Date: Sun, 20 Aug 2006 17:14:42 -0400
Reply-To: Jim Groeneveld <jim2stat@YAHOO.CO.UK>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Jim Groeneveld <jim2stat@YAHOO.CO.UK>
Subject: Re: F-tests and t-tests to chi-square
Hi Ted,
It is not entirely clear what you want to do, but it is my impression that
you are thinking along impossible or rather unjustified ways. All kinds of
tests, whether it be F-tests, t-tests, CHiSq and many more compute their
own statistics. These are more or less independent of each other and can
not be "converted" into one another (only the t and F value have some
mathematical correspondence). Each statistic is then interpreted in terms
of probability and the only comparison that can be done between statistics
is in terms of significance (the so called p-value).
So, to make a long story short, I don't think you can "convert" F and t
values into CHiSq and back.
Regards - Jim.
--
Jim Groeneveld, Netherlands
Statistician, SAS consultant
home.hccnet.nl/jim.groeneveld
On Sun, 20 Aug 2006 13:13:53 -0400, Ted Barker <ted.barker@GMAIL.COM> wrote:
>This may be a simple question, but I do not have a book that explicitly
>deals with the topic.
>
>I am trying to convert F-tests and t-tests to chi-square.
>
>I need to combine multiple imputed datasets where ANOVAs are the mode of
>analysis. Most software, such as PROC MIANANLYZE and NORM, appear geared
>towards regression models. I want to find a solution for ANOVAs, particurly
>for t-tests via LSMEANS, underlying significant interactions (or main
>effects) for categorical variables.
>
>I noticed Paul Allison has a COMBCHI macro for SAS. I was thinking a
>potential solution would be to output F-tests and t-tests of interest from,
>let's say, 5 imputed models, convert to CHISQ, then aggregate with Dr.
>Allison's MACRO, then convert back to the f tests and t-tests. This way I
>might even be able to compute effect size statstistics from the combined F
>and t's.
>
>Any thoughts on this? Any suggestions would be great, particularly in
regard
>to F and t to Chi square.
>
>Best regards,
>
>Ted
>
>--
>Ted Barker, Ph.D
>Department of Psychology
>University of Montreal
| 
