s-news
[Top] [All Lists]

Re: lme question

To: Frank Lawrence <Cougar@psu.edu>
Subject: Re: lme question
From: Spencer Graves <spencer.graves@pdf.com>
Date: Fri, 27 Aug 2004 16:21:10 -0400
Cc: s-news@lists.biostat.wustl.edu
In-reply-to: <000001c48c5f$4b7cfa40$641a7680@hhdev.psu.edu>
References: <000001c48c5f$4b7cfa40$641a7680@hhdev.psu.edu>
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax)
1. I don't know SAS. What's FIML? Is that like imputing missing values? If yes, why do you want to do that? That was a great thing to do when you needed balance in a data set in order to be able to get an answer. However, there is much less rationale for that today. "lme" uses maximum likelihood or restricted maximum likelihood and does not impute missing values to get an answer. If you want to impute the values, you can use "predict.lme". 2. Are you familiar with Pinheiro and Bates (2000) Mixed-Effects Models in S and S-Plus (Springer)? Doug Bates is among the leading experts in the world on the best ways to approach this kind of problem. He and his graduate students, including Jose Pinheiro, developed and wrote "lme", and continue making seminal contributions in this area. I have been amply rewarded for the time I have spent with this and other things he has written.
     hope this helps.  spencer graves

Frank Lawrence wrote:

I did not receive a reply the first time I posted this question so I am
re-submitting it.

I notice that lme does not tolerate missing data on the response variable in
the same way that SAS Proc Mixed handles missing data.  Is there a way to
invoke FIML in lme?

Respectfully,

Frank Lawrence

--------------------------------------------------------------------
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>
  • lme question, Frank Lawrence
    • Re: lme question, Spencer Graves <=