The Statistical Analysis of Roll Call

Wedevelop aBayesian estimation procedure for spatialmodels of roll call voting. We show how a Bayesian approach to roll call analysis overcomes shortcomings and idiosyncracies of NOMINATE (some of which are not widely recognized). Our Bayesian approach (a) applies to any legislative setting, irrespective of size, legislative extremism, or the number of roll calls available for analysis; (b) provides a mechanism for directly incorporating auxiliary information as to the dimensionality of the underlying policy space, the identity of extremist legislators, key votes and the evolution of the legislative agenda; (c) lets us integrate measurement of legislative preferences with the analysis of those preferences