Abstract
Measuring attitude toward homosexuality with well-validated instruments is a critical step for a better understanding of the anti-lesbian/gay public prejudice (formerly known as homophobia). The Attitudes Toward Lesbians and Gay Men Scale-Short Version (ATLG-S) is one of the most known measures of attitude toward lesbians and gay men. This study aimed to test the factorial structure, construct validity, discriminant validity, and reliability of the Romanian version of the ATLG-S. Participants were 496 Romanian graduates, master’s students, and undergraduates from five universities. The sample included 392 women and 104 men aged 18 to 47 years. Participants completed the translated version of ATLG-S along with other measures of religiosity and openness (as a domain of personality stable traits). We used confirmatory factor analysis to test the internal validity of the ATLG-S. The hypothetical model with a single latent factor and all ATLG-S items showed an unsatisfactory statistical fit. Results indicated a slight improvement in the statistical fit of this second model, but still unsatisfactory. Suggestions for improving the model led us to correlate three pairs of items, resulting in a third model with a satisfactory statistical fit (GFI =0.953; TLI = 0.953; CFI = 0.969; RMSEA = 0.085). The internal consistency of the nine-item version of ATLG-S (ATLG-S-9) was excellent (α = 0.91). Correlational analysis revealed a positive and moderate association (r = 0.48; p < 0.001) between ATLG-S-9 score and religiosity. The correlation with openness was negative, but modest (r = - 0.14; p < 0.01). Comparative data are also presented.

Keywords: Lesbians; gay men; attitude; measurement; Romanian version of ATLG-S; psychometric properties

DOI: doi:10.29081/JIPED.2021.25.1.02