Analyzing Ecological data - Practical Solutions when Things Get Complicated - (In prep)  Zuur, Ieno & Saveliev

The original plan was to call this book: "Shit happens". But obviously, this is an inappropriate title for a scientific textbook. However, it does represent the type of data sets that are analyzed in this book; zero inflated data with random effects, zero inflated data with spatial correlation, zero inflated data with temporal correlation, data sets with large number of collinear explanatory variables, binary data with spatial correlation, data measured in transects, etc., etc. The book only contains case studies (with R code). It will be available towards the end of 2010.

 

Update December 2009

The first two chapters are nearly ready. Chapter 5 is on zero inflated models for 2-way nested marmot data. We have multiple observations (on the number of young) per animal, multiple animals in a colony, and multiple colonies. And lots of zeros! We show how to fit ZIP GLMM for 2-way nested data in R, how to apply a model selection, and we derive equations for the correlation between observations from the same animal, and from different animals from the same colony. We also give equations for the correlation in Poisson GLMMs, and in binomial GLMMs. A very challenging data set!

Chapter 3 is on GAMs, and whether you should model a density with a Gaussian distribution, or with a negative binomial GLM plus the offset.

These chapters are in the final stages of writing and we already started with Chapter 4; it is on zero inflated beetle time series. Hence, zero inflation plus temporal auto-correlation. We also started with Chapter 2, which deals with polar bear data and mixed effects modeling.

We still expect to finish the book by the end of 2011.