Using data from nearly 5,000 votes cast by more than 400 judges in courts of last resort from all 50 states, we investigate whether there is a relationship between a state's judicial retention method and the likelihood that a judge votes to join a precedent-overturning majority. We find that relative to judges retained by institutions such as judicial commissions or state legislatures, those retained via either partisan or retention elections are significantly more likely to join majorities that overturn precedent. Most of this effect is due to behavior in high-profile cases that garner media attention. We find little evidence that an impending election moderates these effects. Finally, we find no evidence that judges retained via nonpartisan elections treat precedent differently than their institutionally retained colleagues.

The study of U.S. state supreme courts has been significantly constrained by a lack of available data. To remedy this deficiency, this article introduces an original data set of every state supreme court ruling from 1995 through 2010. We utilize automated textual analysis to search the text of thousands of state supreme court decisions and extract critical information on each case. This automated coding approach produces reliable measures of state supreme court decision making when compared with data collected by human coders. We present trends in docket size, dissent rates, and legal issues being heard in front of the court. This new data set will offer scholars numerous opportunities to expand our knowledge of judicial politics in the American states.