How to Report Regression Results in APA Format

APA Reporting Template

Use this template to report your regression results. Replace the bracketed placeholders with your values.

Simple Linear Regression

A simple linear regression was conducted to predict [outcome variable] from [predictor variable]. The regression model was statistically significant, F(1, [df_residual]) = [F-value], p = [p-value], $R^2$ = [R-squared]. [Predictor] significantly predicted [outcome], B = [unstandardized B], SE = [standard error], $\beta$ = [standardized beta], t([df]) = [t-value], p = [p-value]. For every one-unit increase in [predictor], [outcome] [increased/decreased] by [B value] units.

Multiple Linear Regression

A multiple linear regression was conducted to predict [outcome variable] from [predictor 1], [predictor 2], and [predictor 3]. The overall regression model was statistically significant, F([df_regression], [df_residual]) = [F-value], p = [p-value], $R^2$ = [R-squared], adjusted $R^2$ = [adjusted R-squared]. Together, the predictors accounted for [percentage]% of the variance in [outcome]. [Predictor 1] ($\beta$ = [beta], p = [p-value]) and [predictor 2] ($\beta$ = [beta], p = [p-value]) were significant predictors, while [predictor 3] ($\beta$ = [beta], p = [p-value]) was not.

Worked Example

Scenario: A researcher tested whether hours of practice and self-efficacy predicted performance on a music audition ($N = 85$).

Results:

• Overall model: $F(2, 82) = 24.31, p < .001, R^2 = .37, \text{adjusted } R^2 = .36$
• Practice hours: $B = 1.42, SE = 0.38, \beta = .41, t(82) = 3.74, p < .001$
• Self-efficacy: $B = 0.87, SE = 0.31, \beta = .31, t(82) = 2.81, p = .006$

APA Write-Up:

A multiple linear regression was conducted to predict audition performance from hours of practice and self-efficacy. The overall regression model was statistically significant, F(2, 82) = 24.31, p < .001, $R^2$ = .37, adjusted $R^2$ = .36. Together, the two predictors accounted for 37% of the variance in audition performance. Hours of practice was a significant predictor ($\beta$ = .41, p < .001), as was self-efficacy ($\beta$ = .31, p = .006). For every additional hour of weekly practice, audition scores increased by 1.42 points, holding self-efficacy constant. Practice hours was the stronger predictor based on standardized coefficients.

Reporting Checklist

• [ ] Named the type of regression (simple linear, multiple linear, hierarchical)
• [ ] Stated the outcome variable and all predictor variables
• [ ] Reported the overall model F test with both degrees of freedom
• [ ] Reported $R^2$ (and adjusted $R^2$ for multiple regression)
• [ ] Reported the exact p-value for the overall model
• [ ] For each predictor, reported B (unstandardized), SE, $\beta$ (standardized), t, and p
• [ ] Interpreted the direction and meaning of key coefficients
• [ ] Identified which predictors were significant and which were not
• [ ] Mentioned assumption checks (linearity, normality of residuals, homoscedasticity, multicollinearity)
• [ ] Reported VIF values if multicollinearity was assessed
• [ ] Used italics for statistical symbols (F, p, B, t, SE)

Common Mistakes

1. Confusing B and $\beta$ — B is the unstandardized coefficient (in the original units of the variables). $\beta$ is the standardized coefficient (used to compare relative importance of predictors). Report both.
2. Omitting the overall model test — Always report the F test and $R^2$ for the full model before reporting individual predictors.
3. Not reporting standard errors — SE for each coefficient is needed for readers to evaluate precision and compute confidence intervals.
4. Interpreting non-significant predictors as "having no effect" — A non-significant predictor may still contribute to the model. Say it "was not a statistically significant predictor," not that it "had no effect."
5. Forgetting adjusted $R^2$ — For multiple regression, always report adjusted $R^2$ because $R^2$ increases with every added predictor regardless of its usefulness.
6. Ignoring multicollinearity — If predictors are highly correlated with each other, report VIF values. VIF > 5 (or > 10, depending on the convention) suggests a problem.

Non-Significant Results

If your overall model is not significant:

A multiple linear regression was conducted to predict audition performance from hours of practice and self-efficacy. The overall regression model was not statistically significant, F(2, 82) = 1.87, p = .161, $R^2$ = .04, adjusted $R^2$ = .02. Neither practice hours ($\beta$ = .14, p = .224) nor self-efficacy ($\beta$ = .10, p = .381) significantly predicted audition performance.

If the overall model is significant but an individual predictor is not:

Self-efficacy was not a significant predictor of audition performance, $\beta$ = .09, t(82) = 0.78, p = .438, after controlling for hours of practice.

Results Table Format

Regression Coefficients Table

Note. $R^2$ = .37, adjusted $R^2$ = .36. CI = confidence interval.

Model Summary Table (for Hierarchical Regression)