Roll-call analysis: cohesion and agreement, Snyder-Groseclose party influence, logistic ideal points

RollcallCohesion: per party and vote the Rice index |Y - N| / (Y + N) and the Hix-Noury-Roland agreement index, per legislator the participation rate, and pairwise agreement. PartyInfluence: ideal points from the first principal component of lopsided votes, then per close vote the linear probability model y = a + b x + g party. LogitIdealPoints: penalised joint ML of logistic(alpha_j + beta_j x_i) by alternating Newton steps with mean(x) = 0, sd(x) = 1. Votes are 1, 0 or NA.

Usage

RollcallCohesion(votes, party)

PartyInfluence(votes, party_dummy, lopsided = 0.65)

LogitIdealPoints(votes, prior_sd = 5, max_iter = 500, tol = 1e-10)

Arguments

votes: Legislators by votes matrix (1, 0, NA).
party: Party labels.
party_dummy: 0/1 party indicator.
lopsided: Winning-side share defining a lopsided vote.
prior_sd: Prior standard deviation of bill parameters.
max_iter, tol: Iteration limit and tolerance.

Value

A list.

References

Rice, S. A. (1925). Political Science Quarterly 40, 60-72. Hix, S., Noury, A. and Roland, G. (2005). BJPS 35, 209-234. Snyder, J. M. and Groseclose, T. (2000). AJPS 44, 193-211. Clinton, J., Jackman, S. and Rivers, D. (2004). APSR 98, 355-370.

Examples

RollcallCohesion(rbind(c(1, 1), c(1, 0), c(0, 0)), c("a", "a", "b"))$rice
#> $a
#> [1] 1 0
#> 
#> $b
#> [1] 1 1
#>