A study of the reliability and validity of the Holmes alcoholism scale

The Holmes Am Scale was found to have low internal consistency reliability when a Likert-type response format was employed in both a college and in an alcoholic sample (N = 116). Comparison of AM scores between the alcoholic and college groups revealed a significant point-biserial correlation between Am scores and group; alcoholics had the higher mean score. Discriminant analysis resulted in 76.3% correct classification of respondents into their respective group (alcoholic vs. college) on the basis of Am scores. In addition, Am scores were found to be related modestly to self-reports of quantity and frequency of alcohol use in the college sample. Although these results provide support for the validity of the Am scale, its reliability is yet to be established.