Abstract
Effective interventions require an understanding of the behaviors and cognitions that facilitate positive change as well as the development of psychometrically sound measures. This paper reports on the psychometric properties of the Temptations to Try Alcohol Scale (TTAS), including factorial invariance across different subgroups. Data were collected from 3565 6th grade RI middle school students. Structural equation modeling was used to determine the appropriate factorial invariance model for the 9-item TTAS. The measure consists of three correlated subscales: Social Pressure, Social Anxiety, and Opportunity. Three levels of invariance, ranging from the least to the most restrictive, were examined: Configural Invariance, which constrains only the factor structure and zero loadings; Pattern Identity Invariance, which requires factor loadings to be equal across the groups; and Strong Factorial Invariance, which requires factor loadings and error variances to be constrained. Separate analyses evaluated the invariance across two levels of gender (males vs. females), race (white vs. black) ethnicity (Hispanic vs. Non-Hispanic) and school size (small, meaning <200 6th graders, or large). The highest level of invariance, Strong Factorial Invariance, provided a good fit to the model for gender (CFI: .95), race (CFI: .94), ethnicity (CFI: .94), and school size (CFI: .97). Coefficient Alpha was .90 for Social Pressure, .81 for Social Anxiety, and .82 for Opportunity. These results provide strong empirical support for the psychometric structure and construct validity of the TTAS in middle school students.