Confused by Mixed ANOVA in SPSS? Step-by-Step Guide

You open SPSS, look at your dataset, and realize that none of the statistical tests you already know seem to fit.

Your dissertation compares two or more groups.

You measured the same participants at two or more time points.

Your supervisor expects a clear answer to one question:

Did the groups change differently over time?

A paired t-test only examines change within one group. An independent samples t-test only compares groups at one time point. A one-way ANOVA ignores the repeated nature of your data. If you run the wrong test, you risk writing incorrect conclusions and undermining your dissertation.

Mixed ANOVA solves this problem.

Mixed ANOVA combines repeated measures and group comparisons in one model. It tests:

1. Whether scores changed over time.
2. Whether groups differed overall.
3. Whether one group changed more than another.

That third result, the interaction effect, often provides the most important finding in intervention-based dissertations.

This guide explains exactly how to:

• Decide whether mixed ANOVA fits your study.
• Structure your data correctly.
• Run the analysis in SPSS.
• Interpret every major output table.
• Write the results in APA style.
• Avoid common mistakes that cost students days of unnecessary work.

If you have already spent hours staring at SPSS output and wondering what the interaction effect means, this article will help you move forward.

What Is Mixed ANOVA?

Mixed ANOVA is a statistical test used when your study includes:

• At least one within-subjects factor (repeated measurements such as pre-test, post-test, and follow-up).
• At least one between-subjects factor (independent groups such as treatment and control).
• One continuous dependent variable.

The term mixed refers to the combination of two different types of factors in one analysis.

Researchers also use these names:

• Mixed factorial ANOVA
• Split-plot ANOVA
• Two-way mixed ANOVA
• Repeated measures factorial ANOVA

SPSS runs mixed ANOVA through Analyze → General Linear Model → Repeated Measures.

If you need a refresher on repeated designs, read this guide on Repeated Measures ANOVA in SPSS. If you want to understand multi-group comparisons, see Factorial ANOVA in SPSS.

When Should You Use Mixed ANOVA?

Use mixed ANOVA when your research design includes both repeated observations and independent groups.

Example 1: Nursing Dissertation

A student measures pain scores:

• Before treatment
• After treatment

Across:

• Intervention group
• Control group

Example 2: Education Dissertation

A researcher compares exam scores:

• Beginning of semester
• End of semester

Across:

• Online students
• Classroom students

Example 3: Psychology Dissertation

A doctoral candidate measures anxiety:

• Session 1
• Session 2
• Session 3

Across:

• CBT group
• Waitlist group

Example 4: Organizational Research

A student measures employee engagement:

• Baseline
• Three months later

Across:

• Leadership training group
• Non-training group

In every case, mixed ANOVA answers whether the pattern of change differs between groups.

Why Dissertation Students Get Stuck

Mixed ANOVA looks intimidating because it combines several statistical ideas at once.

Students often struggle with:

• Choosing the correct statistical test.
• Structuring data correctly.
• Understanding assumptions.
• Interpreting the Time × Group interaction.
• Writing clear results for Chapter Four.

Many students run the procedure successfully but still ask:

• Which p-value should I report?
• What does partial eta squared mean?
• Which effect actually answers my hypothesis?
• Why does SPSS show several tables that look similar?

These questions are normal. Mixed ANOVA becomes manageable once you understand the logic behind the test.

What Research Question Does Mixed ANOVA Answer?

Mixed ANOVA answers a practical question:

Did the outcome change over time, and did that change differ across groups?

Suppose your hypothesis states that a training program improves productivity.

If both groups improve equally, the training may not deserve credit.

If the treatment group improves far more than the control group, the intervention likely worked.

The interaction effect captures that difference.

Understanding Within-Subjects and Between-Subjects Factors

Within-Subjects Factor

A within-subjects factor represents repeated measurements collected from the same participant.

Examples:

• Pre-test and post-test
• Baseline, 1 month, and 3 months
• Low, medium, and high workload conditions

Between-Subjects Factor

A between-subjects factor divides participants into separate groups.

Examples:

• Treatment vs control
• Male vs female
• Public vs private schools

Dependent Variable

The dependent variable contains the outcome you want to analyze.

Examples:

• Anxiety score
• Productivity score
• Satisfaction score
• Blood pressure

Types of Effects in Mixed ANOVA

Mixed ANOVA produces three core findings.

Main Effect of Time

This effect tests whether scores changed across repeated measurements.

Question answered: Did participants improve or decline overall?

Main Effect of Group

This effect tests whether groups differed overall.

Question answered: Did one group score higher on average?

Interaction Effect (Time × Group)

This effect tests whether the amount of change differed between groups.

Question answered: Did the treatment group improve more than the control group?

For most dissertations, the interaction effect matters most because it directly evaluates the hypothesis.

2×2 Mixed ANOVA Explained

A 2×2 mixed ANOVA contains:

• Two groups
• Two time points

Example

This design is the most common mixed ANOVA used in dissertations.

Two-Way Mixed ANOVA

A two-way mixed ANOVA includes:

• One within-subjects factor
• One between-subjects factor

The term two-way refers to the two factors in the model.

Three-Way Mixed ANOVA

A three-way mixed ANOVA adds another factor.

Example

This design allows you to examine whether the treatment effect differs by gender.

Assumptions of Mixed ANOVA

Check these assumptions before interpreting your results.

1. Independence of Observations

Each participant must belong to only one group.

2. Normality

The dependent variable should be approximately normally distributed in each group.

Use the guides on Normality Test in SPSS and Kolmogorov-Smirnov Test in SPSS to verify this assumption.

3. Homogeneity of Variance

Groups should have similar variances.

SPSS reports Levene’s Test.

4. Sphericity

When you have three or more repeated measurements, the variances of the difference scores should be similar.

SPSS reports Mauchly’s Test.

How to Structure Your Data in SPSS

Each row should represent one participant.

Where:

• Group: 0 = Control, 1 = Treatment
• PreTest and PostTest: Repeated measurements
• Outcome: Productivity score

Before running the analysis, clean your dataset and verify coding accuracy. The tutorials on SPSS Data Entry and Data Cleaning in SPSS explain how to prepare a reliable dataset.

Step-by-Step: How to Run Mixed ANOVA in SPSS

Step 1: Open the Repeated Measures Dialog

Click:

Analyze → General Linear Model → Repeated Measures

Step 2: Define the Within-Subjects Factor

Enter:

• Within-Subjects Factor Name: Time
• Number of Levels: 2

Click Add, then Define.

Step 3: Assign Repeated Measures Variables

Move:

• PreTest to Time 1
• PostTest to Time 2

Step 4: Add the Group Variable

Move the Group variable into the Between-Subjects Factors box.

Step 5: Request Descriptive Statistics and Effect Sizes

Click Options and select:

• Descriptive Statistics
• Estimates of Effect Size
• Homogeneity Tests

Step 6: Request a Profile Plot

Click Plots:

• Place Time on the Horizontal Axis.
• Place Group as Separate Lines.

Click Add.

Step 7: Run the Analysis

Click OK.

SPSS will generate the complete mixed ANOVA output.

How to Read the Output Without Getting Overwhelmed

SPSS may produce several tables, but only a few require close attention.

1. Descriptive Statistics
2. Levene’s Test
3. Mauchly’s Test (if applicable)
4. Tests of Within-Subjects Effects
5. Tests of Between-Subjects Effects
6. Estimated Marginal Means and Profile Plot

Focus on these tables first.

Descriptive Statistics Table

This table shows the means and standard deviations for each group at each time point.

Example:

The treatment group shows a much larger increase.

These values often help you explain the interaction effect in plain language.

Levene’s Test

Levene’s Test evaluates whether group variances are similar.

Interpretation

• p > .05: Assumption satisfied.
• p < .05: Variances differ.

If the assumption fails, interpret the results cautiously and mention the issue in your dissertation.

Mauchly’s Test of Sphericity

This table appears only when you have three or more repeated measurements.

Interpretation

• p > .05: Sphericity assumption satisfied.
• p < .05: Use Greenhouse-Geisser corrected results.

Tests of Within-Subjects Effects

This table contains:

• Main effect of Time
• Time × Group interaction

Example Output

Interpretation

Scores changed significantly over time, and the amount of change differed between groups.

The intervention group improved more than the control group.

Tests of Between-Subjects Effects

This table reports the main effect of Group.

Example

Interpretation

The groups differed significantly when averaged across all time points.

Which Result Answers Your Hypothesis?

If your hypothesis predicts that an intervention changes outcomes over time, focus on the Time × Group interaction.

This effect tells you whether the treatment group changed more than the control group.

If the interaction is not significant, the intervention did not produce a statistically detectable difference in change.

Interpreting Partial Eta Squared

Partial eta squared measures effect size.

Common benchmarks:

• .01 = Small effect
• .06 = Medium effect
• .14 = Large effect

For example, a partial η² of .36 indicates a large effect.

How to Interpret the Profile Plot

The profile plot displays the means over time.

What to Look For

• Parallel lines suggest little or no interaction.
• Diverging lines suggest a strong interaction.
• Crossing lines suggest a complex interaction.

Use the plot to communicate the practical meaning of your findings.

Complete Interpretation Example

A doctoral student evaluates whether a resilience training program improves nurse burnout scores.

Results

• Time: p < .001
• Group: p = .041
• Time × Group: p < .001

Plain-Language Interpretation

Burnout scores changed significantly over time. The treatment and control groups also differed overall. Most importantly, the treatment group experienced a much larger reduction in burnout than the control group.

Research Conclusion

The resilience training program significantly reduced burnout.

APA Style Reporting Example

A mixed ANOVA examined the effects of treatment group (intervention vs. control) and time (pre-test vs. post-test) on productivity scores. The analysis revealed a significant main effect of time, F(1, 38) = 35.82, p < .001, partial η² = .48. The main effect of group was also significant, F(1, 38) = 5.32, p = .028, partial η² = .12. The interaction between time and group was significant, F(1, 38) = 21.44, p < .001, partial η² = .36, indicating that participants in the intervention group improved more than those in the control group.

To strengthen your write-up, review Formatting SPSS Tables in APA Format and How to Write Up a Dissertation Analysis Using SPSS.

Common Mistakes That Delay Dissertation Progress

Running Separate t-Tests

This approach ignores the interaction and increases Type I error.

Misreading the Main Effect of Time

Improvement over time does not prove that the intervention caused the change.

Ignoring Assumptions

Unchecked assumptions weaken the credibility of your findings.

Using the Wrong Data Layout

SPSS requires repeated measurements in separate columns.

Reporting Only p-values

Always include F-values, degrees of freedom, and effect sizes.

Writing Statistical Jargon Without Explanation

Translate the findings into clear dissertation language.

Mixed ANOVA vs Repeated Measures ANOVA

Use repeated measures ANOVA when you have one group measured repeatedly.

Use mixed ANOVA when you have repeated measurements and independent groups.

Mixed ANOVA vs One-Way ANOVA

One-way ANOVA compares independent groups at a single time point.

Mixed ANOVA analyzes group differences and change over time simultaneously.

For background, read the step-by-step guide on One-Way ANOVA in SPSS.

Mixed ANOVA vs MANOVA

If you analyze several dependent variables at once, you may need MANOVA instead.

See What Is MANOVA?.

What If Your Data Violate Assumptions?

If assumptions fail, you still have options.

• Use corrected degrees of freedom when sphericity fails.
• Consider transformations if the distribution is highly skewed.
• Use linear mixed models when missing data are substantial.

The key is to explain your decisions clearly.

Sample Size Considerations

Mixed ANOVA does not require a fixed minimum sample size.

The appropriate sample depends on:

• Expected effect size
• Number of groups
• Number of repeated measurements
• Desired statistical power

Estimate sample size before data collection using power analysis and formulas such as the Cochran Formula for Sample Size.

Real-World Dissertation Scenario

A DBA student investigates whether leadership coaching increases employee engagement.

Design

• Group: Coaching vs No Coaching
• Time: Baseline and 12 Weeks
• Outcome: Engagement Score

Findings

The coaching group increased from 58 to 76, while the control group increased from 57 to 60.

The interaction effect was statistically significant.

Interpretation

Leadership coaching produced a meaningful improvement in engagement.

This conclusion directly answers the research objective.

Practical Checklist Before You Submit Chapter Four

Use this checklist to avoid last-minute corrections.

• Confirm that mixed ANOVA matches your design.
• Verify data coding.
• Check assumptions.
• Interpret the interaction first.
• Report effect sizes.
• Include a profile plot.
• Write a plain-language conclusion.
• Format tables according to APA guidelines.

Conclusion

Mixed ANOVA in SPSS provides the correct solution when your dissertation compares groups across repeated measurements.

The analysis answers three essential questions: whether scores changed over time, whether groups differed overall, and whether one group changed more than another.

The interaction effect usually carries the greatest importance because it directly tests whether your intervention worked.

Once you structure the dataset correctly and focus on the key output tables, mixed ANOVA becomes far less intimidating.

If you have been stuck trying to decide which p-value matters or how to explain the interaction effect, you now have a practical roadmap for completing the analysis and writing a strong Chapter Four.

Frequently Asked Questions

1. What is mixed ANOVA in SPSS?

Mixed ANOVA tests repeated measurements and group differences in the same model.

2. When should I use mixed ANOVA?

Use it when the same participants are measured more than once and belong to independent groups.

3. What is the most important result?

The Time × Group interaction usually provides the key finding.

4. What is a 2×2 mixed ANOVA?

It includes two groups and two repeated measurements.

5. Can I use three or more time points?

Yes. SPSS will also test the sphericity assumption.

6. What does a significant interaction mean?

It means the groups changed differently over time.

7. What if Mauchly’s test is significant?

Use the Greenhouse-Geisser corrected results.

8. What if my data are not normally distributed?

Mixed ANOVA often remains robust with adequate sample sizes, but you should report the issue.

9. What if participants have missing values?

Extensive missing data may require linear mixed models.

10. Can I analyze Likert scale totals?

Yes. Composite scores often function as continuous variables. See Likert Scale Analysis Help.

11. How do I report mixed ANOVA in APA style?

Report the F-statistic, degrees of freedom, p-value, and partial eta squared.

12. What sample size do I need?

The required sample depends on expected effect size and desired power.

13. Which assumption matters most?

Check normality, homogeneity of variance, independence, and sphericity when applicable.

14. What if my supervisor says the interpretation is unclear?

Translate the statistical results into a direct answer to the hypothesis.

15. How does mixed ANOVA differ from repeated measures ANOVA?

Repeated measures ANOVA analyzes one group over time, while mixed ANOVA also compares independent groups.

16. Can I use mixed ANOVA for survey research?

Yes, provided the outcome variable is continuous and measured repeatedly.