The study compares correlates for a number of different kinds of delinquent acts, ranging from minor and status offenses to serious personal and property offenses; it is hypothesized that the strength of the relationships will be stronger for the more serious offenses. Further, correlations when the delinquent behavior measure is a continuous variable are compared with correlations when the dependent variable is a dichotomized one comprised of the two most extreme ends of the frequency continuum; it is hypothesized that the correlations will be higher for the dichotomized extreme types than for the continuous variable. A self-report questionnaire was administered to 3,065 male and female juveniles attending grades 7-12 in seven school districts in three Midwestern States. Within each school district, schools whose students represented demographic characteristics of the community were selected. For the serious offenses of the type most comparable to those showing up in official data, the study does not show correlations with sociodemographic variables found in police and court data. These findings are more in line with the findings from previous self-report studies than with the findings in Elliot and Ageton (1980). Tabular data are provided from the study, along with two notes and seven references.