Paired Samples T Test - Running, Interpreting, and Reporting Paired-Sample T-Test Results using SPSS

A paired-samples t test is the go-to statistical tool when the same people are measured twice—before and after an intervention—and the goal is to determine whether their continuous outcomes changed in a statistically meaningful way. The method is also known as a repeated-measures technique: it compares each respondent’s “before” score with their own “after” score, rather than comparing two independent groups. That pairing matters because it controls for stable individual differences, making the test more sensitive to real change caused by training, equipment, campaigns, or other interventions.

In the SPSS workflow, the analysis starts with two columns of data: one for the “before” measurements and one for the “after” measurements. After opening SPSS and going to Analyze → Compare Means → Paired-Samples T Test, the “before” variable is assigned to Variable 1 and the “after” variable to Variable 2. Once the test runs, SPSS produces multiple tables, but the key ones for interpretation are the Pairwise Sample Statistics table (descriptive summaries) and the Pairwise Samples Test table (the inferential results). The correlation table between before and after scores is not required for the main significance decision.

Using the example provided, 50 respondents are measured on a continuous outcome (student scores). The mean score rises from 6.2 before training to 12.56 after training. The paired-samples t test then quantifies whether that increase is likely to reflect more than random variation. The reported mean difference is 6.445, with a standard deviation of 4.42 for the difference scores. The t statistic is greater than the common two-tailed threshold of 1.96, and the result is treated as statistically significant—meaning the intervention is associated with a real improvement (not just noise). Because the after mean is higher than the before mean, the change is an increase.

Significance alone doesn’t tell researchers how large the improvement is, so the analysis also calls for an effect size calculation. The transcript uses an ETA-squared approach based on the t value and sample size (N = 50), applying the formula t² / (t² + N − 1). The computed ETA-squared is about 0.72, which Cohen’s 1988 guidelines classify as a large effect. In practical terms, that suggests a substantial shift in students’ ability after SPSS training.

Finally, reporting guidance is emphasized: results should be presented in a properly formatted table rather than copied directly from SPSS output. A template is provided for a paired-samples t test write-up, including the intervention description, the before and after means and standard deviations, the t statistic, the p value (noted as < .001), the mean increase, and a 95% confidence interval that excludes zero (reported as ranging from −7.59 to −5.29 in the example). The combination of a significant p value and a confidence interval that does not cross zero supports the conclusion that the intervention produced a meaningful improvement, with a large effect size indicating the magnitude is not trivial.