How to Report an ANCOVA in APA 7 (With Example)

An ANCOVA (analysis of covariance) compares group means on an outcome while statistically controlling for a continuous covariate, often a pre-test or a baseline measure. The write-up trips people up because you report adjusted means, not the raw group means, and you have to report the covariate's effect too. This guide gives you the exact APA 7 format, a worked example you can copy, and the mistakes reviewers catch.

What you need before you write a single word

An APA 7 write-up of a one-way ANCOVA needs these values from your output:

1. The adjusted means (estimated marginal means) and their standard errors for each group, after controlling for the covariate.
2. The F value for the main effect of the independent variable, with its two degrees of freedom.
3. The exact p value for that main effect.
4. An effect size, usually partial eta squared.
5. The covariate's effect: its own F, df, and p, to show it was worth adjusting for.
6. The result of the homogeneity-of-regression-slopes check (see assumptions below).

The APA 7 format template

Report the adjusted main effect, and mention the covariate:

After adjusting for [covariate], a one-way ANCOVA showed a significant effect of [independent variable] on [dependent variable], F(df1, df2) = X.XX, p = .XXX, partial η² = .XX. The covariate, [covariate], was significantly related to [dependent variable], F(1, df2) = X.XX, p = .XXX.

boltAll 53 templates, ready to paste with the italics set: the Rescue Vault, $27 →

Formatting rules reviewers actually check:

• Italicize the statistical symbols: M, SD, F, p, SE.
• Report both degrees of freedom for each F in parentheses: F(2, 86).
• No leading zero on p or on partial η² (neither can exceed 1).
• Round F to two decimals; report p to two or three decimals, and if it is below .001, write p < .001.
• Report the adjusted (estimated marginal) means, not the unadjusted group means, because the adjusted means are what the ANCOVA actually tested.

A worked example

Suppose you compared post-test scores across three teaching methods (30 students per group) while controlling for pre-test score.

• Covariate (pre-test): F(1, 86) = 34.10, p < .001.
• Main effect (teaching method), adjusted: F(2, 86) = 6.72, p = .002, partial η² = .14.
• Adjusted post-test means: Lecture M = 72.10 (SE = 1.30), Flipped M = 78.40 (SE = 1.30), Self-paced M = 74.60 (SE = 1.30).

Written up in APA 7, that becomes:

After adjusting for pre-test score, a one-way ANCOVA showed a significant effect of teaching method on post-test score, F(2, 86) = 6.72, p = .002, partial η² = .14. The covariate, pre-test score, was significantly related to post-test score, F(1, 86) = 34.10, p < .001. Pairwise comparisons of the adjusted means (Bonferroni-corrected) indicated that the flipped classroom group (M = 78.40, SE = 1.30) scored significantly higher than the lecture group (M = 72.10, SE = 1.30), p = .002.

The APA 7 table (recommended for ANCOVA)

Present the adjusted means and their standard errors, not the raw descriptives:

Note. Means are adjusted for pre-test score (covariate). The effect of teaching method was significant, F(2, 86) = 6.72, p = .002, partial η² = .14.

Mistakes reviewers catch

• Reporting the unadjusted means. An ANCOVA tests the adjusted (estimated marginal) means; those are the ones to report and compare.
• Not reporting the covariate's effect. Show that the covariate was related to the outcome, otherwise the adjustment was pointless.
• Skipping the homogeneity-of-slopes check. If the slopes differ across groups, ANCOVA is not appropriate (see below).
• Using a covariate affected by the treatment. The covariate should be measured before the intervention, or be unaffected by it, or the adjustment removes part of the effect you care about.
• A leading zero on p or partial η². APA drops it for values that cannot exceed 1.
• Reporting only one degree of freedom. Each F needs both.

Before you report: did the test's assumptions hold?

A one-way ANCOVA assumes everything a one-way ANOVA does, plus two extra conditions:

• Homogeneity of regression slopes. The relationship between the covariate and the outcome is the same in every group. Check the interaction between the independent variable and the covariate; it should be non-significant. If it is significant, ANCOVA is not appropriate.
• A linear relationship between the covariate and the outcome, and a covariate that is measured reliably and before the treatment.
• The standard ANOVA assumptions: independence, normality of residuals, and homogeneity of variance.

See the assumptions guide for how to check each one, and the effect size guide for partial eta squared.

Let KyroStat do the write-up for you

Adjusted means, standard errors, a covariate effect, and the homogeneity-of-slopes check are a lot to assemble correctly. KyroStat runs the ANCOVA on your data, checks the assumptions including homogeneity of slopes, and hands you the finished APA 7 sentences, a table of adjusted means, the plots, and the underlying Python or R code. Upload your spreadsheet, and the report is done in seconds.

Run an ANCOVA yourself, in the browser

The numbers above are an illustration. Here is a real one you can run right now — no R installation, no account. ToothGrowth ships with R: tooth length for 60 guinea pigs given vitamin C by two delivery methods (supp) at three doses. Dose is the covariate.

# ANCOVA: does supplement type affect tooth growth,
# after adjusting for dose?
model <- aov(len ~ dose + supp, data = ToothGrowth)
summary(model)

Running that gives:

Df Sum Sq Mean Sq F value   Pr(>F)    
dose         1 2224.3  2224.3  123.99 6.31e-16 ***
supp         1  205.4   205.4   11.45   0.0013 ** 
Residuals   57 1022.6    17.9                     
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Every number in the write-up above comes from output like this. If you want the mechanics rather than the reporting, the free R lesson on this walks through it step by step, and the playground is there whenever you want to try something.

Frequently asked questions

Why report adjusted means instead of the raw group means? Because the ANCOVA compares groups after removing the covariate's influence. The adjusted (estimated marginal) means are the values the test actually evaluated.

What is the homogeneity-of-regression-slopes assumption? It means the covariate relates to the outcome the same way in each group. You check it with the group-by-covariate interaction; a significant interaction means ANCOVA is not the right model.

Can I use a post-test as the covariate? No. The covariate should be a baseline measured before the treatment (such as a pre-test), so the adjustment does not remove part of the treatment effect.

Which effect size do I report for ANCOVA? Partial eta squared for the adjusted main effect, reported without a leading zero, for example partial η² = .14.

What if ANCOVA is not quite my design? Without a covariate, report a one-way ANOVA; with two factors, a two-way ANOVA; and when the same people are measured more than once, a repeated-measures ANOVA. If the residuals are far from normal (check with Shapiro-Wilk), the rank-based Kruskal-Wallis test is the usual fallback. For the effect size itself, see eta squared vs partial eta squared.

Which related guides might I need? Often these: mixed ANOVA, post hoc tests and Levene's test.