The premise of this paper is that while the comparative study of courts can address some vitally important questions in judicial politics, these gains will not be secured without a valid and reliable measure of judge preferences that is comparable within and across courts. Party affiliation of judges is a commonly used but weak substitute that suffers from pronounced equivalence problems. We develop a contextually based, party-adjusted surrogate judge ideology measure (PAJID) and subject this measure to an extensive array of validity tests. We also consider the measure's stability in predicting judge behavior over the course of the judicial career. As the results illustrate, PAJID offers a valid, stable measure of judge preferences in state supreme courts that is demonstrably superior to party affiliation in analyses of judicial decision-making across areas of law and across 52 state high courts.