R help - Sep 2007 - lmer giving negative, or no, estimated standard errors

R Users,

 

Emine Bayman sent this out earlier and we do not think it went through.
Appologies if it did. 

 

We want to fit GLMM with lmer with binomial data and a one-way random

effects model (overall mean is a fixed effect and there are random
effects for each binomial). 

We are using the Laplace method. We are simulating multiple data sets
and use the

Laplace method with control = list (usePQL = FALSE)). For most data sets
it works well, 

but for some we get an error message 

(Error in if (any(sd < 0)) return("'sd' slot has negative
entries") :

missing value where TRUE/FALSE needed) 

In these cases we get an estimate for the fixed effect but do not get an

estimate for se.fixef (its negative).  If we change the method to PQL or
change 

control = list (usePQL = TRUE) 

then we get estimates for both fixed effect and se.fixef, but in the

example below the estimates of fixed effects are different for 3 of 4

cases - and the standard errors. Could someone please help us? 

 

Which, if any, of the estimates for

the intercept is the best one? And why are they different?

 

I think the problem occurs because the MLE for the between group
variance

is zero and so the estimate becomes negative. The different estimates

may have to do with pooling or not pooling the between and within sum of
squares

to estimate both the fixed effect and its standard error - but we can
find no

documentation for this.

 

Thanks,  Can anyone help us please?  Is there any documentation that
might help?

 

Kathryn Chaloner

Dept of Biostat

Univ Iowa

kathryn-chaloner@uiowa.edu

And Emine Bayman, PhD student.

emine-unlu@uiowa.edu

 

Here is one example data set with very different estimates for the fixed

effect and its standard error:

 

install.packages("nlme")

install.packages("Matrix")

install.packages("lme4")

install.packages("car")

library(nlme)

library(Matrix)

library(lme4)

library(car)

library(arm)

y <- c(14, 9, 19, 12, 10, 12, 8, 11, 15, 4, 14, 13, 8, 3)

n <- c( 20, 20, 20, 20, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18)

center <- seq(1:14)

 

example1 <- lmer(cbind( y, n - y) ~ 1   + ( 1 | center) ,

family = binomial, niter = 50, method = "Laplace", control = list

(usePQL = FALSE))

fixef(example1)

se.fixef(example1)

 

example2 <- lmer(cbind( y, n - y) ~ 1   + ( 1 | center) ,

family = binomial, niter = 50, method = "Laplace", control = list

(usePQL = TRUE))

fixef(example2)

se.fixef(example2)

 

example3 <- lmer(cbind( y, n - y) ~ 1   + ( 1 | center) ,

family = binomial, niter = 50, method = "PQL",control = list (usePQL 
TRUE))

fixef(example3)

se.fixef(example3)

 

example4 <- lmer(cbind( y, n - y) ~ 1   + ( 1 | center) ,

family = binomial, niter = 50, method = "PQL",control = list (usePQL 
FALSE))

fixef(example4)

se.fixef(example4)


	[[alternative HTML version deleted]]