Peer influence or projection bias? Predicting respondent delinquency with perceptual measures of peer delinquency in 22 samples.

INTRODUCTION: The current study was designed to determine whether the peer influence effect is the direct result of respondents projecting their own delinquency onto peers in a process referred to as projection bias.

METHODS: Data from 22 gender-homogeneous U.S. and British samples (Ns = 154 to 4098) of individuals under the age of 17 at the time of initial assessment were used to create partial correlations between perceptual measures of peer delinquency and subsequent respondent offending, controlling for concurrent respondent offending.

RESULTS: Significant partial correlations surfaced between peer delinquency and subsequent respondent offending in 21 out of 22 samples, for a significant pooled effect size (rp1o2.o1 ) of 0.158 (95% CI = 0.124‒.193). Although the peer‒offending relationship, controlling for prior offending, was moderated by data recentness-with less recent samples (>16 years old) achieving a higher pooled effect size than more recent samples (≤16 years old)-a small but significant pooled effect size was recorded for both more and less recent samples.

CONCLUSIONS: These results are inconsistent with projection bias as a complete explanation for the peer influence effect and suggest that perceived peer delinquency has a demonstrable and meaningful effect on future offending in adolescents. Despite the lack of support for the projection hypothesis, there is still a need to include additional variables in the analyses (i.e., direct measures of peer delinquency and differential respondent opportunities to observe peer behavior) while entertaining alternative explanations of the perceived peer delinquency‒future offending relationship (i.e., hearsay and rejecting delinquency/non-delinquency).