A study was conducted to investigate the feasibility of using IRT equating for the GRE Subject Test in Mathematics. Two forms of the test were equated using the three-parameter logistic (3PL) model, and the results were compared to the results of the Tucker equating procedure currently used operationally, as well as to equipercentile equating. In addition, the two basic assumptions of the 3PL model were investigated to determine whether they were reasonable within the context of the GRE Mathematics Test. These assumptions were that of unidimensionality, which was investigated using nonlinear, full-information factor analysis, and that the item response function follows a logistic form with three parameters, which was investigated using two different goodness-of-fit procedures. The results of the study indicate that the GRE Mathematics Test is reasonably unidimensional, and that the three-parameter logistic model does yield reasonable fit to the data. It was found that both the IRT and the equipercentile equating procedures yielded somewhat different results than the Tucker method. The results for the IRT procedure and the equipercentile procedure were quite similar.