s-news
[Top] [All Lists]

treating co-variates that do not obey the ordinality assumption in PO mo

To: "s-news@lists.biostat.wustl.edu" <s-news@lists.biostat.wustl.edu>
Subject: treating co-variates that do not obey the ordinality assumption in PO models
From: Pedram Sendi <psendi@uhbs.ch>
Date: Thu, 22 May 2003 15:54:08 +0200
Encoding: 67 TEXT
Organization: IKE
Reply-to: "psendi@uhbs.ch" <psendi@uhbs.ch>
Dear All

I am analysing the effect of a stable partnership (p.stable) adjusted for 
other co-variates on the ability to work (abilityrr) in HIV infected 
patients. Abilityrr has three ordinal categories (0,1,2) and I am modeling 
this using a proportional odds model using Harrell's  Design library. 
Abilityrr does not seem to operate in an ordinal fashion with respect to 
sex (m,f) and log.rna (continous variable). How would I need to treat these 
co-variates in the model? Many thanks for your support!!!

Best wishes,
Pedram


Pedram Sendi, MD, DSc
Institute for Clinical Epidemiology & Division of Infectious Diseases
Institut fur klinische Epidemiologie & Abteilung fur Infektiologie
Basel University Hospital
Kantonsspital Basel
Hebelstrasse 10, 3rd Floor
CH-4031 Basel
Switzerland
Phone: +41 61 265 31 02
Fax: +41 61 265 31 09
E-Mail: psendi@uhbs.ch


-----Original Message-----
From:   Frank E Harrell Jr [SMTP:fharrell@virginia.edu]
Sent:   Wednesday, May 14, 2003 12:48 PM
To:     psendi@uhbs.ch
Cc:     s-news@lists.biostat.wustl.edu
Subject:        Re: [S] comparing lrm models

On Wed, 14 May 2003 12:02:38 +0200
Pedram Sendi <psendi@uhbs.ch> wrote:

> Dear all
>
> I have two binary logistic regression models created with the lrm command 

> from the Design library. Is there any analogon to the anova command for
> comparing glm models when using the Design library? As I understood, the
> anova.Design command cannot be used to compute LR statistics but Wald
> statistics.
>
> Thanks for any help.
> Pedram
>
> Pedram Sendi, MD, DSc
> Institute for Clinical Epidemiology & Division of Infectious Diseases
> Institut fur klinische Epidemiologie & Abteilung fur Infektiologie
> Basel University Hospital
> Kantonsspital Basel
> Hebelstrasse 10, 3rd Floor
> CH-4031 Basel
> Phone: +41 61 265 31 02
> Fax: +41 61 265 31 09
> E-Mail: psendi@uhbs.ch
>

Assuming the models are nested, lrtest(fit1,fit2) should do it.  This also 
assumes that the same sample sizes were used in both fits.

---
Frank E Harrell Jr              Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine  http://hesweb1.med.virginia.edu/biostat



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