t Test, Independent Samples

Uses

The independent samples t test is appropriate whenever the researcher wants to know whether two population group means are different and when the observations in each of the groups are independent of the observations in the other group. In some cases, these groups are composed from naturally occurring populations. For example, a researcher might be interested in comparing the spatial ability of men and women. In this case, all men and women in the country (or at his or her university, etc.) would comprise the two populations (one of men and one of women). The researcher would then construct a sample of men and of women, calculate the sample means, and conduct an independent samples t test to [Page 1552]determine whether any differences between the sample means are real or caused by chance.

In many cases, however, the populations are not naturally occurring but instead are determined by some element of the study. This is particularly common in experimental research, where the researcher manipulates a variable (stress in the preceding example) and assigns participants to one of the two resulting groups. In this case, the groups that result from the manipulation do not reflect actual known populations. Instead, they reflect hypothetical infinite populations that result from the experimental manipulation.

...