A regression analysis produces a lot of output, and APA 7 asks you to report only part of it, in a particular format. This guide walks through a multiple regression from the output to the sentence and the table. The numbers below are illustrative, not from a real study.
Suppose you predicted exam scores from weekly study hours and prior GPA in a sample of 120 students. The software reports an overall model with R² = .34 and F(2, 117) = 30.14, p < .001, and one row per predictor.
State the model, the fit, and what the predictors did, in a short paragraph with the table doing the detail. For example: "A multiple regression predicted exam score from weekly study hours and prior GPA. The model was significant, F(2, 117) = 30.14, p < .001, and explained 34% of the variance in exam scores (R² = .34). Study hours (B = 1.20, 95% CI [0.59, 1.81], p < .001) and prior GPA (B = 8.40, 95% CI [4.24, 12.56], p < .001) were both significant predictors." Say what a coefficient means in the units of your variables: each additional weekly study hour was associated with a 1.20-point higher exam score, holding GPA constant.
Give one row per predictor, with the unstandardized coefficient B, its standard error, the 95% confidence interval for B, the standardized coefficient β if you want to compare predictors, and the p-value. For the example:
Put the model's R² and F in a note under the table or in the text. Use horizontal rules only, with no vertical lines, a numbered title above the table and a descriptive title in italics, as APA 7 table format requires. If you entered predictors in steps, add a row or a note for the change in R² at each step.
A reader should know that you looked. Say briefly how you checked linearity, independence of errors, equal spread of residuals, roughly normal residuals, and collinearity among predictors. Variance inflation factors under about 5 are a common rule of thumb, though the right cutoff depends on your field. If an assumption failed and you used robust standard errors or a transformation, say so. Report the sample size actually used, after any missing data were dropped.
The common mistake is reporting only whether predictors were significant. Readers need the size of each effect in the original units, how uncertain it is, and how much of the outcome the model explains. If your table gives those, the paragraph in the text can stay short. Our free regression APA formatter turns your output into the sentence and the table in the right format.
If your model is more complicated than this, with interactions, clustered data or a non-normal outcome, and you are not sure how to report it, send us the output. Working out what to report is usually a short conversation.
One email whenever we publish something new. No spam, unsubscribe anytime.
{{ subscribeErrorMsg }}