R help - Oct 2005 - R-help Digest, Vol 32, Issue 26

r-help at stat.math.ethz.ch on Wednesday, October 26, 2005 at 6:00 AM -0500
wrote:

Ronaldo,
Try Harold's suggestion. The df still won't agree, because lmer (at
least in its current version) just puts an upper bound on the df. But that
should be OK, because all those t tests are approximations anyways, and you can
get better confidence
intervals (credible intervals, whatever) by using the mcmcsamp() function that
works with lmer()
alan
>
"Doran, Harold" <HDoran at air.org>
responded:
>
>
>>There is an issue with implicit nesting in lmer. In your lme() model you
nest
>>block/irrigation/density/fertilizer. In lmer you need to do something
like
>>(I dind't include all of your variables, but I think the makes the
point)
>>
>>lmer(yield~irrigation*density*fertilizer+(1|fertilizer:density)+(1|density),
data)
>>
>>Which notes that fertilizer is nested in density. 
>>
>>Try this and then compare the results. 
>
"Ronaldo Reis-Jr." <chrysopa at gmail.com>,
wrote:
>
>>I make the correct model with aov, lme do compare with lmer.
>
>>But I cant make a correct model in lmer. Look that the aov and lme
results are
>>similars, but very different from lmer. In aov and lme is used the
correct DF
>>for each variable, in lmer it use a same DF for all? Denom=54.
>
--
Alan B. Cobo-Lewis, Ph.D.		(207) 581-3840 tel
Department of Psychology		(207) 581-6128 fax
University of Maine
Orono, ME 04469-5742     		alanc at maine.edu

http://www.umaine.edu/visualperception

In addition to the response below, Doug Bates has talked about this on
this list previously. I did
> RSiteSearch('bates degrees of freedom lmer')
The first one that came up has Doug's response to this question as well

Harold
 

-----Original Message-----
From: r-help-bounces at stat.math.ethz.ch
[mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Alan Cobo-Lewis
Sent: Wednesday, October 26, 2005 8:53 AM
To: r-help at stat.math.ethz.ch
Subject: Re: [R] R-help Digest, Vol 32, Issue 26

r-help at stat.math.ethz.ch on Wednesday, October 26, 2005 at 6:00 AM -0500
wrote:

Ronaldo,
Try Harold's suggestion. The df still won't agree, because lmer (at
least in its current version) just puts an upper bound on the df. But
that should be OK, because all those t tests are approximations anyways,
and you can get better confidence intervals (credible intervals,
whatever) by using the mcmcsamp() function that works with lmer() alan
>
"Doran, Harold" <HDoran at air.org>
responded:
>
>
>>There is an issue with implicit nesting in lmer. In your lme() model 
>>you nest block/irrigation/density/fertilizer. In lmer you need to do 
>>something like (I dind't include all of your variables, but I think 
>>the makes the point)
>>
>>lmer(yield~irrigation*density*fertilizer+(1|fertilizer:density)+(1|den
>>sity), data)
>>
>>Which notes that fertilizer is nested in density. 
>>
>>Try this and then compare the results. 
>
"Ronaldo Reis-Jr." <chrysopa at gmail.com>,
wrote:
>
>>I make the correct model with aov, lme do compare with lmer.
>
>>But I cant make a correct model in lmer. Look that the aov and lme 
>>results are similars, but very different from lmer. In aov and lme is 
>>used the correct DF for each variable, in lmer it use a same DF for
all? Denom=54.
>
--
Alan B. Cobo-Lewis, Ph.D.		(207) 581-3840 tel
Department of Psychology		(207) 581-6128 fax
University of Maine
Orono, ME 04469-5742     		alanc at maine.edu

http://www.umaine.edu/visualperception

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html