eiwild: eiwild

In eiwild: Ecological Inference with individual and aggregate data

Description

eiwild means Ecological Inference With Individual Level Data.

It enables to estimate the inner cells of an RxC-table with aggregate data with a number of precincts and/or individual level data for a few precincts.

The assumptions are based on Rosen et al's (2001) on the Multinomial-Dirichlet- Model and the Wakefield's (2004) paper on combining individual level data for 2x2-tables. In the master thesis of Thomas Schlesinger (2013) he expanded the 2x2-case to the RxC-case and implemented it in this package. Some of the functions are based on the equivalent functions of the eiPack-package.

It is a hierarchical Bayesian model which uses MCMC-Algorithms to calculate the estimations. Therefore it has all the pros and cons of these methods.

The typical workflow consists of

1. Formatting your data in the accepted way. Loading topleveldat and inspecting the data.frame's aggr and indi will give you a general idea.

2. Running one or multiple times:

   1. function tuneVars

   2. function indAggEi

3. Analyzing results with plot.eiwild, comPlot, summary.eiwild, plotResult

For more examples look at the appropriate functions

Author(s)

Thomas Schlesinger

References

<<to be added>>

eiwild documentation built on May 2, 2019, 6:43 a.m.