University
How to interpret SPSS output: t-test, chi-square, correlation and regression
October 31, 2026
Running a test in SPSS takes a few clicks. Understanding the output tables is the hard part. Here's how to read the four tests students use most, and what to write in chapter four.
The p-value, in one paragraph
Most tests give a significance value, shown as "Sig." in SPSS. If it is below your chosen level, usually 0.05, the result is statistically significant: it's unlikely to be due to chance alone. If it is 0.05 or above, you don't have enough evidence of a difference or relationship. SPSS shows very small values as .000; report these as p < .001.
Independent samples t-test
Use it to compare the average of two groups, for example the test scores of male and female students.
- Group Statistics shows each group's mean and standard deviation: report these.
- In the Independent Samples Test table, first look at Levene's test. If its Sig. is above .05, read the "Equal variances assumed" row; otherwise read "Equal variances not assumed".
- In that row, read t, df and Sig. (2-tailed).
Report it like this: "Female students (M = 64.2, SD = 8.1) scored higher than male students (M = 59.7, SD = 9.3), t(198) = 3.65, p < .001."
Chi-square test of independence
Use it to see whether two categorical variables are related, for example gender and choice of course.
- The Crosstabulation table shows the counts in each group.
- In the Chi-Square Tests table, read the Pearson Chi-Square row: the value, df and Asymp. Sig.
- Check the note under the table: if many cells have an expected count below 5, the result may not be reliable, and Fisher's exact test is often used for 2×2 tables.
Correlation (Pearson's r)
Use it to measure the relationship between two continuous variables, for example study hours and exam scores. The value r runs from −1 to +1: the sign shows the direction and the size shows the strength. As a rough guide, about .1 is weak, .3 moderate and .5 or more strong. Report r, the sample size and the p-value, and remember that correlation doesn't prove one thing causes the other.
Linear regression
Use it to see how well one or more variables predict an outcome.
- Model Summary: R² is the share of the variation in the outcome explained by your predictors (for example, R² = .42 means 42%).
- ANOVA table: if Sig. is below .05, the model as a whole predicts the outcome.
- Coefficients table: for each predictor, B is the change in the outcome for each one-unit increase, Beta lets you compare predictors, and Sig. shows which predictors are significant.
Writing it up
Don't paste raw SPSS tables into your project. Build clean tables in the format your department uses, give the key numbers, and then explain in plain words what each result means for your research question or hypothesis. That explanation is what examiners look for.