The Supreme Court in Flux--Always

President George W. Bush may have the opportunity to appoint several justices to the U.S. Supreme Court in the next few years. What are the odds that new appointees would vote as the President anticipates? Not very good, according to UW Professor Kevin Quinn. 

Quinn, working with Andrew Martin of Washington University in St. Louis, has completed a statistical analysis of the voting behavior of Supreme Court justices over the past five decades, to see how individual justices’ voting patterns have changed over time and what factors may have influenced those changes. The project combines statistics and politics—a perfect mix for Quinn, who is an assistant professor of political science and a core member of the Center for Statistics in the Social Sciences. 

Quinn and Martin first tracked the voting behavior of each Supreme Court justice over time, using a fairly simple model of judicial preferences that ranged from very liberal to very conservative. Then they came up with an “ideal point” on that scale for each justice, reflecting how he or she would most likely vote if there were no constraints. “We found that a lot of justices’ ‘ideal points’ have not only moved quite a bit over time, but rather quickly,” says Quinn. “And in some cases, they continue to change.”

Next the researchers explored the extent to which the cases reaching the Supreme Court might influence how liberal or conservative a justice appears to be. “If the cases reaching the court become more liberal, that might lead to the appearance of more conservative behavior on the part of the justices when in fact their policy preferences have not changed at all,” explains Quinn. 

Quinn and Martin tracked the policy content of more than 3,200 cases that reached the Supreme Court since the 1950s and the resulting votes on their merits. They found large fluctuations, with a more liberal agenda in the early 1960s and the late 1970s and a more conservative agenda at other times. 

“The Supreme Court today looks much more conservative than the court in the 1950s if one does not control for the types of issues facing the court,” says Quinn. “Once those are factored in, the current court is shown to be less conservative than in the 1950s.” 

And individual justices? “What is clear is that, in many cases, looking at a justice’s early voting behavior tells you less about his or her voting behavior five or ten years into the future than you might think,” says Quinn.

Quinn and Martin’s research recently earned them the 2001 Harold Gosnell Prize for the best work in political methodology presented at any political science conference during the preceding year. Next on their agenda? A study of how justices’ votes may be influenced by the actions of their colleagues on the bench. 

“Our past research assumes no strategic interdependence among justices,” explains Quinn. “It assumes that they vote their preferences with no concern for the preferences of the other justices. But in fact, their votes may be influenced by anticipation of how other justices will vote. We’re developing a model to study that.” 

Quinn finds Supreme Court research particularly intriguing because, unlike most statistical work related to voting, the focus is on indivi-dual voting patterns rather than large blocs of voters. “We get to read historical pieces about the idiosyncracies of the justices,” he says, “and tie in qualitative information with our quantitative data. It makes it much more interesting.”