Herbert Kritzer has posted this draft on SSRN.  Here is the abstract;

Judicial elections, particularly elections to state supreme courts, have generated substantial commentary over the last decade. Observers have asserted that elections have become more hotly contested, nastier, more expensive, and more divisive. This paper focuses on the question of whether elections are increasingly partisan. Using a data set of all statewide partisan and nonpartisan elections since 1946 plus a subset of retention elections from that same period, patterns of partisanship in voting are examined. The results of the analysis show that in some circumstances voting patterns in state supreme court elections have become more partisan. Elections that are explicitly partisan have not changed. The change has occurred in a subset of the states employing nonpartisan elections. The states that have experienced this change are those where the state supreme court has been called upon to decide questions where the issues align with partisan divisions (death penalty, same-sex marriage, abortion, property rights, tort reform). Where strong partisan patterns have occurred in retention elections these patterns typically arise in the wake of a controversial decision dealing with one of the partisan-aligned issues. The paper concludes by observing that if courts deal with issues that activate partisan divisions, it is not surprising that voters, given the opportunity, will respond in a way that reflects those same partisan divisions. Importantly, the retention elections associated with the so-called “merit” system do not escape this pattern when the justices of the courts who must stand for retention decide divisive issues.