Many researchers have suggested that the main cause of item bias is the misspecification of the latent ability space. That is, items that measure multiple abilities are scored as though they are measuring a single ability. If two different groups of examinees have different underlying multidimensional ability distributions and the test items are sensitive to these differences, any scoring scheme that does not reflect all of the skills in the interaction of the items and examinees (the complete latent space) will likely produce item bias. Insight is offered the testing practitioner concerning the difference between item bias and item impact and how each relates to item validity. These concepts are addressed from a multidimensional item response theory perspective. Two detection procedures, the Mantel-Haenszel (as modified by P. W. Holland and D. T. Thayer) and the Simultaneous Item Bias procedure of R. Shealy and W. Stout are used to illustrate the detection of item bias. It is concluded that empirically, two or more items will always produce multidimensionality, and as such, their parameters need to be estimated using multidimensional models. One table and 12 figures illustrate the discussion. (SLD)