The Prediction of Drug Use through Discriminate Analysis from Variables Common to Potential Secondary School Dropouts

Abstract

A group of 78 secondary school students who satisfied selected criteria as dropouts was interviewed to determine the following: employment status of parents, the presence or absence of an adult male in the home, the self-image of the students, the perceived ability of the students to communicate with their parents, the sex of the students, and drug use by the students. An effort was made to assure the students of the confidentiality of their responses. Discriminate analysis was used to produce a regression equation to predict drug use from the five predictor variables. The best weighted combination of the five variables yielded a multiple R of .63 with drug use. The presence or absence of an adult male in the home contributed most to the prediction system. The next best predictors were (b) perceived ability to communicate with parents and (c) self-image. The relationships between all three of these variables and drug use were significant at the .01 level. Sex was the only variable that was not a significant predictor of drug use.