An application of latent trait theory in analyzing the field test results of a mathematics proficiency test

The application of one of the latent trait models, the normal ogive model to a large scale test is not numerous. The purpose of present study is to find out information provided by the normal ogive model, using the results of the multiple-choice items of the field test of the Tennessee Proficiency test which have four alternatives. The results of 933 sixth- and seventh-grade students were used in this study. The principal factor solution using tetrachoric correlation was employed to ascertain the dimensions. Item characteristic function and information function following the normal ogive model were produced. Maximum likelihood estimates of abilities of 933 students were computed. The distractor analysis (plausibility function) was employed. Information provided by different types of objectives and test items were discussed.