Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Numerous t-tests

From   Roger Newson <[email protected]>
To   [email protected]
Subject   Re: st: Numerous t-tests
Date   Wed, 26 Oct 2005 17:15:26 +0100

At 16:51 26/10/2005, Michael wrote:
"Ron�n Conroy" <[email protected]> wrote:

-regress- is a lot faster than t-test, and the output is easier to read, IMO. The coefficient and its confidence interval represent the difference in means and its CI.
While I agree that regress is faster and often preferred to t-tests, one should be aware that regress will assume equal variances for the two groups identified by the dummy variable. That assumption may or may not be what you want.
-regress- has a -robust- option to handle unequal variances. However, it is a valid point that (unlike -ttest-) -regress- does not use a Sattertwhaite-corrected degrees of freedom formula, so there is the danger that confidence intervals will be over-conservative at small sample sizes.

For what it's worth, my own -metaparm- package (downloadable from SSC) can be used with -regress- and -parmest- to produce Satterthwaite confidence intervals for differences between means. I sometimes use it for calculating confidence intervals for a ratio of geometric means, allowing for unequal coefficients of variation.

I hope this helps.


Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
Division of Asthma, Allergy and Lung Biology
King's College London

5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom

Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
or 020 7848 6605 International +44 20 7848 6605
Email: [email protected]

Opinions expressed are those of the author, not the institution.

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index