Estimate an ordered probit model.

ordered-probit-model.R

# replication of Hill, Moore, and Mukherjee (2013)

# load packages
library(arm)

# load data
d <- read.csv("http://crain.co/am-files/data/hmm.csv")

# examine data
names(d)
xtabs(~ d$tort)

# make tort and *ordered factor*
d$tort <- as.ordered(d$tort)

# create variable
d$logHROsec <- log(1 + d$HROsecretariatlocation)

# estimate the ordered probit
m <- polr(tort ~ logHROsec + WoverS + cat_rat + allterror + lngdp + lnpop, 
          data = d, method = "probit")
display(m)