What are the assumptions about the pre- and post-tests when the t-statistic is used?
The paired-samples t test assumes that each subject’s score is independent of any other subject’s score and that the difference scores are normally distributed. The assumption of independent observations between subjects is controlled by the study design. When the assumption of normality is not met, the paired-samples t test can be invalid. For each evaluation program, a sample size of 30 or more teachers should be large enough to meet the normality assumption and the paired-sample t test can be used validly. For sample sizes with fewer than 30 teachers, the spreadsheet uses a nonparametric statistical test called the Wilcoxon signed ranks test.
