APA 7 Statistics Reporting
Report the number, in the right notation, with the correct italics and decimals. Everything that follows is APA 7.
① Report it exactly like this
| Analysis | Report as |
|---|---|
| t-test | t(df) = 2.02, p = .047, d = 0.45 |
| ANOVA | F(dfb, dfw) = 5.62, p = .005, η² = .11 |
| Chi-square | χ²(df, N = 120) = 6.31, p = .012, V = .23 |
| Correlation | r(28) = .42, p = .021 |
| Regression (b) | b = 0.52, SE = 0.14, β = .28, t = 3.71, p < .001 |
| Regression (model) | R² = .34, F(2, 97) = 25.1, p < .001 |
| Means | M = 3.45, SD = 0.82 |
Correlation df = N − 2. ANOVA df: k−1 between, N−k within.
② In a sentence
Sleep-deprived participants recalled fewer words (M = 6.2, SD = 2.5) than rested participants (M = 7.4, SD = 2.8), t(78) = 2.02, p = .047, d = 0.45, 95% CI [0.01, 0.89].
The order: the finding in plain words → descriptives → test statistic → p → effect size. The stats go after the finding, not instead of it, and the interval goes with the effect size. Save “what it means” for the Discussion. Full patterns: Writing §3.2.
③ Italics: slanted or upright?
Italic (Latin letters):
t, F, r, p, M, SD, N, n, df, d, b, SE, z, U, R², V.
Upright (Greek letters & numerals):
η², χ², β, ρ, α, ω², and every digit.
Rule of thumb: Latin letter → italic; Greek letter → upright. The subscripts and the numbers themselves are never italic.
④ The p-value, done right
• Report the exact p to two or three decimals: p = .03, p = .048.
• Below .001, write p < .001.
• No leading zero: p = .04, never 0.04 (it can't exceed 1).
• Two-tailed unless you say otherwise.
• A p-value does not show how large an effect is. Pair it with an effect size and its interval: card ⑥.
⑤ Decimals & leading zeros
| Quantity | Dec. | Lead 0? |
|---|---|---|
| p | 2–3 | ✗ no |
| r, ρ, R², η², V, β | 2 | ✗ no |
| t, F, χ² | 2 | ✓ yes |
| Cohen's d | 2 | ✓ yes |
| M, SD, b, SE | 2* | ✓ yes |
Drop the leading zero only when the value can't exceed 1. *Means/SDs: match the scale's precision. Convert your own numbers with the APA formatter.
⑥ The interval, in brackets
• Form: 95% CI [lower, upper]. State the level the first time, then leave it out.
• It follows the leading-zero rule of the statistic it bounds: d = 0.45, 95% CI [0.01, 0.89], but r(28) = .42, 95% CI [.07, .68].
• Name the quantity when it could be misread: 95% CI of the difference [−2.38, −0.02].
• η², R² and V take 90%, not 95%: their test is one-tailed, and a 90% interval just touches zero when p = .05. So R² = .34, 90% CI [.21, .44].
• In a table, APA uses separate LL and UL columns, without brackets.
The one mistake to never make
p = .000 is never correct. SPSS prints Sig. = .000 because it rounds, but a p-value is never exactly zero. Write p < .001. Also avoid p = 0, “p = ns” and the software's five-decimal default. Report the exact value, not asterisks.