md_bayes_gen: MD Bayes Generalize

In lorenc5/eiCompare: Compares EI, Goodman, RxC Estimates

Description

Tunes and estimates MD Bayes algorithm (ei.MD.bayes). This, combined with md_bayes_table() produces tables of results compatible with EI table of results.

Usage

  md_bayes_gen(dat, form, total_yes=TRUE, total, ntunes = 10, totaldraws = 10000, 
                   seed = 12345, sample = 10000, thin = 100, burnin = 10000, 
                   ret.mcmc = TRUE, ci=c(0.025, 0.975), ci_TRUE=TRUE, produce_draws=FALSE, ...)

Arguments

dat	data.frame() object of just raw candidate vote and raw population counts. Put vote results in first set of columns, put population counts next
form	Formula object, e.g.: cbind(V1, V2, novote) ~ cbind(VtdAVap_cor, VtdBVap_cor, VtdHVap_cor, VtdOVap_cor)
total_yes	Logical, default=TRUE. Include total variable from data? Usually when data are stored in percents
total	character, total variable column name
ntunes	Numeric. How much to tune tuneMD. Default = 10
totaldraws	Numeric. Number of total draws from MD. Default = 10000
seed	Numeric. Default = 12345
sample	Numeric. Default = 10000
thin	Numeric. Default = 10
burnin	Numeric. Default = 10000
ret.mcmc	Logical. Default = TRUE
ci	numeric vector of credible interval (low/high), default is 95 percent= c(0.025, 0.975)
ci_TRUE	Logical, default = TRUE. Include credible intervals in reported results.
produce_draws	Logical, default is FALSE. Produces two-item list of table and md.bayes() mcmc draws (for additional testing and analysis)
...	Additional arguments passed to tuneMD() and ei.MD.bayes()

Value

List object of length 1 (when produce_draws=FALSE). List object of length 2 (when produce_draws=TRUE). First item is list of race x candidate tabular results, with mean, SE, and credible intervals. Second item is mcmc draws.

Author(s)

Loren Collingwood <[e-mail]>

References

eiPack, King et. al. (http://gking.harvard.edu/eiR)

Examples

  
  # TOY DATA EXAMPLE
  canda <- c(10,8, 10, 4, 8)
  candb <- 20-canda
  white <- c(15, 12, 18, 6, 10)
  black <- 20 - white
  toy <- data.frame(canda, candb, white, black)
  
  # Generate formula for passage to ei.reg.bayes() function #
  form <- formula(cbind(canda,candb) ~ cbind(black, white)) 
  
  # Then execute md_bayes_gen(); not run here due to time
  md_bayes_gen(toy, form, total_yes=FALSE, ntunes=1, thin=1, totaldraws=100, sample=10, burnin=1)
  
  # Add in mcmc drawrings #
  drawrings <- md_bayes_gen(toy, form, total_yes=FALSE, ntunes=1, thin=1,totaldraws=100,sample=10,burnin                             =1, produce_draws=TRUE)
  head ( drawrings$draws )

lorenc5/eiCompare documentation built on June 5, 2019, 5:18 p.m.