How to report a Pearson correlation in APA style (7th ed.)
A correlation is the shortest result to report and the easiest to report badly. Here is what the sentence needs, when a table is better, a worked example, and six mistakes reviewers flag.
What the sentence must contain
The coefficient r with its degrees of freedom (N − 2) or the sample size, the p-value, the direction in words (positively, negatively), and — APA 7 asks for it — a confidence interval. Say what was correlated with what; "the variables were correlated" names nothing.
The pattern
Age was positively correlated with score, r(58) = .31, p = .015, 95% CI [.06, .52]. For a null result: Age and satisfaction were not significantly correlated, r(58) = .12, p = .36, 95% CI [−.14, .36].
Numbers in the pattern are placeholders.
Drop the leading zero for r and p — neither can exceed 1. Two decimals for r, three for p (or p < .001).
When a table is better
Three or more variables give three or more coefficients. Put them in a correlation matrix — variables down the side, r in the cells, significance marked with asterisks explained in a note, and M and SD in the first columns — and let the text point to the table rather than list every pair.
A worked example
On a synthetic sample of 60 respondents, age was positively correlated with score, r(58) = .26, p = .041, 95% CI [.01, .49]. In the same sample, satisfaction and score were built to move together, r(58) = .87 — a reminder that a very large r can be a feature of how the data were made, not a discovery.
These numbers were produced by the engine that computed the matrix; the same sample is in the six-step walkthrough.
Before you trust r
Pearson's r measures a straight-line relationship. Look at the scatterplot first: a curve, a ceiling, or two or three extreme cases can produce an r that describes nothing. For ordinal data or clear outliers, report Spearman's rho and say why.
Six mistakes reviewers flag
- r without N or df — r = .31 from 20 people and from 2,000 are different claims.
- "Weak" or "strong" from the number alone — say the field’s conventions or give the CI; the label is not the evidence.
- Causal verbs — "age increases score." A correlation supports "was associated with," nothing more.
- A leading zero on r or p — .31, not 0.31.
- r² reported as "explains 10% of the variance" with no interval — an estimate with no uncertainty attached.
- Listing ten pairs in prose — that is a table.
Where the numbers come from
In Statory the sentence is generated from the same computation that produced the matrix, so the two cannot disagree. Every engine is checked line by line against established statistics software and the comparison is published — see the accuracy page.
The same engine runs inside ChatGPT and Claude when you paste a table there.