Reporting the results of a Repeated Measures ANOVA (Analysis of Variance) accurately and effectively is crucial for clear communication of your research findings. This guide provides a step-by-step approach to ensure your report is both statistically sound and easily understandable. We'll cover everything from the basics of what to include to advanced considerations for complex designs.
Understanding Repeated Measures ANOVA
Before diving into reporting, let's briefly recap what Repeated Measures ANOVA is. This statistical test is used when you have repeated measures on the same subjects. That is, the same participants are measured multiple times under different conditions or at different time points. This contrasts with a between-subjects ANOVA where different participants are assigned to each condition. The key advantage of a repeated measures design is that it controls for individual differences, leading to greater statistical power.
Key Elements to Include in Your Report
Your report should clearly and concisely communicate the following information:
1. Descriptive Statistics
Begin by presenting descriptive statistics for each condition or time point. This typically includes:
- Means: The average score for each condition.
- Standard Deviations: A measure of the variability of scores within each condition.
- Number of Participants: The total number of participants included in the analysis.
Example:
"Participants (N = 30) completed a cognitive test under three conditions: Control (M = 25, SD = 5), Treatment A (M = 28, SD = 4), and Treatment B (M = 32, SD = 6)."
2. Assumptions of Repeated Measures ANOVA
Before presenting the results of the ANOVA, you must verify that the assumptions of the test have been met. These include:
-
Sphericity: This assumption states that the variances of the differences between all pairs of conditions are equal. If sphericity is violated, you may need to adjust the degrees of freedom using corrections like Greenhouse-Geisser or Huynh-Feldt.
-
Normality: While Repeated Measures ANOVA is relatively robust to violations of normality, it's good practice to check the distribution of your data. Histograms and Q-Q plots can be helpful.
Example:
"Sphericity was assessed using Mauchly's test (χ² = 2.5, p = .28). The assumption of sphericity was not violated. Visual inspection of the data suggested no major deviations from normality." (Or, if violated: "Mauchly's test indicated a violation of sphericity (χ² = 15.2, p < .001). Greenhouse-Geisser corrected degrees of freedom were used for subsequent analyses.")
3. Inferential Statistics: The ANOVA Results
This is the core of your report. You need to provide the following:
- F-statistic: The F-ratio calculated by the ANOVA.
- Degrees of Freedom (df): This has two components: df between subjects and df within subjects (or df error). If sphericity was violated, report the corrected df.
- P-value: The probability of obtaining the observed results if there were no effect.
Example:
"Repeated measures ANOVA revealed a significant effect of treatment on cognitive performance, F(2, 58) = 5.7, p = .006, η² = .16. The partial eta-squared (η²) indicates a medium effect size."
4. Post Hoc Tests
If your ANOVA is significant, you'll need to conduct post hoc tests (e.g., Bonferroni, Tukey, Sidak) to determine which specific conditions differ significantly from each other.
Example:
"Post hoc comparisons (Bonferroni corrected) revealed that Treatment B significantly differed from both the Control condition (p = .01) and Treatment A (p = .03). There was no significant difference between the Control and Treatment A conditions (p = .45)."
5. Effect Size
Report an effect size measure. Partial eta-squared (η²) is commonly used for ANOVA.
Example: (This is already included in the previous ANOVA result example)
6. Interpretation
Conclude with a clear and concise interpretation of your findings in relation to your research question or hypothesis. Relate your findings to the existing literature where possible.
Reporting Repeated Measures ANOVA with Multiple Factors
If your design includes more than one independent variable (a factorial design), you'll need to report the results for each main effect and any interactions. The same principles described above apply, but you'll have multiple F-statistics, df, and p-values to report.
Software for Repeated Measures ANOVA
Most statistical software packages (e.g., SPSS, R, SAS, Jamovi) can perform Repeated Measures ANOVA. The specific output may vary depending on the software, but the core elements described above should always be reported.
By following this comprehensive guide, you can effectively and accurately report the results of your Repeated Measures ANOVA, ensuring clear and impactful communication of your research findings. Remember to always tailor your report to your specific research question and audience.
