In the alcohol literature, researchers have found that cognitive expectations about how alcohol will affect people are potent predictors of individual alcohol use. These expectations tend to be stronger predictors of drinking than other variables, such as demographics or familial history of alcoholism. People who expect, for example, that alcohol will make them happy, friendly, attractive, etc., drink more alcohol than people who expect that alcohol will make them clumsy, unattractive, and sleepy. These cognitive expectations can be measured in primary school children, and appear to be a product of both societal attitudes toward alcohol intoxication and individual experiences with the drug. We developed the Addiction Research Center Drug Expectancy Questionnaire (ARCDEQ) to extend this research to drugs other than alcohol. We included tobacco as a comparison or control drug because we thought it unlikely to show the magnitude of subjective expectations associated with the other drugs. The ARCDEQ has 46 items for each of five drugs (i.e., alcohol, cocaine, heroin, marijuana, and tobacco). Thirty six of the items referred to expectations about acute effects of the drug, such as feeling """"""""anxious"""""""" or feeling """"""""bold."""""""" These items were balanced to include those on each end of continuums represented by """"""""good"""""""" vs. """"""""bad"""""""" effects, stimulant vs. depressant effects, enhancing vs. impairing effects, and prosocial vs. hostile effects. Ten items were also included for each drug to assess longterm expectations about how chronic use will affect them, such as """"""""enjoy life more"""""""" or """"""""get sick."""""""" Each item used a Likert scale with descriptive anchors of """"""""unlikely"""""""" to """"""""likely"""""""", with """"""""neither"""""""" in the center of the scale. We have administered this questionnaire in computerized form to 86 abstinent, light, and heavy drug users. To date, 75 men and 11 women have answered the ARCDEQ. Twenty subjects answered the questionnaire twice to assess test-retest reliability. For each drug, we divided subjects based on whether they were naive, light, or heavy users of that particular drug, based on quantity-frequency items from the ARCDEQ itself, and the ARC Drug Use Survey, administered separately. We will report on preliminary analyses of this data in which we eliminated some items and developed empirically-derived scales for data reduction and reporting. We consider expectations about abused drugs to be important for two reasons. First, they suggest motives for excessive drug use and may predict who escalates from experimental to heavy use. Second, they have direct relevance for prevention and psychological treatment of drug abuse, namely, to develop interventions to change these expectancies.
