ABSTRACT
Because we are now looking at the difference between two sample means rather than a sample mean and a population mean, statisticians have had to identify how this new statistic is distributed-what its mean and standard error are-so that an inferential statistic can be used to evaluate it. In fact, there are different versions of the test, depending on whether the design is between-or within-subjects. The complete equations for the tests are shown in Appendix VI, where worked examples are given. All the versions of the test use the t-distribution and so the same probability table can be used as for the one-group t-test.
