Using the Paired Samples T-Test Function
The paired samples t-test in EZAnalyze allows you to determine the statistical significance of a difference between two paired means. This test is the most appropriate when you have a repeated measures design where the data are from a single group of cases exposed to two measures over time, as in a pretest-posttest design. Since this is an Advanced feature of EZAnalyze, it will not be discussed at length here. Feel free to consult a statistics textbook or the Internet for information on what a paired t-test is and how it works. If you are already familiar with paired samples t-tests and want EZAnalyze to crunch the numbers for you, read on!
To conduct this analysis, your data will need to have at least two "paired" variables - each variable should have similar data. Think of each of these as a dependent variable. What this test does is tell you if the mean of the first variable is significantly different from mean of the second variable.
To conduct an independent samples t-test, select "Advanced" from the EZAnalyze menu in Excel, then choose "T-Tests" and "Paired Samples".
In the "Paired Samples T-Test" dialog box, select your paired variables - one variable under "Variable 1" and another under "Variable 2" and click OK. If you are running this test on a pre-post test experimental design, you will want to use Variable 1 for your pretest variable, and Variable 2 for your posttest variable.
A note about your variables. These variables should be on the same metric - for example, GPA for social studies and science, or the mean scores of a variable coded from 1-5. This procedure will not be accurate if your variables are not on the same metric - for example, comparing GPA and SAT scores is not appropriate in a paired samples t-test.
When you click OK, a results report will be printed on a separate sheet for your review. (click on "results report" for information on how to interpret this analysis)