Data for the study were collected between 1991 and 1994 from an evaluation of five Treatment Alternatives to Street Crime programs. The sample included the first 522 offenders who provided all relevant data during intake, and most offenders had extensive criminal histories. Self-control was measured using a 23-item self-report index. Two proxy variables were used to measure criminal opportunity, gender and crime-involved friends. Crime measures included number of criminal acts of force and number of criminal acts of fraud reported during a 6-month recall period by the 522 offenders. Results showed self-control was lower among offenders reporting more crimes of force and fraud, but the variance explained by self-control was low in each case. The relation between self-control and fraud crimes depended on criminal opportunity, but the relation between self-control and force crimes was not. Implications of the findings for the general theory of crime are reviewed. 31 references and 4 tables