The usefulness of a latent trait model designed for use with multidimensional test data was investigated in two stages. The first stage consisted of generating simulation data to fit the multidimensional extension of the two-parameter logistic model, applying the model to the data, and comparing the resulting estimates with the known parameters. The unidimensional two-parameter logistic model was also applied to these data. In addition to comparing the estimated parameters with the true parameters, the fit of the unidimensional and multidimensional models to the data were compared. The second stage of the study employed real response data. Items were selected from various subtests of a larger test that had been administered to a large sample in such a way as to simulate shorter tests with varying numbers of subtests. The unidimensional and multidimensional models were applied to these data, and the resulting estimates were used to evaluate the fit of the models to the data. The fit of the two models to the data were then compared and it was determined that the multidimensional model more adequately modeled the real data than did the unidimensional model. (Author/PN)