Measuring adolescent alcohol outcome expectancies.

The psychometric properties and construct validity of the Comprehensive Effects of Alcohol (CEOA) questionnaire were compared with those of the Alcohol Expectancy Questionnaire--Adolescent version (AEQ-A) in relation to adolescent alcohol consumption. Both measures of adolescent alcohol expectancies were found to be internally reliable and temporally stable. Alcohol use was significantly associated with subjective evaluations for Cognitive and Behavioral Impairment and Self-Perception on the CEOA and with expected effects for Cognitive and Motor Impairment and Changes in Social Behavior on the AEQ-A. Compared with the AEQ-A, the CEOA explained more variance in quantity (28%) and a similar variance in frequency (15%) for adolescent alcohol use (AEQ-A quantity = 20%, frequency = 15%). Whereas the general content and psychometric properties of the 2 measures are markedly similar, the Likert response format, shorter length, and assessment of both expected effects and subjective evaluations with the CEOA may offer measurement advantages over the AEQ-A.