RESUMO
Researchers are increasingly administering tests developed and validated in paper format via the Internet. Yet, the equivalence between paper and Internet concerning administration of tests is not typically demonstrated. We evaluated the reliability, factorial and external validity, and measurement equivalency of the Internet version of the 20-item Toronto Alexithymia Scale (TAS-20; Bagby, Parker, & Taylor, 1994; Bagby, Taylor, & Parker, 1994; Lumley et al., 2007) vis-à-vis the paper version. Participants (N = 621) completed the TAS-20 either on the Internet or on paper. Reliability and item-to-scale homogeneity were evaluated for each format. We used confirmatory factor analyses (CFAs) to evaluate factorial validity and used CFA-based factorial invariance procedures to determine measurement equivalency. Alpha coefficients and mean interitem correlations (MICs) were adequate for the full-scale TAS-20 Internet and paper versions and the difficulty identifying feelings (DIF) and difficulty describing feelings (DDF) factor scale test scores; in both formats, alpha and MICs were poorer for externally oriented thinking (EOT) factor test scores compared to scores for the DIF and DDF. The fit of the 3-factor structure of the TAS-20 was adequate for both formats. Factorial invariance across formats was also demonstrated; mean scores for the total scale and each factor scale were not different across formats. Correlations with the domain and facet scales of the Revised NEO Personality Inventory (Costa & McCrae, 1992b) were mostly similar across formats. The Internet and paper versions of the TAS-20 are comparably reliable and valid. An Internet version of the TAS-20 expands opportunities to collect data and permits generalizing of results across studies using the different modes of administration.
