Date: Sat, 27 Apr 2002 14:37:01 -0400
Reply-To: Monique Morrissey <morrissey-duffy@MINDSPRING.COM>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Monique Morrissey <morrissey-duffy@MINDSPRING.COM>
Content-Type: text/plain; charset="iso-8859-1"
I am running a multivariate regression which has failed the Breusch-Pagan and White tests for heteroskedasticity.
I am using (base) SPSS 10.0 for Windows, and I don't know if the program has the capacity to deal with my problem. I know that SPSS 10.0 has a weighted least squares option, but I can't figure out what to do if the squared residuals are a function of several of the independent variables.
I might be willing to purchase an advanced version of SPSS that can do heteroskedasticity-robust estimators, but the guy I spoke to at SPSS didn't know what heteroskedasticity was, and the specs I looked at weren't helpful either.
I am a new subscriber to SPSSX-L, and I apologize if I am asking a question that has been asked before or is too basic for this list. (I tried searching the archives, but was not successful for some reason.)