ALCOHOL ATTITUDE SCALE FOR COLLEGE STUDENTS

The main purpose of this study was to develop a valid and reliable tri-dimensional attitude instrument for the measurement of college students' attitudes toward alcohol use and abuse. Nine hypotheses were proposed which mainly related to the internal consistency criteria and the external criterion for each item as well as the total instrument. A review of the literature indicated three conceptual areas. A table of specifications was developed based on the identified areas. A pool of statements relevant to the table of specifications was collected from a variety of sources from which a tentative instrument was developed and reviewed by experts for clarity of items. From the pool of items, two preliminary parallel forms (A & B) were constructed and administered randomly among 400 college students. The data obtained were subjected to the following statistical techniques: (1) Pearson correlation for criteria of internal consistency; (2) t-test for discriminatory power of each item between upper and lower groups in terms of their attitude score; (3) t-test for discriminatory power of each item between alcohol drinkers and non-drinkers; and (4) distribution of item responses. Based on the results of the statistical analyses, 54 items were selected for the final instrument. The final instrument was administered among 900 college students in the midwestern United States. The final data obtained was subjected to the same statistical analyses utilized at the preliminary stage plus factor analysis. The following results were obtained: (1) the null hypotheses related to the criteria of internal consistency of each item and subscale were rejected, (2) out of 54 items, 50 discriminated significantly between the drinker and non-drinker groups in terms of their attitude scores, (3) male students scored significantly higher than female students, (4) four factors identified were related to the different content areas rather than to the three components of attitudes, (5) the measure of reliability of the total instrument was .96 using split-half and .94 using Cronbach alpha. It was concluded that the alcohol attitude scale is a valid instrument in the population for which it was designed.