THREE TYPES OF EVIDENCE SUPPORT THE HYPOTHESIS: COMPARISONS OF THE EXTENT OF GROUP DELINQUENCY AS ESTIMATED BY SELF-REPORT AND OFFICIAL DATA; REPORTS FROM LAW ENFORCEMENT OFFICIALS OF THEIR OWN BEHAVIOR; AND TESTS OF THE IMPACT OF GROUP VERSUS INDIVIDUAL OFFENSES UPON THE RATE OF JUVENILE ARRESTS. A REVIEW OF THE SUPPORT FOR THE HYPOTHESIS SUGGESTS, FOR EACH TYPE OF EVIDENCE, ALTERNATIVE EXPLANATIONS FOR THE FINDINGS. THIS STUDY REPORTS THE RESULTS OF A SURVEY OF HIGH SCHOOL STUDENTS AND 562 QUESTIONNAIRES COMPLETED BY MALE RESPONDENTS FORM THE BASIS FOR THE ANALYSIS. NINETEEN SELF-REPORTED DELINQUENCY ITEMS WERE SUBJECTED TO A CLUSTER ANALYSIS PROGRAM FOR THE QUESTIONNAIRE, AND A PATH ANALYSIS WAS MADE BASED UPON TOTAL INDICES. RESULTS OF THE PATH ANALYSIS SUGGESTED THAT THE GROUP HAZARD HYPOTHESIS MAY BE MORE APPLICABLE TO THE PROBABILITY OF APPREHENSION THAN TO THE PROBABILITY OF ARREST. ONLY IN THE THEFT BEHAVIORS DOES THE GROUP HAZARD HYPOTHESIS FIND CONSISTENT SUPPORT IN THE STUDY. THE RESULTS SUGGEST THAT THE APPLICABILITY OF THE GROUP HYPOTHESIS APPEARS TO BE MORE STRONGLY SUPPORTED IF IT IS RECONCEPTUALIZED AS BOTH AN APPREHENSION AND AN ARREST HAZARD. TABULAR DATA, NOTES, AND REFERENCES ARE PROVIDED. (MJW)