Skip to content
Prev 4884 / 7420 Next

modelling zero inflated continuous data using a two stage model in glmmpql - question about implementing the delta method

Thanks for your assistance Scott,


Option 1- I think this option may be more within reach of my  
statistical competence! So by my understanding, "following the  
parameter extimates uncertainty through", is equivilent to the  
confidence interval for fixed effect level x being a product of the  
the estimate CIs for both models ie

deltamodelUpperCIfor=probability model upperCI*positivemodelupperCI

Option 2- Hmmm, this might be a bit tricky for me, I did some google  
searches on bootstrapping nlme and glmmpql models and I couldnt find  
any examples of this type of thing for those packages. Nonetheless I  
would like to give this method a try.Any advice on r-packages,  
books/online resources, examples of coding would be much appreciated.

In terms of marginal vs conditional im happy with conditional.

Thanks again for your help

Phil

Quoting Scott Foster <scott.foster at csiro.au>:
Philip Harrison MSc
PhD student (Fisheries Ecology)
Department of Biology
University of Waterloo
200 University Avenue West
Waterloo, Ontario, Canada
N2L 3G1
Cell:226-808-2309
Email:pharriso at uwaterloo.ca