data{
  int N;          // number of voters
  int vote[N];    // vote: 0 (Clinton), 1 (Bush)
  int income[N];  // 1-5 income scale
}
parameters{
  real alpha;    // intercept
  real beta;     // income coefficient
}
model{
  for (n in 1:N){
    vote[n] ~ bernoulli_logit(alpha+income[n]*beta);
                // likelihood          
  }
  alpha ~ normal(0, 10);  // intercept prior
  beta ~ normal(0, 2.5);  // income prior
}