Differential association and substance use: Assessing the roles of discriminant validity, socialization, and selection in traditional empirical tests

Although the correlation between personal and perceived peer substance use remains among the strongest in criminology, the discriminant validity of personal and perceived peer measures remains to be formally tested via confirmatory factor analysis. Further, only limited research has attempted to discern whether substance users seek out similar others rather than being influenced by the substance use that they perceive among their peers. Finally, research has yet to isolate, via panel analysis, the reciprocal relationship between personal substance use and perceived peer attitudes. The present study addresses each of these issues using National Youth Survey data. Results reveal that personal substance-related behavior and perceived peer behavior/attitudes bear only minimal discriminant validity and that, as predicted by Gottfredson and Hirschi’s General Theory of Crime, selection provides a better explanation of their correlation than does socialization.

A Comparison of Two Peer-Referenced Assessment Techniques with Parent and Teacher Ratings of Social Skills and Problem Behaviors

The relationship of parent and teacher ratings of social skills and problem behaviors, using the Social Skills Rating System with two peer-referenced measures (the Peer Nomination Technique and the Pupil Evaluation Inventory), were examined with a sample of regular education fifth- and sixth-grade students. The differential magnitudes of parent and teacher ratings with sociometric reports were examined as indicators of convergent and discriminant validity. The relationships among scales of prosocial and problem behaviors were found to be logically consistent. Teacher ratings of social skills, problem behaviors, and academic competence showed moderate correlations with both peer measures. Parent ratings of social skills and problem behaviors were similarly related, but at somewhat lower magnitudes.