[R-sig-ME] MLM help - longitudinal, (overdispersed) count data

Rebecca Friesdorf rfriesdo at gmail.com
Tue Apr 3 01:04:26 CEST 2018


Hi there,

I’m looking for some help analyzing multi-level (longitudinal) data with
the following characteristics: 1) autoregressive structure (daily diary
study), 2) dependent variable is a count (overdispersed Poisson/negative
binomial distribution). I have multiple level 1 and level 2 predictors. I
have done preliminary analyses in HLM 7 and SPSS, but they don’t seem to
have built-in options for data with both of my two characteristics (correct
me if I am wrong here). Based on these analyses I have significant fixed
and random effects for time and my two other level 1 predictors (and not a
whole lot going on with my level 2 predictors).

Which R package would recommend for my analysis (glmmADMB, nlme)? Perhaps
someone has examples/papers/PDFs that outline, in a practical way, how I
would set up my code (or their own code from a past study)? I had ignored
(omitted) missing cases in my initial attempts to do some of these analyses
in the glmmadmb and nlme package, but will also need to figure that out, so
if you have specific recommendations for how to handle those that would be
much appreciated also.

I only know basics with R so the more instructions/explanation the better.

Thank you!

-- 
*Rebecca Friesdorf*
*Ph.D. Candidate *
*Social Psychology*

*WILFRID LAURIER UNIVERSITY*
*Office: *N2068
*Email*: frie3750 at mylaurier.ca

	[[alternative HTML version deleted]]



More information about the R-sig-mixed-models mailing list