Validation of multidimensional scaling-based modeling of alcohol expectancies in memory: Age and drinking-related differences in expectancies of children assessed as first associates

Background: As evidence has accumulated that alcohol expectancies mediate t he effects of other drinking antecedents, attempts to understand the mechan ism by which expectancies influence behavior have focused on modeling memor y processes. Previous expectancy work, however, has used relatively indirec t approaches to retrieve and model information stored in memory. By using t he method most recommended by memory researchers for directly obtaining unc ontaminated memory contents, we assessed children's expectancies and relate d findings to empirically modeled organization and activation of expectanci es in memory based on scaled instruments. Methods: Individual interviews were conducted with 462 children in 2nd thro ugh 5th grades, and surveys were completed by 1003 children in 3rd, 6th, 9t h, and 12th grades. Interviews and surveys consisted of a measure designed to retrieve participants' first expectancy associate to an alcohol prompt a nd several drinking quantity/frequency questions. Results: Older and higher drinking children were more likely to report posi tive expectancies as their first associate to an alcohol prompt. Age and dr inking-based findings were consistent with organizational structure, dimens ion emphasis shift, and paths of association identified by prior multidimen sional scaling techniques. Conclusions: Consumption of alcohol among children corresponded to accessib ility of positive expectancies in memory. In addition, the use of multidime nsional scaling to study the organization and activation of alcohol expecta ncies in memory was validated.