13.07.2015 Views

Sourcebook of Criminal Justice Statistics, 1983 - Bureau of Justice ...

Sourcebook of Criminal Justice Statistics, 1983 - Bureau of Justice ...

Sourcebook of Criminal Justice Statistics, 1983 - Bureau of Justice ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Research design and proceduresN'PENDIX 10Student Drug Use in America, 1975-1982-­Survey methodology and definitions af termsNJTE: The following is excerpted from Lloyd D. Johnston, Jerald G. Bachman, and Patrick M.O'Malley, Student Drug Use, Attitudes and Beliefs National Trends 19!5-1982, U.S.Department <strong>of</strong> Health and Htxl1an Services, Notional Institute on Drug Abuse (Washington, D.C.:U.S. Government Printing Office, 1982), pp. 3-7; and information provided by the Source.The basic research design involves annual data collections fromhigh school seniors during the Spring <strong>of</strong> each year, beginning with theclass <strong>of</strong> 1975. Data collection takes place in approximately 125 to 140public and private high schools selected to provide an accurate crosssection <strong>of</strong> high school seniors throughout the United States.One limitation in this design is that it does not include in thetarget population those young men and women who drop out <strong>of</strong> highschool before the last few months <strong>of</strong> their senior year (j.e., prior to thesurvey)--between 15 to 20 percent <strong>of</strong> each age cohort. The omission<strong>of</strong> high schaol dropouts does introduce biases in the estimation <strong>of</strong>certain characteristics <strong>of</strong> the entire age group; however, for mostpurpo£es, the small proportion <strong>of</strong> dropouts sets outer limits on the bias.Further, .since the bias from missing dropouts should remain just aboutconstant from year to year, their omission should introduce little biasinto the various types <strong>of</strong> change being estimated for the majority <strong>of</strong>the population. . ..Sampling procedures--The procedure for securing a nationwidesample <strong>of</strong> high school seniors is a multistage one. Stage I is theselection <strong>of</strong> particular geographic areas; Stage 2 is the selection <strong>of</strong> oneor more high schools in each area; and Stage 3 is the selection <strong>of</strong>seniors within each high school.Stage 1--The geographic areas used in this study are theprimary sampling units (PSUS) developed by the Sampling Section<strong>of</strong> the Survey Research Center for use in the Center's nationwideinterview studies. These consist <strong>of</strong> 74 primary areas throughoutthe coterminous United States. In addition to the 12 largestmetropolitan areas, containing about 30 percent <strong>of</strong> the Notion'spopulation, 62 other primary areas are included: lOin theNortheast, 18 in the North Central area, 24 in the South, and 10in the West. Because these same PSUs are used for personalinterview studies by the Survey Research Center, local fieldrepresentatives can be assigned to administer the data collectionsin practically 011 schools.Stage 2--ln the major metropolitan areas more than onehigh school is <strong>of</strong>ten included in the sampling design; in mast othersampling areas 0 single high school is sampled. In all cases, theselections <strong>of</strong> high schools are made such that the probability <strong>of</strong>drawing a school is proportionate to the size <strong>of</strong> its senior class.The larger the senior class (according to recent records), thehigher the selection probability assigned to the high school.When a sampled school is unwilling to participate, a replacementschool os similar to it as possible is selected from the sameg~ographical orea.Stage 3--Within each selected schoo!, up to about 400seniors may be included in the data collection. In schools withfewer than 400 seniors, the usual procedure is to include all <strong>of</strong>them in the data collection. In larger schools, a subset <strong>of</strong> seniorsis selected either by randomly sampling classrooms or by someother random method that IS convenient for the school ond judgedto be unbiased. Sample weights are assigned to each respondentso as t.: ,' voriations inselection probabilities occurring at the earlier stages <strong>of</strong> sampling.The three-stage sampling procedure described above yielded thenumber <strong>of</strong> participating schools and students presented in Table I.Table IClass <strong>of</strong>1975 1976 1977 1978 1979 1980 1981 1982Total number<strong>of</strong> schools 125 123 124 131 131 127 12!l 137Public III 108 108 III III 107 109 116Privote 14 15 16 20 20 20 19 21Students porticipating: a~ber 15,791 16,678 18,436 18,924 16,662 16,524 18,267 18,661Percent 78 77 79 83 82 82 81 83Each school (except for half <strong>of</strong> those in the 1975 data collection)was asked to participate in two data collections, thereby permittingreplacement <strong>of</strong> half <strong>of</strong> the total sample <strong>of</strong> schools each year. Onemotivation for requesting that schools participate for 2 years isadministrative efficiency; it is a costly and time-consuming procedureto secure the cooperation <strong>of</strong> schoolS, and a 2-year period <strong>of</strong>participation cuts down that effort substantially. Another importantadvantage is that whenever an appreciable shift in scores from one654graduating class to the next is observed, it is possible to checkwhether the shift might be attributable to some differences in thenewly sampled schools. This is done simply by repeating the analysisusing only the 60 or so schools that participated both years. Thus far,the half-sample appraach has worked quite well; an examination <strong>of</strong> drugprevalence data from the "matched half-samples" showed that the halfsamples<strong>of</strong> repeat schools yielded drug prevalence trends that werevirtually identical to trends based on all schools.Definitions <strong>of</strong> termsCollege plans--Respondents were asked the question "How likelyis it that you will do each <strong>of</strong> the following things after high school?"One <strong>of</strong> the alternatives listed is "gradUate from college (four yearprogram)." Seniors responding "definitely won't" or "probably won't"were assigned to the college plans category "none or under 4 years."Students responding "probably will" or "definitely will" were assigned tothe college plans category "complete 4 years."Region--States grouped as "I'hrtheast" (Census classifications <strong>of</strong>New England and Middle Atlanticl: Maine, New Hampshire, Vermant,Massachusetts, Rhode Island, Connecticut, New York, New Jersey, andPennsylVania. States grouped as "North Central" (Census classifications<strong>of</strong> East North Central and West North CentraD: Ohio, Indiana, Illinois,Michigan, Wisconsin, Minnesota, Iowa, Missouri, North Dakota, SouthDakota, Nebraska, and Kansas. States grouped as "South" (Censusclassifications <strong>of</strong> South Atlantic, East South Central and West SouthCentraJ): Delaware, Maryland, District <strong>of</strong> Columbia, Virginia, WestVirginia, North Carolina, South Carolina, Georgia, Florida, Kentucky,Tennessee, Alabama, Mississippi, Arkansas, Louisiana, Oklahoma, andTexas. States grouped as "West" (Census classifications <strong>of</strong> Mountainand Pacific): Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona,Utah, Nevada, Woshington, Oregon, and California.Population density:"Large SMSAs" inclL'des the 12 largest StandardMetropolitan Statistical Areas (SMSA) as <strong>of</strong> the 1970 census: NewYork, Los Angeles, Chicago, Philadelphia, Detroit, Son Francisco,Washington, Boston, Pittsburgh, Sf. Louis, Baltimore, andCleveland.''Other SMSAs" include all other Standard MetropolitanStatistical Areas excluding the 12 above. Except in the NewEngland States, an SMSA is a county or group <strong>of</strong> contiguouscounties that contains at 1"I1.y: household respondents, self-respondents,and proxy respondents.I. Household respondent--A household respondent waschosen to answer questions that pertained to the entirehousehold. Items asked <strong>of</strong> the household respondent included suchquestions as whether the residence was owned or rented and totalfamily income. In addition, the household respondent was asked aseries <strong>of</strong> household "screen" questions, designed to elicitinformation concerning crimes against the household such asburglary and auto theft.The interviewer was instructed to interview aknowledgeable adult household member as the householdrespondent; that is, one who appeared to know--or who couldreasonably be expected to know--the answers to the householdquestions. For the data presented in this SOURCEBOOK, thisusually was the head <strong>of</strong> the household or the spouse <strong>of</strong> the head<strong>of</strong> the household. If it became apparent that the particularhousehold member being interviewed was unable to answer thesequestions, a more knowledgeable respondent was found, orarrangements were made to call back when a knowledgeablerespondent was available.2. Self-respondent--Questions that pertained to individualvictimization were asked <strong>of</strong> each household member 14 years <strong>of</strong>age or older. The questions asked <strong>of</strong> each Individual memberinclude personal characteristics and whether they were victimizedin some way during the preceeding 6 months.3. Proxy respondent--Information about each householdmember aged 12 and 13 was obtained by a proxy; that is, thequestions for these persons were asked <strong>of</strong> the householdrespondent or some other knowledgeable adult household member.Proxy respondents were also utilized for those household memberswho were physically or mentally unable to answer the individualquestions, as well as for those household members who weretemporarily absent and not expected to return within theenumeration period.Commercial interviews--For the commercial portion <strong>of</strong> thesurvey, either the owners or the managers <strong>of</strong> the commercl alestablishment were to be interviewed. If the owner or manager wa.not available at an establishment, the interviewer asked for the nam,~and telephone number so that he/she could make an appointment for aninterview. If the owner or manager was temporarily absent for theentire Interview period, or was ill for the entire period, or was toobusy, or if the interviewer could not obtain an interview for some otherreason, the interview was conducted with the assistant manager, an655accountant, the senior sales-clerk, the nurse, secretary, receptionist,or some other employee who was knowledgeable about the business.The conmercial survey was suspended as <strong>of</strong> 1977.QuestionnairesI-husehold portion--The questionnaire used in the householdsurvey had four parts. The first portion <strong>of</strong> the questionnaire completedby the interviewer consisted <strong>of</strong> such items as basic householddemographic data and pertinent information concerning noninterviews.The second portion, asked <strong>of</strong> the household respondent, was designed toobtain--for the purpose <strong>of</strong> describing the household--characteristics <strong>of</strong>the household members i2 years <strong>of</strong> age or older as well as to elicitgeneral information concerning crimes committed against the householdas a whole during the reference period. Items :ncluded in this sectionIncluded such "screen" questions as: "During the last 6 months, didanyone break into or somehow illegally get into your home, garage, oranother building on your property?" and "Did anyone steal or try tosteal, or use your motor vehicle without permission?"The third portion <strong>of</strong> the household questionnaire was asked <strong>of</strong>each household member 12 years <strong>of</strong> age or older, or a proxy. Itconsisted <strong>of</strong> items <strong>of</strong> person characteristics, such as educationalattainment and marital status, as well as individual "screen" questions,desi gned to elicit whether or not the respondent had been the victim <strong>of</strong>a specific crime during the reference period. For example, thefollowing questions were asked: "Did anyone take something directlyfrom you by using force, such as by a stickup, mugging or threat?" and"Did anyone beat you up, attack you or hit you with something, such asa rock or bottle?"The final portion <strong>of</strong> the household questionnaire, the CrimeIncident Report, was used to gather detailed information about crimesreported in either the household screen section (portion II) or theindividual screen section (portion III). One incident report was filledout for each incident reported in answer to a screen question. Forexample, if a respondent said that her purse was snatched once andthat she had been beaten up twice, three Crime Incident Reports--onefor each seporate incident--were completed.Commercial--The commercial victimization questionnaire hadthree primary components. The first contained questions that enabledthe interviewer to classify the type <strong>of</strong> establishment, non interviewreasons, gross income, ownership, and number <strong>of</strong> employees. Thesecond portion contained "screen" questions regarding whether anyburglaries and/or robberies had occurred during the reference periodand, if so, how many. This part ion also included questions regardinginsurance coverage and security measures used by the establishment.The final portion <strong>of</strong> the commercial questionnaire consisted <strong>of</strong>robbery and burglary incident reports. These detailed reports werefiled on every robbery and burglary incident reported in the screenquestionnaire. Seporate reporting forms were used for robbery andburglary incidents for the survey period from July 1972 throughDecember 1972. Beginning in January 1973, the same i nci dentreporting form was used to record details <strong>of</strong> both robbery and burglaryincidents.Changes in the 1979 Household QuestionnaireIn 1979, a revised questionnaire was used in the householdsurvey. To be able to make data comparable across years,specifications were given for reformatting the recponses from the newquestionnaire to coincide with the old survey questionnaire. Tablesconstructed for SOLRCEBOCK are based on the format <strong>of</strong> the old surveyquestionnaire. However, for Tables 3.6 and 3.14, it was not alwayspossible to reformat the new respanse categories (and hence columndistributions) to those from the old questionnaire. The result is thatthe tables presented in this edition <strong>of</strong> SOURCEBOOK are notcomparable to those presented in earlier editions. To aid the reader,the table below compores the response codes for these variables fromthe 1978 and revised 1979 questionnaires.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!