Education statistics guide
How to Report ANOVA in APA 7
ANOVA reports should identify the effect, F(df1, df2), p value, effect size, and post-hoc comparisons when needed.
Updated:
Problem
ANOVA tables contain multiple effects. Without naming the effect and follow-up comparison, readers cannot know what differed.
Who should use this
Use this for papers, research reports, class statistics assignments, and SPSS ANOVA write-ups.
Formula and concept
Common APA fragment: F(df1, df2) = value, p = value, eta squared = value.
One-way ANOVA reports one factor. Two-way ANOVA usually reports two main effects and one interaction. If the interaction is significant, interpret it before relying on main effects.
Use eta squared or partial eta squared according to your course or journal. Label the effect size clearly.
This page helps with calculation and reporting, but users should still confirm their research design, assumptions, and statistical interpretation.
Step by step
- Confirm the ANOVA type and research design.
- Find the effect row and copy F, df1, df2, and p.
- Add an effect size such as eta squared or partial eta squared.
- If the omnibus F is significant with three or more groups, report post-hoc or planned comparisons.
Worked example
One-way example: Teaching method had a significant effect on posttest scores, F(2, 87) = 6.42, p = .003, η² = .13. Tukey comparisons showed the cooperative learning group scored higher than the lecture group. Two-way example: The interaction between method and grade was significant, F(2, 114) = 4.18, p = .018, partial η² = .07.
Common mistakes
- Reporting only p without F and degrees of freedom.
- Ignoring a significant interaction in two-way ANOVA.
- Not explaining which groups differed after a significant omnibus test.
- Mixing eta squared and partial eta squared without labels.
Editorial and reproducibility review:
Turn the guide into an auditable decision
This section reworks the guide’s one-way example (F(2, 87) = 6.42) and two-way example (interaction F(2, 114) = 4.18) to check that "significant" is not read as "every group differs."
Reproduce the one-way example and confirm what the omnibus F does and does not tell you
The example reports F(2, 87) = 6.42, p = .003, η² = .13, followed by "Tukey comparisons showed the cooperative learning group scored higher than the lecture group." Self-check: the omnibus F only establishes that at least one group differs from another somewhere among the three groups — it does not by itself say which pair. If the write-up stopped at F(2, 87) = 6.42, p = .003, a reader would know a difference exists but not where, which is exactly why the Tukey sentence has to follow it, not replace it.
Reproduce the two-way example and confirm which effect size label applies
The two-way example reports the interaction as F(2, 114) = 4.18, p = .018, partial η² = .07 — not plain η². Build a parallel case: a region × material interaction of F(1, 96) = 3.85, p = .028, partial η² = .04. Write it out in full: "The interaction between region and material was significant, F(1, 96) = 3.85, p = .028, partial η² = .04." This checks that you are labeling the effect size consistently as partial η² for the two-way design, matching the guide’s own example rather than switching to unlabeled η².
Self-check the guide’s own warning about what a significant ANOVA means
The guide states directly that a significant ANOVA "means at least one difference exists" — not that every group differs from every other group. Apply this to the one-way example: F(2, 87) = 6.42, p = .003 tells you at least one of the three teaching methods differs from another, and only the follow-up Tukey comparison narrows that down to "cooperative learning is higher than lecture." Writing "all three methods differed significantly" from the omnibus F alone would overstate what the test actually showed.
Completion criteria
- Reproduced the one-way example and can explain that F(2, 87) = 6.42, p = .003 alone shows a difference exists but not where it is.
- Reproduced the two-way interaction example and used partial η² consistently, matching the guide’s own labeling rather than switching to unlabeled η².
- Would not conclude "every group differs" from a significant omnibus F without a follow-up comparison.
- Confirmed whether the design is one-way or two-way before deciding which effect-size label (η² vs partial η²) applies.
Recommended tools
Related guides
FAQ
- Does a significant ANOVA mean every group differs?
- No. It means at least one difference exists, so follow-up comparisons are often needed.
- How do I write df1 and df2?
- APA style usually writes F(df1, df2), such as F(2, 87) = 6.42.
- How many results are in a two-way ANOVA?
- Usually two main effects and one interaction, depending on the model and research question.
- Can I report partial eta squared only?
- Yes, if that is the required effect size. Label it clearly.
Next step
Generate an APA ANOVA sentence, then check the SPSS effect row and post-hoc output.