Understanding SPSS Regression

A visual guide to interpreting Model Summaries, ANOVA, and Coefficient tables in statistical analysis.

Learning Goals

Learn to read the three main SPSS regression tables: Model Summary, ANOVA, and Coefficients.
Translate statistical output into plain language.
Write the fitted regression equation and interpret coefficients.

Model Summary Table

This table provides a high-level overview of how well your model fits the data.

R Correlation between observed and predicted values.

R² Proportion of variance in the dependent variable explained by the predictors.

Adjusted R² Corrected R² accounting for predictor count. Essential for comparing models.

Std. Error Average distance that observed values fall from the regression line.

Example Interpretation

R² = .42, Adjusted R² = .40
→ About 40% of the variation in job satisfaction is explained by the predictors in the model.

ANOVA Table

The ANOVA table reports the overall significance of the model. Keep your eyes on the F-statistic and its Sig. value.

F-test

Tests whether the model with predictors explains significantly more variance than a model with none.

p-value (Sig.)

If < .05, the model is statistically significant overall.

Important Constraint

A significant F-test does not mean every predictor is significant. It only confirms that the model, as a whole, works better than chance.

Coefficients Table

This is the most detailed part of the output. It shows the regression equation and tests each predictor.

Metric	Meaning
B (Unstandardized)	Slope in original units. "For each extra training hour, satisfaction increases by 0.3 points."
Beta (Standardized)	Slope in standardized units. Useful for comparing which predictor is "stronger".
t & Sig. (p)	Tests if this specific predictor contributes significantly.
Constant	The Intercept. Expected value of Y when all predictors = 0.

y = α + β₁x₁ + β₂x₂ + … + ε

Interpretation

B = 0.25, p < .01
Each additional unit of teaching support is associated with a 0.25 increase in job satisfaction.

Reporting Tip

Use B for real-world interpretation.
Use Beta to compare predictor strength.
Always include t-values & p-values.

Putting it all together

Combine information from all three tables to write your report:

Model Fit (R²) + Significance (ANOVA) + Equation (Coefficients)

"The regression model was significant, F(3, 96) = 12.45, p < .001, explaining 38% of variance in satisfaction (Adjusted R² = .38). Teaching support (B = .25, p < .01, β = .45) was the strongest predictor, followed by workload (B = –.18, p < .05, β = –.28)."

Quick Check-in

Adjusted R² accounts for the number of predictors in the model. It imposes a penalty for adding useless predictors, whereas plain R² will always increase (or stay the same) when you add variables. It is the better metric for comparing models.

It tests whether the model as a whole explains significantly more variance in the dependent variable than a null model (a model with no predictors).

It means there is a significant negative relationship. Since it is standardized, for every 1 standard deviation increase in the predictor, the dependent variable decreases by 0.30 standard deviations.