Page 1 of 1

Test of Equality for Paired Samples

Posted: Mon May 23, 2016 8:22 am
by robert downes
What is the formula used for equality of means testing for a group of two series?

I can replicate the results for the Welch t-test. I'm interested in why, when running the tests and specifying "common sample" the results are identical as not selecting "common sample". I have calculated the test statistic for paired sample by hand which I verified with excel's & R's calculation. The same for the common t-test. However, EViews gives the same t-stat for common sample or not. Is there a way in EViews to use the t-stat for paired samples to test difference-in-means?

Using EViews 9.5

Re: Test of Equality for Paired Samples

Posted: Tue May 24, 2016 9:49 am
by EViews Glenn
Sorry, but EViews doesn't have a built-in paired sample test. It's obviously trivial to calculate by hand and we rarely get requests. I'll put it on the list of things to consider.

Re: Test of Equality for Paired Samples

Posted: Tue May 24, 2016 10:10 am
by robert downes
Thanks for the reply.

After rereading the EViews Users Guide, it seems that I misinterpreted the "common sample" option for "tests of equality" as an option to specify a paired sample.