Simulates roll calls from a spatial probit model (normal vector uniform on the
sphere, cutting point U(-0.5, 0.5), P(yea) = Phi(beta (x.n - c)), Philox
draws) or takes observed votes, and finds each roll call's optimal cutting
line or plane exactly (candidate normals perpendicular to pairs in two
dimensions, through triples in three), reporting errors and APRE.
Usage
OptimalCuttingLines(ideals, votes = NULL, n_votes = 0, beta = 5, seed = 0)
Arguments
- ideals
Ideal points (vector or matrix with 1 to 3 columns).
- votes
Optional 0/1/NA vote matrix (legislators by roll calls).
- n_votes
Number of roll calls to simulate.
- beta
Signal-to-noise of the simulated votes.
- seed
Philox seed.
Value
list(votes, normals, cuts, errors, classification, apre, and
true_normals, true_cuts when simulated).
References
Poole, K. T. (2000). Nonparametric unfolding of binary choice data.
Political Analysis 8, 211-237. Poole, K. T. (2005). Spatial Models of
Parliamentary Voting. Cambridge University Press.
Examples
OptimalCuttingLines(c(-1, -0.5, 0.4, 1), votes = matrix(c(0, 0, 1, 1)))$apre
#> [1] 1
