The purpose of this study was to develop an IRT model that would enable the estimation of decision indices based on composite scores. The composite scores, defined as a combination of unidimensional test scores, were either a total raw score or an average scale score. Additionally, estimation methods for the normal and compound multinomial models were explained, and the results from the three measurement models were compared. Finally, the impact on the magnitude of the decision indices based on cut score placement, and score metric was investigated. Both consistency and accuracy indices were estimated. The decision consistency indices included the coefficient of agreement and Cohen's kappa. The decision accuracy indices included the accuracy rate and the false-negative and false-positive error rates. In order to investigate the adequacy and behavior of these decision indices, an empirical dataset was employed, and four composites were created. Four cut scores, located at the 25th, 50th, 80th and 90th percentiles, were used simultaneously and separately, and the trends in the magnitudes of the indices were examined. First, the IRT model estimates for the coefficient of agreement, Cohen's kappa (1960), and the accuracy rate were the largest compared to the normal and compound multinomial models. In turn, the IRT error-rate estimates were smallest relative to the other two models. Second, when the cut score was placed at the 50th percentile, the coefficient of agreement and the accuracy rate were the lowest and Cohen's kappa and the error rates were the greatest compared to the cut scores at the 25th, 80th, and 90th percentiles. The opposite was true for the cut score located at the 90th percentile; the coefficient of agreement and the accuracy rate were the highest and Cohen's kappa and the error rates were the least compared to the cut scores at the 25th, 50th, and 80th percentiles. Finally in the comparison across score metrics, the raw score composites produced a larger coefficient of agreement, Cohen's kappa and accuracy rate and the smallest error rates relative to the scale score metric. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]