By Damon Verial – Updated August 30, 2022

KamiPhotos/iStock/GettyImages

Before applying linear regression or other parametric techniques, it’s critical to verify that your data exhibit a linear relationship between the predictor (x) and outcome (y). A linear relationship follows the form y = c x, where c is a constant. If this assumption is violated, regression estimates can be biased and inference unreliable. SPSS offers intuitive tools to visually assess linearity, making it a staple in applied statistics.

Step 1 – Import Your Data

Enter your observations into SPSS’s Data Editor, or open an existing .sav file via File → Open. Each case should occupy one row, with variables in columns.

Step 2 – Launch the Scatterplot Dialog

Navigate to Graphs → Legacy Dialogs → Scatter/Dot. This opens the Scatterplot dialog box.

Step 3 – Select the Simple Scatterplot Option

Choose Simple Scatter and click Define.

Step 4 – Assign Variables

In the Define Simple Scatter dialog, drag your predictor to the X‑Axis slot and the outcome to the Y‑Axis slot. By convention, place the variable of primary interest on the y‑axis. Click OK to generate the plot.

Step 5 – Evaluate the Plot for Linearity

Inspect the scatterplot. A roughly oval cloud of points indicates a linear relationship. Patterns such as curves, clusters, or a fan shape suggest nonlinearity, implying that the data may not satisfy regression assumptions. In such cases, consider transformations, non‑linear models, or alternative analysis methods.

References

• Using SPSS Syntax – Jacqueline Collier, 2010