The Use and Validity of Home Language Surveys in State English ...

Recommendations

Info

EVEA Project Deliverable Use and Validity of HLS in ELPA Systemsproposed entering multiple continuous variables measuring language dominance on the HLS todetermine their relative contribution to a model predicting language and conceptual development. Sheargued this method would counter validity problems faced when otherwise attempting to groupstudents by standardized measures. Gonzalez found that ratings on the HLS better predicted children’sverbal than non-verbal conceptual development, and better predicted children’s performance inSpanish than in English. Although other measures of language proficiency were used in this study, thereare no reported comparisons of dominance ratings on the HLS to those obtained from the teacherratings or standardized tests.Gutiérrez-Clellen and Kreiter (2003) developed a parent questionnaire to determine the extent to whichthe child’s years of exposure to a language, language(s) spoken in the home, as well as language(s)spoken in other settings, are related to a child’s grammatical performance on spontaneous narrativesamples using a wordless picture book. Results revealed the parent-reported exposure variablescombined significantly to predict grammatical performance in Spanish. Percentage of overall exposureto Spanish at home accounted for 26% of the variance in grammatical utterances. However, none of theexposure variables were significant predictors of performance in English.The authors also related parents’ ratings of their child’s language proficiency with the child’sgrammatical performance in English and in Spanish to determine the extent to which parents can aid inidentifying their child’s language status. Results revealed a high correlation between parent ratings ofthe child’s Spanish and the child’s actual use of grammatical Spanish utterances in the narrative task (r =.75, p < .0001). A moderate correlation was also found between parent ratings of the child’s English andthe child’s actual grammatical performance in English (r = .32, p < .05). The authors concluded thatparent ratings can be used to accurately determine the child’s language status. However, given thecontext of this paper, we caution that the ratings and exposure reports were not especially successful atdetermining proficiency with English—the children’s second language—compared with Spanish, thechildren’s first language.A parent survey was recently developed by Reese, Thompson, and Goldenberg (2008) to collectinformation about the language used with the child in specific contexts (e.g., parent speaking to thechild, literacy activities) as well as child language use with other adults in the home. The authors alsoconducted parent interviews during which parents were asked to indicate the language heard by thechild in various contexts (e.g., at the park; by the babysitter) in addition to the language most commonlyheard by the child. Validity and reliability information was not reported. Nevertheless, parent reportsfrom the surveys and interviews provided a rich description of the children’s language environment andshowed a great deal of variability within and across the 14 communities in the sample. Findingsindicated while children primarily used Spanish with adults, they spoke much more English amongthemselves than with adults.In a follow-up study, Reese and Goldenberg (2008) related the variability found in the reports ofchildren’s home language environment with their literacy development. The authors focused theirinvestigation on links between literacy-related language background variables (i.e., frequency of readingin English and in Spanish with the child, the child’s reading language, and the parents’ reading language)and the Woodcock Language Proficiency Battery-Revised in English and Spanish (WLPB-R; Woodcock,1991; Woodcock & Muñoz-Sandoval, 1995). Results that revealed English as the child’s reported readinglanguage and English as the parents’ reported reading language were each associated with higher Basic12
EVEA Project Deliverable Use and Validity of HLS in ELPA SystemsReading and Passage Comprehension scores on the WLPB-R. Furthermore, reported frequency ofreading to children in English was associated with higher WLPB-R Passage Comprehension scores.Of the literature surveyed for the present review, only two studies provided information regardingvalidity or reliability testing of the survey instrument or both (Duursma, Romero-Contreras, Szuber,Proctor, & Snow et al., 2007; Leseman & de Jong, 1998). Given that Leseman and de Jong (1998) studiedDutch, Surinamese and Turkish families in inner-cities in the Netherlands, the language backgroundquestionnaire they developed reflected the majority language of the Netherlands—Dutch. The studyreport focused on the internal consistency of items on the language background questionnaire and there-test reliability of the measures administered at three separate home visits. Cronbach’s alpha was .95for the measure at the first visit, .91 for the second visit, and .93 for the final visit. The intercorrelationsbetween measures at each visit were about .90, suggesting the language background measuremaintained re-test reliability. In this sample, reported home language was related to a number ofbackground characteristics including socioeconomic status, parents’ informational and recreationalliteracy and children’s vocabulary development at ages 4 and 7, as well as decoding and readingcomprehension at age 7.Duursma et al. (2007) used the Parent Interview and Response Questionnaire (PIRQ; developed inconjunction with the National Institute of Child Health and Human Development and the Center forApplied Linguistics) to elicit information from the parents of 5 th grade English language learners abouthome language use and exposure as well as literacy practices in English and in Spanish. Items on thePIRQ loaded on two factors: home language and parental help (support for oral discourse, literacy andschool-related tasks). These two scales were then tested for internal consistency; Cronbach’s alpha forthe home language scale was .93. To test validity of the PIRQ, the authors examined correlationsbetween items on the home language scale and children’s English and Spanish WLPB-R letter wordidentification and picture vocabulary. Relevant for this review, home language was correlated with bothletter-word identification (r = .36, p < .01) and picture vocabulary (r = .67, p < .01) in English; the authorsconcluded that the PIRQ is a valid instrument for examining the influence of language background onstudents’ vocabulary. Subtleties emerged when looking across groups of students receiving instructionin English versus instruction in Spanish, with English vocabulary best predicted by support for Englishliteracy by parents for those instructed in English, and by the combination of paternal preference forEnglish and student gender for those instructed in Spanish.The amount of Spanish and English used in the home was also studied by Lindholm-Leary and Hernandez(2009) as part of a study of the range of background factors impacting dual-language, grade 4-8 studentachievement in English and Spanish language arts. Their results were disaggregated by student languagestatus provided by the schools (presumably based on the state ELPA) and show the relationshipbetween language(s) spoken in the home and classification of students into EL, Reclassified FluentEnglish Proficient (RFEP) and (Initial Fluent) English Proficient (EP). The findings suggest that there is nosimple association between the language(s) spoken in the home and a students’ language status inschool. While 26% of EL students hear only Spanish in the home, perhaps contrary to expectations andstereotypes, students hearing mostly Spanish are not overwhelmingly the EL students (51%) rather theyare the RFEP students (61% of them) and even 16% of EP students hear mostly Spanish in the home.What characterized the EP students from the other students is the fact that they never hear onlySpanish in the home, and certainly hear mostly English, although very few heard only English (16%).13
Page 10 and 11: EVEA Project Deliverable Use and Va
Page 19: EVEA Project Deliverable Use and Va

The Use and Validity of Home Language Surveys in State English ...

Create successful ePaper yourself

Delete template?

Save as template?