Social determinants of alcohol consumption: The effects of social interaction and model status on the self-administration of alcohol.

R. Lorraine Collins State University of New York at Stony Brook George A. Parks and G. Alan Marlatt University of Washington Two studies were conducted to assess variables related to the social determinants of alcohol consumption. In Study 1, moderate- and heavy-drinking male under- graduates were paired with confederates who behaved in a sociable or unsociable manner while modeling either light or heavy consumption. Modeling occurred in the sociable conditions but not in the unsociable conditions, where subjects tended to drink heavily. In Study 2, a similar group of subjects was exposed to one of three social status conditions crossed with light versus heavy consumption. The results indicated a modeling effect in all social status conditions. These studies provide further support for the existence of a modeling effect that can be disrupted by a lack of rapport between drinking partners. This latter finding has implications for the etiology of problem drinking because it suggests that increased alcohol consumption may serve as a strategy for coping with aversive social interactions. Initial research on the effect of modeled consumption rates on social drinking (Caudill & Marlatt, 1975) suggested that heavy-drink- ing men tended to match the consumption of their drinking partner whether his con- sumption was heavy or light. Subsequent examinations of this phenomenon replicated these findings in laboratory analogue drinking tasks such as the taste-rating task (Cooper, Waterhouse, & Sobell, 1979; Hendricks, So- bell, & Cooper, 1978; Lied & Marlatt, 1979; Watson & Sobell, 1982), natural bar settings (Reid, 1978), and seminaturalistic bar settings (Caudill & Lipscomb, 1980). Much of the research concerning the mod- eling of alcohol consumption has focused on manipulating characteristics of the model, including the nature of the social interaction between the model and the subject. The effects of manipulating social interaction are unclear. In the Caudill and Marlatt study, model's drinking rate (light vs. heavy) and a