Did anxiety fall from pretest to posttest to follow-up? Do participants rate three product designs differently? When the same people are measured on three or more occasions or conditions, a series of paired t-tests inflates the error rate, and the correct test is a repeated measures ANOVA. Its SPSS output contains many tables and four versions of the main F-test, and students often report the wrong row. This guide explains when to use the test, its assumptions, the eight steps in SPSS, how to read the output, and how to report it in APA style.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Statistical Analysis service. Sources last checked: November 2026.
What is a repeated measures ANOVA? A repeated measures ANOVA is a test that compares the means of three or more related measurements taken from the same participants, such as scores at several time points or under several conditions. It removes stable differences between individuals from the error term, which makes it more powerful than comparing independent groups of the same size.
When should you use a repeated measures ANOVA?
Use it when:
- the dependent variable is continuous (interval or ratio, or a scale score treated as such);
- the same participants provide every measurement (a within-subjects factor);
- there are three or more levels; with two levels the paired samples t-test gives the same answer.
You can add a between-subjects factor (for example, intervention and control groups measured at three time points). That is a mixed or split-plot ANOVA, run in the same SPSS dialog.
What are the assumptions?
| Assumption | How to check | If violated |
|---|---|---|
| Continuous dependent variable | Measurement level | Use the Friedman test |
| No extreme outliers in any level | Boxplots of each measurement | Check data entry; run with and without |
| Approximate normality of each level (or of differences) | Shapiro-Wilk, Q-Q plots; robust with moderate samples | Friedman test or transformation |
| Sphericity | Mauchly’s test in the output | Greenhouse-Geisser or Huynh-Feldt correction |
Sphericity means the variances of the differences between every pair of levels are equal. It is only an issue with three or more levels, and it is the assumption most specific to the repeated measures ANOVA.
📱 Ask a statistician about your data — Send a short description of your instrument and sample size on WhatsApp and get a suitable analysis plan. Every analysis comes with the raw SPSS file and a results report.
How do you run a repeated measures ANOVA in SPSS in 8 steps?
Your data must be in wide format: one row per participant and one column per measurement (for example anx_pre, anx_post, anx_follow).
- Go to Analyze > General Linear Model > Repeated Measures.
- In Within-Subject Factor Name, type a name such as time; in Number of Levels, type 3; click Add.
- Optionally type a Measure Name (for example anxiety), then click Define.
- Move the three columns, in order, into Within-Subjects Variables. Move any grouping variable into Between-Subjects Factor(s) if you have one.
- Click Plots, put time on the horizontal axis, click Add, then Continue.
- Click EM Means, move time to Display Means for, tick Compare main effects, and choose Bonferroni as the confidence interval adjustment.
- Click Options and tick Descriptive statistics and Estimates of effect size.
- Click Paste to save the syntax, then run it.
How do you read the output?
SPSS prints several tables. Read them in this order:
- Descriptive Statistics: means and standard deviations at each level.
- Mauchly’s Test of Sphericity: if p > .05, sphericity can be assumed; if p < .05, it is violated. The table also gives the epsilon values.
- Tests of Within-Subjects Effects: four rows for the same effect. Choose one:
| Situation | Row to report |
|---|---|
| Mauchly’s test not significant | Sphericity Assumed |
| Violated, Greenhouse-Geisser epsilon below 0.75 | Greenhouse-Geisser |
| Violated, epsilon 0.75 or above | Huynh-Feldt (or Greenhouse-Geisser, more conservative) |
| Lower-bound | Rarely used; very conservative |
- Pairwise Comparisons: Bonferroni-adjusted differences between each pair of levels, with confidence intervals.
- Multivariate Tests and Tests of Within-Subjects Contrasts: optional; the multivariate approach does not assume sphericity, and contrasts test linear or quadratic trends.
The correction does not change the F value; it multiplies the degrees of freedom by epsilon, which changes the p-value. Report the corrected degrees of freedom, usually with two decimals.
How do you calculate and interpret effect size?
With Estimates of effect size ticked, SPSS reports partial eta squared in the within-subjects table. Common benchmarks (Cohen) are .01 small, .06 medium and .14 large. For a mixed design, report partial eta squared for the main effects and the interaction. Report confidence intervals for the pairwise differences as well.
How do you report a repeated measures ANOVA in APA style?
The figures below are a hypothetical example (36 participants, three time points), not real data:
A repeated measures ANOVA showed that anxiety scores differed across the three time points. Mauchly’s test indicated that the assumption of sphericity was violated, χ²(2) = 7.10, p = .029, so Greenhouse-Geisser corrected results are reported (ε = .81). Anxiety changed significantly over time, F(1.62, 56.70) = 12.40, p < .001, ηp² = .26. Bonferroni-adjusted comparisons showed that anxiety was lower at posttest (M = 41.2, SD = 7.9) than at pretest (M = 47.5, SD = 8.3), p < .001, with no significant change between posttest and follow-up.
Add a line chart of the means with error bars, and a table of means and standard deviations if there are many conditions.
What changes in a mixed design?
If you moved a grouping variable into Between-Subjects Factor(s), the output adds Levene’s test for each measurement, Box’s M test, and a Tests of Between-Subjects Effects table. The key result is usually the time by group interaction in the within-subjects table: it tells you whether the change over time differed between groups. If the interaction is significant, examine simple effects (the effect of time within each group) rather than interpreting the main effects alone, and plot separate lines for each group.
What mistakes should you avoid?
- Reporting the Sphericity Assumed row when Mauchly’s test is significant.
- Running multiple paired t-tests without correction instead of the omnibus test.
- Entering the data in long format, which the GLM Repeated Measures dialog does not accept.
- Ignoring missing values: SPSS drops any participant with a missing measurement (listwise), which can reduce the sample sharply.
- Writing p = .000 instead of p < .001.
- Omitting effect sizes.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
What is the non-parametric alternative to a repeated measures ANOVA?
The Friedman test compares three or more related measurements using ranks. Follow it with Wilcoxon signed-rank tests with a Bonferroni adjustment for pairwise comparisons.
What if Mauchly’s test cannot be computed?
With only two levels, sphericity is always met, and SPSS shows no test. With very small samples the test has low power, so many researchers report the Greenhouse-Geisser correction regardless.
Can I use a repeated measures ANOVA with missing data?
SPSS GLM removes participants with any missing measurement. If many are lost, a linear mixed model (Analyze > Mixed Models > Linear) can use all available data.
How many participants do I need?
Run a power analysis in G*Power under ANOVA: Repeated measures, within factors, entering the expected effect size, the number of measurements and the correlation among them.
Is a mixed ANOVA the same test?
It extends a repeated measures ANOVA by adding a between-subjects factor. The key result is often the time by group interaction.
Sources
- Field, A. (2018). Discovering statistics using IBM SPSS Statistics (5th ed.). SAGE.
- Mauchly, J. W. (1940). Significance test for sphericity of a normal n-variate distribution. The Annals of Mathematical Statistics, 11(2), 204-209. https://doi.org/10.1214/aoms/1177731915
- Greenhouse, S. W., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24(2), 95-112. https://doi.org/10.1007/BF02289823
- Huynh, H., & Feldt, L. S. (1976). Estimation of the Box correction for degrees of freedom from sample data in randomized block and split-plot designs. Journal of Educational Statistics, 1(1), 69-82. https://doi.org/10.3102/10769986001001069
- American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.). https://apastyle.apa.org/
- Girden, E. R. (1992). ANOVA: Repeated measures. SAGE.
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.