[R-sig-ME] Fwd: same old question - lme4 and p-values

[Douglas Bates]

---------- Forwarded message ----------
From: Douglas Bates <bates at stat.wisc.edu>
Date: Fri, Apr 4, 2008 at 7:54 AM
Subject: Re: same old question - lme4 and p-values
To: andreas.nord at zooekol.lu.se



On Fri, Apr 4, 2008 at 5:24 AM,  <andreas.nord at zooekol.lu.se> wrote:
 > Dear Prof. Bates,

 >  I've recently switched to using R for my analyses, and I find the
lme4 package to be extremely helpful. I have read your explanation
(posted on the mailing list) of why you choose not to display
p-values. Unfortunately, most of the journals I publish in require
that I include p-values, which is why I have to find a way of
calculating them from the lmer output. However, not being a trained
statistician I have some difficulties following your recommendations
given in the explanatory text. In other words, after having fitted my
model, I am not at all sure on what to do in order to obtain p-values
(or similar).

 >  I am sorry to have to bother you with a question I know you have
already answered many times, but perhaps you would be so kind as to
give me some hints on how to proceed.

 I understand your situation.  Statisticians have created the "every
 question of scientific interest must be answered by a p-value" monster
 and now it turns on us.  Nevertheless I am reluctant to give advice on
 p-values in lme4 because apparently I don't know how to do it
 correctly.

 May I send a copy of this reply to the
 R-SIG-Mixed-Models at r-project.org mailing list? ("SIG" == "Special
 Interest Group")? (I ask your permission to send the copy because I am
 quoting your original question.)   Some who subscribe to that mailing
 list may have the courage to wade into this swamp and offer their
 advice.