anominate: alpha-NOMINATE: Ideal Point Estimator

Description Usage Arguments Value Author(s) References See Also Examples

View source: R/anominate.R

Description

Fits the Bayesian (MCMC-based) version of Poole and Rosenthal's NOMINATE model described in Carroll, Lewis, Lo, Poole and Rosenthal, “The Structure of Utility in Spatial Models of Voting,” American Journal of Political Science 57(4): 1008–1028. The method estimates the structure of political actors' utility functions from roll call data with the inclusion of a separate parameter denoted as alpha. alpha values of 1 indicate normal (Gaussian) utility, while alpha values of 0 indicate quadratic utility.

Usage

1
2
anominate(rcObject, dims=1, nsamp=1000, thin=1, burnin=500, minvotes=20, 
	lop=0.025, polarity=1, random.starts=TRUE, verbose=FALSE, constrain=FALSE)

Arguments

rcObject

An roll call matrix of class rollcall, from Simon Jackman's pscl package

dims

Number of dimensions to estimate

nsamp

Total number of iterations for the sampler. nsamp divided by thin must be larger than burnin.

thin

Thinning interval

burnin

Number of iterations to be discarded

minvotes

Minimum number of votes required for a legislator to be included in the analysis

lop

A proportion between 0 and 1, the cut-off used for excluding lopsided votes, expressed as the proportion of non-missing votes on the minority side. The default, lop=0.025, eliminates votes where the minority is smaller than 2.5 percent

polarity

A vector specifying the row number of the legislator(s) constrained to have a positive (i.e., right-wing or conservative) score on each dimension

random.starts

If TRUE, initial values for the legislator and bill parameters are randomly drawn from a uniform distribution. If FALSE, the W-NOMINATE estimates are used as the initial values

verbose

If TRUE, the progress of the sampler at each 100th iteration is printed to the screen

constrain

If TRUE, this constrains alpha=1 and does not estimate it. This option should be used if more than one dimension is being estimated, which makes the method equivalent to a Bayesian implementation of Poole and Rosenthal's (1997) NOMINATE model.

Value

A list with the following elements:

alpha

An object of class mcmc with the sampled values of the alpha parameter

beta

An object of class mcmc with the sampled values of the beta parameter

legislators

A object of class mcmc with the sampled values of the legislator ideal points, with each dimension stored in a separate list (e.g., the first dimension coordinates are stored in legislators[[1]], the second dimension coordinates in legislators[[2]], etc.)

yea.locations

A object of class mcmc with the sampled values of the Yea locations (midpoint - spread in W-NOMINATE) for each vote, with each dimension stored in a separate list (e.g., the first dimension coordinates are stored in yea.locations[[1]], the second dimension coordinates in yea.locations[[2]], etc.)

nay.locations

A object of class mcmc with the sampled values of the Nay locations (midpoint + spread in W-NOMINATE) for each vote, with each dimension stored in a separate list (e.g., the first dimension coordinates are stored in nay.locations[[1]], the second dimension coordinates in nay.locations[[2]], etc.)

wnom.result

An object of class nomObject with the W-NOMINATE results

Author(s)

Christopher Hare, Royce Carroll, Jeffrey B. Lewis, James Lo, Keith T. Poole and Howard Rosenthal

References

Carroll, Royce, Jeffrey B. Lewis, James Lo, Keith T. Poole and Howard Rosenthal. 2013. “The Structure of Utility in Spatial Models of Voting.” American Journal of Political Science 57(4): 1008–1028.

Poole, Keith T. and Howard Rosenthal. 1997. Congress: A Political-Economic History of Roll Call Voting. New York: Oxford University Press.

See Also

'summary.anominate','plot.anominate','densplot.anominate','traceplot.anominate'.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
data(sen111)

### This command conducts estimates, which we instead load using data()
#sen111_anom <- anominate(sen111, dims=1, polarity=2, nsamp=200, thin=1,
#	 burnin=100, random.starts=FALSE, verbose=TRUE)

data(sen111_anom)

summary(sen111_anom)

## Graphical summaries
plot(sen111_anom)
densplot.anominate(sen111_anom)
traceplot.anominate(sen111_anom)

anominate documentation built on May 29, 2017, 1:01 p.m.