plot.predict.ideal {pscl}R Documentation

plot methods for predictions from ideal objects

Description

Plot classification success rates by legislators, or by roll calls, using predictions from ideal.

Usage

plot.predict.ideal(x, type = c("legis", "votes"),...)

Arguments

x an object of class predict.ideal.
type string; one of legis or votes.
... further arguments passed to or from other methods.

Details

type="legis" produces a plot of the “percent correctly predicted” for each legislator/subject (using the classification threshold set in predict.ideal) against the estimated ideal point of each legislator/subject (the estimated mean of the posterior density of the ideal point), dimension at a time. If the legislators' party affiliations are availble in the rollcall object that was passed to ideal, then legislators from the same party are plotted with a unique color.

type="votes" produces a plot of classification rates for each roll call, by the percentage of legislators voting for the losing side. The x-ordinate is jittered for clarity.

Value

After drawing plots on the current device, exits silently returning invisible(NULL).

Author(s)

Simon Jackman jackman@stanford.edu

See Also

predict.ideal ideal

Examples

data(s109)
id1 <- ideal(s109,
             d=1,
             normalize=TRUE,    ## local identification in 1d
             store.item=TRUE,   ## need this for predictions
             maxiter=500,       ## short run for demo purposes
             burnin=100,
             thin=10) 
phat <- predict(id1)
plot(phat,type="legis")
plot(phat,type="votes")

[Package pscl version 1.03 Index]