How to Report a One-Way ANOVA in APA 7
Academic Decoder (@academicdecoder)
Created by a PhD-qualified lecturer · Updated
The short answer
A one-way ANOVA tests whether the means of three or more independent groups differ. In APA 7 you report each group's mean and standard deviation, then F with its two degrees of freedom, the exact p-value and an effect size (usually η²). If the ANOVA is significant, you add the post hoc comparisons that show which groups differ.
This guide is for undergraduate and master's students writing a dissertation results chapter. It uses one research scenario throughout: sixty students were randomly allocated to three self-study formats (Lecture recordings, Worked examples or Practice quizzes, 20 per format), studied on their own for two weeks, then sat the same exam. These are simulated data for illustration, the same example as the free SPSS Output Explorer.
When to use a one-way ANOVA
- One outcome measured on a continuous scale, such as an exam score.
- One grouping variable with three or more groups, such as study format.
- Different people in each group. If the same people were measured under every condition, use a repeated-measures ANOVA instead.
With only two groups, the independent-samples t-test is usually reported instead. With two grouping variables, such as study format and year group, you need a factorial (two-way) ANOVA. Running several t-tests in place of one ANOVA raises the chance of a false positive.
Assumptions and important checks
- Independent observations. Each person appears in one group only, and one person's score does not affect another's. This comes from your design, not from a test.
- Roughly normal scores within each group. The ANOVA copes well with mild departures, especially when the groups are similar in size and not very small. With small, clearly skewed groups, the Kruskal-Wallis test is the usual alternative.
- Similar variances across groups, checked with Levene's test (Tests of Homogeneity of Variances in SPSS).
If Levene's test is significant
A significant Levene's test suggests the group variances are not equal. The standard ANOVA assumes they are, so report Welch's ANOVA instead (in SPSS, tick Welch under Options; it appears in the Robust Tests of Equality of Means table) and use Games-Howell for the post hoc comparisons, because it does not assume equal variances. Welch's test changes the second df, which usually has decimals: report it to two decimal places, and say that you used Welch's ANOVA.
In the worked example Levene's test was not significant, F(2, 57) = 0.77, p = .466, so the standard ANOVA and Tukey's HSD fit these data. For comparison, Welch's test on the same data gives F(2, 37.70) = 8.35, p = .001. The conclusion is the same here, but with unequal variances and unequal group sizes the two tests can disagree, which is why the version you report matters.
How to read the SPSS output
This is the worked example as it would appear in SPSS. The highlighted values are the ones you report.
Simulated SPSS output · layout modelled on SPSS 29
| N | Mean | Std. Deviation | Std. Error | |
|---|---|---|---|---|
| Lecture recordings | 20 | 59.60 (highlighted: a value you report) | 7.214 (highlighted: a value you report) | 1.613 |
| Worked examples | 20 | 68.55 (highlighted: a value you report) | 8.095 (highlighted: a value you report) | 1.810 |
| Practice quizzes | 20 | 67.85 (highlighted: a value you report) | 8.958 (highlighted: a value you report) | 2.003 |
| Total | 60 | 65.33 | 8.972 | 1.158 |
| Mean | Std. Deviation | |
|---|---|---|
| Lecture recordings | 59.60 (highlighted: a value you report) | 7.214 (highlighted: a value you report) |
| Worked examples | 68.55 (highlighted: a value you report) | 8.095 (highlighted: a value you report) |
| Practice quizzes | 67.85 (highlighted: a value you report) | 8.958 (highlighted: a value you report) |
| Sum of Squares | df | Mean Square | F | Sig. | |
|---|---|---|---|---|---|
| Between Groups | 991.033 | 2 (highlighted: a value you report) | 495.517 | 7.515 (highlighted: a value you report) | .001 (highlighted: a value you report) |
| Within Groups | 3758.300 | 57 (highlighted: a value you report) | 65.935 | ||
| Total | 4749.333 | 59 |
| df | F | Sig. | |
|---|---|---|---|
| Between Groups | 2 (highlighted: a value you report) | 7.515 (highlighted: a value you report) | .001 (highlighted: a value you report) |
| Within Groups | 57 (highlighted: a value you report) |
| (I) Format | (J) Format | Mean Difference (I-J) | Std. Error | Sig. | 95% Confidence Interval | |
|---|---|---|---|---|---|---|
| Lower Bound | Upper Bound | |||||
| Lecture recordings | Worked examples | −8.950* (highlighted: a value you report) | 2.568 | .003 (highlighted: a value you report) | −15.129 | −2.771 |
| Practice quizzes | −8.250* (highlighted: a value you report) | 2.568 | .006 (highlighted: a value you report) | −14.429 | −2.071 | |
| Worked examples | Practice quizzes | .700 (highlighted: a value you report) | 2.568 | .960 (highlighted: a value you report) | −5.479 | 6.879 |
| (I) − (J) Format | Mean Difference (I-J) | Sig. |
|---|---|---|
| Lecture recordings − Worked examples | −8.950* (highlighted: a value you report) | .003 (highlighted: a value you report) |
| Lecture recordings − Practice quizzes | −8.250* (highlighted: a value you report) | .006 (highlighted: a value you report) |
| Worked examples − Practice quizzes | .700 (highlighted: a value you report) | .960 (highlighted: a value you report) |
- Descriptives show the pattern before any statistics: Lecture recordings scored lowest, and the other two formats were close together.
- F(2, 57) = 7.52, p = .001 is the overall (omnibus) test. With three groups and 60 students, the df are 3 − 1 = 2 between groups and 60 − 3 = 57 within groups. It shows the three means are not all equal, but not which ones differ.
Reading the post hoc comparisons
A significant F only tells you that at least one group mean differs. Post hoc comparisons test each pair of groups while controlling the overall chance of a false positive, which separate t-tests would not. SPSS lists each pair twice, once in each direction; you only need each pair once. For the worked example (Tukey's HSD):
Lecture recordings − Worked examples
- Mean difference
- −8.95
- p
- .003
- 95% CI of the difference
- [−15.13, −2.77]
Lecture recordings − Practice quizzes
- Mean difference
- −8.25
- p
- .006
- 95% CI of the difference
- [−14.43, −2.07]
Worked examples − Practice quizzes
- Mean difference
- 0.70
- p
- .960
- 95% CI of the difference
- [−5.48, 6.88]
- Each mean difference keeps the order of the comparison (first group minus second), so −8.95 means Lecture recordings scored 8.95 points lower than Worked examples.
- The confidence interval gives a range of plausible values for each population difference. The first two intervals do not include zero, which agrees with their significant p-values.
- The non-significant pair matters too. Worked examples and Practice quizzes did not differ significantly, which is as important to the conclusion as the two differences that did.
Key values to report, and where to find them
You need the group descriptives, five values from the ANOVA itself, and one decision (which version of the test you used). Most values come directly from your software, although some need an extra option ticked.
M and SD for each group
- What it tells the reader
- The pattern of scores and their spread
- SPSS
- Descriptives table (tick Descriptive under Options)
- jamovi
- Descriptives table (One-Way ANOVA, Additional Statistics)
- R
psych::describeBy()
F
- What it tells the reader
- How large the differences between group means are, relative to the variation within groups
- SPSS
- ANOVA table, column F
- jamovi
- F
- R
summary(aov()), F value
The two df
- What it tells the reader
- Between groups (k − 1) and within groups (N − k)
- SPSS
- ANOVA table, df: Between Groups and Within Groups
- jamovi
- df1 and df2
- R
- Df (the factor row and the Residuals row)
p
- What it tells the reader
- Whether the group means differ significantly overall
- SPSS
- ANOVA table, Sig.
- jamovi
- p
- R
- Pr(>F)
η²
- What it tells the reader
- How much of the variation in the outcome is associated with group membership
- SPSS
- ANOVA Effect Sizes table (SPSS 27+, Eta-squared), or Between Groups ÷ Total Sum of Squares
- jamovi
- ANOVA analysis, Effect Size: η²
- R
effectsize::eta_squared()
Post hoc comparisons
- What it tells the reader
- Which pairs of groups differ, by how much
- SPSS
- Multiple Comparisons table (Post Hoc, e.g. Tukey)
- jamovi
- Post Hoc Tests (Tukey or Games-Howell)
- R
TukeyHSD()
Which test you used
- What it tells the reader
- Whether equal variances were assumed
- SPSS
- Tests of Homogeneity of Variances; Robust Tests of Equality of Means (Welch)
- jamovi
- Fisher's (equal variances) or Welch's (unequal variances)
- R
oneway.test():var.equal = TRUE(Fisher) or the default (Welch)
Which effect size to report
Eta squared (η²) is the most common effect size for a one-way ANOVA. It is the between-groups sum of squares divided by the total sum of squares: the proportion of variation in the outcome associated with group membership. In the worked example, η² = .21: about 21% of the variation in exam scores is associated with study format, a large effect by Cohen's benchmarks (.01 small, .06 medium, .14 large). The p-value cannot tell you this.
- Partial eta squared is the same number in a one-way ANOVA, because there is only one factor. SPSS labels it "partial" when you run the ANOVA through General Linear Model. Use one name consistently.
- Omega squared (ω²) corrects for the upward bias in η², so it is slightly smaller: .18 for the worked example, against η² = .21. Some departments prefer it.
- Like p, values that cannot exceed 1 are written without a leading zero: η² = .21, not 0.21.
Worked example · One-way ANOVA
What does the ANOVA tell us?
Three study formats were compared using simulated exam-score data.
A one-way ANOVA showed a statistically significant difference in exam scores between study formats, F(2, 57) = 7.52, p = .001, η² = .21.
A short example for learning, not a complete, dissertation-ready results paragraph.
This tells us that at least one group mean differs. It does not tell us which groups differ. To answer that, we need the post-hoc comparisons.
What else needs reporting?
A complete write-up may need group descriptive statistics, post-hoc comparisons, effect-size interpretation and appropriate qualification of the findings.
Your statistical results, explained and ready to write up.
Already completed your analysis in SPSS, jamovi, JASP, R or another supported statistical program? Enter the key values from your output into Academic Decoder to understand, visualise, report and defend your findings.
- APA 7 write-upA structured results paragraph using the values you enter.
- Plain-English explanationUnderstand what your statistical result means, including what the effect size tells you.
- Graphs & visualsA relevant chart of your result, downloadable as PNG or SVG where supported.
- Discussion sentence startersDevelop your discussion in your own words.
- Defend This ResultPractise examiner-style questions and understand the limits of your conclusions.
- Save as Draft or FinalOrganise your results and return to them while you write your dissertation.
- My Chapter 4Turn your saved results into an editable Word draft, with a reporting checklist.
- Viva preparationDownload examiner-style practice questions and suggested answers to revise from.
The Decoder doesn't run the statistical analysis. You complete your analysis in your own software, then enter the relevant values from your output.
£15.99 one-time · All 26 supported analyses · Unlimited decodes · No subscription
Means & group comparisons (9)
- Independent t-test
- Paired t-test
- One-Way ANOVA
- Factorial Two-Way ANOVA
- Repeated-Measures ANOVA
- Mixed ANOVA
- ANCOVA
- MANOVA
- MANCOVA
Non-parametric (4)
- Mann-Whitney U Test
- Wilcoxon Signed-Rank Test
- Kruskal-Wallis
- Friedman Test
Categorical (2)
- Chi-Square Test
- Chi-Square Goodness-of-Fit
Correlation & regression (7)
- Pearson Correlation
- Spearman Correlation
- Simple Linear Regression
- Multiple Linear Regression
- Binary Logistic Regression
- Multinomial Logistic Regression
- Ordinal Logistic Regression
Survival analysis (3)
- Kaplan-Meier
- Log-Rank Test
- Cox Regression
Reliability (1)
- Cronbach's Alpha
Reporting a non-significant result
A non-significant ANOVA is reported with the same values, without post hoc comparisons. Simulated data: Lecture recordings M = 64.10, SD = 8.20; Worked examples M = 67.30, SD = 8.40; Practice quizzes M = 66.20, SD = 8.70; 20 per group.
A one-way ANOVA showed no statistically significant difference in exam scores between study formats, F(2, 57) = 0.74, p = .480, η² = .03.
A short example for learning, not a complete, dissertation-ready results paragraph.
- Say the data did not provide sufficient evidence of a difference. Do not say the formats "were the same" or "had no effect"; a non-significant test does not show that.
- Do not run post hoc tests to look for a difference after a non-significant ANOVA. Comparisons you planned before collecting data are different: test them as your analysis plan sets out.
- Do not describe the highest mean as the "best" format. Without a significant result, the differences between the means are descriptive only.
- Still report η². Here it is .03, a small effect, which tells the reader how large the observed differences were.
APA 7 formatting rules
| Rule | Correct | Incorrect |
|---|---|---|
| Italicise statistical symbols | F, p, M, SD, η² | F, p, M, SD, η² (not italic) |
| Give both df, between groups first | F(2, 57) = 7.52 | F(57) = 7.52 or F(57, 2) = 7.52 |
| Report the exact p-value, to two or three decimals | p = .001 | p < .05 |
| Very small p-values use "less than" | p < .001 | p = .000 |
| No leading zero for values that cannot exceed 1 | p = .001, η² = .21 | p = 0.001, η² = 0.21 |
| Keep the leading zero for values that can exceed 1 | mean difference = 0.70 | mean difference = .70 |
| Round most statistics to two decimals | F(2, 57) = 7.52 | F(2, 57) = 7.515 |
| Name the post hoc test | Post-hoc comparisons using Tukey's HSD indicated… | Post-hoc tests showed… |
| Use a true minus sign for negative values | mean difference = −8.95 | mean difference = -8.95 (hyphen) |
SPSS prints "Sig. = .000" when p is very small. That never means p is zero, so report it as p < .001.
Common mistakes
- Writing p = .000. Report p < .001 instead.
- Stopping at the overall F. A significant ANOVA does not say which groups differ. Report the post hoc comparisons as well.
- Running several t-tests instead of an ANOVA. Each extra test raises the chance of a false positive. The ANOVA with post hoc tests controls that.
- Giving only one df, or the two in the wrong order. Write F(between, within): F(2, 57).
- Reporting p without an effect size. Add η² (or the effect size your department expects) and say which one it is.
- Ignoring a significant Levene's test. Report Welch's ANOVA and Games-Howell, and say so.
- Claiming more than the design allows. Groups you did not assign, such as year groups or job roles, can only be described as differing. A cause needs a randomised design like the worked example, described in your method section.
- Calling a non-significant result "no difference". Say there was insufficient evidence of a difference.
You can report your result. But can you explain it?
Supervisors and examiners rarely ask you to recite the numbers. They ask what your result allows you to claim, and why you chose your test, effect size and interpretation. Could you answer these in a sentence, about your own result?
- “Your ANOVA is significant. Which groups actually differ, and how do you know?”
- “Why did you use a post hoc test rather than running separate t-tests?”
- “Can you say the study format caused the difference in exam scores?”
Defend This Result
This guide doesn't give model answers. In the Result Decoder, Defend This Result helps you practise answering questions like these about your own result, including the evidence that supports your conclusions and their important limitations. It can't predict the questions you'll be asked.
Practise with my own resultFrequently asked questions
Do I need post hoc tests if my ANOVA is significant?
With three or more groups, yes, unless you planned specific comparisons in advance. A significant ANOVA shows that the group means are not all equal. It does not show which groups differ. Post hoc tests, or planned contrasts, answer that.
Which post hoc test should I use?
Tukey's HSD is the usual choice when the group variances are similar. Games-Howell is designed for unequal variances. Bonferroni suits a small number of planned comparisons. Name the test you used in your results paragraph.
Should I report eta squared or partial eta squared?
In a one-way ANOVA they are the same number, because there is only one factor. Use the name your software gives, and keep it consistent through your results chapter. Some departments prefer omega squared, which is slightly smaller and less biased; check what yours expects.
What if SPSS shows Sig. = .000?
Report p < .001. SPSS rounds very small p-values to .000, but the p-value is never exactly zero.
Can I run an ANOVA with only two groups?
Yes. It gives the same p-value as an independent-samples t-test with equal variances assumed. Most researchers report the t-test for two groups, so use the ANOVA when you have three or more.
More statistics guides
How to Report an Independent-Samples t-Test in APA 7
What to include, where to find each value in SPSS, jamovi and R, worked examples (including Welch's test), formatting rules and common mistakes.
How to Report a Paired-Samples t-Test in APA 7
Before-and-after scores from the same people: what to report, where to find each value, a worked example, which Cohen's d to use, and common mistakes.
How to Report a Chi-Square Test of Independence in APA 7
Two categorical variables: what to report, where to find each value, the expected count check, Cramér's V, a worked example and common mistakes.
Not sure this is the right test?
Answer a few questions about your study and the free Test Navigator suggests which test fits.