The independent-samples t test is appropriate whenever two means drawn from independent samples are to be compared. The variable used to form the groups may already exist; however, a cut point on a continuous variable can be provided to dynamically create the groups during the analysis. As with all t tests, the independent-samples t test assumes that each sample mean comes from a population that is reasonably normally distributed, especially with respect to skewness. Test variables with extreme or outlying values should be carefully checked; boxplots can be used for this.

  • Use the Explore procedure or the One-Sample Kolmogorov-Smirnov Test procedure to test the assumption of normality.
  • Use the Runs Test procedure to check the assumption that the value of the test variable is independent of the order of observation.
  • If your grouping variable has more than two groups, try the One-Way ANOVA procedure.
  • If your test variables do not satisfy the assumptions of the two-sample t test, try the Mann-Whitney-Wilcoxon tests in the Two-Independent-Samples Tests procedure.