Cumulative empirical evidence suggests that the majority of offenses for which juveniles are apprehended involved more than one offender. Evidence that supports this claim has generally been interpreted as support for the theoretical assertion that "delinquency is predominantly a group phenomenon." Most of the studies reported in the literature, however, are based exclusively on official records (either police or juvenile court records). This raises the issue of whether the proportion of offenses committed by groups would also be high if other sources of data were used. The current self-report study used groups of respondents to represent various points on an official delinquency continuum: official nondelinquents, persistent community offenders, and incarcerated offenders. The relationship between "group violation rates" (proportion of self-reported offenses that were committed in a group context) and other characteristics of offenses were examined. An analysis was made of the relationships between group violation rates, seriousness of offense, frequency of violation, frequency of arrest, and arrest rates. High group violation offenses tended to be serious and less frequent and had higher arrest rates than offenses committed by an individual. 5 tables and 70 references