What I want to work on / think about / play with while visiting CREEM
Speaker: Philip Dixon (Professor, Department of Statistics, Iowa State University, Ames IA)
In this informal talk, I will describe some of the projects and ideas that I want to do while at CREEM. These include:
- Analysis of overdispersed count data. This is motivated by my involvement in a very large weed management study. Counts of weed seeds in farm soils are highly overdispersed. The analysis of the data raises issues of incorporating overdispersion into mixed models, model selection that accounts for overdispersion, and diagnostics for the appropriate type of overdispersion.
- Incorporating biological detail into Integrated Population Models. This is motivated by my work on modeling trends in Mourning Dove populations. The first versions of our model were derived from state-space time-series models. Recent banding (ringing) studies have allowed more biologically-based process models, but there is considerable uncertainty in estimated vital rates, so the biological models do not add much. How much better estimates of vital rates do we need?
- Modeling marked point patterns with multiple marks per location. This is motivated by a recent school board election in Ames. My interest is in estimating the geographic extent of support for one particular group of candidates. Additional details will be much clearer with pictures.