The potential of the Rasch model to develop scores, on a ratio scale, suitable for interindividual comparisons, from intact groups with disparate distribution characteristics was investigated. The specific problems studied were: (1) the effects of skewed test score distributions on the ability parameter of the Rasch measurement model; (2) the effects of group size on the ability parameter of the Rasch measurement model; (3) the interactive effects of skewed test score distributions and group size on the ability parameter of the Rasch measurement model; and (4) the effects of skew and total group size on the standard errors of estimate of item log easiness estimates. The data for the study in the form of item responses were randomly selected from 120,000 students who participated in the Florida State-Wide Testing Program in September 1971. The 90-item mathematics test was selected. Thirty-five raw score distributions characterized by seven levels of skew and five group sizes were constructed. Group responses were submitted to a computer program which estimated the model's parameters according to a maximum likelihood procedure. Results of the study indicated that the estimates derived from the Rasch measurement model were not independent of the group used to produce them. Differences were minimal in the middle score range, but large in low and high score range. Eleven tables present the study data. (Author/DB)