Dear R-experts,
Quite new to R on this end, so I hope toy can help me out here:
I am running version 2.7.1 on Windows Vista. I have small dataset which consists of:
# NestID: nest indicator for each chicken. Siblings sharing the same nest have the same nest indicator.
# Chick: chick indicator consisting of a unique ID for each single chick.
# Year: 1, 2.
# ClutchSize: 1-, 2- , 3-eggs.
# HO: hatching order within each clutch (1, 2, 3 [first, second and third-hatched chick]).
# SibComp: sibling competence: present/ absent (0, 1)
# Death10: death at ten days post-hatch (0, 1)
In order to account for lack of independence at the nest level (many
chicks are nested in nest... aha!), I'd like to run a GLMM with random slopes and intercepts for nests.
Using lmer, my model for survival at 10 days, for example, would read (or not!) as follows:
model1 <- lmer(Death10 ~ HO + ClutchSize + Sibcomp + Year + (HO+0|NestID) + (1|NestID), family=binomial, 1)
summary(model)
The output give me the following error:
Error en mer_finalize(ans, verbose) : q = 416 > n = 198
1. Whay does this error mean? I was told that my model fails to converge.
2. If so, is there any solution to a lack of convergence?
3. Maybe I could use a model with only random intercept (See below)?
model2 <-lmer(Death10 ~ HO + ClutchSize + Sibcomp + Year + (1|NestID), family=binomial, 1)
I look forward to receiving your two cent?s worth.
Best, Luciano
____________________________________________________________________________________
?Busc? desde tu celular!
Yahoo! oneSEARCH ahora est? en Claro
http://ar.mobile.yahoo.com/onesearch
naive questions about model building and error output
2 messages · Luciano La Sala, Douglas Bates
1 day later
On Fri, Dec 19, 2008 at 1:28 PM, Luciano La Sala
<lucianolasala at yahoo.com.ar> wrote:
Dear R-experts, Quite new to R on this end, so I hope toy can help me out here: I am running version 2.7.1 on Windows Vista. I have small dataset which consists of: # NestID: nest indicator for each chicken. Siblings sharing the same nest have the same nest indicator. # Chick: chick indicator consisting of a unique ID for each single chick. # Year: 1, 2. # ClutchSize: 1-, 2- , 3-eggs. # HO: hatching order within each clutch (1, 2, 3 [first, second and third-hatched chick]). # SibComp: sibling competence: present/ absent (0, 1) # Death10: death at ten days post-hatch (0, 1) In order to account for lack of independence at the nest level (many chicks are nested in nest... aha!), I'd like to run a GLMM with random slopes and intercepts for nests. Using lmer, my model for survival at 10 days, for example, would read (or not!) as follows: model1 <- lmer(Death10 ~ HO + ClutchSize + Sibcomp + Year + (HO+0|NestID) + (1|NestID), family=binomial, 1) summary(model) The output give me the following error: Error en mer_finalize(ans, verbose) : q = 416 > n = 198 1. What does this error mean? I was told that my model fails to converge.
In my overly cryptic notation it means you have too many random effects to be able to estimate them from the available data. Your model has a total of 416 random effects but only 198 observations from which to estimate the parameters.
2. If so, is there any solution to a lack of convergence?
I think you will need to simplify the model.
3. Maybe I could use a model with only random intercept (See below)?
model2 <-lmer(Death10 ~ HO + ClutchSize + Sibcomp + Year + (1|NestID), family=binomial, 1)
Yes, that seems more reasonable.
I look forward to receiving your two cent's worth. Best, Luciano
____________________________________________________________________________________ ?Busc? desde tu celular! Yahoo! oneSEARCH ahora est? en Claro http://ar.mobile.yahoo.com/onesearch _______________________________________________ R-sig-mixed-models at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models