In current Basic Statistics module of correlation matrix, you can obtain the numeric results but not the scatterplot of the partial correlation. A workaround of plotting partial correlation through stepwise linear regression model builder will be explained in this article.

In this example, Poverty.sta data from Statistica example data folder is used to analyze the partial correlation between "POP_CHNG" and "PT_POOR" after controlling "AGE".

1. Open Poverty.sta data from Statistica example dataset folder

2. Select "Statistics | Advanced Models | Stepwise Model Builder | Linear Regression"

3. Click "Select variables" to choose "POP_CHNG" as the Dependent variable, "PT_POOR" and "AGE" as Continuous Predictors

4. After selecting those variables, "PT_POOR" and "AGE" will be listed as Marginal variables and "POP_CHNG" will be listed as Dependent variable. Click "Full sample".

5. Wait till "PT_POOR" and "AGE" are displayed in the "Marginal Results Table", select "Age" and then click "Add variables"

6. Wait till "AGE" is moved to the "Model Results Table", click "Graphs" under the "Marginal Results"

7. Three graphs will be generated, select the "Partial Regression Plot" 

8. The default fitting line type for this partial regression plot is LOWESS, continue the following steps to change to a different line type

9. Double click the graph background to access "Graph Options" dialog

10. Under "Plot | Fitting", select your desired line type from the "Fit Type" drilldown (e.g. Linear for linear fitting) and click "OK"

11. To add statistics on the graph, double click graph background again to access "Graph Options" dialog

12. Under "Plot | Scatterplot", click "Statistics" to select statistics in the "Statistics" dialog box and then close the dialog

13 .Click "OK" in the "Graph Options" dialog, a partial correlation scatterplot with your selected fitting type and statistics will be produced.