To whom it may concern:
I have a question about how to appropriately conduct an lmer analysis for
negative binomially distributed data. I am using R 2.2.1 on a windows machine.
I am trying to conduct an analysis using lmer (for non-normally distributed
data and both random and fixed effects) for negative binomially distributed
data. To do this, I have been using maximum likelihood, comparing the full
model to reduced models (containing all but one effect, for all effects).
However, for negative binomially distributed data, I need to estimate the
parameter theta. I have been doing this by using a negative binomial glm of the
same model (except that all the effects are fixed), and estimating mu as the
fitted model like so:
model_1 <-glm.nb(y~x1+x2+x3, data = datafilename)
mu_1 <- fitted(model_1)
theta_1 <- theta.ml(y, mu_1, length(data), limit = 10, eps =
.Machine$double.eps^0.25, trace = FALSE)
Then, I conduct the lmer, using the estimated theta:
model_11 <-lmer(y~x1+x2+(1|x3), family = negative.binomial(theta = theta_1,
link = “log”), method = “Laplace”)
First, I wondered if this sounds like a reasonable method to accomplish my
goals.
Second, I wondered if the theta I use for reduced models (nested within
model_11) should be estimated using a glm.nb with the same combination of
variables. For example, should a glm.nb with x1 and x3 only be used to estimate
theta for an lmer using x1 and x3?
Third, I wish to test for random effects of one categorical variable with 122
categories (effects of individual). For this variable, the glm.nb (for
estimating theta) does not work--it gives this error message:
Error in get(ctr, mode = "function", envir =
parent.frame())(levels(x), :
orthogonal polynomials cannot be represented accurately enough for 122
degrees of freedom
Is there any way that will allow me to accurately estimate theta using this
particular variable (or without it)? Or should I be using a Poisson
distribution (lognormal?) instead, given these difficulties?
If anyone has advice on how to properly conduct this test (or any references
that might tell me in a clear way), I would be very grateful. Also, please let
me know if I should provide additional information to make my question clearer.
Please respond to me directly, as I am not subscribed to this list.
Thank you very much,
Tracy S. Feldman
Postdoctoral Associate, the Noble Foundation, Ardmore, OK.
__________________________________________________
[[alternative HTML version deleted]]
I think you should use glmm.admb.>library(glmmADMB) >?glmm.admbglmm.admb package:glmmADMB R Documentation Generalized Linear Mixed Models using AD Model Builder Description: Fits mixed-effects models to count data using Binomial, Poisson or negative binomial response distributions. Zero-inflated versions of Poisson and negative binomial distributions are available. 2006/7/28, Tracy Feldman <tracysfeldman at yahoo.com>:> To whom it may concern: > > I have a question about how to appropriately conduct an lmer analysis for negative binomially distributed data. I am using R 2.2.1 on a windows machine. > > I am trying to conduct an analysis using lmer (for non-normally distributed data and both random and fixed effects) for negative binomially distributed data. To do this, I have been using maximum likelihood, comparing the full model to reduced models (containing all but one effect, for all effects). However, for negative binomially distributed data, I need to estimate the parameter theta. I have been doing this by using a negative binomial glm of the same model (except that all the effects are fixed), and estimating mu as the fitted model like so: > > model_1 <-glm.nb(y~x1+x2+x3, data = datafilename) > mu_1 <- fitted(model_1) > theta_1 <- theta.ml(y, mu_1, length(data), limit = 10, eps = .Machine$double.eps^0.25, trace = FALSE) > > Then, I conduct the lmer, using the estimated theta: > > model_11 <-lmer(y~x1+x2+(1|x3), family = negative.binomial(theta = theta_1, link = "log"), method = "Laplace") > > First, I wondered if this sounds like a reasonable method to accomplish my goals. > > Second, I wondered if the theta I use for reduced models (nested within model_11) should be estimated using a glm.nb with the same combination of variables. For example, should a glm.nb with x1 and x3 only be used to estimate theta for an lmer using x1 and x3? > > Third, I wish to test for random effects of one categorical variable with 122 categories (effects of individual). For this variable, the glm.nb (for estimating theta) does not work--it gives this error message: > Error in get(ctr, mode = "function", envir = parent.frame())(levels(x), : > orthogonal polynomials cannot be represented accurately enough for 122 degrees of freedom > Is there any way that will allow me to accurately estimate theta using this particular variable (or without it)? Or should I be using a Poisson distribution (lognormal?) instead, given these difficulties? > > If anyone has advice on how to properly conduct this test (or any references that might tell me in a clear way), I would be very grateful. Also, please let me know if I should provide additional information to make my question clearer. > > Please respond to me directly, as I am not subscribed to this list. > > Thank you very much, > > Tracy S. Feldman > > Postdoctoral Associate, the Noble Foundation, Ardmore, OK. > > __________________________________________________ > > > > [[alternative HTML version deleted]] > > > > ______________________________________________ > R-help at stat.math.ethz.ch 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. > > >-- ?????? Department of Sociology Fudan University
Tracy Feldman <tracysfeldman <at> yahoo.com> writes:> > To whom it may concern: > > I have a question about how to appropriately conduct an lmer analysis fornegative binomially distributed> data. I am using R 2.2.1 on a windows machine. > > I am trying to conduct an analysis using lmer (for non-normally distributeddata and both random and fixed> effects) for negative binomially distributed data. To do this, I have beenusing maximum likelihood,> comparing the full model to reduced models (containing all but one effect, forall effects). However, for> negative binomially distributed data, I need to estimate the parameter theta.I have been doing this by> using a negative binomial glm of the same model (except that all the effectsare fixed), and estimating mu> as the fitted model like so: >\I haven't tried it, but you could also consider using a Poisson-lognormal (rather than neg binomial, which is Poisson-gamma) distribution, which might make this all work rather well in lmer: www.cefe.cnrs.fr/esp/TBElston_Parasitology2001.pdf cheers Ben Bolker