The use of a three-parameter logistic model for applying latent trait theory has become more popular because of the availability of computer programs. The program Logist 5 can be used not only for the item parameter estimation in the three-parameter logistic model, but also in the two parameter logistic model by setting the third parameter equal to zero. In the present study, comparison was made between the estimated item parameters obtained by assuming the normal ogive model and those obtained by Logist 5 assuming the two parameter logistic model. In some cases, item parameter estimation was also made by Logist 5 assuming the three-parameter logistic model, and comparison was extended to those results also. Data were obtained from the performance of fifth through ninth graders in Japan who took a battery of thirteen tests, Shiba's Word/Phrase Comprehension Tests. Results from four tests were chosen for this study. Resultant estimated item parameters were compared with those obtained using the tetrachoric methods. Results showed the mean and standard deviation of the distribution of the maximum likelihood estimate seem to be somewhat different from those of the distribution theta, and a certain systematic tendency was observed. (LMO)