The youths all lived in Arizona and were interviewed in 1990 after obtaining parental permission. The families all spoke English. The analysis focused on three aspects of mutual interaction between siblings: warmth, conflict, and frequency of contact with mutual friends. The study hypothesis was that sibling relationship variables would condition the behavioral similarity of the younger and older sibling. For both drug abuse and delinquency, this prediction was confirmed for warmth and mutual friends in that siblings who reported warmer mutual relationships or greater contact with mutual friends were more alike behaviorally. Social class, parental drug abuse, or child-rearing styles did not explain these effects. Findings suggested that, given the existence of sibling effects, the strength of shared familial influences of other origins must be revised downward. Tables and 17 references (Author abstract modified)