APA 7 rules this generator applies
- Italicise statistical symbols: t, F, p, M, SD, r, N, U. Greek letters such as χ and η are italicised here as most style guides show them.
- No leading zero for values that cannot exceed 1: p = .032, r = .32, η² = .25. Keep it for values that can: d = 0.78, t = 0.95.
- Exact p-values to three decimals, except below .001, which is written p < .001. Never write p = .000.
- Two decimals for test statistics, means and standard deviations. Degrees of freedom are whole numbers, except for Welch’s test, where they are given to two decimals.
- Always include an effect size where one exists (d, η², V, r), so readers can judge how large the effect is, not only whether it is significant.
Templates for each test
| Test | How the statistics are written |
|---|---|
| Independent or paired t-test | t(38) = 3.10, p = .004, d = 0.98 |
| Welch’s t-test | t(29.37) = 1.95, p = .061, d = 0.61 |
| One-way ANOVA | F(2, 27) = 4.56, p = .020, η² = .25 |
| Chi-square test | χ²(2, N = 200) = 7.20, p = .027, Cramér’s V = .19 |
| Pearson correlation | r(48) = .32, p = .023 |
| Mann–Whitney U | U = 23.5, z = −2.03, p = .043, r = .41 |
| Wilcoxon signed-rank | T = 3, p = .010 |
Where to find the numbers
Every value the generator needs is in standard software output. If you do not have them yet, the free calculators produce them: the t-test calculator (t, df, p), the effect size calculator (Cohen’s d), the ANOVA calculator (F, both df, p, η² and Tukey post-hoc tests), the chi-square calculator (χ², df, p, Cramér’s V), the correlation calculator and the Mann–Whitney calculator.
Common reporting mistakes
- Reporting a one-tailed p-value without saying so. The generator flags p-values that do not match the statistic, which often catches this.
- Writing “p = .000” because the software rounded. Use p < .001.
- Leaving out the direction of the effect. Say which group was higher; the generator does this from the means or medians.
- Calling a non-significant result “proof of no difference”. “No significant difference” means the data were not strong enough to show one, not that the groups are equal.
For how to describe the analysis itself (software, assumptions checks, corrections), see how to document statistical analyses in a report or thesis.
Frequently asked questions
How do you report a t-test in APA style?
Give the test, the means and standard deviations of the groups, then t with its degrees of freedom in brackets, the p-value and an effect size, for example: t(38) = 3.10, p = .004, d = 0.98. Italicise the statistical symbols.
Why is there no zero before the decimal point in p = .004?
APA style drops the leading zero for values that cannot be greater than 1, such as p-values, correlations, eta squared and Cramer's V. Values that can exceed 1, such as t, F, d and means, keep it.
How do I report a very small p-value in APA?
Report exact p-values to two or three decimals, but write p < .001 for anything smaller than .001. Never write p = .000.
How do I report Welch's t-test degrees of freedom?
Welch's test has fractional degrees of freedom. Report them to two decimals, for example t(29.37) = 1.95.
Is this generator a substitute for the APA manual?
No. It follows the common APA 7 conventions for reporting results, but your instructor or journal may have additional requirements, and you should always check the numbers against your software output.