Intro to SPSS > Paired Samples T Test in SPSS
About the Paired Samples T Test in SPSS
The purpose of the paired samples t test in SPSS is to find statistically significant differences between the means of two sets of related observations. For example, you may want to find out if a depression treatment works. You measure a group of patients for their depression levels (the “pretest”), then assign some type of intervention, like counseling. Some time later, you measure the group’s depression levels again. Hopefully, the mean depression level will have gone down: the paired samples t-test can help quantify this.
The null hypothesis for the test is that the differences between the two population means in zero (no difference). It’s often used in pretest-posttest scenarios.
Assumptions
Assumptions are similar to other t-tests:
- Dependent scale variable.
- Observations should be independent.
- Dependent variable should look approximately normal.
- No outliers for the dependent variable.
Paired Samples T-Test SPSS: Steps
Follow these steps to run a paired samples t-test in SPSS:
- Click analyze
- Select compare means
- Click paired samples t-test
- Click the “pre-test” group, then click the center arrow to move the variable to the right.
- Click the “post-test” group, then click the center arrow to move the variable to the right.
- Click OK.
Reading the SPSS Paired T Test Output
The first thing you’ll want to look at is the means to see if they are different. This is shown on the image below:
Next, check to see that the calculated sample mean falls into the confidence interval:
Finally, check your p-value—the probability the observed value was due to chance. This is shown to the far right in the above figure. If this figure is smaller than your alpha level, then you can reject the null hypothesis that the means are equal.
Note: SPSS runs the paired samples t test at a 95% confidence level / 5% alpha level. You can change this by clicking “Options” in the paired samples t-test window:
Subscribe to my YouTube channel for more SPSS and general statistics tricks and tips!