Out of 60 school districts in a Midwestern State approached to participate in the study, 55 agreed to participate. These districts contained 20,366 students. High schools were randomly selected from each quadrant to be representative of the respective populations in those quadrants. A computer-generated simple random sample of 3,395 students was developed from those schools. The sample proved to be representative of the State population. Analyses were based on 3,335 students. Participants completed two self-report questionnaires, one during their homeroom period and another during an equivalent study period later in the day. As one measure of self-control, nine behaviors were dichotomized (0=no, 1=yes) and considered as indicators of low self-control. The Grasmick et al. (1993) cognitive scale was also used to measure self-control. Property and person offenses committed by the participants were measured with items of the delinquency scale developed by Nye and Short, which has been used extensively in research, yielding reports of high reliability and validity. The most noteworthy finding was that a behavioral measure and a widely used cognitive measure of self-control were significantly related to 4 forms of delinquency when 19 other factors were simultaneously considered in the analyses. Despite equivocal evidence for the potency and universality claims of the general theory of self-control, there is support for a more modest claim about the relative importance of self-control in explaining certain forms of delinquency. 5 tables and 86 references