Reviewers rarely reject a paper because the analysis was wrong — they reject it because the analysis was reported wrong: missing effect sizes, inconsistent decimals, tables that duplicate text. This checklist covers the APA 7 conventions that journals actually enforce, with worked examples for the most common tests.
The universal reporting formula
Nearly every inferential result follows the same skeleton: descriptives → test statistic with degrees of freedom → exact p value → effect size → plain-language interpretation. Italicise statistical symbols (M, SD, t, F, p, d), report p to three decimals (use p < .001 below that), and drop the leading zero for statistics that cannot exceed 1 (p, r, η²).
Worked examples by test
| Test | Template sentence |
|---|---|
| Independent t-test | Experimental participants (M = 78.4, SD = 6.2) outperformed controls (M = 71.1, SD = 7.0), t(58) = 3.41, p = .001, d = 0.88. |
| One-way ANOVA | Anxiety differed across the three programmes, F(2, 87) = 5.12, p = .008, η² = .11; Tukey tests showed… |
| Correlation | Self-efficacy correlated positively with achievement, r(118) = .42, p < .001. |
| Multiple regression | The model explained 34% of variance in burnout, F(3, 196) = 33.6, p < .001, R² = .34; workload was the strongest predictor (β = .41, p < .001). |
| Chi-square | Preference was associated with gender, χ²(2, N = 240) = 9.84, p = .007, Cramér's V = .20. |
Effect sizes are not optional
APA 7 (§3.7) states that effect sizes and confidence intervals are needed for readers to appreciate the magnitude of findings. In practice, most editors now treat a missing effect size as a mandatory revision.
Tables and figures: the three rules
- No duplication: a result lives in the text or a table, never both in full.
- Self-contained: every table needs a number, an italicised title, defined abbreviations and probability notes (* p < .05, ** p < .01).
- Horizontal lines only: APA tables use top, header and bottom rules — no vertical lines, no zebra shading.
What reviewers complain about most
- Exact p values replaced by 'p < .05' everywhere — report exact values.
- Effect sizes reported but never interpreted against benchmarks.
- Assumption checks (normality, homogeneity) silently skipped.
- Degrees of freedom missing or inconsistent with the reported N.
- Rounding chaos: means to one decimal in tables, two in text.
Frequently Asked Questions
Should I report non-significant results?
Yes — fully, with the same template (statistic, df, exact p, effect size). Selective reporting is a questionable research practice and reviewers actively look for it.
How many decimal places does APA 7 require?
Two decimals for most statistics (means, SDs, t, F), three for p values, and no leading zero for statistics bounded by 1 such as p, r and η².
Can confidence intervals replace p values?
They complement rather than replace them in most journals. Report both where space allows: t(58) = 3.41, p = .001, d = 0.88, 95% CI [0.35, 1.41].
Does Celsus write the results section for me?
We draft your results section to APA 7 or your target journal's house style from your own analysis output, deliver the formatted tables and figures, and document our contribution so authorship and ethics stay clean.