Random versus Clinical Samples:

Abstract

This study included a clinical sample (N = 127) and a stratified random sample (N = 114) of adolescents (ages 13-17). Multiple regression was performed with delinquency as the dependent variable and family dynamics, parental alcohol use, and type of sample as the independent variables. Interaction terms were constructed by multiplying each regressor by type of sample (coded 0 = random sample, 1 = clinical sample). The interaction terms for three family dynamics variables (organization, intellectual-cultural orientation, and moral-religious emphasis) were statistically significant, indicating that the relationship between these regressors and delinquency differed significantly across type of sample. Consequently, within-sample regressions were performed. Family conflict was the only statistically significant predictor of delinquency in the stratified random sample (R' = .21). Intellectual-cultural orientation, organization, and father's alcohol use were significant predictors of delinquency in the clinical...