Tom S. Clark (Emory University Department of Political Science) and Benjamin Lauderdale (Ph.D. candidate, Princeton University Department of Politics) have posted “Locating Supreme Court Opinions in Doctrine Space” on SSRN, see here. As an initial note, this paper is very dense and technical as it is clearly intended to be published in a political science journal. But I am going to write about it anyway because it could turn out to be an important paper in the field of judicial politics. The authors start with the (in my opinion, correct) claim that most models of Supreme Court decision-making and opinion-writing are focused on the actual outcome of a case (affirm, reverse, etc.), and then that outcome is labeled as a conservative or liberal result. As the authors argue, this binary categorization largely ignores the substantive content (i.e., the reasoning) of Supreme Court opinions, which is one of the most important reasons for the Court to write a reasoned opinion. The superficial labeling of opinions as conservative or liberal can undermine models finding that the author of the opinion, median Justice, or median member have the most influence over the content of Supreme Court opinions. I find this paper significant because it continues to bridge the gap between legal and political science scholars, as the former ordinarily places greater emphasis on opinion content and the latter on judicial outcomes (at least up until now). However, there are potential weaknesses to this paper, as the authors of the piece properly note. The authors’ model, for example, places opinions along a doctrinal space or ideological continuum based on the precedents cited in an opinion. The authors eliminate purely procedural citations, but the model does not take into account that “some opinions may be narrower or broader than others in their application of precedent,” or that the “citation data generating process may not be entirely driven by doctrinal concerns,” both of which I find to be pretty significant limitations of their model. Ultimately, the authors conclude that the data supports the theory that the median of the opinion coalition, rather than the median Justice or opinion author, is most influential in determining the estimated location of the majority opinion. As I said, this is a difficult but potentially important paper as a theoretical matter.
Continuing the theme of political science research in this round-up, Anna J. Harvey (New York University Department of Politics) and Michael J. Woodruff (New York University Department of Politics) have posted “Confirmation Bias in the United States Supreme Court Judicial Database” on SSRN, see here. Many political science models of Supreme Court decisionmaking rely on the extensive database compiled by Harold Spaeth, see here, which has been updated a number of times by other scholars and Harold since its public release in the late 1980s, see here. Each case in the database contains 247 pieces of information, roughly grouped into six categories, see here. Professors Harvey and Woodruff take issue with the data coding of outcome variables, arguing that there is a bias in the coding based on the expectations of the coder–that is, a decision is more likely to be labeled as liberal if it was issued by a liberal Court (e.g., the Warren Court). In other words, the authors argue that “that cases are disproportionately assigned issue codes that tend to lead to judgment codes confirmatory of expectations about the ideological character of the judgments typically issued by the deciding Court.” The authors then connect the confirmation bias to one area where they believe it matters: the effect of congressional preferences on the Court’s judgments. There are obviously some serious weaknesses to the Spaeth database, as these authors and Carolyn Shapiro (Chicago-Kent Law School) have noted. Nonetheless, so long as those weaknesses are taken into account, the Spaeth database is an enormously useful and important source of data about the Supreme Court. I, along with numerous other scholars, use it often.