Heteroscedasticity

#1
Hello, I've done a search on this but have not found any similar threads.

Using EVIEWS White test I have identified heteroscedasticity in my multi-variate regression model. I understand that one way to adjust for this is to perform a GLS. I am unsure of how to transform my model to this however. If someone could walk me through each stage (preferably with reference to either excel or spss) that would be massively appreciated. My statistical knowledge is not particularly advanced.

Alternatively, eviews offers an application called "white heterscedasticity-consistent standard error". Would this be an equally valid way to adjust for heteroscedasticity?

Thanks,
Epsilon.
 

dtae

New Member
#2
I use Stata, so you'll have to do a bit of translating of this advice.
I test all of my regression models for heteroscedasticity using the Breusch–Pagan/Cook–Weisberg test [hettest command in stata]. If they show significant deviations from homoscedasticity, I use robust standards errors [reg y x1, robust in stata]. I'm not sure what the SPSS equivalent is, but searching for robust standard errors, or robust regression might get you somewhere.

D