In a situation where the population distribution of latent trait scores can be estimated, the ordinary maximum likelihood estimator of latent trait scores may be improved upon by taking the estimated population distribution into account. In this paper empirical Bayes estimators are compared with the liklihood estimator for three samples of 300 cases within the context of the Rasch model. The empirical Bayes estimates varied more than the likelihood estimates, due to the fact that not only item parameters, but also population distribution parameters had to be estimated from the sample data, the largest difference being 0.09 for a score equal to 20. The results were based on a computer simulation for which the model and distributional assumptions were known to be correct. For real data, there is a question of appropriateness of the distributional assumptions--apart from the question of fit of the Rasch Model. The normality assumption can be tested by means of the test of fit proposed by Andersen and Madsen. If the normality assumption turns out to be inadequate, other distributions may be fitted to the data using an approach similar to that proposed in this paper. (Author/RL)