Date: Wed, 12 Oct 2011 17:03:21 -0500
Reply-To: Warren Schlechte <Warren.Schlechte@TPWD.STATE.TX.US>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Warren Schlechte <Warren.Schlechte@TPWD.STATE.TX.US>
Subject: Re: logistic Regression Standard Error
Content-Type: text/plain; charset="us-ascii"
I think what Robin was trying to say (he can correct me if he wants) is
that the standard errors are derived from the variance-covariance
matrix. That matrix is estimated by inverting a matrix called the
Hessian, which in turn is composed of the partial second derivatives of
the log likelihood with respect to the parameters.
In other words, doing it by hand is difficult, except in trivial
situations (i.e., 2 parameters); that's what we use computers and
numerical approximations for.
SAS's Proc Logistic documentation
(http://www.okstate.edu/sas/v8/saspdf/stat/chap39.pdf) may provide some
clarification, assuming you've had sufficient math.
Warren Schlechte
-----Original Message-----
From: Robin R High [mailto:rhigh@UNMC.EDU]
Sent: Wednesday, October 12, 2011 2:27 PM
Subject: Re: logistic Regression Standard Error
"An Introduction to Generalized Linear Models" by Annette Dobson has a
good example of the computations in the chapter on binary variables,
though it helps to do your math stat homework and look into the unified
explanations of statistical models that belong to the exponential family
and also have basic knowledge of calculus (derivatives) and matrix
inversion
Robin High
UNMC
From:
satish Kumar <satish.abx@GMAIL.COM>
To:
SAS-L@LISTSERV.UGA.EDU
Date:
10/11/2011 09:12 PM
Subject:
logistic Regression Standard Error
Sent by:
"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
Hi, I'm trying to learn and apply logistic regression using SAS. Can
someone
please help me how standard error estimates are calculated using in
logistic
regression. I tried several books including Hosmer and lemeshow also
Allison
but was not successful. If you can point me to an example or hand
calcuations
that would be great.
Regards
Satish
