I want to run a multiple regression analysis using SPSS. I have used the Mahalanobis d square method to find outliers. However my question is, do I add the dependent variable to the list of independent variables when looking for outliers, or is it just the independent variables that need to be scanned for outliers? (I have around 7 independent variables and 1 dependent)
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
3
You might go either way, but since there are specific methods to look for outliers in the dependent variable after the regression is fitted (for instance, with the so-called 'deleted residuals') I would check regressors for abnormal observations first, then look for outliers in the response when the regression is fitted.
-
$\begingroup$ Outliers in X-space are often called leverage points, search this site! $\endgroup$ Dec 16, 2020 at 18:30
-
$\begingroup$ Yes, I know, but I understood points with high leverage were not the primary concern of the asker. But true, looking for leverage points would be a good way to catch most anything one would consider an abnormal observation in the regressor space. $\endgroup$ Dec 16, 2020 at 21:45
-
$\begingroup$ Well, I found the Q unclear, and am not so sure about the OP's intentions ... I think the Q needs a more elaborate answer, covering uncertain interpretations ... $\endgroup$ Dec 16, 2020 at 22:41