topic Re: Correct for Heteroskedasticity with PROC REG in SAS Programming

Correct for Heteroskedasticity with PROC REG

Quantopic — Thu, 15 Mar 2018 11:52:35 GMT

Hi all SAS users,

I'm running a linear regression model by using PROC REG (v. SAS 9.4).

The model did not overcome the test about heteroschedasticity and I want to correct for the presence of heteroschedastic errors.

Browsing on the internet I didn't find anything useful to do that.

Can you provide a way, like, for instance, a statement, to do it?

Thanks for the help!

Miracle — Thu, 15 Mar 2018 12:01:33 GMT

Maybe try to log-transform your outcome variable if that helps?

Quantopic — Thu, 15 Mar 2018 13:56:00 GMT

Thanks for replying @Miracle, but I cannot do that!

I need to correct it by applying the heteroscedasticity-consistent parameter estimates.

I can doit by using another PROC too.