s-news
[Top] [All Lists]

Re: reference value in lrm

To: jose Bartolomei <surfprjab@hotmail.com>
Subject: Re: reference value in lrm
From: Frank E Harrell Jr <f.harrell@vanderbilt.edu>
Date: Mon, 27 Nov 2006 09:38:01 -0600
Cc: s-news@lists.biostat.wustl.edu
In-reply-to: <BAY105-F3851E46CAB10F8914F2874B6E60@phx.gbl>
References: <BAY105-F3851E46CAB10F8914F2874B6E60@phx.gbl>
User-agent: Icedove 1.5.0.7 (X11/20061013)
jose Bartolomei wrote:
Dear S users,

I am conducting a logistic regression (lrm) using the Design library.

I have a categorical dependent variable and I will like to change de reference value in the model.

There is no need to change the reference value when fitting the model. All estimates should come from summary( ), contrast( ), predict( ), and for those it is very easy to use any category of your choice as the reference value. summary( ) defaults to the modal category but do ?summary.Design to see how to use others quite easily.

Frank Harrell


In addition I will like to exclude one of the categories included in the dependent variable.

Which are the commands within the lrm model for that?

Thanks,

Jose


--
Frank E Harrell Jr   Professor and Chair           School of Medicine
                     Department of Biostatistics   Vanderbilt University

<Prev in Thread] Current Thread [Next in Thread>