There are two types of independent-samples t tests: pooled variances and separate variances. The former is used when the two population variances are similar to one another, whereas the latter is used when the variances are significantly disparate. The rationale for having these two options is that when
two samples’ variances are similar, they can safely be combined (pooled) into a single estimate of the population variance. When they are markedly unequal, however, they must be mathematically manipulated before being combined. You will not be able to tell merely by looking at two samples’ variances whether you should use a pooled-variance or separate-variance approach, but that is fine. In this book, you will always be told which one to use. When we get to SPSS, you will see that this program produces results from both of these tests, along with a criterion to use for deciding between them. (More on this later.) You will, therefore, always be able to figure out which type of test to use.
Pooled variances: The type of t test appropriate when the samples are independent and the population variances are equal.
Separate variances: The type of t test appropriate when the samples are independent and the population variances are unequal.