DIFFERENTIAL ASSOCIATION AND MARIJUANA USE: A CLOSER LOOK AT SUTHERLAND (WITH A LITTLE HELP FROM BECKER)*

Based on Sutherland's differential association theory and Becker's early research on marijuana use, a contingency model estimating the exact probability of getting high on marijuana under various associational and motivational conditions is specified and tested. Data from surveys at two universities fit this model closely. Predicted first-order interactions and nonlinear effects of motivational balance and peer association are statistically significant and generate highly precise estimates of the probability of getting high. These results suggest that linear main-effects models employed in previous research on differential association processes do not adequately reject the complex casual structure of Sutherland's theory. In addition, this study raises serious questions about claims that differential association theory is untestable and has been made outdated by social learning theory.