how to test the equality of sub-sample variances ?

fan · Unread post by **fan** » Tue Oct 11, 2016 10:38 am

Dear Tom,

Could you kindly let me know how I can test the equality of sub-sample variances in Rats? Thanks

TomDoan · Unread post by **TomDoan** » Tue Oct 11, 2016 10:56 am

Subsample variances of what?

fan · Unread post by **fan** » Wed Oct 12, 2016 9:32 pm

TomDoan wrote:Subsample variances of what?

Thank you for reply. For example, I would like to test the equality of the variances of economic growth during recessions and expansions.

TomDoan · Unread post by **TomDoan** » Wed Oct 12, 2016 10:18 pm

In the context of what type of model? Is there a reason you can't just do a LR test?

fan · Unread post by **fan** » Thu Oct 13, 2016 10:21 am

TomDoan wrote:In the context of what type of model? Is there a reason you can't just do a LR test?

Hi Tom, I am not using any model yet, just simple descriptive statistics for samples from different periods.

TomDoan · Unread post by **TomDoan** » Thu Oct 13, 2016 10:44 am

You do have a model; y(t)=mu(i)+noise for i=1,2 depending upon the subsample. Just use a Goldfeld-Quandt test.

fan · Unread post by **fan** » Thu Oct 20, 2016 10:15 am

TomDoan wrote:You do have a model; y(t)=mu(i)+noise for i=1,2 depending upon the subsample. Just use a Goldfeld-Quandt test.

thank you for the reply

The RATS Software Forum