Multiple Linear Regression in SPSS

This tutorial shows how to do multiple linear regression in SPSS. Different from simple linear regression, multiple linear regression has more than 2 or more independent variables (i.e., multiple Xs).

In this tutorial, we test if students’ writing scores and math scores can be used to predict reading scores.

Read = b0 + b1 Write + b2 Math

This dataset in this tutorial has been used in some other tutorials online (See UCLA website and another website). You can download the SPSS SAV file here.

Steps of Multiple Linear Regression in SPSS

Step 1: Select the menu of “Linear”

“Analyze”, then “Regression”, then “Linear.”

Simple Linear Regression in SPSS: Select menu of "Linear"
Multiple Linear Regression in SPSS: Select the menu of “Linear”

Step 2: Define variables

You can drag “read” as “Dependent” and “write” and “math” in “Block 1 of 1.” Then, you can hit OK to see the output.

Multiple Linear Regression in SPSS: Define variables
Multiple Linear Regression in SPSS: Define variables

Step 3: Interpret the output

The following is the key output of simple linear regression from SPSS.

Part 1 output
Part 1 output

The p-values for both “write” and “math” are < 0.001, which indicates that both of them are significant predictors. In other words, writing and math scores can be used as a predictor for reading scores.

We can also write out the model statement as follows, namely update the constant and regression coefficients. It means that, for writing scores to increase by 1 unit, reading scores will increase by 0.33 units. Further, for math scores to increase by 1 unit, reading scores will increase by 0.52 units.

Read = b0 + b1 Write + b2 Math = 7.54 + 0.33 Write + 0.52 Math


Further Reading

In addition, we can see that R2 is 0.496, which is greater than the R2 when it is a simple linear regression model that has only one predictor of write (0.356). For how to do simple linear regression, please refer to another tutorial.