Purpose: To examine the correlations between multiple risk behaviors in adolescent populations to document the extent to which problem behaviors are intercorrelated and to identify factors associated with variations in these correlations.
Methods: Studies from 1977 through the end of 1999 that included two or more problem behaviors in adolescents were identified by literature searches using the PsychLit database, Social Sciences Citation Index, manual journal searches and "ancestry" approaches. The behaviors studied were alcohol use, marijuana use, illicit drug use, cigarette smoking, general deviant behavior, and sexual activity. Included studies reported correlation coefficients between variables.
Results: Across all studies, the mean correlation between any two pairs of problem behaviors was 0.35, with a standard deviation of 0.28. This suggests that, on average, about two-thirds of the variation in problem behavior is the result of unique rather than common causes. The magnitude of the correlations varied as a function of the age of the adolescent, with lower correlations being evident for older adolescents. In addition, the magnitude of the correlation varied as a function of when the study was conducted, with studies of past generations showing stronger connections between risk behaviors than current generations.
Conclusions: The data suggest that there is considerably more unique variation in classic adolescent problem behaviors than common variation.