|
Methods To address this issue, data were drawn from the three longitudinal projects of OJJDP's Program of Research on the Causes and Correlates of Delinquency: the Rochester Youth Development Study, the Denver Youth Survey, and the Pittsburgh Youth Study. All three projects used prospective longitudinal designs that followed the same individuals from childhood or early adolescence through early adulthood. Overall, the three projects selected probability samples totaling 4,000 urban youth. At each site, the youth and a primary caretaker were interviewed separately in private settings at established intervals. The specific designs of the projects have been reported in other OJJDP publications, especially Browning et al. (1999). In these studies, delinquency was measured by self-reports of involvement in a variety of delinquent behaviors ranging from petty theft to aggravated assault; youth also indicated their use of illegal drugs ranging from marijuana to heroin. In the analysis that follows, the responses for self-reported delinquency and drug use were cumulated over a 4-year period covering middle adolescence. For the Rochester project, family transitions were counted by comparing family structure in adjacent interviews with boys and girls between the ages of 13 and 17 (up to a maximum of eight transitions using 6-month interviews). For example, if a youth lived with both biological parents during the first interview and with the biological mother only at the second interview, a transition occurred. If the mother's partner subsequently moved into the household, a second transition occurred. A similar strategy was used in the Denver project for boys and girls ages 14 to 18 for the sample members who were born in 1974 (up to a maximum of four transitions using annual interviews). In the Pittsburgh project, which included only boys, retrospective reports of the number of changes in caretaker status occurring between the ages of 11 and 15 were obtained from the parent respondents for the seventh-grade cohort (no predetermined maximum number of transitions). The maximum possible number of transitions varied across the three sites because of these differences in design and measurement strategies.
| ||||
