multiple service systems that low-income families arelikely to navigate. It provides a wide-ranging look at service use (and service nonuse) in a diversegroup of low-income families, including publicly funded healthcare and income supplements as wellas a network of smaller, voluntary prevention and early intervention services. Our findings arestrengthened by the use of multiple data sources and mixed methods, involving the ability to linkindividual-level administrative data across service systems and the opportunity to look in depth at theservice experiences of a small sample of the study families.At the same time, the data sources used in this study, including mothers’ self-reports on service useand service records in administrative data systems, had limitations. The database for the MCHP wasan important source of information on the types of services families received in the system and thekinds of referrals made to providers outside the system for this study. However, there is a need foradditional longitudinal data and data system capabilities to understand how families enter and leavethe system over time as their families grow. Additional sources of information on the location ofservices, community needs for services, and the outcomes of referrals would assist funders andproviders of services in planning and monitoring the systems’ ability to ensure families follow up andget connected to the services they need.Chapin Hall at the University of Chicago 145

Chapin Hall at the University of Chicago 146

BibliographyAbidin, R. R. (1995). Parenting stress index (3rd ed.). Lutz, FL: Psychological Assessment Resources,Inc.Akinbami, L. (2006). The state of childhood asthma, United States, 1980–2005. Advance Data, 381, 1–24.American Academy of Pediatric Dentistry. According to the American Academy of Pediatric Dentistry(http://www.aapd.com), children should first visit a dentist when they get their first tooth or nolater than their first birthday. Pediatricians and childcare providers might also encourage mothers’use of dental care. Florida participates in a “Head Start Dental Home Initiative,” which mightinfluence families whose children are enrolled in Head Start and Early Head Start.Andreassen, C., & Fletcher, P. (2007). Early childhood longitudinal study, birth cohort (ECLS–B)psychometric report for the 2-year data collection (NCES 2007–084). National Center forEducation Statistics, Institute of Education Sciences, U.S. Department of Education. Washington,DC. Retrieved June 12, 2008, from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2007084Antle, B. F., Frey, A., Barbee, A., Frey, S., Grisham-Brown, J., & Cox, M. (2008). Child care subsidy andprogram quality revisited. Early Education & Development, 19, 560–573.Attree, P. (2005). Parenting support in the context of poverty: A meta-synthesis of the qualitativeevidence. Health and Social Care in the Community, 13, 330–337.Balaji, A. B, Claussen, A. H., Smith, D. C., Visser, S. N., Morales, M. J., & Perou, R. (2007). Socialsupport networks and maternal mental health and well-being. Journal of Women’s Health, 16,1386–1396.Beals, D., De Temple, J., & Dickinson, D. (1994). Talking and listening that support early literacydevelopment of children from low-income families. In D. Dickinson (Ed.), Bridges to literacy:Children, families, and schools (pp. 19–42). Cambridge, MA: Blackwell.Bornstein, M. H. (1991). Approaches to parenting in culture. In M. H. Bornstein (Ed.), Culturalapproaches to parenting (pp. 3–19). Hillsdale, NJ: Lawrence Erlbaum and Associates.Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic status and child development. Annual Review ofPsychology, 53, 371–399.Bradley, R., & Corwyn, R. (2007). Externalizing problems in fifth grade: Relations with productiveactivity, maternal sensitivity, and harsh parenting from infancy through middle childhood.Developmental Psychology, 43, 1390–1401.Chapin Hall at the University of Chicago 147

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design.Cambridge, MA: Harvard University Press.Bronfenbrenner, U. (1986). Ecology of the family as context for human development: Researchperspectives. Developmental Psychology, 22, 723–742.Brooks-Gunn, J. (2003). Do you believe in magic?: What we can expect from early childhoodintervention programs. Social Policy Report, 17, 3–4.Burchinal, M. R., Peisner-Feinberg, E., Pianta, R., & Howes, C. (2002). Development of academic skillsfrom preschool through second grade: Family and classroom predictors of developmentaltrajectories. Journal of School Psychology, 40, 415–436.Castro, D. C., Bryant, D. M., Peisner-Feinberg, E. S., & Skinner, M. L. (2004). Parent involvement inHead Start Programs: The role of parent, teacher and classroom characteristics. Early ChildhoodResearch Quarterly, 19, 413–431.Chase, R., Arnold, J., Schauben, L., & Shardlow, B. (2005). Childcare use in Minnesota: 2004 statewidehousehold childcare survey. St. Paul, MN: Wilder Research.Chase-Lansdale, P. L., Valdovinos D'Angelo, A., & Palacios, N. A. (2007). A multidisciplinaryperspective on the development of young children in Mexican American immigrant families. In M.H. Bornstein, K. Deater-Deckard & J. E. Lansford (Eds.), Immigrant families: Multidisciplinaryviews on the 21st century. New York: Guilford Press.Chang, Y., Fine, M. A., Ispa, J., Thornburg, K. R., Sharp, E., & Wolfenstein, M. (2004). Understandingparenting stress among young, low-income, African-American, first-time mothers. EarlyEducation and Development, 15, 265–282.Chernoff, J. J., Flanagan, K. D., McPhee, C., & Park, J. (2007). Preschool: first findings from the thirdfollow-up of the early childhood longitudinal study, birth cohort (ECLS-B) (No. NCED 2008-025).Washington, DC: National Center for Educational Statistics, Institute of Education Sciences, U.S.Department of Education.Children’s Services Council (2007). Policy brief on declines in participation in Florida’s KidCareprogram declined statewide and in Palm Beach County between 2004 and 2006Children’s Services Council. (2008). State of the child in Palm Beach County: Early care andeducation―school readiness update. Retrieved July 16, 2008, fromhttp://cache.trustedpartner.com/docs/library/000238/ECESchool_Readiness2008.pdfChildren’s Services Council. (2010). State of the child: 2010 school readiness update Palm BeachCounty. Retrieved April 8, 2011 fromhttp://cache.trustedpartner.com/docs/library/000238/CORRECTEDSchoolReadiness.pdfCovering Kids and Families. (2005). One in 4 uninsured children in Florida had no medical care for theentire year. Retrieved May 14, 2008, from http://www.coveringkidsandfamilies.orgDaro, D., Howard, E., Tobin, J. & Harden, A. (2005). Welcome Home and Early Start: An assessment ofprogram quality and outcomes. Chicago: Chapin Hall at the University of Chicago, 142 pp.De Feyter, J. J. & Winsler, A. (2009). The early developmental competencies and school readiness oflow-income, immigrant children: Influences of generation, race/ethnicity, and national origins.Early Childhood Research Quarterly, 24, 411–431.Dickinson, D. K. (Ed.). (1994). Bridges to literacy: Children, families, and schools. Cambridge, MA:Blackwell Publishers.Chapin Hall at the University of Chicago 148

Dinan, K. A. (2005a). Federal polices restrict immigrant children’s access to key public benefits:Children in low-income immigrant families policy brief. Retrieved June 5, 2008, from the NationalCenter for Children in Poverty website: http://www.nccp.org/pub_fpr05.htmlDinan, K. A. (2005b). State policies can promote immigrant children’s economic security: Children inlow-income immigrant families policy brief. Retrieved June 5, 2008, from the National Center forChildren in Poverty website: http://www.nccp.org/pub_spc05.htmlDuncan G. J. & Brooks-Gunn, J. (2000). Family poverty, welfare reform, and child development. ChildDevelopment, 71, 188–196.Ecocultural Scale Project (1997). The ecocultural family interview manual. Unpublished manuscript, LosAngeles: Ecocultural Scale Project.Edin, K. & Lein, L. (1997). Making ends meet: How single mothers survive welfare and low-wage work.New York: Russell Sage Foundation.Epstein, J. L. (2001). School, family, and community partnerships. Boulder, CO: Westview Press.Federal Interagency Forum on Child and Family Statistics. (2008). America’s children in brief: Keynational indicators of well-being. Washington, DC: U.S. Government. Printing Office. RetrievedJuly 18, 2008, from Forum on Child and Family Statistics website:http://www.childstats.gov/americaschildren/phenviro.aspFlorida Department of Health Vital Statistics Annual Reports (2006). Retrieved April 17, 2008, fromhttp://www.flpublichealth.comFreeman, D. (1996). “To take them at their word”: Language data in the study of teachers’ knowledge,Harvard Educational Review, 66, 732-761.Fuller, B., Holloway, S. D., & Liang, X. (1996). Family selection of childcare centers: The influence ofhousehold support, ethnicity, and parental practices. Child Development, 67, 3320–3337.Glaser, B. & Strauss, A. (1967). The discovery of grounded theory: Strategies for qualitative research.New York: Aldine.Gomby, D. S. (2005). Home visitation in 2005: Outcomes for children and parents.(Invest in KidsWorking Paper No. 7). Retrieved June 13, 2008, from Committee for Economic Developmentwebsite: http://www.ced.org.Hart, B. & Risley, T. R. (1995). Meaningful differences in the everyday experience of young Americanchildren. Baltimore: Paul H. Brookes.Henly, J. R. & Lyons, S. (2000). The negotiation of childcare and employment demands among lowincomeparents. Journal of Social Issues, 56, 683–706.Henly, J. R., Danziger, S. K. & Offer, S. (2005). The contribution of social support to the material wellbeingof low-income families. Journal of Marriage and the Family, 67, 122–140.Herbst, C. M. (2008). Who are the eligible non-recipients of child care subsidies? Children and YouthServices Review, 30, 1037–1054.Herbst, C. M & Tekin, E. (2009). Child care subsidies and child development. (NBER Working Paper14474), National Bureau of Economic Research, Inc.Hernandez, D. J., Denton, N. A., Macartney, S. E. (2007). Children in immigrant families—The U.S. and50 states: National origins, language, and early education. Washington, DC: Child Trends.Chapin Hall at the University of Chicago 149

Hill, N. E., & Craft, S. A. (2003). Parent-school involvement and school performance: Mediatedpathways among socioeconomically comparable African American and Euro-American Families.Journal of Educational Psychology, 95, 74–83.Hirshberg, D., Huang, D.S. & Fuller, B. (2005). Which low-income parents select child care? Familydemand and neighborhood organizations. Children and Youth Services Review, 27, 1119–1148.Horn, I. B., Brenner, R., Rao, M., & Cheng T. L. (2006). Beliefs about the appropriate age for initiatingtoilet training: Are their racial and socioeconomic differences? Journal of Pediatrics, 149, 165–168.Howes, C. (1997). Children’s experiences in center-based child care as a function of teacher backgroundand adult: child ratio. Merrill-Palmer Quarterly, 43, 404–425.Howes, C. & Smith, E. (1995). Relations among child care quality, teacher behavior, children’s playactivities, emotional security, and cognitive activity in child care. Early Childhood ResearchQuarterly, 10, 381–404.Hundley, K. (2008, March 7). State changes tale: Jobs disappearing. St. Petersburg Times.Jeynes, W. H. (2003). A meta-analysis: The effects of parental involvement on minority children’sacademic achievement. Education and Urban Society, 35, 202–218.Henry J. Kaiser Family Foundation. (2008). Florida: Health insurance coverage of children 0–18, states(2007–2008), U.S. (2008). Retrieved April 15, 2010, fromhttp://www.statehealthfacts.org/profileind.jsp?ind=127&cat=3&rgn=11Karp, N. (2006). Designing models for professional development at the local, state and national levels. InZaslow, M., & Martinez-Beck, I. (Eds.) Critical issues in early childhood professionaldevelopment (pp. 225–230). Baltimore: Brookes.King, M. (2007). Immigrants in the U.S. healthcare system; five myths that misinform the Americanpublic. Center for American Progress.Kinukawa, A., Guzman, L., & Lippman, L. (2004). National estimates of childcare and subsidy receiptfor children ages 0 to 6: What can we learn from the national household education survey?Washington, DC: Child Trends. Retrieved July 27, 2007, fromhttp://www.childtrends.org/Files/NHES_Research_brief_10_19_04.pdfKoppelman, J. (2002). Federal child care funding for low-income families. (National Health PolicyForum, Issue Brief no. 780). Washington, DC: George Washington University.Landry, S. H., Smith, K. E., Swank, P. R., Assel, M. A., & Vellet, S. (2001). Does early responsiveparenting have a special importance for children’s development or is consistency across earlychildhood necessary? Developmental Psychology, 37, 387–403.Landry S. H., Smith K. E., & Swank P. R. (2006). Responsive parenting: establishing early foundationsfor social, communication, and independent problem-solving skills. Developmental Psychology,42, 627–642.Lawrence, S., & Kreader, J. L. (2005). Predictors of child care subsidy use. New York: Child Care &Early Education Research Connections.Lee, B. J., George, R., Reidy, M., Kreader, J. L., Georges, A., Wagmiller, R. L., Jr., et al. (2004). Childcare subsidy use and employment outcomes of TANF mothers during the early years of welfarereform: A three-state study. Chicago: Chapin Hall at the University of Chicago.Chapin Hall at the University of Chicago 150

Leventhal, T., Brooks-Gunn, J., McCormick, M. C., & McCarton, C. (2000). Patterns of service use inpreschool children: Correlates, consequences, and the role of intervention. Child Development, 71,802–819.Liang, X., Fuller, B., & Singer, J. D. (2000). Ethnic differences in child care selection: The influence offamily structure, parental practices, and home language. Early Education Research Quarterly, 15,357–384.Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Newbury Park, CA: Sage Publications, Inc.Lippman, L., Vandivere, S., Keith, J., & Atienza, A. (2008). Childcare use by low-income families:Variations across states. Washington, DC: Child Trends. Retrieved July 2, 2008, fromhttp://www.childtrends.org/Files//Child_Trends-2008_07_02_RB_ChildCareLowIncome.pdfLowe, E. D. & Weisner, T. (2004). “You have to push it”―who’s going to raise your kids?: Situatingchildcare and childcare subsidy use in the daily routines of lower income families. Children andYouth Services Review, 26, 143–171.Lyons, S., Karlstrom, M., & Haywood, T. (2007). Implementation study of the comprehensive servicesprogram of Palm Beach County, Florida: Final report. Chicago: Chapin Hall at the University ofChicago.Martinez-Schallmoser, L., MacMullen, N. & Telleen, S. (2003). The effect of social support andacculturation on postpartum depression in Mexican-American women. Journal of TransculturalNursing, 14, 329–333.McLoyd, V. C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53,185–204.Meidel, W. T., & Reynolds, A. J. (1999). Parent involvement in early intervention for disadvantagedchildren: Does it matter? Journal of School Psychology, 37, 379–402.Miles, M. B., & Huberman, M. A. (1994). Qualitative data analysis: An expanded sourcebook. ThousandOaks, CA: Sage Publications.National Center for Education Statistics (NCES), U.S. Department of Education. (2003). Early childhoodlongitudinal study birth cohort.National Center for Education Statistics (NCES), U.S. Department of Education. (2007).Nationalestimates of participation in afterschool programs.National Research Council. (1998). Preventing reading difficulties in young children. Snow, C. E., Burns,M.S., & Griffin, P. (Eds.) Washington, DC: National Academy Press.National Research Council and Institute of Medicine. (2000). From neurons to neighborhoods: Thescience of early childhood development: executive summary. Committee on Integrating theScience of Early Childhood Development; Shonkoff, J. P, & Phillips, D. A. (Eds). Washington,D.C.: National Academy Press.NICHD Early Child Care Research Network. (2000). Characteristics and quality of child care for toddlersand preschoolers. Applied Developmental Science, 4, 116–135.NICHD Early Childcare Research Network. (2007). Age of entry to kindergarten and children’s academicand socioemotional development. Early Education & Development, 18, 337-368.Nihira, K., Weisner, T. S., & Bernheimer, L. P. (1994). Ecocultural assessment in families of childrenwith developmental delays. American Journal on Mental Retardation, 98, 551–566.Chapin Hall at the University of Chicago 151

Olds, D. L., Kitzman, H., Hanks, C., Cole, R, Anson, E., Sidora-Arcoleo, K., et al. (2007). Effects ofnurse home visiting on maternal and child functioning: Age-9 follow-up of a randomized trial.Pediatrics, 120, 832–845.Patton, M. (2002). Qualitative research and evaluation methods. Thousand Oaks, CA: Sage Publications,Inc.Pianta, R. C., Belsky, J., Vandergrift, N., Houts, R., & Morrison, F. (2002). Classroom effects onchildren’s achievement trajectories in elementary school. American Educational Research Journal,45, 365–397.Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the generalpopulation. Applied Psychological Measurement, 1, 385–401.Raikes, H., Luze, G., Brooks-Gunn, J., Raikes, H. A., Pan, B., Tamis-LeMonda, C., et al. (2006). Motherchildbook reading in low-income families: Correlates and outcomes during the first three years oflife. Child Development, 77, 924–953.Reynolds, A. J., Ou, S., & Topitzes, J. W. (2004). Paths of effects of early childhood intervention oneducational attainment and delinquency: A confirmatory analysis of the Chicago Child-ParentCenters. Child Development, 75, 1299–328.Shen, J., Tackett, W., & Ma, X. (2009). Second Evaluation Report for Palm Beach County QualityImprovement System (QIS). Kalamazoo, MI: Western Michigan State University.Shlay, A. B., Weinraub, M., Harmon, M., & Tran, H. (2004). Barriers to subsidies: Why low-incomefamilies do not use child care subsidies. Social Science Review, 33, 134–157.Snow, C. E.; Barnes, W. S.; Chandler, J.; Goodman, I. F.; & Hemphill, L. (1991). Home and schoolinfluences on literacy. Cambridge, MA: Harvard University Press.Snow, C. E., Burns, M. S., & Griffin, P. (Eds.) (1998). Preventing reading difficulties in young children.Washington, DC: National Academy Press.Spielberger, J., Haywood, T., Schuerman, J., Michels, L., & Richman, H. (2005). Evaluation of theChildren’s Behavioral Health Initiative, Palm Beach County, Florida: Final report. Chicago:Chapin Hall at the University of Chicago.Spielberger, J., Schuerman, J., Lyons, S., & Haywood, T. (2006). The Palm Beach County family study:Baseline report. Chicago: Chapin Hall at the University of Chicago.Spielberger, J., Lyons, S., Gouvea, M., Haywood, T., & Winje, C. (2007). The Palm Beach County familystudy: Second annual report. Chicago: Chapin Hall at the University of Chicago.Spielberger, J., Rich, L., Gouvea, M., Winje, C., Scannell, M., Harden, A., et al., (2009). The Palm BeachCounty family study: Third annual report. Chicago: Chapin Hall at the University of Chicago.Spielberger, J., Rich, L., Winje, C., & Scannell, M. (2010). The Palm Beach County family study: Fourthannual report. Chicago: Chapin Hall at the University of Chicago.Tandon, D. (2004). Access to prenatal care in Palm Beach County: Findings and recommendations.Baltimore: Johns Hopkins University School of Medicine.Uno, D., Florsheim, P., & Uchino, B. (1998). Stress, social support, and parenting practices among whiteand Latino adolescent mothers. Journal of Youth and Adolescence, 27, 587–607.U.S. Department of Health and Human Services Centers for Disease Control and Prevention (2005).Chapin Hall at the University of Chicago 152

Weisner, T. S. (1984). A cross-cultural perspective: Ecocultural niches of middle childhood. In AndrewCollins (Ed.), The elementary school years: Understanding development during middle childhood(pp. 335–369).Washington, DC: National Academy Press.Weisner, T. (1997).The ecocultural project of human development: Why ethnography and its findingsmatter. Ethos, 25, 177–190.Weisner, T. (2002). Ecocultural understanding of children’s developmental pathways. HumanDevelopment, 45, 275–281Westat, Chapin Hall Center for Children, James Bell Associates. (2002). Evaluation of familypreservation and reunification programs final report. Washington, DC: U.S. Department of Healthand Human Services..Chapin Hall at the University of Chicago 153

Appendix A: Longitudinal StudySample and MethodsResearch Questions, Sample, and MethodsThe longitudinal study examines the use and effects of a broad array of health, educational, childcare, andpsychosocial services received by families with young children in four targeted geographic areas (TGAs)of Palm Beach County, starting with the birth of a child (referred to as the “focal child”). One purpose ofthe study was to describe the children and families in the TGAs who use various services and examinechild and family risk factors, service needs, services received, and experiences using services. In addition,the study has explored 1) patterns of service use over time and relations between early service use andlater service use, and 2) relationships between service use patterns and selected indicators of familyfunctioning and child development.The study was originally conceived to follow a cohort of children born in 2004 or 2005 for a period of 8or 9 years—or until the focal children were in third grade. However, the study was later modified tofollow these children until they reached kindergarten. Below are the primary research questions for thestudy.What are baseline demographic characteristics of families of young children in the TGAs?What services are available and used by families of young children in the TGAs?Are there patterns of service use? What combinations, if any, of services occur most frequently? Howdoes service use change over time?Who are the children and families who use social services in the TGAs? What is the relation betweenchild and family characteristics and service use (i.e., number and types of services used)? Whatcharacteristics determine service use?Chapin Hall at the University of Chicago 154

What are pathways to service use? What are inhibiting and facilitating factors in use of services?What are the potential effects of service use? How is service use related to child outcomes such asphysical, socio-emotional, and behavioral health, school readiness, and school success; and to familyfunctioning and abuse/neglect rates?To address these questions, the longitudinal study uses a mixed-methods approach to gather a widevariety of information about the characteristics and needs of families the system is intended to serve andhow families use services. Primary methods include the following:An analysis of administrative data on service use and key outcomes of all children born in the TGAsand in the county during 2004 and 2005 and who remain in the county at various data collectionpoints during a 6-year period. To date, we have collected and analyzed available administrative datafrom Department of Health Vital Statistics, the Maternal Child Health Partnership (MCHP) database(Right Track and now FOCiS), the Department of Children and Families (DCF) database on reportsof child abuse and neglect, and the Childcare Enhanced Field System (EFS) database on child subsidyuse.A longitudinal survey of the service use experiences of a sample of families with young children inthe TGAs, employing annual in-person interviews with a baseline sample of 531 mothers of newbornchildren and brief telephone interviews with the same parents about 6 months after each in-personinterview for 5 years, supplemented by administrative data on service use and child and familyoutcomes for 8 years.A 3-year qualitative study involving in-depth interviews and observations of a small subsample offamilies to enhance what is learned through analysis of the household survey and administrative dataabout service use, motivations to use services, and how services fit into families’ lives. Six waves ofinterviews were conducted with a small sample of mothers drawn from the survey sample; and oneinterview was conducted with a sample of 16 male caregivers, most of whom were the biologicalfather of the focal child.These three primary sources of data for the longitudinal study—administrative records, householdsurveys, and in-depth qualitative interviews—are designed to complement and supplement one another.For example, data from the structured interview sample provide additional information to help us identifycorrelates of child and family outcomes and better understand relations among services, child and familyindicators, and outcomes found in the administrative data. At the same time, administrative data provideadditional information about service use and outcomes for mothers in the interview sample. Informationgathered from the embedded qualitative study extends what we learn from both administrative data andChapin Hall at the University of Chicago 155

the structured interview study about families’ day-to-day lives, how service use does or does not fit intotheir lives, and other family characteristics and activities that affect family functioning and childdevelopment. Additional information on the methodology of each of the three study components ispresented below. 75Administrative RecordsThe sample for the administrative data component of the study is a population cohort of all families inPalm Beach County who had a child born between January 1, 2004, and December 31, 2005, with a focuson the subsample of families living in the TGAs at the time of their child’s birth. Through data-sharingagreements established with several agencies in Palm Beach County, including the Department of Health(DOH), Department of Children and Families (DCF), the Children’s Services Council, the School Districtof Palm Beach County, and Boys Town South Florida, we receive and analyze periodic exports of data onbirth cohort characteristics, service use, and child and family outcomes.There are four main sources of administrative data for the longitudinal study. First, the DOH VitalStatistics database provides information on the use of prenatal care, birth outcomes, and maternaldemographic characteristics. Second, the FOCiS database (formerly called Right Track) for the MCHPsystem information on prenatal and postnatal assessments, names of agencies providing maternal-childhealth services, types of services, and dates of prenatal and postnatal services related to the births of thefocal children in 2004 and 2005; and subsequent children born in 2006 and 2007. (See Table A- 1 for alist of the types of Maternal Child Health Partnership services (that are defined as treatment) recorded inthe FOCiS database.) Third, the DCF HomeSafenet database supplied information on reports of childabuse and neglect starting in 2004 through 2006; a new system, the Florida Safe Family Network,provided data from 2007 through 2009. Finally, the EFS database supplies information on the use ofchildcare subsidies. The fourth report drew largely from data in the FOCiS and DOH Vital Statistics datasystems.In this, the final year of the study (2009–2010), we continued to gather information from these datasources for all families who remain in Palm Beach County as well as for the study families. In addition,we also plan to obtain and analyze information from the school district database on children’s schoolreadiness and behaviors and academic progress in kindergarten; these results will be presented in a laterreport.75 Reviews of literature on barriers to and facilitators of service use, including findings about service use from evaluation reportsof individual programs funded by the Children’s Services Council (CSC) of Palm Beach County during the period of the study,for example, the Children’s Behavioral Health Initiative, Comprehensive Services, and the Family & Community Partnership, aswell as other research on services in Palm Beach County.Chapin Hall at the University of Chicago 156

Table A- 1. FOCiS Codes for MCHP Services Defined as TreatmentActivity Code Code Description Program a10 Nutrition WHIN11 Breastfeeding Peer Counseling WHIN12 Translation Services WHIN13 Smoking Cessation WHIN14 General Health Education WHIN15 Case Management WHIN16 Depression Screening WHIN16 Prenatal Depression Screening HS17 Depression Counseling WHIN18 Adolescent Pregnancy Prevention WHIN19 Pregnancy/Childbirth 4 WHIN20 Parenting Skill Building/Education WHIN21 Male Support Services WHIN40 Nutrition WHIN41 Breastfeeding Peer Counseling WHIN42 Translation Services WHIN43 Smoking Cessation WHIN44 General Health Education WHIN45 Case Management WHIN46 Depression Screening WHIN46 Interconceptional Depression Screening HS47 Depression Counseling WHIN48 Adolescent Pregnancy Prevention WHIN49 Pregnancy/Childbirth 4 WHIN50 Parenting Skill Building/Education WHIN51 Male Support Services WHIN3102 Participant Need Assessment HS3215 Initial Assessment HS3215 Initial Assessment WHIN3320 Care Coordination-Face To Face WHIN3320 Care Coordination: Face to Face HS3321 Care Coordination-Not Face To Face WHIN3321 Care Coordination: Not Face to Face HS3322 Initial Family Support Planning HS3322 Initial Family Support Planning WHIN3323 Update Family Support Planning HS3323 Update Family Support Planning WHIN4501 Nutrition Assessment Counseling HS6515 Oral Translation Services HS6516 Written or Assisted Translation HS8002 Psychosocial Counseling HS8004 Parenting Support and Education HS8006 Childbirth Education HS8008 Breast Feeding Education and Support HS8013 Interconceptional Education and Counseling HS8026 Smoking Cessation Counseling HSa HS=Healthy Start program; WHIN=Women’s Health InitiativeChapin Hall at the University of Chicago 157

The analysis of administrative data depends on accurately and reliably linking individual-level identifiersin data records across multiple service system databases. Linking is done through a process ofprobabilistic record matching, which calculates the probability that two records belong to the same personby using multiple pieces of identifying information. The most uniform common identifiers include name,Social Security number, date of birth, gender, race/ethnicity, and zip code, but these vary acrossadministrative data systems. Because of data entry errors, administrative data systems may have errors inclient-identifying information, which can result in more than one system-unique ID for the same person.Probabilistic matching technology corrects for these “duplication” errors by linking records in differentsystems based on the probability of similar identifiers representing data records for the same person.Household SurveysAnother component of the study involves more intensive data collection on a sample of 531 families withchildren born in the TGAs during 2004 and 2005. The primary source of information on these families isa structured survey conducted twice a year, once in person and once by telephone; this information issupplemented by data obtained from administrative records. The sample for this part of the study wasrecruited through two maternal health programs in Palm Beach County, the Healthy Mothers/HealthyBabies Coalition and Healthy Start/Healthy Families, between July 2004 and November 2005. 76 Thebaseline study report (Spielberger, Schuerman, Lyons & Haywood, 2006) provides other informationabout the recruitment process and recruitment issues for the study.The sample was stratified along two dimensions. First, we assumed that families with more risk factorswere more likely to have contact with services. Thus, we attempted to develop the sample so that abouthalf would be families identified as “high risk” (children at high risk of poor outcomes or families at highrisk of dysfunction) on a hospital screen or home assessment. Second, because the Glades TGA issparsely populated and historically more transitory than other areas of the county, 77 we wanted to ensurethat the sample was large enough to make reasonable estimates of its characteristics. Hence, the samplewas structured according to the proportions shown in Table A- 2.76 We followed this approach as an alternative to obtaining permission to sample from birth records for several reasons. First, wedid not know whether we would be able to attain Institutional Review Board (IRB) approval under the Health InsurancePortability and Accountability Act (HIPAA) to use protected health information for sample identification and recruitment. Wealso were following the precedent of an earlier study about access to prenatal care in Palm Beach County (Tandon, 2004), whichrecruited and interviewed newly delivered mothers in maternity wards. Finally, we recognized that because of their experienceworking with mothers in the TGAs, the hospital liaisons and nurses might be more trusted by potential respondents than otherrecruiters.77 In contrast to the other three TGAs, the Glades TGA in the western part of the county is a large, but sparsely populatedagricultural area that includes several migrant families who harvest sugarcane, citrus fruit, and other crops. According to datafrom the 2000 Census, percentages of families with children under the age of 18 living in poverty is higher in the Glades (46%)than in the other TGAs (25%) (CSC, May 2006). These percentages compare with an overall percentage of 13 percent for theproportion of all families with children under the age of 18 who live in poverty in the county as a whole.Chapin Hall at the University of Chicago 158

Table A- 2. Interview Study Sampling PlanTGA % High Risk % Not High Risk %TotalNon-Glades 40 40 80Glades 10 10 20Total 50 50 100Data for identifying families as “high risk” and “not high risk” were based on screens and assessmentsconducted at various points during pregnancy, at birth, and after birth and recorded in an administrativedata system for the Maternal Child Health Partnership. 78 At the time of the study, mothers who receivedprenatal services from the Department of Health were usually given an assessment of risk during thosecontacts. Many mothers also were given the Healthy Start Infant Risk Screen in the hospital shortly aftergiving birth. On a scale from 0 to 10, a score of 4 or higher was considered an indicator of possible risk.Mothers who scored 4 or higher were encouraged to accept a home visit from a Healthy Start nurse.Mothers who received lower scores were not offered a home visit, but could request one. 79 Subsequently,mothers who were visited at home were assessed again and designated as having services levels of E, 1, 2,or 3 on the basis of scored risk assessments, observations by nurses (and perhaps other medicalpersonnel), and clinical judgments of nurses. Mothers who were assigned levels 2 and 3 were thought toneed more frequent or more varied services and thus were loosely referred to as “high risk.” 80In addition to risk status and TGA, mothers were to have recently given birth and to have custody of thatchild to be included in the study. The baby did not have to be a first-born child, however, to qualify themother for study participation. Other selection criteria included a maternal age and language. Mothers hadto be 16 years and speak one of the three main languages spoken in Palm Beach County—English,Spanish, or Haitian Creole.Each year, mothers were interviewed in person for about an hour to an hour and a half in their homes oranother location. Six months after an in-person interview, they participated in a brief telephone interview78 Known as Right Track when we began the study, the data system is now called FOCiS.79 At the time we began recruiting, the Healthy Start program included a universal home visiting component for all newlydelivered mothers. In spring 2005, the program changed to target mothers most in need of and more likely to use services.Mothers in a “special low-risk” group are also offered a home visit if they are younger than 19, new to the county, have deliveredtheir first child with no or only late-term prenatal care, have no identified pediatrician, have difficulty bonding with their baby, orseem to lack social support (personal communication with Tanya Palmer and Christine Walsh, 2005).80 Clearly, none of these measures of risk is perfect in identifying children and families that may need services. The screenings,assessments, and level designations all involve judgments, which are of unknown reliability and validity. In addition, not allmothers receive prenatal services. Not all are given the in-hospital screen. Some mothers refuse the assessment. Hospital liaisonsare not present on weekends, and although other hospital personnel are supposed to visit mothers who give birth on weekends,those contacts do not always happen. Nor do all mothers agree to the postpartum visit by nurses. Those mothers who do agree arenot always visited for a variety of reasons. Thus, the selectivity at each of these stages is likely to be biased and cannot beassumed to be random.Chapin Hall at the University of Chicago 159

lasting 20 to 30 minutes. The baseline in-person interviews were conducted soon after recruitment whenthe focal child was between 1 and 6 months of age. A telephone interview occurred 6 months later whenthe focal child was between 7 and 12 months old. The third interview or second in-person interviewoccurred when the focal child was between 13 and 18 months of age. The third in-person interviewoccurred when the focal child was between 24 and 30 months old. The fourth in-person interviewoccurred when the focal child was between 36 and 42 months old. The fifth in-person interview occurredwhen the focal child was between 48 and 54 months of age. A final, ninth interview or the fifth telephoneinterview was conducted with a small sample of families when their focal children were between 54 and60 months of age. All of these interviews were conducted by trained interviewers employed by Westat, alarge survey research firm headquartered in Rockville, Maryland.Topics for all five rounds of in-person interviews are listed in Table A- 3. Interview topics weredeveloped by Chapin Hall researchers with input from CSC and Westat research staff and with referenceto protocols used in other large-scale evaluations and studies of service use, children’s development, andfamily functioning. 81Before we finalized the consent and protocols for the baseline interview, we piloted them in spring 2004with a small sample of mothers of infants residing in the TGAs of Palm Beach County (copies of thestructured interviews are available from the authors). The interview topics covered a wide range in aneffort to develop a complete description of the demographic backgrounds of families and the variety ofother factors that are likely to affect the well-being and development of their children and familyfunctioning. In addition to the availability and use of formal services, these factors include family livingconditions, physical and mental health, medical care, parenting practices, and access to informal andcommunity support. In subsequent years of the study, the content of the structured interviews hasremained largely the same so that change over time can be measured. However, some revisions have beenmade in selected items to reflect the developmental stage of the focal child, changing family orcommunity circumstances, or new information about service availability.81 Other studies include the national evaluation of Family Preservation and Reunification Programs (Westat, Inc., Chapin Hall, &James Bell Associates, 2002), the Early Childhood Longitudinal Study, Birth Cohort (U.S. Department of Education, 2006), andthe evaluation of the Cayuhoga County Early Childhood Initiative (Daro, Howard, Tobin, & Harden, 2005).Chapin Hall at the University of Chicago 160

Table A- 3. Household Survey Topics over TimeTopicsHouseholdCompositionFamily ContextAdequacy ofLivingArrangementsHealth,Insurance,Prenatal Care,ChildDevelopmentFamilyConcerns andUse of FormalServicesChildcare andAfterschoolArrangementsSocial SupportParentingPracticesPublic Supportsand FamilyIncomeMaternalFunctioningYear/Focal Child’s AgeYear 1/0–6 months Year 2/12–18 mos. Year 3 /24–30 Year 4/36–42 mos. Year 5/48–54 mos.mos.Description of all household members including name, relationship to theSame as Year 1 Same as Year 2 Same as Year 3 Same as Year 4respondent, age, gender, race, student and employment status, number of hoursworked per week, and educational levelFamily structure and social/economic context, including children who do not live in Same as Year 1; Similar to Year 2; Same as Year 3; Same as Year 4the home, the children’s father(s), marital status, employment, language spoken in added question on added literacy and added items onthe home, access to transportation, living conditionsmaternal nativity significant family child supportevent itemsConditions of family’s housing—electricity, plumbing, heating/air conditioning, Same as Year 1 Same as year 2 Same as Year 3 Same as Year 4cooking appliances, lack of basic household furnishings, and overcrowding(1) Health of and medical care for children and adults; (2) prenatal care; (3) healthinsurance coverage for children and the parent’s knowledge of public healthinsurance programs for children; and (4) where medical care is usually obtained andbarriers to medical care(1) Concerns in the past year about (a) meeting family’s basic needs (b) child(ren)’shealth and development; (2) frequency of concern; (3) whether they tried to get helpfrom a community agency or program; (4) whether they received help; (5)satisfaction with the agency or program; and (6) contacts with any other services(1) For preschool children, type of care arrangement, time in care, cost of care, andsatisfaction with the arrangement; and (2) for school-age children, participation inorganized activities or afterschool programs, and satisfaction with afterschoolarrangements(1) Support parents receive from spouse or partner, relatives and friends, places ofworship, and other community supports such as medical, school, and social workprofessionals; (2) satisfaction with help from these supporters, and (3) accessibilityand adequacy of supportSimilar to Year 1;added subsequentpregnancy itemSimilar to Year 2,with childdevelopment itemsadjusted for 2-yearoldchildrenSimilar to Year 3with childdevelopment itemsadjusted for 3-yearoldchildrenSimilar to Year 4with childdevelopment itemsadjusted for 4-yearoldchildrenSame as Year 1(added birthcontrol and familyplanning to SectionC)Same as Year 2 Same as Year 3 Same as Year 4Same as Year 1 Same as Year 2, Same as Year 3 Same as Year 4modified to getsubsidy use for allRespondents usingSimilar to Year 1interview items(dropped separatereligious items;added affection andconflict)Maternal and paternal positive and negative parenting practices Same as Year 1,with some newitemsReceipt of public transfers such as Food Stamps, rent vouchers, TANF, WIC, andSocial Security, and to estimate their total household income from all sources(1) Center for Epidemiological Studies Depression Scale (CES-D) to assessdepressive symptoms in previous week; and (2) use of alcoholic beverageschildcareSimilar to Year 2;dropped 2 itemsSame as Year 2,added frequencyscaleSame as Year 1 Same as Years 1and 2Same as Year 1,plus the ParentingStress Inventory(PSI)Same as Year 2,added questions onsmokingSame as Year 3 Same as Year 4Same as Year 3 Same as Year 4Same as Year 3 Same as Year 4Same as Year 3 Same as Year 4Chapin Hall at the University of Chicago 161

Sample Response RatesTable A- 4 presents the number of mothers who have completed each phase of the longitudinal studyinterviews to date. A total of 444 mothers, or 84 percent of the baseline sample of 531, completed thesecond in-person interview; and 399 (75%) completed the third-year in-person interview. A total of 390(73%) of the mothers participated in all three waves of in-person interviews; 355 (67%) completed thefourth-year in-person interview; and a total of 335 (63%) completed all four waves of in-personinterviews. The final in-person interview was completed by a total of 353 (66%) and 310 (58%)completed all five in-person interviews. As shown in Table A- 5, the most frequent reason for sampleattrition is mothers moving out of the study area or being unable to contact. Only a small percentage ofmothers have declined to participate in follow-up interviews or left the study for other reasons.A regression analysis of the mothers who completed the fifth-year interview indicates that this attrition isalmost random. The analysis included the following as explanatory variables: living in the Glades, beinghigh risk, having been born in the U.S., being married, being black, being Hispanic, having a high schooldiploma, yearly income, mother’s age at the first interview, and scoring high on the depression scale. Inyear 5, the only significant variables were ethnicity and education level at the baseline interview. In year5, black mothers were almost 3 times more likely than mothers of white or other ethnic backgrounds to beinterviewed in the fifth year. Additionally, mothers who, at the time of the first the baseline interview, hada high school diploma were almost twice as likely to be interviewed at year 5 as compared to those whodid not have a high school diploma at the baseline interview.Table A- 4. Sample and Response Rates for Longitudinal Study over TimeInterview Wave Data Collection Period Sample RetainedN% a1. Baseline In-person Interview July 2004–Nov. 2005 5312. First 6-Month Telephone Jan. 2005–May 2006 484 913. Second In-person Interview July 2005–Nov. 2006 444 844. Second 6-Month Telephone Jan. 2006–May 2007 432 815. Third In-person Interview July 2006–Dec. 2007 399 756. Third 6-Month Telephone Jan. 2007–May 2008 387 737. Fourth In-person Interview July 2007–Dec. 2008 355 678. Fourth 6-Month Telephone Jan. 2008–May 2009 357 679. Fifth In-Person Interview July 2008–Dec 2009 353 6610. Fifth 6-Month Telephone b Jan 2009–Sept 2009 181 34aA total of 702 mothers were recruited for the study, and 589 were selected for the original sample. Response rate is based on therate for all eligible cases, which excludes mothers who no longer live in the study area, whose children were too old for thesample, who had lost custody of their children, or who were ineligible for another reason. Attempts are made to contactrespondents for two consecutive time periods before they are dropped from the study sample pool.bThe study was terminated on September 30, 2009, before all of the fifth-year 6-month telephone surveys could be completed.Chapin Hall at the University of Chicago 162

Table A- 5. Reasons for Sample Attrition over TimeInterview WaveData CollectionPeriod1. Baseline In-person July 2004–Nov. 2005SampleReason for AttritionLossMoved/NotRefused Other aLocatedN % N % N % N %2. First 6-Month Phone Jan. 2005–May 2006 47 9 35 75 10 21 2 43. Second In-person July 2005–Nov. 2006 87 16 61 70 20 23 6 74. Second 6-Month Phone Jan. 2006–May 2007 100 19 67 67 22 22 11 115. Third In-person July 2006–Dec. 2007 132 25 94 71 28 21 10 86. Third 6-Month Phone Jan. 2007–May 2008 144 27 110 76 27 19 7 57. Fourth In-person July 2007–Dec. 2008 176 33 134 76 32 18 10 68. Fourth 6-Month Phone Jan. 2008–May 2009 174 33 125 72 34 20 15 99. Fifth In-Person July 2008–Dec 2009 178 34 137 77 30 17 11 6a “Other” includes reasons such as “final not home, maximum contacts,” and death of mother or child.Qualitative InterviewsWe added a qualitative study of a subsample of families in the spring of 2006 to provide a more in-depthand complete understanding of service use and other study topics. Using a mixed-sampling plan, werandomly selected 58 English- and Spanish-speaking mothers from the full study sample; 51 motherswere located and agreed to participate in either the first or second qualitative interview. Because thequalitative study started a year after the larger study, and we wanted to interview mothers when theirchildren were young, we limited the sample pool to mothers whose babies were born in 2005. We alsoexcluded Haitian Creole–speaking mothers from the qualitative sample because they are a smallproportion of the larger sample, and we did not have resources to hire a Creole-speaking interviewer.Thus, we divided the sample pool by initial risk level and then sampled Glades and non-Glades mothersin proportion to their representation in the larger study.Qualitative interviewers met with families twice a year to conduct in-depth, semistructured interviews thatlast about 90 minutes. 82 All interviews were tape-recorded with the permission of mothers, transcribed,and validated to confirm the accuracy of the transcription. In the case of interviews conducted in Spanish,translation was carried out concomitantly with transcription. Interviewers also wrote detailed summarynotes of the information collected during the interview and their observations of the home andneighborhood environment, parent-child interactions, and child behavior.82 The qualitative interviews are conducted by four trained graduate students from Florida Atlantic University and FloridaInternational University. All have had previous experience with qualitative methodology, and two are fluent in Spanish.Chapin Hall at the University of Chicago 163

The qualitative interviews were based on the Ecocultural Family Interview (EFI) framework 83 anddesigned to provide a more complete understanding of how families use, experience, and view services intheir daily lives and how family and community contexts influence children’s development and schoolreadiness. They examined families’ perspectives on the following topics: daily routine and householdinformation; beliefs, goals, and practices about child rearing; experiences with educational and childcareservices; work, economic well-being, and use of income support programs; use of healthcare and socialservices; and mobility and neighborhood characteristics. At the same time, the interview format alsoencouraged mothers to bring up and discuss topics that were of greatest concern to them. Additional datawere collected through observations about the home environment, parent-child interactions, and childbehavior during the home visits and recorded in summary field notes. Thus, the interview guides for eachwave of qualitative interviews were very similar, but each interview varied from mother to mother andfrom one wave to another.We used the “grounded theory” approach to qualitative analysis (e.g., Glaser & Strauss 1967; Lincoln &Guba 1985; Miles &Huberman 1994; Patton 2002), which builds theory based on a systematic approachto coding, usually termed “the constant comparative method.” We examined interview transcripts andsummary notes line by line to see what ideas and patterns the data reflected, and we developed codes forthe data based on the ideas and patterns. 84 Although the interview topics provided an initial guide foranalysis, we also looked for other themes and meanings that were not in the original protocol to emerge.In the coding process, we tried to capture both the representational meaning, or the content of what wassaid, and the presentational meaning, or how it was said, that is, mothers’ use of language and narrativestyle (Freeman, 1996). As we identified concepts, we compared and contrasted them with previouslyidentified concepts and grouped similar concepts together in categories. Throughout the process, wewrote memos to document relationships among concepts and categories and emerging ideas and patterns.Our goals in the analysis and interpretation of the qualitative data were to both understand the experiencesof mothers as individuals in the context of their daily circumstances and to compare and synthesize thenarratives of all of the mothers to form descriptions of “typical” or “composite patterns” of different kindsof service users. This occurred through a simultaneous process of deduction and induction. In particular,83 Ecocultural Scale Project (1997), The Ecocultural Family Interview Manual (Los Angeles, CA: Ecocultural Scale Project); T.S. Weisner (1984), Ecocultural niches of middle childhood: A cross-cultural perspective, in W.A. Collins (Ed.), DevelopmentDuring Middle Childhood (, NJ: LEA Press); K. Nihira, T. S. Weisner, & L. P. Bernheimer (1994), Ecocultural assessment infamilies of children with developmental delays, American Journal on Mental Retardation, 98, 551–566; T. Weisner (1997).Theecocultural project of human development: Why ethnography and its findings matter. Ethos, 25, 177–190.84 A qualitative data software program, Atlas.ti, is used to facilitate the systematic analyses and coding of the interviews.Chapin Hall at the University of Chicago 164

we sought to identify the ecological and cultural factors that shape mothers’ decisions to use (or not use)services.Additional information on the qualitative methodology and findings based on an analysis of the first threewaves of qualitative data can be found in the third-year study report (Spielberger et al., 2009). In thatreport, we particularly focused on the barriers to and facilitators of families’ use of various types ofservices—healthcare, economic supports, childcare, and social services—and how services fit into thecontext of their daily lives. In addition, we looked for emerging themes in their social support andparenting practices and beliefs because as we read the data about various types of services, developedcodes, and analyzed the data, informal social support and parenting practices were emerging as importantthemes related to service use. Additional findings, based on the analysis of the remaining three waves ofqualitative interviews with mothers and additional interviews with male caregivers, will be presented inlater study reports.Chapin Hall at the University of Chicago 165

Appendix B: Additional Data onPBC and TGA Birth Cohorts,2004–2009Table B- 1. Risk Screen Scores of Mothers with Newborns, 2004–2009 aScreen Score in FOCiS Database “At Risk” Screen Score bSampleN% All% All % MothersNnMothersMothers Screened2004–2005 Birth Cohort 29,620 21,556 73 6,858 23 32Non-TGAs 17,185 11,515 67 2,455 14 21TGAs 12,435 10,041 81 4,403 35 442006 Birth Cohort 15,434 11,798 76 3,962 26 34Non-TGAs 8,718 6,176 71 1,422 16 23TGAs 6,716 5,622 84 2,540 38 452007 Birth Cohort 15,414 12,338 80 4,054 26 33Non-TGAs 8,785 6,657 76 1,471 17 22TGAs 6,629 5,681 86 2,583 39 452008 Birth Cohort 14,943 12,562 84 4,052 27 32Non-TGAs 8,578 6,972 81 1,532 18 22TGAs 6,365 5,590 88 2,520 40 452009 Birth Cohort 13,897 10,409 75 2,977 21 29Non-TGAs 8,053 5,785 72 1,124 14 19TGAs 5,844 4,624 79 1,853 32 402004–2009 Birth Cohort 89,308 68,663 77 21,903 25 32Non-TGAs 51,319 37,105 72 8,004 16 22TGAs 37,989 31,558 83 13,899 37 44a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. The birth of twins, triplets, and other multiples are counted as onebirth event. This data reports Maternal Child Health Partnership risk screening that is slightly different than was reported in theyear 3 report due to updated score data.b “At Risk” indicates a score of 4 or above on the Healthy Start risk screen.Chapin Hall at the University of Chicago 166

Table B- 2. Characteritics of TGA Mothers with Maternal Child Health Partnership Risk Screens, 2004–2009 aTGA 2004–2005 TGA 2006 TGA 2007 TGA 2008 TGA 2009MCHPMCHPMCHPMCHPMCHPMaternalBirthBirthBirthBirthBirthCharacteristic b “At“At“At“At“AtEventsRisk” d EventsRisk” d EventsRisk” d EventsRisk” d EventsRisk” d(n=12,435) (n=4,403) (n=6,716) (n=2,540) (n=6,629) (n=2,583) (n=6,365) (n=2,520) (n=5,844) (n=1,853)% % % % % % % % % %Unmarried mother 57 83 60 84 61 84 62 85 62 86Teen mother atchild’s birth14 23 14 23 14 22 13 21 12 18< HS education 35 56 38 62 37 61 36 60 36 65Black c 36 51 35 45 37 46 37 48 38 47Hispanic c 39 39 43 47 42 46 42 44 39 45Foreign-born 47 49 48 53 48 53 48 51 46 52Low birth weight 9 15 9 15 10 15 10 15 10 16a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “at risk” and mothers residing in theGlades TGA. This data reports Maternal Child Health Partnership risk screening that is slightly different than was reported in the year 3 report due to updated score data. The birthof twins, triplets, and other multiples are counted as one birth event.b Does not include mothers who responded “I don’t know” to any of the Vital Statistics questionsc To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.d “At Risk” indicates a score of 4 or above on the Healthy Start risk screen.Chapin Hall at the University of Chicago 167

Table B- 3. MCHP and Chapin Hall Mean Risk Index for TGA Birth Cohorts aSampleMCHP Risk Index for Births withRisk Screen Score in FOCiSChapin Hall Risk Index for WholePopulationMean Risk Index σ Mean Risk index b σ2004–2005 Birth Cohort 2.0 2.10 2.4 1.73Non-TGAs 1.5 1.91 1.9 1.61TGAs 2.7 2.13 3.0 1.682006 Birth Cohort 2.1 2.12 2.3 1.79Non-TGAs 1.6 1.93 1.9 1.67TGAs 2.8 2.15 3.0 1.752007 Birth Cohort 2.2 2.08 2.4 1.79Non-TGAs 1.6 1.91 1.9 1.67TGAs 2.8 2.10 2.9 1.772008 Birth Cohort 2.1 2.06 2.4 1.79Non-TGAs 1.6 1.86 2.0 1.70TGAs 2.8 2.09 3.0 1.732009 Birth Cohort 2.1 2.10 2.4 1.78Non-TGAs 1.6 1.95 2.0 1.70TGAs 2.8 2.08 3.0 1.742004–2009 Birth Cohort 2.1 2.10 2.4 1.68Non-TGAs 1.6 1.91 3.0 1.72TGAs 2.7 2.12 1.9 1.66a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This table reports Maternal Child Health Partnership risk screeningthat is slightly different than was reported in the year 3 report due to updated score data. The birth of twins, triplets, and othermultiples are counted as one birth event.b Because not all mothers in the county were screened at birth, we calculated a risk index based on their number of 11demographic and health characteristics recorded in Vital Statistics. These were the following: (1) no or late prenatal care, (2)mother does not have high school diploma or GED, (3) mother is not married, (4) mother age 19 or lower at birth, (5) mother notborn in U.S., (6) mother received WIC while pregnant, (7) mother smoked, (8) mother had medical complications (other thanprevious C-section), (9) mother had delivery complications, (10) baby’s weight less than 2500 grams, and (11) baby’s gestationalage 36 weeks or less. Mothers whose responses were “I don’t know” to any of these questions were not coded as “at risk” for thatquestion.Chapin Hall at the University of Chicago 168

Figure B- 1. Characteristics of Mothers with Newborns in Palm Beach County aa Source: Vital Statistics. Sample numbers exclude mothers who gave birth in Palm Beach County but were residents of othercounties and mothers who responded “I don’t know” to any item. Mothers who had multiple births were counted once.bTo be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.c Because not all mothers in the county were screened at birth, we calculated a Chapin Hall risk index based on their number of11 maternal demographic and health characteristics recorded in Vital Statistics. These were the following: (1) no or late prenatalcare, (2) no high school diploma or GED, (3) not married, (4) age 19 or less at birth, (5) not born in U.S., (6) used WIC whilepregnant, (7) smoked, (8) prior medical complications (other than previous C-section), (9) delivery complications, (10) baby’sweight less than 2500 grams, and (11) baby’s gestational age 36 weeks or less. Mothers who responded “I don’t know” to any ofthese items were not coded as “at risk” for that item.Chapin Hall at the University of Chicago 169

Figure B- 2. Characteristics of Mothers with Newborns in Non-TGAs aa Source: Vital Statistics. Sample numbers exclude mothers who gave birth in Palm Beach County but were residents of othercounties and mothers who responded “I don’t know” to any item. Mothers who had multiple births were counted once.bTo be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.cChapin Hall risk index (see note in Figure B- 1).Chapin Hall at the University of Chicago 170

Figure B- 3. Characteristics of Mothers with Newborns in TGAs aa Source: Vital Statistics. Sample numbers exclude mothers who gave birth in Palm Beach County but were residents of othercounties and mothers who responded “I don’t know” to any item. Mothers who had multiple births were counted once.bTo be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.c Chapin Hall risk index (see note in Figure B- 1).Chapin Hall at the University of Chicago 171

Table B- 4. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 aSampleBirthEvents bIn MCHP FOCiSData System cIn MCHP FOCiSActivity File dMCHP TreatmentActivity eN n n % n %2004–2005 Birth Cohort 29,620 24,725 83 20,132 68 7,956 27Non-TGAs 17,185 13,535 79 10,171 59 2,926 17TGAs 12,435 11,190 90 9,961 80 5,030 402006 Birth Cohort 15,434 13,298 86 9,388 61 4,357 28Non-TGAs 8,718 7,152 82 4,228 48 1,595 18TGAs 6,716 6,146 92 5,160 77 2,762 412007 Birth Cohort 15,414 13,631 88 9,385 61 4,384 28Non-TGAs 8,785 7,460 85 4,313 49 1,608 18TGAs 6,629 6,171 93 5,072 77 2,776 422008 Birth Cohort 14,943 13,724 92 9,049 61 4,324 29Non-TGAs 8,578 7,699 90 4,237 49 1,679 20TGAs 6,365 6,025 95 4,812 76 2,645 422009 Birth Cohort 13,897 12,623 91 7,917 57 3,149 23Non-TGAs 8,053 7,111 88 3,788 47 1,295 16TGAs 5,844 5,512 94 4,129 71 1,854 322004–2009 Birth Cohort 89,308 78,001 87 55,871 63 24,170 27Non-TGAs 51,319 42,957 84 26,737 52 9,103 18TGAs 37,989 35,044 92 29,134 77 15,067 40a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership service usethat is slightly different than was reported in the year 3 report due to updated activity data.bThe birth of twins, triplets, and other multiples are counted as one birth event. Of the birth events in Palm Beach County from2004 through 2009, 1.9 percent had more than a single birth; 1.5 percent of TGA birth events had more than a single birth.cThis is not a measure of how many mothers were served by the HB system, but rather how many families had at least somecontact with the HB system.d This is a measure of how many families were served at least once by the HB system.e This is a measure of how many families were served with a treatment activity, as defined by correspondence between CSC andChapin Hall as the following activity codes: 0,11,12,13,14,15,16,17,18,19,2021,40,41,42,43,44,45,46,47,48,49,50,51,3320,3321,3322,3323,4501,6515,6516,8002,8004,8006,8008, 8013, 8026.Chapin Hall at the University of Chicago 172

Figure B- 5. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 aa Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership service usethat is slightly different than was reported in the year 3 report due to updated activity data.bThe birth of twins, triplets, and other multiples are counted as one birth event. Of the birth events in Palm Beach County from2004 through 2009, 1.9 percent had more than a single birth; 1.5 percent of TGA birth events had more than a single birth.cThis is not a measure of how many mothers were served by the HB system, but rather how many families had at least somecontact with the HB system.d This is a measure of how many families were served at least once by the HB system.e This is a measure of how many families were served with a treatment activity, as defined by correspondence between CSC andChapin Hall as the following activity codes: 0,11,12,13,14,15,16,17,18,19,2021,40,41,42,43,44,45,46,47,48,49,50,51,3320,3321,3322,3323,4501,6515,6516,8002,8004,8006,8008, 8013, 8026.Chapin Hall at the University of Chicago 173

Figure B- 6. Percentage of PBC Mothers in MCHP FOCiS Data System, 2004–2009 aa Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership service usethat is slightly different than was reported in the year 3 report due to updated activity data.bThe birth of twins, triplets, and other multiples are counted as one birth event. Of the birth events in Palm Beach County from2004 through 2009, 1.9 percent had more than a single birth; 1.5 percent of TGA birth events had more than a single birth.cThis is not a measure of how many mothers were served by the HB system, but rather how many families had at least somecontact with the HB system.d This is a measure of how many families were served at least once by the HB system.e This is a measure of how many families were served with a treatment activity, as defined by correspondence between CSC andChapin Hall as the following activity codes: 0,11,12,13,14,15,16,17,18,19,2021,40,41,42,43,44,45,46,47,48,49,50,51,3320,3321,3322,3323,4501,6515,6516,8002,8004,8006,8008, 8013, 8026.Chapin Hall at the University of Chicago 174

Table B- 5. Logistic Regression Predicting Likelihood of Receiving Treatment aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 1.1 NSMother used WIC while pregnant 2.1 ***Mother used tobacco while pregnant b 1.3 ***Late, unknown, or no prenatal care c 1.2 ***Mother obtained less than high school diploma 2.1 ***Mother unmarried 2.2 ***Mother younger than 20 1.9 ***Maternal health problem during pregnancy d 1.2 ***Child complications at delivery e 1.4 ***Low birth weight f 1.0 NSFocal child born premature 1.0 NSGlades residency 2.1 ***Race/nativityBlack—Foreign-born 0.7 ***Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born 0.3 ***Black—U.S.-born 0.3 ***Hispanic—U.S.-born 0.3 ***White/other—U.S.-born 0.2 ***Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 0.9 ***Focal child was born in 2006 0.9 *Focal child was born in 2007 0.9 **Focal child was born in 2008 0.9 ^Focal child was born in 2009 0.6 ***χ 2 (22, N=106,327) 28466.70Pseudo R 2 0.2349a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c “Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationseVentilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complicationsf Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 175

Table B- 6. Characteristics of Mothers in the TGA Birth Cohorts aTGA 2004–2005 TGA 2006 TGA 2007 TGA 2008 TGA 2009 2004–2009 Birth CohortMaternalCharacteristic b BirthEvents(n=12,435)MCHPTreatment(n=5,030)BirthEvents(n=6,716)MCHPTreatment(n=2,762)BirthEvents(n=6,629)MCHPTreatment(n=2,776)BirthEvents(n=6,365)MCHPTreatment(n=2,645)BirthEvents(n=5,844)MCHPTreatment(n=1,854)BirthEvents(n=37,989)MCHPTreatment(n=15,067)% % % % % % % % % % % %Unmarried mother 57 72 60 73 61 75 62 75 62 76 60 74Teen mother 14 21 14 21 14 21 13 19 12 20 13 21< High schooleducation35 52 38 57 37 58 36 58 36 54 36 55Black c 36 38 35 32 37 36 37 36 38 39 36 36Hispanic c 39 51 43 58 42 56 42 55 39 51 41 54Foreign-born 47 60 48 63 48 61 48 61 46 57 47 61Low birth weight 9 12 9 11 10 10 10 11 10 11 10 11a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “at risk” and mothers residing in theGlades TGA. This data reports Maternal Child Health Partnership service use that is slightly different than was reported in the year 3 report due to updated activity data. The birthof twins, triplets, and other multiples are counted as one birth event.b Does not include mothers who responded “I don’t know” to any of the vital statistics questionsc To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.Chapin Hall at the University of Chicago 176

Figure B- 7. Characteristics of Mothers in TGA Birth Cohort aaSource: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversampling ofmothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership risk screeningthat is slightly different than was reported in the year 3 report due to updated score data. The birth of twins, triplets, and othermultiples are counted as one birth event.b To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.cTreatment activity was defined by correspondence between CSC and Chapin Hall as the following activity codes: 10, 11, 12, 13,14, 15, 16, 17, 18, 19, 20, 21, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 3320, 3321, 3322, 3323, 4501, 6515, 6516, 8002,8004, 8006, 8008, 8013, 8026.Chapin Hall at the University of Chicago 177

Figure B- 8. Characteristics of Mothers in TGA Birth Cohort with MCHP Service Treatment aa Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership risk screeningthat is slightly different than was reported in the year 3 report due to updated score data. The birth of twins, triplets, and othermultiples are counted as one birth event.b To be consistent with the racial/ethnic categories in the survey, mothers who identified as Haitian are coded as black.c Treatment activity was defined by correspondence between CSC and Chapin Hall as the following activity codes: 10, 11, 12,13, 14, 15, 16, 17, 18, 19, 20, 21, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 3320, 3321, 3322, 3323, 4501, 6515, 6516, 8002,8004, 8006, 8008, 8013, 8026.Chapin Hall at the University of Chicago 178

Table B- 7. Characteristics of TGA Birth Cohort by Type of Services in MCHP System aTGA population with TreatmentMaternal CharacteristicActivityBirth With MCHP Unmarried Teen < HSForeignbornweightLow birthNBlack HispanicEvents Treatment Mother Mother Education% % % % % % % % %2004–2005 12,435Care coordination only b 2,676 22 53 65 18 44 37 49 62 10Intensive care coordination c 2,354 19 47 78 25 61 39 53 57 13Family support planning d 1,176 9 23 79 26 56 46 44 49 142006 6,716Care coordination only b 1,440 21 52 67 17 50 31 58 66 10Intensive care coordination c 1,320 20 48 79 25 65 34 57 60 12Family support planning d 638 9 23 114 42 89 51 74 80 202007 6,629Care coordination only b 1,489 22 54 69 16 51 33 57 65 9Intensive care coordination c 1,287 19 46 82 27 65 39 55 57 12Family support planning d 633 10 23 83 30 63 45 47 48 122008 6,365Care coordination only b 596 9 23 61 12 31 39 40 52 7Intensive care coordination c 2,049 32 77 80 21 66 34 59 64 12Family support planning d 453 7 17 84 27 61 52 39 43 152009 5,844Care coordination only b 639 11 34 66 13 46 28 56 63 8Intensive care coordination c 1,215 21 66 82 24 59 44 48 54 12Family support planning d 283 5 15 84 27 63 46 44 47 14All Years 37,989Care coordination only b 6,840 18 45 66 16 46 34 53 62 9Intensive care coordination c 8,225 22 55 80 24 63 38 55 59 12Family support planning d 3,183 8 21 81 28 61 45 46 49 14a Source: Vital Statistics and FOCiS (2004–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “at risk” and mothers in the Glades TGA. Data on MCHPservices differ slightly from year 3 report due to updated activity data. The birth of twins, triplets, and other multiples counted as one birth event.b Includes mothers who consented received care coordination only and not intensive services. Care coordination includes the following activity codes: 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 40,41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 3320, 3321, 4501, 6515, 6516, 8002, 8004, 8006, 8008, 8013, and 8026 provided by the following agencies: BANK, Healthy Mothers/Healthy Babies Coalition,Healthy Start/Healthy Family Nurses, Oakwood, Parent-Child Center, Center for Family Services, Planned Parenthood, HUGS, and Comprehensive AIDS Program.c Includes mothers who received intensive care coordination under the same activity codes listed above, but these services are provided by the following agencies: Healthy Mothers/Healthy BabiesCoalition, Nurture the Future, NOAH, Families First, Guatemalan Mayan Center, Sickle Cell Foundation, Haitian American Council, Esereh Youth and Family Center, American Lung Association,WHIN, and Minority Development and Empowerment, Inc.d Family support planning (FSP) services includes activity codes 3321 and 3322 provided by any agency.Chapin Hall at the University of Chicago 179

Table B- 8. Logistic Regression Predicting Likelihood of Receiving Intensive Care Coordination aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 2.9 ***Mother used WIC while pregnant 1.1 *Mother used tobacco while pregnant b 1.6 ***Late, unknown, or no prenatal care c 1.4 ***Mother obtained less than high school diploma 2.0 ***Mother unmarried 1.5 ***Mother younger than 20 1.5 ***Maternal health problem during pregnancy d 1.2 ***Child complications at delivery e 1.0 NSLow birth weight f 1.2 **Focal child born premature 1.3 ***Glades residency 1.3 ***Race/nativityBlack—Foreign-born 0.9 **Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born 0.5 ***Black—U.S.-born 1.4 ***Hispanic—U.S.-born 0.8 ***White/other—U.S.-born 0.6 ***Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 1.0 NSFocal child was born in 2006 1.1 NSFocal child was born in 2007 1.0 NSFocal child was born in 2008 5.0 ***Focal child was born in 2009 3.0 ***χ 2 (22, N=24,163) 4251.76Pseudo R 2 0.1613a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c“Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationse Ventilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complications.f Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 180

Table B- 9. Logistic Regression Predicting Likelihood of Receiving Care Coordination Only aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 0.3 ***Mother used WIC while pregnant 0.9 *Mother used tobacco while pregnant b 0.6 ***Late, unknown, or no prenatal care c 0.7 ***Mother obtained less than high school diploma 0.5 ***Mother unmarried 0.6 ***Mother younger than 20 0.7 ***Maternal health problem during pregnancy d 0.8 ***Child complications at delivery e 1.0 NSLow birth weight f 0.8 **Focal child born premature 0.8 ***Glades residency 0.8 ***Race/nativityBlack—Foreign-born 1.1 **0Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born 1.9 ***Black—U.S.-born 0.7 ***Hispanic—U.S.-born 1.3 ***White/other—U.S.-born 1.7 ***Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 1.0 NSFocal child was born in 2006 0.9 NSFocal child was born in 2007 1.0 NSFocal child was born in 2008 0.2 ***Focal child was born in 2009 0.3 ***χ 2 (22, N=24,163) 4251.76Pseudo R 2 0.1613a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c “Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationseVentilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complications.f Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 181

Table B- 10. Use of Enhanced Services by Mothers in TGA Birth Cohorts aType of Enhanced Service2004–2005200620072008N% of Mothersreceiving treatmentnPrenatal bnPostnatal cTranslation 1,327 26 425 1,198Breastfeeding 573 11 124 521Parenting support and education 262 5 63 222Psychosocial/depression counseling 207 4 52 183Nutrition 164 3 68 107Childbirth education 100 2 68 39Interconceptional education/counseling 37 1 4 34Smoking cessation 21 0 10 12Male support services 12 0 - 12Translation 1,292 47 524 1,178Breastfeeding 441 16 120 390Psychosocial/depression counseling 141 5 69 105Parenting support and education 104 4 31 85Nutrition 78 3 40 64Childbirth education 54 2 26 39Interconceptional education/counseling 31 1 6 25Male support services 31 1 9 28Smoking cessation 5 0 5 -Translation 1,314 47 549 1,175Breastfeeding 475 17 122 411Psychosocial/depression counseling 152 5 51 121Parenting support and education 90 3 49 66Nutrition 78 3 46 70Childbirth education 62 2 33 49Male support services 44 2 21 38Smoking cessation 5 0 3 2Interconceptional education/counseling 1 0 - 1Translation 1,092 41 478 933Breastfeeding 199 8 71 156Psychosocial/depression counseling 197 7 97 167Nutrition 75 3 55 62Parenting support and education 74 3 38 58Childbirth education 46 2 37 27Male support services 44 2 25 34Interconceptional education/counseling 25 1 - 25Smoking cessation 3 0 2 2Chapin Hall at the University of Chicago 182

Type of Enhanced Service2009N% of Mothersreceiving treatmentnPrenatal bnPostnatal cTranslation 435 23 259 263Psychosocial/depression counseling 173 9 116 121Breastfeeding 78 4 39 46Parenting support and education 48 3 5 45Nutrition 29 2 18 18Childbirth education 10 1 5 6Male support services 7 0 4 4Interconceptional education/counseling 6 0 - 6Smoking cessation 5 0 3 22004–2009 Birth CohortTranslation 5,535 36 2,254 4,821Breastfeeding 594 4 190 489Psychosocial/depression counseling 889 6 390 714Parenting support and education 1,822 12 487 1,579Childbirth education 283 2 177 163Nutrition 428 3 230 323Interconceptional education/counseling 101 1 10 92Smoking cessation 40 0 24 18Male support services 138 1 59 116a Source: Vital Statistics and FOCiS (2003–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership activity that isslightly different than was reported in the year 3 report due to updated activity data. The birth of twins, triplets, and othermultiples are counted as one birth event. Enhanced services were defined through conversations with CSC and are defined byactivity codes 10, 11, 12, 13, 14, 17, 18, 19, 20, 21, 40, 41, 42, 43, 44, 47, 48, 49, 50, 51, 4501, 6515, 6516, 8002, 8004, 8006,8008, 8013, 8026, provided by any agency.bPrenatal activity includes any activity attributed to the child’s mother occurring 300 days or less before the child’s birth.cPostnatal activity includes any activity attributed to a child’s mother occurring from a child’s birth up to the birth of asubsequent sibling. It also includes any activity attributed to the child.Chapin Hall at the University of Chicago 183

Table B- 11. MCHP Prenatal and Postnatal Services and Referrals, 2004–2009 aCohort Yearand Service Type2004–2005 12,435TGA Population Prenatal Care b Postnatal Care cNMeanDays(SD) n %MeanDays(SD) n %Any treatment activity 5,030 16.3 24.0 3,230 64 5.0 8.0 4,413 88 9.9 16.8Care coordination only d 2,676 4.1 3.7 1,472 55 1.9 2.6 2,248 84 2.2 2.5Intensive care coordination e 2,354 30.1 29.2 1,759 75 8.6 10.2 2,166 92 21.5 24.6Family support planning 1,176 2.3 3.0 626 53 0.7 1.3 961 82 1.6 2.2Enhanced services f 1,830 4.5 6.9 587 32 0.7 1.7 1,639 90 3.8 6.52006 6,716Any treatment activity 2,762 14.9 21.6 1,762 64 4.9 8.4 2,403 87 10.0 16.5Care coordination only d 1,429 4.4 4.9 782 55 2.2 3.3 1,186 83 2.2 3.2Intensive care coordination e 1,333 26.4 26.4 981 74 7.9 10.8 1,220 92 18.5 20.4Family support planning 639 2.1 1.4 321 50 0.7 0.8 534 84 1.4 1.1Enhanced services f 1,559 4.7 6.7 659 42 1.3 3.0 1,395 89 3.4 5.12007 6,629Any treatment activity 2776 14.1 20.5 1,734 62 4.8 7.9 2,451 88 9.3 15.6Care coordination only d 1,488 4.1 3.9 855 57 2.2 2.9 1,224 82 1.9 2.2Intensive care coordination e 1,288 25.6 25.2 879 68 7.8 10.4 1,227 95 17.8 19.6Family support planning 633 1.9 1.4 383 61 0.8 0.8 477 75 1.2 1.1Enhanced services f 1,569 4.7 6.6 675 43 1.3 2.5 1,384 88 3.4 5.42008 6,365Any treatment activity 2,645 11.9 16.8 1,831 69 4.8 7.3 2,149 81 7.1 12.3Care coordination only d 596 4.7 3.8 551 92 3.6 2.9 349 59 1.1 1.9Intensive care coordination e 2,049 14.1 18.4 1,280 62 5.2 8.1 1,800 88 8.9 13.4Family support planning 453 1.8 1.3 263 58 0.7 0.8 332 73 1.1 1.1Enhanced services f 1,329 5.2 7.2 604 45 1.5 3.3 1,136 85 3.6 5.32009 5,844Any treatment activity 1,854 9.5 12.2 1,412 76 4.8 6.4 1,182 64 4.7 8.2Care coordination only d 639 5.1 4.3 611 96 4.2 3.4 266 42 0.9 1.7Intensive care coordination e 1,215 11.8 14.2 801 66 5.1 7.5 916 75 6.6 9.5Family support planning 283 1.5 0.8 129 46 0.5 0.6 201 71 1.0 0.8Enhanced services f 661 4.7 6.2 388 59 2.1 3.6 438 66 2.6 4.4All Years 37,989Any treatment activity 15,067 14.0 20.7 9,969 66 4.9 7.8 12,59 84 9.1 16.28Care coordination only d 6,828 4.3 4.1 4,271 63 2.4 3.0 5,273 77 1.9 2.5Intensive care coordination e 8,239 22.1 25.0 5,700 69 7.0 9.6 7,329 89 15.1 19.8Family support planning 3,184 2.0 2.1 1,722 54 0.7 1.0 2,505 79 1.4 1.6Enhanced services f 6,948 4.7 6.8 2,913 42 1.3 2.8 5,992 86 3.5 5.6a Source: Vital Statistics and FOCiS (2003–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “atrisk” and mothers in the Glades TGA. Results on MCHP services differ slightly from year 3 report due to updated activity data. Twin, triplet, andother multiple births are counted as one birth event. For readability, standard deviations for means are not included; these are available uponrequest.b Prenatal activity includes any activity attributed to the child’s mother occurring 300 days or less before the child’s birth.c Postnatal activity includes any activity attributed to a mother (or child) occurring from a child’s birth up to the birth of a subsequent sibling.d “Care coordination only” days are counts of treatment days for mothers who only receive care coordination not intensive care coordination.e “Days of service” for intensive care coordination births counts all days of treatment activity (whether intense or not) for births that receive atleast one day of intensive care coordinationf Enhanced services were defined through conversations with CSC and are defined by activity codes 10, 11, 12, 13, 14, 17, 18, 19, 20, 21, 40, 41,42, 43, 44, 47, 48, 49, 50, 51, 4501, 6515 , 6516, 8002, 8004, 8006, 8008, 8013, 8026, provided by any agency.MeanDaysChapin Hall at the University of Chicago 184(SD)

Table B- 12. Logistic Regression Predicting Likelihood of Receiving Enhanced Services aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 1.3 NSMother used WIC while pregnant 1.0 NSMother used tobacco while pregnant b 1.0 NSLate, unknown, or no prenatal care c 1.1 **Mother obtained less than high school diploma 2.1 ***Mother unmarried 1.4 ***Mother younger than 20 0.9 *Maternal health problem during pregnancy d 1.2 ***Child complications at delivery e 1.2 **Low birth weight f 1.1 ^Focal child born premature 1.1 NSGlades residency 1.1 ***Race/nativityBlack—Foreign-born 0.4 ***Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born 0.2 ***Black—U.S.-born 0.1 ***Hispanic—U.S.-born 0.2 ***White/other—U.S.-born 0.1 ***Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 3.2 ***Focal child was born in 2006 5.2 ***Focal child was born in 2007 5.1 ***Focal child was born in 2008 3.8 ***Focal child was born in 2009 2.0 ***χ 2 (22, N=24,163) 7259.51Pseudo R 2 0.2595a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c “Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationseVentilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complications.f Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 185

Table B- 13. Outside Service Referrals for 2004–2009 TGA Cohorts Who Received TreatmentServices a 2004–2005 TGA 2006 TGANumberof BirthswithReferralsN=5,030Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNumberof BirthswithReferralsN=2,762Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNutrition and baby supplies 842 434 561 522 239 361Transportation 462 210 344 437 185 326Parenting or health education 398 146 305 320 112 252Medical referrals 323 133 217 188 56 145Counseling 214 87 154 132 70 78Financial services 217 92 146 99 38 71Access to care: insurance 231 126 116 129 71 67Educational services 320 122 218 178 60 138Housing 193 81 130 81 42 50Childcare 181 28 160 120 22 108Employment services 55 17 40 40 9 31Immigration services 32 9 24 15 8 7Domestic violence 32 6 28 22 7 17Teenage-targeted programs 30 14 16 18 5 14Other 754 283 614 587 199 490Numberof BirthswithReferrals2007 TGA 2008 TGAN=2,776Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNumberof BirthswithReferralsN=2,645Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNutrition and baby supplies 683 311 473 303 200 142Transportation 390 215 248 133 72 89Parenting or health education 409 156 301 172 65 114Medical referrals 220 76 159 132 46 92Counseling 115 61 66 41 25 17Financial services 119 35 97 71 24 54Access to care: insurance 115 62 58 51 24 29Educational services 210 78 156 59 35 30Housing 85 47 48 28 12 17Childcare 108 25 89 30 10 21Employment services 50 14 37 19 8 11Immigration services 11 2 9 7 3 5Domestic violence 28 10 19 4 1 3Teenage-targeted programs 42 13 31 10 7 3Other 545 227 428 170 74 119Chapin Hall at the University of Chicago 186

Numberof BirthswithReferrals2009 TGA All YearsN=1,854Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNumberof BirthswithReferralsN=15,067Numberof BirthswithPrenatalReferrals bNumber ofBirths withPostnatalReferrals cNutrition and baby supplies 195 152 45 2545 1336 1582Transportation 54 15 46 1476 697 1053Parenting or health education 61 16 46 1360 495 1018Medical referrals 50 8 43 913 319 656Counseling 20 10 10 522 253 325Financial services 25 6 19 531 195 387Access to care: insurance 4 1 3 530 284 273Educational services 41 8 33 808 303 575Housing 11 7 5 398 189 250Childcare 18 3 15 457 88 393Employment services 9 2 7 173 50 126Immigration services - - - 65 22 45Domestic violence 4 2 2 90 26 69Teenage-targeted programs - - - 100 39 64Other 95 37 59 2151 820 1710a Source: Vital Statistics and FOCiS (2003–2009). Data for the year 5 study sample were weighted to account for oversamplingof mothers “at risk” and mothers residing in the Glades TGA. This data reports Maternal Child Health Partnership referrals thatare slightly different than were reported in the year 3 report due to updated referral data. The birth of twins, triplets, and othermultiples are counted as one birth event.b Prenatal referrals include any referrals attributed to the child’s mother occurring 300 days or less before the child’s birth.c Postnatal referrals include any referrals attributed to a child’s mother occurring from a child’s birth up to the birth of asubsequent sibling. It also includes any referrals attributed to the child.Chapin Hall at the University of Chicago 187

Table B- 14. Logistic Regression Predicting Likelihood of Receiving Family Support Planning aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 2.0 *Mother used WIC while pregnant 1.2 ***Mother used tobacco while pregnant b 2.0 ***Late, unknown, or no prenatal care c 1.2 ***Mother obtained less than high school diploma 1.5 ***Mother unmarried 1.6 ***Mother younger than 20 1.3 ***Maternal health problem during pregnancy d 1.2 ***Child complications at delivery e 1.0 NSLow birth weight f 1.1 NSFocal child born premature 1.2 ***Glades residency 1.3 ***Race/nativityBlack—Foreign-born 1.3 ***Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born 0.6 ***Black—U.S.-born 2.0 ***Hispanic—U.S.-born 1.3 ***White/other—U.S.-born 1.2 **Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 1.0 NSFocal child was born in 2006 1.0 NSFocal child was born in 2007 0.9 ^Focal child was born in 2008 0.6 ***Focal child was born in 2009 0.6 ***χ 2 (22, N=24,163) 1371.26Pseudo R 2 0.0552a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c “Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationseVentilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complications.f Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 188

Table B- 15. OLS Logistic Regression Predicting Days of Treatment Services aPredictor Variable Odds Ratio Sig.Baseline CharacteristicsMother drank alcohol while pregnant 1.4 **Mother used WIC while pregnant 1.4 ***Mother used tobacco while pregnant b 0.4 **Late, unknown, or no prenatal care c 0.2 **Mother obtained less than high school diploma 2.7 ***Mother unmarried 1.5 ***Mother younger than 20 1.8 ***Maternal health problem during pregnancy d 0.7 ***Child complications at delivery e 0.4 ***Low birth weight f 0.2 NSFocal child born premature 0.0 NSGlades residency 1.6 ***Race/nativityBlack—Foreign-born -1.8 ***Hispanic—Foreign-born (excluded category) — —White/other—Foreign-born -2.1 ***Black—U.S.-born -2.0 ***Hispanic—U.S.-born -2.6 ***White/other—U.S.-born -2.2 ***Year of Focal Child’s BirthFocal child was born in 2004 (excluded category) — —Focal child was born in 2005 -0.2 ^Focal child was born in 2006 -0.4 ***Focal child was born in 2007 -0.6 ***Focal child was born in 2008 -1.0 ***Focal child was born in 2009 -1.7 ***Pseudo R 2 0.1004N 106,327a Weights were not applied to this analysis.b Tobacco use does not include those who indicated that they were “unsure” if they used tobacco, but did include those who quitwhile pregnant.c “Late” prenatal care is prenatal care that began in the third trimester.d Diabetes, hypertension, previous preterm birth, previous poor pregnancy outcome, pregnancy result of infertility treatments, or“other” health complicationseVentilator required, newborn given antibiotics, significant birth injury, hyaline membranes disease, NICU admission, seizures,serious neurologic dysfunction, surfactant given to newborn, or “other” delivery complications.f Less than 2500 grams^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 189

Appendix C: Additional Data onYear 5 Study SampleChapin Hall at the University of Chicago 190

Table C- 1. Selected Characteristics of Mothers by Nativity aCharacteristic at Year 5All Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)FOCiS initial risk screen (%)At risk/high need screen score 31 29 33Study Risk Index***Mean (SD) 3.5 (1.5) 3.9 (1.4) 3.1 (1.6)Mother’s race (%)***Hispanic 51 77 20Black, not Hispanic 41 20 65Age of mother***Mean age (SD) 29.0 (5.7) 31.0 (6.0) 26.8 (4.5)Teen mother at focal child’s birth (%)*** 15 7 24Mother’s education (%)***High school graduate/GED (or more) 49 34 68Marital status (%)***Married, living with husband 33 42 22Number of children (%)One 25 27 22Two 32 29 35Employment (%)Mother currently working 46 45 47Husband/partner working (n=229)*** 85 93 73Main language spoken in home (%)***English 50 15 92Spanish 43 73 7Haitian Creole 6 10 1Income-to-need ratio (%)Living at or below poverty threshold 52 56 48Income support (%)WIC*** 52 65 36Food Stamps** 47 38 57Home ownership (%)Mother or other household member owns home 29 29 28a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or *** p < .001.Chapin Hall at the University of Chicago 191

Table C- 2. Selected Health Characteristics of Mothers by Nativity a,bCharacteristic at Year 5All Mothers Foreign-born U.S.-born(N=310) (n=168) (n=142)Health insuranceMother covered*** 45 19 75All or some children covered** 87 83 93Health statusMother’s health good/very good/excellent** 83 77 90Focal child’s health good/very good/excellent* 95 92 98Focal child has special medical needs 19 16 22One or more well-baby check-ups 80 80 81Depression: CES-D score > 16 21 19 22a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Previous versions of this table displayed U.S.-born mothers, foreign-born mothers who had been in the U.S. for 5 or more yearsand foreign-born mothers who had been in the U.S. for less than 5 years. At the time of the year 5 in-person survey, just nineforeign-born mothers had been in the U.S. for less than 5 years.Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Table C- 3. Services, Economic Supports, and Social Support by Nativity a,bCharacteristic at Year 5All Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)Income supportsWIC*** 52 65 36Food Stamps** 47 38 57Earned Income Tax Credit* 35 29 42SSI (Social Security Disability)*** 11 2 21TANF 2 1 3Rent voucher*** 11 2 21Unemployment insurance** 9 4 15Attendance at religious servicesOnce a month or more often 52 56 47Use of formal servicesFive or more** 33 24 43ChildcareFocal child in childcare** 62 54 71Other preschool child in childcare (n=146)** 42 27 56Childcare subsidy for focal child 32 32 31Childcare subsidy for other children (n=74) 42 43 41a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Previous versions of this table displayed U.S.-born mothers, foreign-born mothers who had been in the U.S. for 5 or more yearsand foreign-born mothers who had been in the U.S. for less than 5 years. At the time of the year 5 in-person survey, just nineforeign-born mothers had been in the U.S. for less than 5 years.Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 192

Table C- 4. Social and Emotional Behaviors of Focal Children at Year 4 and Year 5 aSocial-Emotional IndicatorPositive characteristics% Mothers (N=310)Year 4 Year 5Focal child seems happy 100 100Focal child is eager to learn new things 100 99Focal child likes to try new things 100 98Focal child is accepted and liked by other children 99 99Focal child likes playing with other children close to his/her age 99 98Focal child pays attention well 97 94Focal child adjusts easily to new situations 97 93Focal child helps or cooperates with adults 96 97Focal child finishes what he/she is asked to do 93 90Negative characteristicsFocal child is overly active and unable to sit still 79 72Focal child gets upset easily 75 74Focal child worries about things 62 57Focal child does things without thinking 52 43Focal child is aggressive 40 33aData were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Percentagesindicate the total of mothers who responded “sometimes” and “most of the time” on a 3-point scale.*Paired sample t-tests indicated that these year-to-year differences are statistically significant at (p < .05): paysattention4 vs.paysattention5; adjusts4 vs. adjusts5; newthings4 vs. newthings5; active4 vs. active5; aggressive4 vs. aggressive5;withoutthinking4 vs. withoutthinking5.Chapin Hall at the University of Chicago 193

Table C- 5. Social and Emotional Behaviors of Focal Children by Nativity at Year 5 aSocial-Emotional Indicator b All Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)Positive behaviorsFocal child seems happy 99 100 99Focal child is eager to learn new things 99 99 100Focal child likes to try new things 98 99 97Focal child is accepted and liked by other children 99 99 99Focal child likes playing with other children his/her age* 98 100 97Focal child pays attention well 94 95 93Focal child adjusts easily to new situations 94 92 95Focal child helps or cooperates with adults 97 96 98Focal child finishes what he/she is asked to do 90 91 90Negative behaviorsFocal child is overly active and unable to sit still*** 72 83 60Focal child gets upset easily 74 77 71Focal child worries about things*** 57 73 38Focal child does things without thinking* 43 49 35Focal child is aggressive 33 34 32a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bBased on a 3-point scale: “almost never,” “sometimes,” and “almost all the time.” Percentages combine “sometimes” and“almost all the time” responses.Chi-square analysis indicate differences between foreign-born and native-born mothers are statistically significant at *p < .05 or***p < .001.Table C- 6. Communication Skills of Focal Children at Year 4 and Year 5 aCommunication Skill% Mothers atYear 4% Mothers atYear 5(N=310)(N=310)Focal child is a good listener 97 93Focal child speaks clearly so strangers understand 92 92Focal child refers to him/her -self as “I” 87 91Focal child uses appropriate social greetings 79 88Focal child waits his/her turn to talk 73 69Focal child can say his/her first and last name 61 89a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Percentagesindicate the total of mothers who responded “sometimes” and “most of the time” on a 3-point scale.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant: greetings4 vs. greetings5;listener4 vs. listener5; names4 vs. names5.Chapin Hall at the University of Chicago 194

Table C- 7. Communication Skills of Focal Children by Nativity at Year 5 aAll Mothers Foreign-born U.S.-bornCommunication Skills(N=310) (n=167) (n=143)Focal child is a good listener* 93 96 90Focal child speaks clearly so strangers understand^ 92 90 95Focal child refers to him/her -self as “I”* 91 94 87Focal child uses appropriate social greetings^ 88 85 92Focal child waits his/her turn to talk 69 69 68Focal child can say his/her first and last name^ 89 86 92a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Percentages in thistable combine “sometimes” and “almost all the time” responses.Chi-square analysis indicate that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or ***p < .001.Table C- 8. Language Skills over Time by Sex of Focal Child a% All Children % by Sex of ChildHow ChildYear 3 Year 3 Year 4 Year 4 Year 5 Year 5Year 3 Year 4 Year 5CommunicatesBoys Girls Boys Girls Boys Girls(n=305) (n=306) (n=306) (n=167) (n=139) (n=165) (n=140) (n=166) (n=140)Does not talkyet1 0 0 2 0** 1 0 0 0**Mostly 1-wordsentences23 8 2 29 17 10 6 2 22- to 3-wordphrases50 20 5 52 49 21 19 7 3Fairlycomplete, short 23 37 23 17 31 37 36 31 14sentencesLong,complicated 2 35 70 1 4 31 40 60 81sentencesa Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Four to fivemothers did not respond to these questions.Chi-square tests at year 3 and at year 5 indicate that girls are significantly more likely to use complete sentences than boys (p

Table C- 10. Preliteracy and Pre-academic Skills of Focal Children at Year 4 and Year 5 a% Mothers at Year 4 % Mothers at Year 5Preliteracy/Pre-academic Behaviors(N=310)(N=310)Child scribbles or draws on paper 98 99Child looks at picture books on his/her own 98 99Child points to pictures while looking at picture books 97 98Child can count 93 99Child says the names of the colors 90 97Child points to colors when you say the names 84 96Child says the names of the letters 81 90Child pretends to read the words in a book* 76 91Child can tell what is in each picture in a picture book* 74 98Child can draw one or more shapes that you can recognize* 69 82Child points to letters of the alphabet when you say thenames*57 79Child tells you what he/she has drawn* 56 89Child tries to draw shapes, numbers, or letters* 49 88Child can draw one or more letters that you can recognize* 42 67Child draws pictures of people or objects* 39 72Child reads the written words in a book* 9 28a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.” Percentagesindicate the total of mothers who responded “sometimes” and “most of the time” on a 3-point scale.*Paired sample t-tests indicate that the following year-to-year differences are statistically significant at p < .05 or lower:pretends4 vs. pretends5; reads4 vs. reads5; tellpicture4 vs. tellpicture5; pointscolors4 vs. pointscolors5; sayscolors4 vs.sayscolors5; pointsletters4 vs. pointsletters5; saysletters4 vs. saysletters5; counts4 vs. counts5; drawspictures4 vs.drawspictures5; tellinwords4 vs. tellinwords5; triestodraw4 vs. triestodraw5; drawshapesrecognize4 vs. drawshapesrecognize5;drawlettersrecognize4 vs. drawlettersrecognize5.Chapin Hall at the University of Chicago 196

Table C- 11. Emerging Literacy and Pre-academic Skills of Focal Children by Nativity at Year 5 aPreliteracy/Pre-academic Behaviors% AllMothers% Foreignborn% U.S.-born(N=310) (n=167) (n=143)Child scribbles or draws on paper 99 99 99Child looks at picture books on his/her own b 99 99 99Child points to pictures while looking at picture books b 98 99 98Child can count up to 10 c 99 98 99Child says the names of the colors (red, yellow, blue, green) d 97 96 99Child points to colors when you say the names (red, yellow, blue,green) d 96 95 97Child can say most of the names of the letters d *** 90 83 98Child pretends to read the words in a book b 91 90 92Child can tell what is in each picture in a picture book b 98 98 99Child can draw one or more shapes that you can recognize c * 82 76 88Child points to letters of the alphabet when you say the names b ** 79 73 86Child tells you what he/she has drawn 89 86 92Child tries to draw shapes, numbers, or letters 88 86 92Child can draw one or more letters that you can recognize^ 67 62 73Child draws pictures of people or objects 72 69 75Child reads the written words in a book b * 29 23 35a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bPercentages indicate mothers who responded “sometimes” or “almost all the time” (versus “almost never”) on a 3-point scale.cPercentages indicate mothers who responded “yes” (versus “no) to item.dMothers were asked whether child can point or say “all,” “most,” “some,” or “none” of the names, colors, and letters.Chi-square analysis indicates that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 197

Table C- 12. Access to Reading Materials at Home by Nativity at Year 5 aAll Mothers Foreign-born U.S.-bornFocal child has regular access to(N=310) (n=167) (n=143)Books for children 97 96 99Religious books like a bible or prayer book 87 86 87Dictionaries or encyclopedias*** 65 54 77Other books (e.g., novels, biographies or nonfiction)***48 32 66Comic books*** 47 64 26Magazines for children 43 48 37Catalogs 42 40 44Newspapers** 40 32 50Books from the library 39 34 44Magazines for adults (e.g., Newsweek, People, orSports Illustrated)***33 23 46Mean number (SD) of children’s books focal childhas access to in home now, including library24.2 (33.6) 18.9 (27.0) 30.6 (39.1)books**Mean number (SD) of hours focal child watchestelevision and/or movies per day***2.8 (1.6) 2.5 (1.4) 3.1 (1.8)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”Chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Table C- 13. Mothers’ Positive Parenting Practices during Previous 3 Months by Nativity at Year 5 aActivityAll Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)Praised children 99 99 100Took children on errands (e.g., post office, bank, orstore)98 98 98Took children outside for walk or play 96 96 97Encouraged children to read a book^ 95 93 98Read books to children 95 94 96Sang songs with children 92 90 94Told stories to children 90 91 88Did household chores with children*** 79 68 91Played with a game, puzzle, or building toy withchildren*77 71 83Talked to children about a television program* 75 70 81Played with clay, drew pictures, or did other arts andcrafts***74 63 87Took children to library*** 43 30 58Mean (SD) Positive Parenting Score (range: 0–1) b ** 0.92 (.13) 0.90 (.14) 0.95 (.11)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b The mean parenting score is based on only the items included in all surveys so that comparisons could be made over time. Thefollowing items were excluded in calculation of the mean: “played with a game, puzzle, or building toy,” “did household chores,”“talked about a television program,” and “took child to library.”Chi-square analysis indicates that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 198

Table C- 14. Mothers’ Negative Parenting Practices during Previous 3 Months by Nativity at Year 5 aActivityAll Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)Lost temper with children** 60 69 49Found hitting/spanking a good way to get children to listen 25 23 27Got more angry than meant to with children 26 29 22Punished children for not finishing food on plate 9 10 8Blamed children for something not their fault^ 18 14 22Mean (SD) Negative Parenting Score (range: 0–1) b 0.19 (.23) 0.19 (.23) 0.20 (.23)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”bThe mean score uses only items included in all surveys; the item “lost temper with child(ren)” was excluded in calculating themean.Chi-square analysis indicates that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or ***p < .001.Table C- 15. Parent Involvement Activities during Previous 3 Months by Nativity at Year 5 aActivity b All Mothers Foreign-born U.S.-born(N=310) (n=167) (n=143)Talked to teacher about child’s progress (at times other thanparent-teacher conference)75 72 78Helped child with homework 72 68 76Attended parent-teacher conference* 62 70 53Participated in field trip or family event at school*** 44 33 56Attended PTA or other parent meetings 33 33 32Mean (SD) Parent Involvement Score (range: 0–1) c 0.57 (.30) 0.55 (.29) 0.59 (.30)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Only mothers who had children in childcare or school were asked about these activities.c Mean parent involvement score based on all items asked at each time point.Chi-square analysis indicates that differences are statistically significant at ^p < .10, *p < .05, **p < .01, or ***p < .001.Table 16. Social Support by Nativity a,b All Mothers Foreign-born U.S.-bornFrequency of Social Support at Year 5(N=310) (n=167) (n=143)Mean (SD) Mean (SD) Mean (SD)Husband/partner (scale range: 0–7) c 5.9 (0.9) 5.9 (0.8) 5.9 (1.1)Family/friend (scale range: 0–16)*** 4.3 (3.2) 3.5 (2.8) 5.1 (3.4)a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Previous versions of this table displayed U.S.-born mothers, foreign-born mothers who had been in the U.S. for 5 or more years,and foreign-born mothers who had been in the U.S. for less than 5 years. At the time of the year 5 in-person survey, just nineforeign-born mothers had been in the U.S. for less than 5 years.c Husband/partner n=231 (94 of U.S.-born mothers and 136 of foreign-born mothers).One-way ANOVA tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 199

Table C- 17. Number and Most Frequent Types of Focal Children’s Care Arrangements betweenBirth and 5 Years aNumber of Different Arrangements across Nine Time Points b Frequency PercentOne 43 15Child with parent 41 14Child with relative 1 0Child with friend/neighbor 1 0Two 30 11Child with parent and childcare center, Head Start, or prekindergarten 17 6Child with parent and relative 8 3Child with parent and friend/neighbor 3 1Child with parent and multiple/other arrangements 2 0Three 41 14Child with parent and relative 12 4Child with parent and friend/neighbor 7 2Child with parent, relative, and childcare center/Head Start 5 2Child with parent and childcare center/Head Start, pre-k, or family childcare 5 2Child with parent, friend/neighbor, and childcare center/Head Start 3 1Child with parent, relatives, and prekindergarten (school-based) 2 1Child with childcare center/Head Start, and family childcare 2 1Child with parent and other/multiple arrangements 2 1Child with parent, relatives, and friend/neighbor 2 1Child with relative, childcare center/Head Start, and multiple arrangements 1 0Four 43 15Child with parent, relative, and childcare center/Head Start/prekindergarten 12 4Child with parent and relative 4 1Child with parent, childcare center/Head Start, and family childcare 4 1Child with parent, relative, and friend/neighbor 4 1Child with parent and childcare center/Head Start 3 1Child with parent, childcare center/Head Start, and multiple arrangements 3 1Child with parent, relative, friend/neighbor, and childcare center/Head Start 2 1Child with parent and friend/neighbor 2 1Child with parent, relative, and multiple/other arrangements 2 1Child with parent, friend/neighbor, and childcare center/Head Start/prekindergarten 2 1Child with parent, friend/neighbor, and multiple arrangements 1 0Child with parent, relative, and family childcare 1 0Child with relative, childcare center/Head Start, and family childcare 1 0Child with relative, childcare center/Head Start, and prekindergarten (school-based) 1 0Child with relative, childcare center/Head Start, and multiple arrangements 1 0Five 49 17Child with parent, relative, and childcare center/Head Start 9 3Child with parent, relative, friend/neighbor, and childcare center/Head Start 6 2Child with parent, childcare center/Head Start, and family childcare 5 2Child with parent and relative 4 1Chapin Hall at the University of Chicago 200

Number of Different Arrangements across Nine Time Points b Frequency PercentChild with parent, relative, and friend/neighbor 4 1Child with parent and childcare center/Head Start 2 1Child with parent and friend/neighbor 2 1Child with parent, childcare center/Head Start, and prekindergarten 2 1Child with parent, friend/neighbor, and childcare center/Head Start 2 1Child with parent, relative, childcare center/Head Start, and multiple arrangements 2 1Child with parent, childcare center/Head Start, and multiple arrangements 1 0Child with childcare center/Head Start, family childcare, and multiple arrangements 1 0Child with parent and family childcare 1 0Child with parent, childcare center/Head Start, and other arrangements 1 0Child with parent, friend/neighbor, childcare center/Head Start, and prekindergarten 1 0Child with parent, friend/neighbor, childcare center/Head Start, and family childcare 1 0Child with parent, relative, and multiple arrangements 1 0Child with parent, relative, friend/neighbor, childcare center/Head Start, and otherarrangements1 0Child with parent, relative, childcare center/Head Start, and prekindergarten (schoolbased)1 0Child with relative, childcare center/Head Start/prekindergarten (school-based), andmultiple arrangements2 1Six to Eight 78 27Child with parent, relative, and childcare center/Head Start 8 3Child with parent, relative, childcare center/Head Start, family childcare, and multiplearrangements8 3Child with parent, relative, friend/neighbor, childcare center/Head Start, and familychildcare6 2Child with parent, relative, childcare center/Head Start, and family childcare 6 2Child with parent, relative, childcare center/Head Start, and multiple arrangements 5 2Child with parent, relative, friend/neighbor, and childcare center/Head Start 5 2Child with parent, relative, childcare center/Head Start, and prekindergarten (schoolbased)4 1Child with parent and relative 3 1Child with parent and friend/neighbor 3 1Child with parent, relative, and friend/neighbor 3 1Child with parent and childcare center/Head Start 2 1Child with parent, childcare center/Head Start, and family childcare 2 1Child with parent, childcare center/Head Start, family childcare, and multiplearrangements2 1Child with parent, relative, childcare center/Head Start, prekindergarten (schoolbased),and family childcare2 1Child with parent, relative, friend/neighbor, and multiple arrangements 2 1Child with parent, relative, friend/neighbor, childcare center/Head Start,prekindergarten (school-based), and multiple arrangements2 1Child with parent, friend/neighbor, and childcare center/Head Start 1 0Child with parent, relative, and prekindergarten (school-based) 1 0Child with parent, friend/neighbor, and other arrangements 1 0Child with parent, relative, family childcare, and multiple arrangements 1 0Chapin Hall at the University of Chicago 201

Number of Different Arrangements across Nine Time Points b Frequency PercentChild with parent, relative, friend/neighbor, childcare center/Head Start, familychildcare, and multiple arrangements1 0Child with parent, relative, friend/neighbor, family childcare, and multiplearrangements1 0Child with parent, relative, childcare center/Head Start, prekindergarten (schoolbased),and other arrangements1 0Child with parent, relative, childcare center/Head Start, prekindergarten (schoolbased),and multiple arrangements1 0Child with parent, friend/neighbor, and family childcare 1 0Child with parent, childcare center/Head Start, and prekindergarten (school-based) 1 0Child with parent, relative, friend/neighbor, and prekindergarten (school-based) 1 0Child with parent, childcare center/Head Start, prekindergarten (school-based), familychildcare, and multiple arrangements1 0Child with parent, relative, friend/neighbor, childcare center/Head Start, and multiplearrangements1 0Child with relative, friend/neighbor, childcare center/Head Start, family childcare, andother arrangements, and multiple arrangements1 0Child with relative, childcare center/Head Start, family childcare, and multiplearrangements1 0a To perform this analysis, we could not weight the data to account for the oversampling of mothers in the Glades and mothersassessed “at risk.” Twenty-six mothers were not interviewed at one or more of the 6-month telephone follow-up points, but wereinterviewed at all five yearly points. These mothers are included in the other analyses, but are not included in this analysis as wedo not know what type of childcare arrangements their focal child was in for all nine time points.b The nine time points include years 1 (1–6 mos.), 1.5 (6–12 mos.), 2 (12–18 mos.), 2.5 (18–24 mos.), 3 (24–30 mos.), 3.5 (30–36mos.), 4 (36–42 mos.), 4.5 (42–48 mos.), and 5 (48–54 mos.). Some arrangement categories were collapsed in this table.Chapin Hall at the University of Chicago 202

Figure C- 1. Mean Days of MCHP Services for Year 5 Sample by Age of Child Through 2008 aaSource: Vital Statistics and FOCiS (2003–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “at risk” and mothers residing in theGlades TGA.Chapin Hall at the University of Chicago 203

Table C- 18. Days of MCHP Services for Year 5 Sample by Age of Child through 2008 aTotals Prenatal b Postnatal cService Type b 0–3 4–6 7–9 0–3 4–6 7–9 10–12 13–15 16–18Total Prenatal Postnatalmos mos mos mos mos mos mos mos mosCare Coordination Only d% (N=310) 38 17 37 0 2 17 31 23 3 1 0 1 2 3 5Mean days of service e 4.9 1.5 3.3 0.0 0.1 0.7 0.8 0.5 0.1 0.0 - 0.0 0.0 0.1 0.1Maximum days 20 11 11 1 4 10 8 6 4 2 - 1 2 4 7Intensive Care Coordination f% (N=310) 43 32 39 5 9 31 39 38 27 17 12 11 8 6 10Mean days of service 41.1 10.1 31.0 0.2 0.6 8.7 8.5 6.4 4.5 2.3 1.8 1.7 1.1 0.9 2.6Maximum days 204 41 174 9 9 37 32 30 27 25 18 16 20 33 77Enhanced Services% (N=310) 39 11 37 1 1 9 28 13 6 4 3 5 5 3 4Mean days of service 4.8 0.7 4.1 0.0 0.0 0.6 1.3 0.7 0.4 0.3 0.2 0.3 0.2 0.3 0.5Maximum days 51 12 51 1 2 12 7 11 9 7 8 10 5 23 17a Source: Vital Statistics and FOCiS (2003–2009). Data for the year 5 study sample were weighted to account for oversampling of mothers “at risk” and mothers residing in theGlades TGA. This table reports Maternal Child Health Partnership activity that is slightly different than was reported in the year year-3 report due to updated activity data. Thebirth of twins, triplets, and other multiples are counted as one birth event.bPrenatal activity includes any activity attributed to the child’s mother occurring 300 days or less before the child’s birth.cPostnatal activity includes any activity attributed to a child’s mother occurring from a child’s birth up to the birth of a subsequent sibling. It also includes any activity attributed tothe child.d“Days of service” for care coordination only births counts all days of treatment activity received for births that only receive care coordination but not intensive care coordination.eFor readability, standard deviations are not included in this table; they are available upon request.f “Days of service” for intensive care coordination births counts all days of treatment activity (whether intense or not) for births that receive at least one day of intensive carecoordination.19–21mos22–24mos24+mosChapin Hall at the University of Chicago 204

Table C- 19. Characteristics of Mothers by Timing of Receipt of MCHP Treatment Services a,bMaternal CharacteristicBaseline SamplePrenatal Services Only PostnatalNo Servicesor Pre- and Postnatal Services Only(n=63) (n=148) (n=100)%c % %“At risk” (mothers with score)^ 29 47 53Vital Statistics DataBlack, not Hispanic* 58 40 32Hispanic* 32 55 60Foreign-born* 33 59 60Less than high school diploma* 26 55 60Teen mother at focal child’s birth 12 20 15Unmarried mother 67 70 75No prenatal care, care began 9th month, or10 19 28unknown*Baby low birth weight 12 11 15Mother used tobacco or was unsure if she smoked8 3 5during pregnancyMother smoked 9+ cigarettes/day or was unsure if4 2 2she smoked during pregnancyMother used alcohol — — —Baby—abnormal conditions c 0 3 3Survey Data at BaselineGlades 19 11 12Number of children (Mean, SD)^ 2.3 (1.2) 2.0 (1.2) 1.8 (1.1)Teen mother at birth of first child* 40 55 67Lives with husband/partner 60 67 69Mother employed* 22 9 15Living at or below poverty threshold* 55 74 70Owns home (self or household member) 26 17 21Focal child special needs 3 9 11Other children special needs 15 15 7Mother has physical/mental health problem 4 8 7Other household member has health problem 12 10 13Partner support score (Mean, SD)^ 5.7 (1.2) 5.1 (1.6) 5.4 (1.4)Family/friend support score (Mean, SD)* 5.1 (3.7) 3.9 (3.2) 3.9 (3.3)Depression: CES-D score > 16 42 32 31DCF involvement in child’s first year 5 11 10a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b Previous versions of this table displayed U.S.-born mothers, foreign-born mothers who had been in the U.S. for 5 or more years,and foreign-born mothers who had been in the U.S. for less than 5 years. At the time of the year 5 in-person survey, just nineforeign-born mothers had been in the U.S. for less than 5 years.c z-test not completed because at least one column proportion is equal to zero.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: atrisk(1)vs. atrisk(2); atrisk(1) vs. atrisk(3); black(1) vs. black(2); black(1) vs. black(3); hispanic(1) vs. hispanic(2); hispanic(1) vs.hispanic(3); foreignborn(1) vs. foreignborn(2); foreignborn(1) vs. foreignborn(3); diploma(1) vs. diploma(2); diploma(1) vs.diploma(3); prenatalcare(1) vs. prenatalcare(3); teenmother(1) vs. teenmother(3); motherworking(1) vs. motherworking(2);poverty(1) vs. poverty(2).Chapin Hall at the University of Chicago 205

Table C- 20. Characteristics of Mothers Who Received MCHP Services by Number of TreatmentServices Days a 1–4 5–13 14–3536+ ServiceService Service ServiceMaternal CharacteristicDaysDays Days Days(n=84) (n=52) (n=59) (n=54)% b % % %“At risk” (mothers with score) 47 58 44 46Vital Statistics DataBlack, not Hispanic 39 25 38 34Hispanic 58 70 54 55Foreign-born 62 69 56 59Less than high school diploma 58 68 46 60Unmarried mother 64 80 76 69Teen mother at focal child’s birth 19 24 16 16Baby low birth weight 10 9 14 18No prenatal care, care began 9th month, or17 18 18 37unknown^Mother used tobacco or was unsure if she smoked3 5 6 2during pregnancyMother smoked 9+ cigarettes per day or was unsure 0 5 2 2if she smoked during pregnancy cMother used alcohol — — — —Baby-abnormal conditions 2 3 2 2Survey Data at BaselineGlades 13 15 4 11Number of children (Mean, SD)** 1.8 (0.9) 1.7 (0.8) 1.8 (1.2) 2.4 (1.4)Teen mother at birth of first child 40 47 43 39Lives with husband/partner 72 77 59 75Mother employed 14 7 8 7Living at or below poverty* 57 73 80 83Owns home (self or household member)* 31 8 15 8Focal child special needs* 4 11 18 10Other children special needs* 10 1 10 28Mother has physical/mental health problem* 3 1 7 19Other household member has health problem 11 5 11 19Partner support score (Mean, SD) 5.4 (1.3) 5.6 (1.3) 5.2 (1.7) 5.1 (1.7)Family/friend support score (Mean, SD) 4.4 (3.2) 3.5 (2.9) 3.6 (3.4) 3.6 (3.3)Depression: CES-D score > 16 26 26 42 35DCF involvement in child’s first year 5 12 15 13a Data were weighted to adjust for the oversampling of mothers in the Glades and mothers screened “at risk.”b All figures in regular font are percentages; means and standard deviations are italicized.c z-test not completed because at least one column proportion is equal to zero.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower:prenatal(1) vs. prenatal(4); poverty(1) vs. poverty(3); poverty(1) vs. poverty(4); ownhome(1) vs. ownhome(2); ownhome(1) vs.ownhome(4); focalchild(1) vs. focalchild(3); otherchild(1) vs. otherchild(4); otherchild(2) vs. otherchild(4); otherchild(3) vs.otherchild(4); motherhealth(1) vs. motherhealth(4); motherhealth(2) vs. motherhealth(4).Chapin Hall at the University of Chicago 206

Table C- 22. Number of Service Areas Used by Mothers in Glades and Non-Glades TGA overNumber of ServiceYear 1***Time aAreas Used % All Mothers % Glades % Non-Glades(N=310) (n=40) (n=270)0 services used 2 0 21–2 services used 27 13 293–4 services used 53 45 545 or more services used 18 43 14Year 2***% All Mothers % Glades % Non-Glades % Outside TGAs(N=310) (n=38) (n=251) (n=20)0 services used 0 0 0 01–2 services used 38 23 38 703–4 services used 42 33 46 155 or more services used 20 44 16 15Year 3**% All Mothers % Glades % Non-Glades % Outside TGAs(N=310) (n=38) (n=244) (n=29)0 services used 3 0 4 01–2 services used 46 26 51 363–4 services used 34 42 29 615 or more services used 17 32 17 4Year 4**% All Mothers % Glades % Non-Glades % Outside TGAs(N=310) (n=39) (n=235) (n=37)0 services used 5 0 5 131–2 services used 38 15 43 263–4 services used 38 56 34 425 or more services used 20 28 18 18Year-5% All Mothers % Glades % Non-Glades % Outside TGAs(N=310) (n=38) (n=232) (n=41)0 services used 5 0 6 31–2 services used* 25 5 28 283–4 services used 37 47 33 465 or more services used* 34 47 33 23aData were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: 1–2services(2) vs. 1–2services(1); 1–2services(2) vs. 1–2services(3); 5+services(1) vs. 5+services(3).*Chi-square tests indicate statistically significant differences between Glades and non-Glades mothers in year 1 (χ 2 = 20.116, p

Table C- 22. Patterns of Service Use by Nativity in 12 Months Prior to Surveys over Time aYear 1 Year 2 Year 3 Year 4 Year 5Service AreaForeignbornborn born born born born born born born bornU.S.- Foreign-U.S.- Foreign-U.S.- Foreign-U.S.- Foreign-U.S.-(n=167) (n=143) (n=167) (n=143) (n=167) (n=143) (n=167) (n=143) (n=167) (n=143)Healthcare for mother or children 78 71 94 97 90 97* 89 89 84 87Getting enough food (WIC, foodstamps)91 87 87 86 75 61* 74 65 73 70Family planning or birth control — b — 44 20*** 35 20** 26 13** 29 20Dental care 10 34*** 17 38*** 22 31 46 66*** 52 68**Childcare assistance (includes subsidy) 8 29*** 10 46*** 22 36** 22 56*** 33 51**Parenting information 70 77 20 27 8 15* 2 8* 8 8Paying rent or bills 5 17** 2 19*** 3 19*** 3 20*** 3 30***Housing or emergency shelter c 1 10** 6 14* 3 18*** 3 20*** 4 18***Transportation 17 18 15 16 4 11* 1 8** 5 9Legal issues 1 4 1 4 5 6 4 10* 7 6Employment 2 8** 2 9** 2 4 1 9** 3 16***Clothes for children 4 8 2 7 1 4 2 2 4 6Mental health or substance abuse 1 4 0 4* 3 4 5 11 8 13Translating things into English d — — 12 0 8 0 3 0 9 0Mean (SD) of items common over 3.0 3.9 2.9 3.9 2.6 3.4 2.8 3.8 3.4 4.4time(1.2) (2.2)*** (1.3) (2.2)*** (1.5) (2.1)*** (1.9) (2.2)*** (2.3) (2.5)***a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b Item not included in the year 1 survey.c The year 1 survey had one item combining housing and emergency shelter, but were asked as two separate items in all subsequent surveys; they are combined in this analysis.d Only those mothers who do not speak English as their primary language were asked this question (n=146, 150, 153, 142).One-way ANOVA and chi-square tests indicate differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 208

Table C- 23. Number of Services Used by Nativity over Time aYear 1*** Year 2*** Year 3*** Year 4*** Year 5***Services UsedForeignbornborn born born born born born born born bornU.S.- Foreign-U.S.- Foreign-U.S.- Foreign-U.S.- Foreign-U.S.-(n=167) (n=143) (n=167) (n=143) (n=167) (n=143) (n=167) (n=143) (n=167) (n=143)0 services used 2 2 0 0 4 1 6 4 5 51–2 services used 30 23 44 31 53 39 48 26 33 153–4 services used 58 47 47 37 34 34 37 39 39 345 or more services used 10 28 10 32 10 27 10 31 23 46Average services used (SD) 3.0 (1.2) 3.9 (2.2) 2.9 (1.3) 3.9 (2.2) 2.6 (1.5) 3.4 (2.1) 2.8 (1.9) 3.8 (2.2) 3.4 (2.3) 4.4 (2.5)a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”Chi-square tests indicated differences are statistically significant at *p < .05, **p < .01, or ***p < .001.Chapin Hall at the University of Chicago 209

Table C- 24. Mothers’ Baseline and Year 4 Characteristics by Number of Services Reported atMaternal CharacteristicGroup 10 ServicesGroup 21–2 ServicesGroup 33–4 ServicesGroup 45+ Services(n=15) (n=78) (n=113) (n=104)%c % % %Baseline CharacteristicsGlades b * 0 3 16 19U.S.-born* 44 29 43 63Black, not Hispanic* 44 29 34 57Hispanic* 47 65 60 35EducationLess than high school diploma 42 50 60 51High school diploma 58 50 40 49Teen mother at focal child’s birth b * 0 7 21 16MCHP servicesDays of care coordination 91 88 86 86Days of intensive care coordination 21 31 39 44Study risk index (mean, SD) 2.9 (1.6) 3.3 (1.5) 3.7 (1.5) 3.6 (1.6)Year 4 CharacteristicsNumber of childrenOne child* 39 35 30 13Two children^ 44 37 31 38Three or more children* 18 28 39 49Lives with husband/partner b * 100 96 88 77Mother employed* 53 58 51 27Living at or below poverty threshold* 47 49 54 72Own home (self or household member) 26 29 29 23Focal child special needs b 0 12 17 23Other children special needs b 0 14 11 22Mother has physical/mental health problem b^ 0 0 4 10Other household member has health problem* 9 2 6 30Partner support score (Mean, SD) 5.9 (0.6) 6.0 (0.6) 5.7 (1.1) 5.7 (1.2)Family/friend support score (Mean, SD) 4.3 (3.2) 3.3 (2.8) 3.4 (2.4) 3.7 (2.6)Depression: CES-D score > 16 9 23 20 28Stress: PSI/SF score > 86 9 12 12 20DCF involvement in child’s first year b * 0 4 6 19a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b z-test not completed because at least one column proportion is equal to zero.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: glades(2)vs. glades(3); glades(2) vs. glades(4); nativity(2) vs. nativity(4); nativity(3) vs. nativity(4); black(2) vs. black(4); black(3) vs.black(4); hispanic(2) vs. hispanic(4); hispanic(3) vs. hispanic(4); teenmom(2) vs. teenmom(3); children(1) vs. children(4);children(2) vs. children(4); children(3) vs. children(4); partner(2) vs. partner(4); employed(2) vs. employed(4); employed(3) vs.employed(4); poverty(2) vs. poverty(4); poverty(3) vs. poverty(4); motherhealth(3) vs. motherhealth(4); otherhealth(2) vs.otherhealth(4); otherhealth(3) vs. otherhealth(4); dcf(2) vs. dcf(4); dcf(3) vs. dcf(4).Chapin Hall at the University of Chicago 210

Table C- 25. Linear Regression Analysis of Number of Services Used in Year 5 aPredictor Variable Coefficient Sig.Glades .78 NSRace/nativityBlack—U.S.-born .84 ^Black—Foreign-born .76 NSHispanic—U.S.-born -.56 NSHispanic—Foreign-born (excluded category) — —White/other 1.91 *Days of intensive care coordination .87 *Lives with husband/partner at year 4 -1.77 **Number of children at year 4One child aTwo children -.63 ^Three or more children (excluded category) — —Mother employed at year 4 -.33 NSIncome at or below poverty threshold at year 4 .54 NSOther children—special needs at year 4 .85 ^Investigated DCF reports, birth to age 1 -.18 NSConstant 4.64 ***R 2 .27aThis variable dropped out of the analysis.∧p < .10, * p < .05, ** p < .01.Chapin Hall at the University of Chicago 211

Table C- 26. Baseline and Year 4 Characteristics of Mothers Who Changed/Did Not ChangeService Use between Year 2 and Year 5 a Group 1Group 2Decreased No (or Little) ChangeMaternal CharacteristicService Use b in Service UseBaseline CharacteristicsGroup 3IncreasedService Use b(n=50) (n=175) (n=86)% % %Glades 11 14 13U.S.-born 41 46 49Black, not Hispanic 43 37 48Hispanic 52 55 47EducationLess than high school diploma 46 53 59High school diploma 54 47 41Teen mother at focal child’s birth^ 4 17 17MCHP servicesDays of care coordination 88 87 85Days of intensive care coordination 45 36 37Study risk index (mean, SD) 3.4 (1.8) 3.5 (1.4) 3.7 (1.6)Year 4 CharacteristicsNumber of children^One child 37 25 22Two children 35 37 31Three or more children 28 38 47Lives with husband/partner* 88 95 73Mother employed* 55 48 32Living at or below poverty threshold 59 57 60Owns home (self or household member) 23 26 29Focal child special needs 21 13 22Other children special needs 23 13 16Mother has physical/mental health problem 5 3 8Other household member has health problem* 15 8 24Depression: CES-D score > 16 18 25 22a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b Indicates mothers who decrease or increased services use by two or more services.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower:teenmom(1) vs. teenmom(2); children(1) vs. children(3); partner(2) vs. partner(3); momwork(1) vs. momwork(3); momwork(2)vs. momwork(3); othershealth(2) othershealth(3).Chapin Hall at the University of Chicago 212

Table C- 27. Baseline Characteristics of Mothers Who Changed/Did Not Change Service Usebetween Year 1 and Year 5 a Group 1Group 2Decreased No (or Little) ChangeBaseline CharacteristicsService Use b in Service UseGroup 3IncreasedService Use b(n=64) (n=155) (n=91)% % %Percent mothers 21 50 29Average number of services used, year 1 (SD)*** 5.1 (2.5) 3.1 (1.2) 2.8 (1.4)Average number of services used, year 5 (SD)*** 2.3 (2.0) 3.0 (1.3) 6.4 (2.3)Glades 12 14 12U.S.-born* 52 38 55Black, not Hispanic^ 40 35 51Hispanic* 50 59 42EducationLess than high school diploma 48 60 46High school diploma 52 40 54Teen mother at focal child’s birth 11 14 19Healthy Beginnings servicesDays of care coordination 90 84 88Days of intensive care coordination 41 35 40Study risk index (mean, SD)^ 3.1 (1.3) 3.6 (1.4) 3.6 (1.7)a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”b Indicates mothers who decrease or increased services use by two or more services.*Denotes z-test of column proportions (or t-test of means where applicable) statistically significant at p < .10 or lower: nativity(2)vs. nativity(3); black(2) vs. black(3); hispanic(2) vs. hispanic(3).Chapin Hall at the University of Chicago 213

Table C- 28. Changes in Circumstances between Year 3 and Year 4 and Number of Services Usedat Year 5 aAll mothers (N=310)Percent ofMothersNumber ofServices Used inYear 5New baby in year 4 38 4.3 (2.6)**Same number of children 62 3.6 (2.3)Moved two or more times in year 4 2 5.9 (3.6)*Moved once or less 98 3.8 (2.4)Mothers married at year 3 (n=33)Divorced/separated in year 4 13 2.0 (1.7)Still married 87 3.3 (2.5)Mothers with a husband/partner at year 3 (n=230)No husband/partner at year 4 11 4.6 (2.6)*Still with husband/partner 89 3.5 (2.4)Mothers employed at year 3 (n=153)Not employed at year 4 32 4.7 (2.5)**Still employed 68 3.2 (2.3)Mothers with employed husbands/partners at year 3 (n=188)Husband/partner not employed at year 4 7 5.3 (2.8)**Still employed 93 3.3 (2.3)Mothers with no health problem at year 3 (n=292)Health problem at year 4 2 6.5 (2.8)**Still no health problem 98 3.7 (2.3)Focal child with no special needs at year 3 (n=250)Special needs at year 4 10 4.8 (3.1)*Still no special needs 90 3.5 (2.3)Household income above poverty threshold at year 3 (n=144)Household income below poverty threshold at year 4 36 3.6 (2.4)Household income above poverty threshold 64 3.2 (2.3)a Data were weighted to adjust for oversampling of mothers from the Glades and mothers screened “at risk.”Chapin Hall at the University of Chicago 214