Partial correlation
Measures the linear relationship between two numeric variables after removing the influence of one or more control variables from both.
Menu: Analyze → Correlate → Partial…
When should I use it?
- Is study time related to grades once hours of paid work are held constant?
- Does a correlation between ice-cream sales and drownings vanish after controlling for temperature?
What are the assumptions?
- Linear relationships; roughly normal variables without extreme outliers.
- Control variables measured without much error.
How do I run it in StatLab?
- Move the variables to correlate to Variables.
- Move the control variable(s) to Controlling for.
- Tick Zero-order correlations to compare with the uncontrolled correlations.
How do I read the output?
| Output | Meaning |
|---|---|
| Correlation | Partial r between each pair. |
| Significance | p-value (two- or one-tailed). |
| df | N − 2 − number of control variables. |
Worked example
On the practice survey dataset (File → Sample dataset), this command — which the dialog's Paste button also produces — runs the procedure:
PROC glmPartialCorr /vars=(pre post) /controls=(age) /tails=2 /zero=NO.
Output tables: Correlations.
How would I report it in APA style?
Partial correlations controlling for Age (years) (N = 150): Pre-test score and Post-test score: r(147) = .883, p < .001.
Computed by StatLab's validated procedure code when this page was built. The example only illustrates the steps; choose variables that fit your own research question.
What are common mistakes?
- Reading a partial correlation as causal evidence.
- Controlling for a variable that lies on the causal path (a mediator).