Displaying 1 result from an estimated 1 matches for "gm2s".
Did you mean:
gm2
2009 Nov 20
1
different results across versions for glmer/lmer with the quasi-poisson or quasi-binomial families: the lattest version might not be accurate...
...- SIMULATED EXAMPLE
############# here are the commands both for the quasibinomial and quasipoisson:
library(lme4)
set.seed(1)
period<-rnorm(1000)
herd<-rep(1:50,eac=20)
herd.effect<-rnorm(50)
toto<-rbinom(1000,1,exp(period+herd.effect[herd])/(1+exp(period+herd.effect[herd])))
gm2s <- lmer(toto ~ period + (1 | herd), family = quasibinomial)
summary(gm2s)
gm2sL <- lmer(toto ~ period + (1 | herd), family = binomial)
summary(gm2sL)
#now poisson with similar data:
#essai avec jeu de données simulé:
set.seed(2)
toto<-rpois(1000,exp(period+herd.effect[herd]))...