In this study, we analyzed dimensionality for Pettigrew and Meertens' Blatant and Subtle Prejudice Scale (1995). More specifically, we sought to verify two hypotheses: first, that the scale is one-dimensional, and second, that the difference in the items featured in the two subscales (blatant and subtle) is exclu-sively due to their different locations on the scale of the latent variable. Expressing the latter hypothesis in the language of the Item Response Theory measurement approach we expected items from the bla-tant subscale to be more difficult (they require a greater level of prejudice to be positively answered by persons) than those of the subtle subscale. To this end, we used two databases (N 1 = 323; N 2 = 268), which we analyzed through confirmatory factor analysis and the partial credit model. Results con-firmed our hypotheses, and highlighted the usefulness of the Item Response Theory models in analyz-ing prejudice scales.