Over time, American political scientists have conceptualized the work of the judicial branch in several distinct ways, producing a rich literature on judicial decision making. A disproportionate share of that literature focuses on the U.S. Supreme Court. This focus is unsurprising given the Court’s distinct role in the American legal system. The Supreme Court is the court of last resort, and it is the unique province of the Court “to say what the law is” (Marbury v. Madison, 1803). The Supreme Court is also distinct from the other branches of national government. It is the only branch whose members are unelected, its members are the least publicly visible, and their work the most obscure. Perhaps in part for these reasons, the Court enjoys much broader support among the American public than its sister branches. Meanwhile, students of the Court have often asked “What drives Supreme Court decision making?” In this post, we’ll take a look at the traditional legal formalist understanding of Supreme Court decision making and then discuss the attitudinal model which has supplanted it. In a follow-up post I will survey some alternatives to the attitudinal model which purport to offer a more complete picture.
The Good Ole Legal Model
The traditional way of thinking about judicial decision making, legal formalism (or “legalism”), can trace its lineage back to such notable figures as Sir William Blackstone. Legal formalists emphasize the role of legal considerations—adherence to precedent, strict interpretation of relevant legal texts, and the law as a closed logical system—in shaping judicial reasoning. By the early twentieth century, “realists” began pointing out the often indeterminate and nonbinding nature of legal considerations—e.g., the availability of precedents to support either outcome in an adversarial dispute and the hopelessly vague or ambiguous nature of many statutes—and alleged that judicial decision making was more about who was deciding the case than any objective body of law from which outcomes could be logically deduced. These criticisms have the most force when it comes to the Supreme Court because one might expect cases for which there is a straightforward legal answer to be resolved at a lower level if litigated at all. In fact, while scholars have come to agree that ideology plays a much smaller role in the lower federal courts (Zorn & Bowie 2010), and these courts are generally “faithful agents” of Supreme Court doctrine (Haire, Songer, & Lindquist 2003, 154), the Supreme Court itself seems to be another story.
Because of the indeterminate nature of legal considerations at the level of the Supreme Court, a great deal hinges on the motivations of each justice. By the mid-twentieth century, it was to this study that political science research and some empirical legal research soon turned. Broadly, the shift toward this type of inquiry was dubbed “behavioralism”. Behavioralist “models emphasize the role of judges’ social background or personal attributes on judicial decisions” (Heise 2002, 833). The early behavioralists looked for patterns related to the judges’ socio-economic background, race, sex, and prior work experience. However, each of these factors was found to have very limited explanatory power:
Consequently, Dan Bowen’s pronouncement thirty years ago about the efficacy of behavioralism judicial decision-making models still holds force today: “A final inescapable conclusion about the explanatory power of the sociological background characteristics of [judges] is that they are generally not very helpful” (Heise 2002, 835).
In short, studies in this area found little relationship between social background characteristics and judicial decisions (cf. Segal & Spaeth 1993, 231-234). Scholars next began exploring whether a judge’s political ideology could be used to predict judicial decisions, and the attitudinal model was born.
Along Came the Attitudinal Model
Following the realist critique of legal formalism, judicial scholars Jeffrey Segal and Harold Spaeth published The Supreme Court and the Attitudinal Model in 1993, a culmination of work begun decades earlier by C. Herman Pritchett’s The Roosevelt Court (1948), Glendon Schubert’s The Judicial Mind (1976), and crystalized by Spaeth in the intervening years. Although it was Schubert who “first provided a detailed attitudinal model” (Segal & Spaeth 1993, 67), Spaeth’s greatest contribution to the field came in the creation of the U.S. Supreme Court Judicial Data Base in 1990. This data base constituted a major breakthrough in the ability of scholars to reliably and comparably test hypotheses about Supreme decision making. According to Djupe and Epstein:
Prior to the Spaeth Data Base, judicial specialist relied on several publicly-available data sets…or collected their own data. Either way, issues of verification, reliability, and the like went virtually ignored; neither Schmidhauser’s nor Schubert’s, nor Ulmer’s documentation contain any statements about reliability or verification; of the ten articles on judicial decision making in the American Journal of Political Science in the decade prior to the appearance of the Spaeth Data Base (1980-90), only two were attentive to matters of replication or reliability (Epstein, Walker, and Dixon 1989; Gates 1987)—and, in all likelihood, neither would pass muster under contemporary standards (Djupe & Epstein 1998, 1012-13).
The Supreme Court and the Attitudinal Model was indeed the culmination of a tremendously successful research program establishing the prima facie case for the validity of the attitudinal model. However, the work did come under increasing criticism for its lack of a direct test of the influence of legal factors. An updated work, The Supreme Court and the Attitudinal Model Revisited (2002) was Segal and Spaeth’s answer to this criticism. The discussion that follows focuses on key features of the later work.
The model they present is premised upon a simple assertion: “the Supreme Court decides disputes in light of the facts of the case vis-à-vis the ideological attitudes and values of the justices” (Segal & Spaeth 2002, 86), rather than the legal considerations with which most legal academics concern themselves. In their work, Segal and Spaeth attempt first to discredit the legal model and then to establish their own attitudinal model as the best explanation of Supreme Court decision making. We’ll explore some of the technical details of their work and consider some limitations.
Though legalists vary somewhat, argue Segal and Spaeth, they are united by a common belief that law matters in judicial decision making. While legalists scour Supreme Court opinions for insight into the justices’ reasoning, Segal and Spaeth, in behavioralist fashion, insist that the only relevant factor is what the justices do in fact. All of the legal reasoning, cites to precedent, and quibbling over statutory text obscures the fact that justices are merely voting their ideological preferences. In furtherance of this claim, Segal and Spaeth devise a test to determine whether precedent can be observed to exert any “gravitational force” on justices. They devise a test examining the progeny of non-unanimous precedent-setting cases to determine whether justices who dissented in the original case conformed to the precedent in progeny cases despite their previously demonstrated preferences to the contrary. The authors find that “the justices are rarely influenced by stare decisis” and “rather easily avoid supporting precedents with which they disagree” (Id. at 298, 310).
Dismissing many other legalist arguments as non-falsifiable and therefore unworthy of full consideration, Segal and Spaeth then proceed to present the attitudinal model as an alternative explanation of Supreme Court decision making. To test their model, the authors “examine all Supreme Court decisions dealing with the reasonableness of a search or seizure from the beginning of the 1962 term through the end of the 1998 term (N = 217)” (Id. at 316). The measure the effect of several factors on “the decision of the Supreme Court whether or not to exclude evidence or find a search unreasonable. A liberal decision is one that prohibits the use of questionably obtained evidence; a conservative decision is one that admits such evidence” (id.). In measuring attitudes, the authors rely on “newspaper editorials that characterize nominees prior to confirmation as liberal or conservative insofar as civil rights and liberties are concerned” (Id. at 321).
From a predictability standpoint, the authors’ model performs surprisingly well. Specifying the dependent variable as each justice’s vote in the search and seizure cases (N = 1,900), Segal and Spaeth find that attitudes alone “achieved a 70 percent prediction rate, for a 32 percent reduction in error over the justices’ mean of 56 percent” while the facts alone produced a 62 percent prediction rate and 14 percent reduction in error (Id. at 324). “This suggests,” say the authors, “that in predicting votes, one is clearly better off knowing the attitudes of the justices than the facts of the case” (Id. at 325).
While the rate of prediction is significant and lends substantial support to Segal and Spaeth’s arguments, many scholars consider it premature to declare the victory of the attitudinal model. One obvious drawback to this approach is its quite narrow application heretofore. Segal and Spaeth only test their model on a very limited range of cases: non-unanimous Supreme Court civil liberties cases on the merits. All other legal subject areas and cases decided on procedural grounds are necessarily excluded. Unanimous decisions, roughly one-third of all Supreme Court decisions (Epstein, Landes, & Posner 2012), are also excluded. Even on their strongest ground, predictability, the authors must demonstrate that their model can be applied more generally before many other scholars will concede.
Moreover, predictability is not the only test of a model’s worth. A good social science theory not only predicts but explains the phenomenon in question. Several scholars charge that the attitudinal model is overly dismissive of the immense legal-institutional environment in which the justices operate (Black & Owens 2009). Any account of what drives Supreme Court decision making which does not deal seriously with justices as individuals operating within a diverse institutional network is fatally incomplete. “Thus, theories of judicial behavior must become more complex if they are to achieve a higher level of explanation and prediction” (Gibson 1983, 7). To be sure, most alternative theories of Supreme Court decision making take for granted that attitudes play a significant role, but go on to specify ways in which the relevant legal institutions matter (See Gillman 1997a). In my next post I will survey some of those theories.