When investigating differences between two sets of scores, the t test is appropriate. If the two sets of data are from two groups of subjects, then the independent t test is appropriate. If the two sets are from the same subjects, the dependent t test is required. In this paper, the authors describe the use of a third test when part of a data set meets independent t assumptions and part meets the dependent t assumptions. This test is based on the general linear model. Application to a real study is described. (Author/BW)