A Better Kind of High: How Religious Commitment Reduces Drug Use Among Poor Urban Teens
APPENDIX A: STUDY DETAILS
The study uses data from the National Youth Survey (NYS), a longitudinal study of a national probability sample of 1,725 persons aged 11 to 17 in early 1977. After the seven birth cohorts in the sample and their parents were first interviewed in early 1977, adolescent respondents were re-interviewed in each of the subsequent five years. Each of the interviews is known as a “wave.” The study analyzes the data from Waves 3-5, when items on individual religious commitment were included in the survey, and Wave 1 parent data only for the measures of neighborhood disorder, class, and family structure.
We further refined this data because we have to treat neighborhood disorder, which can change over time, as time-invariant because the NYS only contains data for it in the initial, 1977 interviews. Accordingly, the study confines our analysis to those adolescents who did not move during the five-year survey period, and assumes that the neighborhood environment remained the same as that described at the first-wave interview. While the NYS data do not include any direct measure of residential mobility, it did classify people’s place-of-residence by a urban-suburban-rural classification.37 Over the first five waves a total of 220 respondents moved between two of the levels of classification, leaving 1,505 respondents eligible for our analysis, 72% (n = 1,087) of which provided sufficient data for multilevel analysis. The sex (53% male and 47% female) and race (79% white and 21% nonwhite) distributions of the final sample remain the same as those of the original total sample at the first five waves.