You have run your multiple regression in SPSS, Stata or R, and now you are looking at a screen full of numbers. Which ones go into your results chapter, and how should they be written? This guide shows you exactly what to report for a multiple linear regression in APA 7 style, with an example paragraph and table you can adapt.
What to report
A complete regression write-up usually includes:
- The model and its purpose: what you predicted and which predictors you entered.
- Assumption checks: a brief statement that the assumptions were met, or how you dealt with violations.
- Overall model fit: R² (or adjusted R²) and the F test with its degrees of freedom and p value.
- Each predictor: the unstandardised coefficient (B), its standard error (SE), the standardised coefficient (β), the t value, the p value and ideally a 95% confidence interval for B.
- Interpretation: what the results mean in relation to your hypotheses.
Check the assumptions first
Before reporting results, confirm and briefly mention:
- Linearity between predictors and the outcome (scatterplots).
- Independence of errors (for example a Durbin-Watson value close to 2).
- Homoscedasticity (a plot of standardised residuals against predicted values with no clear funnel shape).
- Normality of residuals (histogram or P-P plot).
- No problematic multicollinearity (VIF values below about 5, or tolerance above 0.2).
- No unduly influential cases (for example Cook’s distance below 1).
APA 7 formatting rules for statistics
- Italicise statistical symbols such as B, SE, t, F, p and R². Greek letters such as β are not italicised.
- Report most statistics to two decimal places and p values to two or three decimals.
- Report exact p values (p = .023), except below .001, which is written as p < .001.
- Omit the leading zero for values that cannot exceed 1, such as p and R² (write .32, not 0.32).
- Give degrees of freedom in brackets after F and t.
Example results paragraph
Suppose you surveyed 200 employees and tested whether job autonomy, supervisor support and workload predict work engagement.
A multiple linear regression was conducted to examine whether job autonomy, supervisor support and workload predict work engagement. Preliminary analyses indicated no violations of the assumptions of linearity, normality, homoscedasticity or multicollinearity (all VIFs < 1.6). The model was statistically significant, F(3, 196) = 30.75, p < .001, and explained 32% of the variance in work engagement (R² = .32, adjusted R² = .31). Job autonomy (β = .38, p < .001) and supervisor support (β = .24, p < .001) were significant positive predictors of engagement, whereas workload was not a significant predictor (β = −.09, p = .154). Hypotheses 1 and 2 were therefore supported, and Hypothesis 3 was not supported.
Example regression table
APA tables have a number in bold, a title in italics on the next line, horizontal lines only, and a note below for abbreviations or significance information.
Table 4
Multiple Regression Predicting Work Engagement
| Predictor | B | SE | 95% CI for B | β | t | p |
|---|---|---|---|---|---|---|
| (Constant) | 1.42 | 0.31 | [0.81, 2.03] | 4.58 | < .001 | |
| Job autonomy | 0.41 | 0.07 | [0.27, 0.55] | .38 | 5.86 | < .001 |
| Supervisor support | 0.26 | 0.07 | [0.12, 0.40] | .24 | 3.71 | < .001 |
| Workload | −0.10 | 0.07 | [−0.24, 0.04] | −.09 | −1.43 | .154 |
Note. N = 200. R² = .32, adjusted R² = .31, F(3, 196) = 30.75, p < .001. CI = confidence interval.
The figures above are illustrative. Always take your values from your own output.
Where to find the numbers in SPSS
- Model Summary: R², adjusted R² and Durbin-Watson.
- ANOVA table: the F value, its degrees of freedom (regression and residual) and Sig. (the p value).
- Coefficients table: B, Std. Error, Beta (β), t, Sig., the 95% confidence intervals (if requested) and the collinearity statistics.
Common mistakes
- Reporting “p = .000”. SPSS rounds; write p < .001 instead.
- Pasting raw SPSS output tables into the thesis instead of formatting an APA table.
- Interpreting β as if it were B, or the other way round.
- Claiming causation from cross-sectional survey data.
- Forgetting to state whether each hypothesis was supported.
Want your regression or other analyses checked, run or written up correctly? Our data analysis specialists produce APA-formatted tables and clear explanations of every result.