Citation

Lipton WL, Smith MD. J. Res. Crime Delinq. 1983; 20(2): 199-213.

Copyright

(Copyright © 1983, SAGE Publishing)

DOI

10.1177/002242788302000204

PMID

unavailable

Abstract

Delinquency research is characterized by empirical models which generally afford low degrees of explanatory power. While the small R2s which characterize the discipline may be attributed, in part, to methodological and theoretical problems, part of the explanation for the insufficiency of our models may lie in a failure to consider the statistical issue of suppression. Utilizing data from the first two waves of the Youth in Transition study, an empirical example is offered in an attempt to demonstrate the importance of considering suppressor effects in delinquency research. Beginning with a bivariate correlation between social class and delinquency that is effectively zero, multivariate analyses are conducted in order to determine whether attachment to school serves as a suppressor in this relationship. The results of the analyses reveal that social class is, in fact, related to delinquent involvement but that the relationship is apparent only when controlling for the attachment variables. Ramifications of this finding are discussed and an explanation of how to recognize suppressor effects is offered.