s-news
[Top] [All Lists]

Re: Continuous covariates interpretation in glme

To: "Frank E Harrell Jr" <f.harrell@vanderbilt.edu>
Subject: Re: Continuous covariates interpretation in glme
From: "Ali & Hind Lazrak" <alazrak@telus.net>
Date: Wed, 6 Feb 2008 11:12:22 -0800
Cc: "Splus listserve" <s-news@lists.biostat.wustl.edu>
References: <001301c86863$9c0603b0$6601a8c0@Alazrak> <96C6A984FE81AF4492020E9281CE67A8011924B8@sewinexch00.insightful.com> <000a01c868ed$51aa43f0$6601a8c0@Alazrak> <47AA0075.8070705@vanderbilt.edu>
Hi:

Maybe I should have mentioned that I am fitting a logistic regression.with random intercepts to account for the hierarchical structure of my data. I was thinking that it would be easier to interpret the result as the increase in odds ratio per one standard deviation shift.
Am I wrong?

Hind
----- Original Message ----- From: "Frank E Harrell Jr" <f.harrell@vanderbilt.edu>
To: "Ali & Hind Lazrak" <alazrak@telus.net>
Cc: "Stella Karuri" <swkaruri@insightful.com>; "Splus listserve" <s-news@lists.biostat.wustl.edu>
Sent: Wednesday, February 06, 2008 10:46 AM
Subject: Re: [S] Continuous covariates interpretation in glme


Ali & Hind Lazrak wrote:

Hello
You are right, I wanted to transform the continuous covariates for ease of interpretation. However, by standardize, I meant x-mean(x)/sqrt(var(x).
For newbies like me :
I have found that scale() does the job.
 Thank you Stella.
Hind

That would seem to cloud rather than help the interpretation

Frank Harrell


    ----- Original Message -----
    *From:* Stella Karuri <mailto:swkaruri@insightful.com>
    *To:* Splus listserve <mailto:s-news@lists.biostat.wustl.edu>
    *Sent:* Wednesday, February 06, 2008 9:46 AM
    *Subject:* Re: [S] Continuous covariates interpretation in glme

    Splus does not standardize continous variates – by standardizing I
    believe you mean centering to improve interpretability, (if x is the
    continuous covariate, x*=(x-mean(x))).

     When you specify the formula, you can center your covariate i.e.
    instead of

    glme( y~ x);

    have

    xst<- x-mean(x); glme(y~xst)

     stella

    * From: * s-news-owner@lists.biostat.wustl.edu
    <mailto:s-news-owner@lists.biostat.wustl.edu>
    [mailto:s-news-owner@lists.biostat.wustl.edu] *On Behalf Of *Ali &
    Hind Lazrak
    *Sent:* Tuesday, February 05, 2008 5:57 PM
    *To:* Splus listserve
    *Subject:* [S] Continuous covariates interpretation in glme

     Dear all,

     I ran a generalized mixed effects model to predict a binary outcome
    using the 'glme' procedure. This model has two random subject
    (persons nested within companies) and 5 fixed effects terms for the
    subjects.

    Among these 5 fixed effects variables, 2 are continuous.

    As I am interpreting the results I have a question pertaining to the
    treatment of continuous covariates:

    Does Splus standardize automatically continuous covariates?

     Thank you kindly for your answer.

     Hind



--
Frank E Harrell Jr   Professor and Chair           School of Medicine
                     Department of Biostatistics   Vanderbilt University
--------------------------------------------------------------------
This message was distributed by s-news@lists.biostat.wustl.edu.  To
unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
the BODY of the message:  unsubscribe s-news




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