[R-sig-ME] zero-inflation in glmmADMB

[Highland Statistics Ltd]

> Hello,
>
> I am investigating impacts of forest fragmentation due to gas development
> on songbirds, and have run mixed effects models using lme4.  My response
> variables are species abundances (count data), fixed effects are landscape
> metrics (forest edge density, percent gas cover, percent forest cover), and
> random effects are year and survey point (repeated measures over an 8-yr
> period).
>
> However -- having recently realized that I need to better account for the
> high numbers of zeros in my count data, I decided to try zero-inflated
> poisson models in glmmADMB.
>
> (1) Although many of the species I'm looking at have high numbers of zeros,
> the results of my ordinary poisson models and zeroInflation=TRUE poisson
> models appear to be identical, which seems odd:
>
> *Ex.*:
> CERW.pois <- glmmadmb(CERW.50m ~ Edge.100m + pGAS.100m + pFOR.500m
> + (1|Year) + (1|Point), family="poisson")
>
> CERW.zipois <- glmmadmb(CERW.50m ~ Edge.100m + pGAS.100m + pFOR.500m
> + (1|Year) + (1|Point), zeroInflation=TRUE, family="poisson")
>
> Am I doing something wrong here, or is it common for outputs to look nearly
> identical, with or without zero-inflation?


I suggest that you write out the equations of both models. See where it 
differs. Then stick in the estimated parameters, and especially focus on 
the binary part. That will answer your question.


As to the second question...try using offset(Log(What Ever)) directly in 
the formula...that should work. Note the log!


Alain

>
> (2)  I am also trying to incorporate offsets for detection probability
> (calculated using the QPAD approach outlined in Solymos et al., 2013) --
> but the offset term doesn't appear to be recognized ("unused argument"
> error)
>
> *Ex.*:
> (.... family="poisson", offset=offset(CERW.offs))
>
> *For what it's worth, here is the offset code that works in lme4 (doesn't
> work in glmmADMB):
> (.... family="poisson", offset=*corrections2offset*(CERW.offs))
>
> Any suggestions or advice would be greatly appreciated.
>
> Thanks,
> Laura Farwell
> West Virginia University
>
> 	[[alternative HTML version deleted]]
>








-- 
Dr. Alain F. Zuur

First author of:
1. Beginner's Guide to GAMM with R (2014).
2. Beginner's Guide to GLM and GLMM with R (2013).
3. Beginner's Guide to GAM with R (2012).
4. Zero Inflated Models and GLMM with R (2012).
5. A Beginner's Guide to R (2009).
6. Mixed effects models and extensions in ecology with R (2009).
7. Analysing Ecological Data (2007).

Highland Statistics Ltd.
9 St Clair Wynd
UK - AB41 6DZ Newburgh
Tel:   0044 1358 788177
Email: highstat at highstat.com
URL:   www.highstat.com