-----Original Message-----
From: R-sig-mixed-models [mailto:r-sig-mixed-models-bounces at r-
project.org] On Behalf Of Ben Bolker
Sent: Tuesday, May 7, 2019 12:31 PM
To: r-sig-mixed-models at r-project.org
Subject: Re: [R-sig-ME] A consultation about DF of the result of lmer
It's very hard to say without more information (try e.g. here
<https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-
reproducible-example>
or look at other posts in the list archive
<https://stat.ethz.ch/pipermail/r-sig-mixed-models/>
A first guess is that you have a fixed-effect predictor that's
supposed to be a factor with 6 levels, but is actually being interpreted
as a numeric variable. If that's the case, then either changing it
within the data set
mydata$pred1 <- factor(mydata$pred1)
or doing it on the fly in your model
lmer(response ~ factor(pred1) + ...)
should fix the problem.
On 2019-05-05 9:11 p.m., ??? wrote:
Dear,
I am a graduated student who's topic is ecology. Recently, I am
studying how to establish a liner mixed model to exclude the error
caused by the difference of site by using lme4 in R. I found when I
calculated p value by using the function of "Anova" of car package the
Df is 1. But in fact the data set has 6 levels. Then I also operate
according to the code of reference PDF of lme4 in P52(lmer) without any
change. When run "anova(fm1, fm2)" I found the Df of fm1 also is 1. As I
see if the Df is wrong the p value would be wrong either. I want to know
how to correct my wrong. I will very appreciate your reply.
Best regards,
Rumeng He
[[alternative HTML version deleted]]