I'm trying to calculate the Mahalanobis Distance (to check for outliers) in SPSS for a scale of 25 items. When I use the Regression function and save the Mahalanobis Distance values, everything seems to run correctly and MOST of my cases will have an MD value of some sort. About 10% of my cases do not (the cells are simply blank) and while some I can maybe explain away in terms of missing data for specific cases, there remains a substantial chunk of cases where I can't figure out why the cells for the MD values are blank.

I've tried looking at each of the individual cases and there doesn't seem to be anything strange about how the data was entered for the specific scale items.

Can someone please explain to me what is happening and if there are any potential fixes to this?

  • 1
    $\begingroup$ Can you show your data and syntax? There must be some effect of missing cases or filtering or case weight, if specified. To produce the (squared) Mahalanobis distance for a case, the case must be nonmissing and have positive weight for the X variables and the Y variable. $\endgroup$ – ttnphns Mar 3 '17 at 12:03
  • $\begingroup$ Btw, if you don't want to get involved with some Y variable (for it is not needed really to compute the distance, only Xs are needed) you may also use function !smahalc (get it in "Matrix-End Matrix functions" collection on my web-page) which is to run from matrix session. It will return the same result as Regression. $\endgroup$ – ttnphns Mar 3 '17 at 12:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.