The aim of this study was to develop a new 29-item HIV/AIDS knowledge measure and to examine its psychometric properties for three samples of adults: non-HIV-positive heterosexual people, non-HIV-positive men who have sex with men (MSM), and people living with HIV/AIDS (PLWHA). The authors collected data using an online questionnaire. A total of 9,349 Italian individuals agreed to participate in the study: 694 individuals (7.4%) were PLWHA, 5,232 (56.0%) were HIV negative, and the remaining 3,423 (36.6%) were MSM. Using two-parameter item response theory analysis, a bifactor model was found to be better fitting than a one-factor model or a 12 correlated first-order factor model. Differential item functioning showed evidence of measurement nonequivalence of the instrument for the three samples of adults. The reliability of HIV/AIDS knowledge scale among PLWHA was satisfactory. Criterion-related validity was only achieved among non-HIV-positive heterosexual people, non-HIV-positive MSM, as the HIV/AIDS knowledge scale was related with attitudes toward condom use, condom use with casual partners, unknown HIV status of partner, and HIV stigma. Among non-HIV-positive heterosexual people, the HIV/AIDS knowledge scale adds to the prediction of condom use above that of attitudes toward condom use scale. The HIV/AIDS knowledge scale is especially discriminating at low to medium levels of knowledge.
Keywords: HIV testing; HIV/AIDS; IRT; bifactor; condom use; knowledge.