Skip to content
Prev 5147 / 7420 Next

best choice of GLMM for seed set data

Unfortunately, with a limited range of counts (0 - 4), none of the
classical count distributions (Poisson or negative binomial, or their
zero-inflated counterparts) is likely to fit very well.  You might try
cumulative ordinal logistic regression (without proportional odds
assumption if you can).  I'm currently working on a logistic quantile
regression approach for bounded counts - but the ms is still in prep and
needs review.

Brian

Brian S. Cade, PhD

U. S. Geological Survey
Fort Collins Science Center
2150 Centre Ave., Bldg. C
Fort Collins, CO  80526-8818

email:  cadeb at usgs.gov <brian_cade at usgs.gov>
tel:  970 226-9326


On Wed, Aug 26, 2015 at 12:32 PM, Mariano Devoto <mdevoto at agro.uba.ar>
wrote: