turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Welch's two sample t-test

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

06-01-2017 01:16 PM

I am very new to SAS and still tyring to understand the results tables that are produced. I have been asked to run a Welch's two sample t-test on a data set. I am able to run the process t-test using the 't Tests' under the Tasks and Utilities. However for the Welch's test that refers to an unequal variance, I am unsure where in the code you tell it that your sample has unequal variances. Or is that you refer to 'Pooled' or 'Satterwaite' rows when you need to refer to the numbers produced? I can see that in in one of the result boxes that are produced it calculates variances on 'Pooled' and 'Satterwaite'. Do I look to see which variance is unequal and those will be the Welch's results I should use?

Appreciate the insight.

Accepted Solutions

Solution

06-01-2017
02:08 PM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to vasqj

06-01-2017 01:38 PM

The Satterthwaite test is for unequal variances. Some references cal it the Aspin-Welch-Satterthwaite t-test.

The Getting Started example in PROC TTEST contains an example. It is produced by default, so you do not need to specify any specify an option for it.

All Replies

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to vasqj

06-01-2017 01:38 PM

I don't find a reference to a Welch's t-test but proc ttest supports the Cochran option which is a ttest with unequal variance.

That would just be one additional parameter on the Proc statement such as:

Proc ttest data=mydata cochran;

class classvar;

var analysisvar;

run;

Solution

06-01-2017
02:08 PM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to vasqj

06-01-2017 01:38 PM

The Satterthwaite test is for unequal variances. Some references cal it the Aspin-Welch-Satterthwaite t-test.

The Getting Started example in PROC TTEST contains an example. It is produced by default, so you do not need to specify any specify an option for it.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Rick_SAS

06-01-2017 02:08 PM

Rick thank you. I just need to get a bit more familar with the readouts from the results.