Survey of attitudes toward statistics (SATS): An investigation of its construct validity and its factor invariance by gender

The construct validity of the Survey of Attitudes Toward Statistics (SATS) was investigated by examining its factorial structure and by determining its gender invariance properties. The scale was administered to Social and Political Studies undergraduate students of a large European University. Construct validity evidence for the SATS was assessed using parcelsbased CFA correlations, group differences and multiple regression analysis. The four-factor model was adopted and used for subsequent analysis for both empirical and theoretical reasons. All correlations among subscales were high except for the correlation between the Difficulty and the Value subscales. Male and female students' attitudes toward statistics courses differ slightly in the Affective and the Cognitive competence subscales but not in the Value subscale. In addition, the results showed no differences between male and female students in the expected statistics course achievement.