A multiple regression predicts one continuous outcome from two or more predictors. The output is larger than a simple regression, and APA 7 wants both the overall model and a coefficient row for every predictor, usually as a table. This guide gives you the exact format, a worked example you can copy, and the mistakes reviewers catch.
What you need before you write a single word
An APA 7 write-up of a multiple regression needs two layers.
The model fit:
- R-squared and, because there are several predictors, adjusted R-squared.
- The F statistic with its two degrees of freedom (regression df = number of predictors, residual df = N minus predictors minus 1).
- The p value for the overall model.
For each predictor:
- The unstandardized coefficient (b) and its standard error (SE).
- The standardized coefficient (beta).
- The t value and p value.
- Optionally a 95% confidence interval for b.
The APA 7 format template
Report the model in text, then give the coefficients in a table:
A multiple regression predicting [outcome] from [predictor 1], [predictor 2], and [predictor 3] was significant, F(df1, df2) = X.XX, p = .XXX, adjusted R-squared = .XX. [Predictor 1] was a significant positive predictor, b = X.XX, beta = .XX, p = .XXX, while [predictor 2] was not, b = X.XX, beta = .XX, p = .XXX.
Formatting rules reviewers actually check:
- No leading zero on p, R-squared, adjusted R-squared, or beta. Keep the leading zero on b and the CI.
- Report both degrees of freedom for F, regression first: F(3, 96).
- Say which predictors are significant and in which direction; do not just report the omnibus F.
- Round to two decimals. If p is below .001, write p < .001.
A worked example
Suppose you predicted job performance from conscientiousness, cognitive ability, and years of experience in 100 employees.
- Model: F(3, 96) = 14.20, p < .001, R-squared = .31, adjusted R-squared = .29
Written up in APA 7, that becomes:
A multiple regression predicting job performance from conscientiousness, cognitive ability, and years of experience was significant, F(3, 96) = 14.20, p < .001, adjusted R-squared = .29. Conscientiousness (b = 0.38, beta = .34, p < .001) and cognitive ability (b = 0.29, beta = .27, p = .003) were significant positive predictors; years of experience was not (b = 0.05, beta = .06, p = .48). Together the three predictors explained 29 percent of the variance in performance.
The APA 7 regression table
For multiple regression a coefficient table is expected. APA tables use horizontal rules only:
| Predictor | b | SE | beta | t | p |
|---|---|---|---|---|---|
| (Intercept) | 1.20 | 0.44 | --- | 2.73 | .008 |
| Conscientiousness | 0.38 | 0.10 | .34 | 3.80 | < .001 |
| Cognitive ability | 0.29 | 0.09 | .27 | 3.06 | .003 |
| Years of experience | 0.05 | 0.07 | .06 | 0.71 | .48 |
Note. N = 100. R-squared = .31, adjusted R-squared = .29, F(3, 96) = 14.20, p < .001.
Mistakes reviewers catch
- Reporting R-squared but not adjusted R-squared. With several predictors, report adjusted R-squared, which penalizes adding predictors.
- Interpreting non-significant predictors as effects. If a predictor's coefficient is not significant, do not describe it as influencing the outcome.
- Confusing b and beta. b is in original units; beta is standardized and lets you compare predictors' relative strength.
- Ignoring multicollinearity. If predictors are highly correlated, coefficients become unstable. Check the variance inflation factor (VIF) and mention it if it is a concern.
- A leading zero on p, R-squared, or beta. APA drops it.
- Writing p = .000. Report p < .001.
Before you report: did the model's assumptions hold?
Multiple regression assumes a linear relationship, independent and normally distributed residuals, homoscedasticity, and low multicollinearity among predictors. Check residual plots and VIFs before trusting the model. For a single predictor, see our guide on reporting a simple linear regression.
Let KyroStat do the write-up for you
Formatting a multiple regression, with adjusted R-squared, a full coefficient table, and multicollinearity checks, is where errors creep in. KyroStat fits the model, checks the assumptions and VIFs, and hands you the finished APA 7 sentences, a publication-ready coefficient table, and the underlying Python or R code. Upload your spreadsheet, and the report is done in seconds.
Frequently asked questions
Should I report R-squared or adjusted R-squared? Both are informative, but with multiple predictors adjusted R-squared is the headline value because it accounts for the number of predictors.
How do I report the F test degrees of freedom? Number of predictors first, then residual df: F(3, 96) for three predictors and 100 cases.
What is beta and why report it? Beta is the standardized coefficient. Because it is in standard-deviation units, it lets you compare the relative strength of predictors measured on different scales.
My p value shows as .000. What do I write? Report p < .001. A p value is never exactly zero.