A psychometric analysis of the tobacco use and attitude survey: Methods, results, and recommendations

Research has confirmed that 27% of college students are smokers. To tailor effective tobacco prevention and control policies and practices for students, it is essential to evaluate influential attitudes, behaviors, and tobacco marketing strategies. Unfortunately, few tobacco and college student specific instruments have been developed and validated. Currently, the Tobacco Use & Attitude Survey developed by the BACCHUS Network is widely used on campuses. The purpose of this study was to quantify the psychometric properties of this 65-item instrument containing items designed to evaluate: perceived frequency of tobacco use on campus, self-reported use, reasons for use, attitudes toward smoking, and exposure to tobacco advertising and prevention messages. Several strategies were included in the psychometric analysis of this instrument. A test/retest protocol determined consistency of subject responses. Data was collected from a convenience sample of 40 matched pairs of undergraduate subjects at a Midwestern University. One week elapsed between data collection points. A factor analysis was conducted to evaluate the validity of items and a Cronbach's Alpha test determined internal consistency of the instrument. The factor analysis confirmed several subscales. Internal consistency reliability was moderately high with the exception of one subscale (Cronbach's a=16-.84) (mean =.57), and test-retest (reliability correlation between forms) ranged from .33 to 1.0 (mean = .78). These results suggest that while some items yielded high internal consistency values and test/retest values, several items should be revised to improve the psychometric properties of this instrument.