It is all too rare when books are more than the sum of their individual articles. Fortunately for scholars of judicial politics, both The Politics of Precedent on the U.S. Supreme Court and Judging on a Collegial Court are exactly that. Despite the fact that earlier incarnations of several chapters previously appeared in print, both books significantly improve upon the earlier analyses and make independent and important contributions to the literature on courts. In fact, some of the results presented in Precedent contradict previous research published by the authors (see, for example, p. 123).

Both Precedent and Judging are examples of theoretically based empirical work that moves the literature beyond the attitudinal model. While both books build off the attitudinalist tradition, they both find that other factors matter as well. In Precedent, Thomas Hansford and James Spriggs find that the law itself matters. Justices cannot simply vote their policy preferences; the need to provide legally sound justifications for their decisions means that they are constrained by precedent (though not all precedents constrain equally). In Judging, Virginia Hettinger, Stefanie Lindquist, and Wendy Martinek examine decision making on the U.S. courts of appeals. They find that in addition to attitudinal factors, the unique institutional position of the courts of appeals affects behavior in important ways. Both these books are well grounded in the literature and develop the literature in significant ways as well.

The central question motivating the Hansford and Spriggs book is simple, yet understudied: “when and why the [Supreme] Court develops law through the interpretation of its precedent” (p. 2). Throughout the book they distinguish between two types of precedent interpretation: positive and negative. A “positive” interpretation of precedent occurs when the Court favorably uses a past case in a current case, while a “negative” interpretation occurs when the Court limits, distinguishes, or overrules a past case in a current case. The authors illustrate these different types of interpretations with examples from the Court's history. Theoretically, they argue that the interpretation of precedent is best illustrated by the interaction of both legal and attitudinal factors. That is, precedent is not only a constraint on the justices but also an opportunity for them to influence legal policy. They find that “justices do not change law simply based on their policy preferences or on the existing state of precedent; they do so based on an interactive relationship between these two factors” (p. 130).

There is much to like in this book. The empirical analysis is carefully executed and accessible even to those without training in higher-level statistics. The authors make use of extensive footnotes to be explicit about every coding decision they make. In sum, the analysis is an example of science at its best. Some readers may quibble with how certain concepts are operationalized, but Hansford and Spriggs do an excellent job of justifying their choices, and all the choices made are plausible and theoretically defensible.

However, I do have a minor quibble with the design in Chapter 6. In this chapter, the authors seek to answer the question “When deciding a particular case … how does the Court choose to interpret existing precedents?” (p. 93). In selecting their cases to address this question, they treat the universe of past cases as available precedent for a given case. So, for a case decided on January 22, 1996, “the set of precedents for this … case includes all precedents decided from the start of the 1946 term up to and including January 22, 1996” (p. 96). The authors justify this choice because alternative measures are underinclusive in the range of 10%–27% (pp. 95–96). While this may be so, it seems to me that there must be a better alternative between existing underinclusive measures and inclusion of all the cases (which is certainly overinclusive of relevant precedents). Not surprisingly, variables measuring whether the case was in the same issue area (either broadly or narrowly defined) are statistically significant in the empirical analysis (p. 100). This overinclusiveness also may explain the somewhat divergent results of this chapter, compared with Chapters 4 and 5 (p. 107). The authors are aware of this and argue that it is better to be overinclusive than underinclusive (p. 107), which is undoubtedly true. I just wonder if maybe that is a false choice.

Unlike Precedent,Judging focuses on the U.S. courts of appeals, for years an understudied and largely forgotten institution, though there have been several excellent studies done in recent years. Using data from 1960 to 1996, Hettinger, Lindquist, and Martinek focus on two types of dissensus: horizontal and vertical. When judges on a panel disagree with one another and write either a concurring or dissenting opinion, this is horizontal dissensus; when the courts of appeals reverse a district court decision, this is vertical dissensus (p. 5). While the authors admit that both forms of dissensus are relatively rare (p. 4), they argue that it is important to understand the dynamics of judicial decision making on the courts of appeals because these are effectively the courts of last resort for the vast majority of cases.

Given the fact that courts of appeals “serve as the de facto (if not the de jure) court of last resort in the overwhelming majority of appeals” (p. 13), one wonders why dissent is so rare in these cases, compared to dissent on the U.S. Supreme Court. That is, since the courts of appeals handle 600 cases for every one disposed of by the U.S. Supreme Court (p. 13), why is there not more dissensus on the courts of appeals? What makes them different from the Supreme Court? The authors do a good job of explaining the determinants of dissensus on these courts, but leave open the question of why it is so rare compared to both the Supreme Court and other “effective” courts of last resort, such as state supreme courts. One possible explanation is that many of the cases the courts of appeals handle are routine. However, the authors limit their analysis to published opinions that are “generally issued in the more important policymaking cases” (p. 133, n. 23). Given that the courts of appeals are similar to the Supreme Court on many institutional features—term, method of appointment, the fact that this is the last job for most judges (p. 23)—the reader is left wondering about the relatively low rate of dissensus. Some discussion of this, especially grounded in the literature, would have strengthened the book.

Methodologically, Judging is well executed, and the results are carefully explained so that readers with little or no statistical training will not have problems understanding the analysis. There are two issues regarding the development of the models, though, that readers might be curious about. The first is the fact that the authors limit their analysis to three-judge panels and omit en banc decisions. It seems to me that (at least in Chapters 3 and 5) the authors could have easily included such decisions in their analysis (with a variable denoting such) or conducted separate analyses only on en banc cases. Is dissensus higher in these cases? One would think so, given that en banc cases tend to involve highly salient issues and that circuits tend to be more ideologically heterogeneous than panels. Are the determinants the same? Do they vary by circuit? Questions like these would have fit in nicely with the scope of the book.

Second, and perhaps more significantly, I think the analyses in Chapters 3 and 5 would have benefited from formalization. For example, on page 42, the authors hypothesize that appellate court judges are likely to reverse trial court decisions that deviate from the judges' ideology. On page 98, they find that “[w]hile ideological disagreement influenced an individual circuit judge's decision to dissent, no such relationship emerges in our analysis of the panel decision to reverse the district court judge.” This may be true, but it seems to me that the theory (and thus the model specification) is incomplete without including the Supreme Court (and perhaps the whole circuit). For example, let's say the trial court makes a conservative decision upholding an illegal search. The appeals court is liberal and would like to overturn the trial court. The authors would predict a higher likelihood of reversal here. However, suppose the Supreme Court is conservative and is likely to side with the trial court if they decide the case. How would the appeals court behave now? It seems to me that they are not more likely to reverse given that they are likely to be reversed by the Supreme Court (and presumably, these judges have the same aversion to reversal as lower court judges). What if the Supreme Court were liberal? Now, the appeals court is likely to reverse the trial court. Basically, it seems to me that the courts of appeals would behave differently, depending upon the preferences of the Supreme Court. (The same analysis can also be carried out using the full circuit, since the case could be heard en banc after the panel makes its decision.) Failure to take into account the fact that the courts of appeals are able to be reversed may account for some of the anomalous results. That being said, the analysis in Judging still represents a major step forward in our understanding of judicial behavior on courts other than the U.S. Supreme Court.

In sum, both Precedent and Judging will be of interest to scholars of the judiciary, and they will be cited regularly and routinely assigned in graduate classes in judicial politics. In many ways, they represent the best of the scientific enterprise: Scientific research is a cumulative enterprise, and just as these books built off prior studies, future studies will build off them.