Date:         Sun, 25 Sep 2005 12:51:30 +0200
Reply-To:     Marta García-Granero
              <biostatistics@terra.es>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Marta García-Granero
              <biostatistics@terra.es>
Organization: Asesoría Bioestadística
Subject:      Re: Standard Error of an F-statistic
In-Reply-To:  <0741DA624F60AE43A659DD2264AA4A65FDA7E5@ucsur-8.ucsur.nt.pitt.edu>
Content-Type: text/plain; charset=ISO-8859-1

Hi Bozena,

I think you have a slightly wrong idea of what a standard error is
used for. We can talk about the standard error of a certain statistic
(used like estimators of the corresponding parameters in their
populations of origin), like the mean, a difference of means (like in
the case I answered before), a percentage, a difference in
percentages, an Odds Ratio, a Relative Risk... They reflect the
precision of your measure, and are used both for testing and computing
confidence intervals. You don't normally talk about standard error of
a F, Chi-square, Student's t statistic... they are used just to obtain
the p-value for a given test.

What do you mean by the standard error of the F-statistic for
interaction?. This F-value is used to obtain a test p-value, only for
that.

ZB> Close to the topic, can someone tell me the formula for the
ZB> standard error for the F-statistic for interaction? I ran a RM
ZB> Ancova using GLM and I am not sure how to calculate the standard
ZB> error for the F-value based on my output.

HTH

Marta