Standardized tests are widely used in intellectual disability research, either as dependent or control variables. Yet, it is not certain that their items give rise to the same performance in various groups under study. In the present work, 48 participants with Down syndrome were matched on their raw score on Raven's Colored Progressive Matrices (RCPM) with 48 participants with intellectual disability of undifferentiated etiology and 48 typical children. An item analysis was then conducted using the "transformed item difficulties" method, a statistical approach designed to detect "differential item functioning" between groups. Results showed that the difficulty of items was highly similar for the three groups. It is concluded that RCPM can be used with considerable confidence in comparative studies including typical, Down syndrome and intellectually disabled participants of undifferentiated etiology. Some methodological implications of these findings are discussed. (Contains 3 figures and 1 table.)