glht after lmer with "$S4class-" and "missing model.matrix-" errors
You have a mismatch between glht and the model fit by lmer. A model fit by lmer is an object of an S4 class (that's what the "4" in the package name "lme4" is to indicate). The glht function, or possibly a default method for that generic, is treating the model as an S3 class object. Also, you should pay attention to that warning about the estimate of the variance-covariance matrix of the random effects being singular. You need to reconsider the model before trying to do multiple comparisons. On Fri, Jul 25, 2008 at 4:27 AM, Anna Radtke
<annaradtke2309 at googlemail.com> wrote:
Hello everybody. In my case, calculating multiple comparisons (Tukey) after lmer produced the following two errors:
sv.mc <- glht(model.sv,linfct=mcp(comp="Tukey"))
Error in x$terms : $ operator not defined for this S4 class Error in factor_contrasts(model) : no 'model.matrix' method for 'model' found! What I have done before:
sv.growth <- groupedData(length~meas|box_id,outer=~comp,data=sv.growth) model.sv <- lmer(length~comp+(meas|box_id),data=sv.growth)
Warning message: In .local(x, ..., value) : Estimated variance-covariance for factor 'box_id' is singular
summary(model.sv)
Linear mixed-effects model fit by REML
Formula: length ~ comp + (meas | box_id)
Data: sv.growth
AIC BIC logLik MLdeviance REMLdeviance
1587 1606 -786.4 1605 1573
Random effects:
Groups Name Variance Std.Dev. Corr
box_id (Intercept) 466698.1 683.153
meas 230733.7 480.347 -1.000
Residual 9138.3 95.595
number of obs: 120, groups: box_id, 40
Fixed effects:
Estimate Std. Error t value
(Intercept) 600.90 21.31 28.196
comproot -124.84 30.14 -4.142
compshoot -167.36 30.14 -5.553
compxfull -375.13 30.14 -12.446
Correlation of Fixed Effects:
(Intr) comprt cmpsht
comproot -0.707
compshoot -0.707 0.500
compxfull -0.707 0.500 0.500
Thanks for youR help in advance.
Anna
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.