[R-sig-ME] a problem with lme4_0.999375-0?

[vito muggeo]

In previous message I used arm::se.ranef() to get st.errs of the 
predictions from the lmer fit.
However the problem occurs also without using arm::se.ranef(), and it 
seems to depend on ranef()..

o<-ranef(M1,postVar=TRUE)
sqrt(attr(o[[1]],"postVar")[,,1:5])

regards,
vito

vito muggeo ha scritto:
> Dear all,
> 
> I'm running data from "Gelman&Hill Data Analysis Using regression.. 
> pag259-261" (you can get data by running the code below). In particular 
> I'm interested in getting variances of the predictions of the random 
> intercepts. I get "right" (as compared with those from the book) results 
>   using lme 0.99875-9 (downloaded from CRAN).
> 
> However using the latest lme4_0.999375-0 obtained via
> 
> install.packages("lme4", repos = "http://r-forge.r-project.org")
> 
> the variances of the predictions are wrong..(they are constant 
> regardless of different cluster sizes. You can check this running the 
> code below..
> 
> Many thanks,
> vito
> 
> 
> srrs2 <- 
> read.table("http://www.stat.columbia.edu/~gelman/arm/examples/radon/srrs2.dat",
>     header=T, sep=",")
> 
> mn <- srrs2$state=="MN"
> radon <- srrs2$activity[mn]
> log.radon <- log (ifelse (radon==0, .1, radon))
> floor <- srrs2$floor[mn]       # 0 for basement, 1 for first floor
> n <- length(radon)
> y <- log.radon
> x <- floor
> # get county index variable
> county.name <- as.vector(srrs2$county[mn])
> uniq <- unique(county.name)
> J <- length(uniq)
> county <- rep (NA, J)
> for (i in 1:J){
>    county[county.name==uniq[i]] <- i
> }
> library(arm)
> 
> M1<-lmer(y~1+x+(1|county))
> ranef(M1)[[1]][1:10,] #OK
> 
> se.ranef(M1)[[1]][1:10,] #wrong results: they are constant!!!??!!!
> 
> 
> 
>  > sessionInfo()
> R version 2.6.1 (2007-11-26)
> i386-pc-mingw32
> 
> locale:
> LC_COLLATE=Italian_Italy.1252;LC_CTYPE=Italian_Italy.1252;LC_MONETARY=Italian_Italy.1252;LC_NUMERIC=C;LC_TIME=Italian_Italy.1252
> 
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> 
> other attached packages:
> [1] foreign_0.8-23    car_1.2-7         arm_1.1-1 
> R2WinBUGS_2.1-6   coda_0.13-1       lme4_0.999375-0   Matrix_0.999375-4 
> lattice_0.17-2
> [9] MASS_7.2-38
> 
> loaded via a namespace (and not attached):
> [1] grid_2.6.1
> 
> 

-- 
====================================
Vito M.R. Muggeo
Dip.to Sc Statist e Matem `Vianelli'
Università di Palermo
viale delle Scienze, edificio 13
90128 Palermo - ITALY
tel: 091 6626240
fax: 091 485726/485612