model {
    for (i in 1:N) {
        mu[i] <- theta[1]*education[i] + theta[2]*sex[i] + theta[3]*seedebates[i]
            + theta[4]*importance[i] + theta[5]*involvement[i] + theta[6]*importance[i]*involvement[i] 
            + theta[7]*catholic[i] + theta[8]*partyid[i]
       ### LIKE R: logit P(Y <= k|x) = zeta_k - X*beta ###
       for (j in 1:(Ncat-1)) { logit(Q[i,j]) <- cut[j] - mu[i] } 
       p[i,1] <- Q[i,1]
       for (j in 2:(Ncat-1)) { p[i,j] <- Q[i,j] - Q[i,(j-1)] }
       p[i,Ncat] <- 1 - Q[i,(Ncat-1)]
       close[i] ~ dcat(p[i,1:Ncat])
       E.y[i] <- close[i] - mu[i]
    }
    sd.y <- sd(E.y[])
    for (k in 1:Nvar) { theta[k] ~  dt(0,1,5) }
    for (k in 1:(Ncat-1)) {  cut0[k] ~  dt(0,1,5) }
    cut[1:(Ncat-1)] <- sort(cut0)
}

# "typeofplace"GV "state"GV       "married"EV  "education"OV "familyincome"EV "seedebates"EV
# "efficacy"OV    "involvement"OV "predict"OV  "close"X      "importance"X    "partyid"X     
# "willvote"OV    "sex"EV         "race"EV     "age"EV       "lifecycle"EV    "socialclass"EV 
# "wasmilitary"EV "parentsborn"EV "catholic"EV "church"GV