predict.ideal {pscl}R Documentation

predicted probabilities from an ideal object

Description

Compute predicted probabilities from an ideal object. This predict method uses the posterior mean values of x and beta to make predictions.

Usage

## S3 method for class 'ideal':
predict(object,
                        cutoff=.5,
                        burnin=NULL,
                        ...)

## S3 method for class 'predict.ideal':
print(x,digits=2,...)

Arguments

object an object of class ideal (produced by ideal) with item parameters (beta) stored; i.e., store.item=TRUE was set when the ideal object was fitted
cutoff numeric, a value between 0 and 1, the threshold to be used for classifying predicted probabilities of a Yea votes as predicted Yea and Nay votes.
burnin of the recorded MCMC samples, how many to discard as burnin? Default is NULL, in which case the value of burnin in the ideal object is used.
x object of class predict.ideal
digits number of digits in printed object
... further arguments passed to or from other methods.

Details

Predicted probabilities are computed using the mean of the posterior density of of

x

(ideal points, or latent ability) and

beta

(bill or item parameters). The percentage correctly predicted are determined by counting the percentages of votes with predicted probabilities of a Yea vote greater than or equal to the cutoff as the threshold.

Value

An object of class predict.ideal, containing:

pred.probs the calculated predicted probability for each legislator for each vote.
prediction the calculated prediction (0 or 1) for each legislator for each vote.
correct for each legislator for each vote, whether the prediction was correct.
legis.percent for each legislator, the percent of votes correctly predicted.
vote.percent for each vote, the percent correctly predicted.
yea.percent the percent of yea votes correctly predicted.
nay.percent the percent of nay votes correctly predicted.
party.percent the average value of the percent correctly predicted by legislator, separated by party, if party information exists in the rollcall object used for ideal. If no party information is available, party.percent = NULL.
overall.percent the total percent of votes correctly predicted.
ideal the name of the ideal object, which can be later evaluated
desc string, the descriptive text from the rollcall object passed to ideal

Note

When specifying a value of burnin different from that used in fitting the ideal object, note a distinction between the iteration numbers of the stored iterations, and the number of stored iterations. That is, the n-th iteration stored in an ideal object will not be iteration n if the user specified thin>1 in the call to ideal. Here, iterations are tagged with their iteration number. Thus, if the user called ideal with thin=10 and burnin=100 then the stored iterations are numbered 100, 110, 120, .... Any future subsetting via a burnin refers to this iteration number.

See Also

ideal, summary.ideal, plot.predict.ideal

Examples

data(s109)

## Not run: 
id1 <- ideal(s109, meanzero=TRUE,
             store.item=TRUE)      ## too long for examples
## End(Not run)

id1 <- ideal(s109,
             d=1,
             meanzero=TRUE,
             store.item=TRUE,  ## need this to be TRUE for predict
             maxiter=500,
             burnin=100,
             thin=10)  

phat <- predict(id1)
phat         ## print method

[Package pscl version 1.03 Index]