The Attitudinal Model

I received this request yesterday from Bill Henderson, one of the ELS editors: “Since our audience is not entirely law and courts people familiar with the attitudinalist literature, I was wondering if you would consider a blog entry that summarizes the central findings of your work with Spaeth and the various critiques that have arisen in response.”

Bill: sure.

For those who want the full treatment, the best single source on this is probably The Supreme Court and the Attitudinal Model Revisited (hereafter SCAMR) But for those seeking a quick summary, here goes.

The attitudinal model holds that judges decide cases in light of their sincere ideological values juxtaposed against the factual stimuli presented by the case. Thus, for example, as searches by the police grows more intrusive, first liberal, then moderate, and then conservative judges should become increasingly likely to reject the search.

Numerous behavioral implications devolve from this model. At the case level, a court’s decisions should depend on the factual stimuli in the case. This implication is consistent with, but not unique to, the attitudinal model. At the judge level, differences in judges’ attitudes should influence aggregate levels of their ideological voting. Additionally, at the case-judge level, the votes of particular judges in particular cases should depend on the interaction between the case stimuli and the judges’ attitudes.

While attitudes undoubtedly play a large role throughout the realm of judicial decision making, the likelihood of judges behaving consistently with the attitudinal model will depend on institutional incentives and disincentives for attitudinal behavior. It should be at its highest at the U.S. Supreme Court, which sits at the top of the judicial hierarchy (and thus cannot be overruled by higher courts), which has docket control (and thus can weed out legally frivolous cases that no self-respecting judge could decide only on her ideology), and whose judges have life tenure (and need not fear for job security), and lack ambition for higher office (and thus have little incentive to placate others).

Consistent with the model, I have found that case stimuli readily explains the votes of the Supreme Court in search and seizure cases (1984 American Political Science Review), while others have replicated case-stimuli findings across a variety of different issue areas. As noted, though, such results are consistent with both attitudinal and other models of judicial decisions.

At the judge level, using an exogenous (i.e., independent) measure of the justices’ ideology -- one derived from editorials on the nominees subsequent to their nomination but prior to their confirmation -- I find a strong relationship between their ideology and the liberalism of their aggregate voting behavior once on the Court (r = .78) (updated from Segal and Cover 1989 APSR).

Finally, as for the juxtaposition between the justices’ ideology and case stimuli, the model well predicts the Court’s search and seizure decisions, where the combination of the justices’ attitudes and a series of case stimuli leads to a model that predicts 71% of the justices’ votes correctly.

One set of criticisms of this work came from rational choice theorists, who argue that the justices cannot vote their sincere preferences because they must be concerned about getting overturned by Congress. Pablo Spiller has provided the best systematic empirical support for this position, but there are problems with his analyses (see SCAMR, ch. 8). To be fair, though, I also made some mistakes in my initial foray into this area.

Recent work suggests that the impact of attitudes may not be homogenous across cases. See the recently published article by Isaac Unah and Ange-Marie Hancock, and a sophisticated multi-level modeling manuscript by Brandon Bartels.

Other critics of the original Supreme Court and the Attitudinal Model correctly contended that while we may have demonstrated the important role that attitudes play, our critique of the limited role of law at the Supreme Court was, at best, anecdotal. At the time, Spaeth and I argued that the impact of law could not be tested systematically because “the various modes of legal decision making cannot be operationalized evenhandedly.” I think we spoke too soon.

In an attempt to take our critics seriously, Spaeth and I examined the influence of precedent on Supreme Court decision making, using as our starting point the notion by Judge Jerome Frank that "Stare decisis has no bite when it means merely that a court adheres to a precedent that it consid­ers correct. It is significant only when a court feels con­strained to stick to a former ruling although the court has come to regard it as unwise or unjust" (United States v. Shaughnessy 1955, p. 719). Thus, when Justice Stewart reversed himself and accepted the right to privacy in Eisenstadt v. Baird (1972), we have a prima facie case that the justice was influenced by the precedent established in Griswold. Tests using this standard find that stare decisis influences the decisions of Supreme Court justices only about 10% of the time (though more than that, as predicted, in the least salient of the Court’s cases).  We did not expect to see a mechanical or deterministic response to such precedents, but rather, some meaningful level of gravitational movement toward the position established in the precedent, and that we did not find. The article version of this project (Segal and Spaeth 1996 American Journal of Political Science) resulted in a forum on the issue that attracted interesting responses by Lee Epstein, Stefanie Lindquist and several others. The book drew an extended and typically thoughtful response from Howard Gillman in Law and Social Inquiry (Spring 2001). Howard proposed a postpositive meaning of law that requires only the subjective belief of the judge that s/he is acting in good faith. Whether such a requirement is falsifiable or not remains to be seen. But even if we could use ethnographic methods to truly know the mindset of the judge, the universal human characteristic of motivated reasoning, that is, the ability to convince ourselves of that which we wish to believe (in addition to Eileen Braman’s work on this in the 2006 Journal of Politics, Milt Lodge and Chuck Taber have an article on this in the July 2006 American Journal of Political Science), makes this far too low a hurdle for virtually anyone to get tripped up by.

Comments

[William Henderson]

Many thanks for this exchange. I read Spaeth and Segal as a 3L in my law and social science class (the same semester, in fact, as my Con Law Equal Protection / Sub Due Process class). The clouds literally parted; everything in Con Law made a lot more sense.

It is follow-up work like my Indiana colleague Eileen Braman that, I think, has the opportunity to eventually find its way into law school class rooms.

Finally, as a former federal appellate law clerk, I observed both jurisprudential and and attitudinal tendencies, though much more of the former than the later. Most cases have a statute or case on point and don't tap into intense preferences. Just reading the briefs and writing the bench memo, most competent clerks can predict on the outcome of 80 to 90 percent of the cases. In my experience (a sample of one), only about 10 percent of the *published* cases have sufficient legal uncertainty to invite judical lawmaking (though some jurists, I admit, are more activist than others).

My former professor said it best: former Circuit Court clerks tend to believe in jurisprudence while former SCOTUS clerks inevitably inch closer to a legal realist position.

[Eileen Braman]

Sean, knowing how to blog is one thing -- knowing how to post links is another altogether. The Journal of Politics piece Jeff mentioned (May 2006) is the one on threshold decision making. The analogy paper (coauthored with Tomas E. Nelson) is currently 'under review' and not available on line. I'd be happy to send you a copy of the latest iteration.

[Sean Wilson]

Eileen:

.. could you post a link to your research? I'd like to put it on my website on the "new research" page. (If you do not want to post it here, perhaps you could email it to me).

[Eileen Braman]

Thanks Jeff for mentioning the JOP piece.

I'd just like to mention for any legal academics who may have been told the model is strictly "at odds" with legal conceptions of decision making it is not necessarily so -- although that IS a big misconception.

My research has actually evolved via an effort to reconcile legal and attitudinal accounts of decision making. I'm basically interested in how legal decision makers reach conclusions consistent with their preference in the context of accepted decisional norms. Specifically, I've done this in a series of experiments looking at how attitudes influence analogical perception (how lay and legally trained decision makers perceive precedents related to pending litigation) and experiments looking at "threshold" decisions in complex (multidimensional) cases.

In each of my studies there is evidence of attitudinal influence AND legal constraint -- but quite honestly more attitudinal influence than I was expecting as a former law student and practitioner all too familiar with the legal socialization process.

I guess my point is there is great potential to do work at the intersection of these perspectives. In my next project I hope to test the conventional wisdom about the differential constraining influence of rules vs. standards on legal discretion where decision makers are asked to make similar judgments under different decision rules.

I wish I had heard more about this stuff as a law student -- then again maybe its better I didn't (I have tons of research questions now). Still I think legal academics who do not know about this model or dismiss it summarily are doing themselves (and perhaps ther students) a significant disservice. I can't believe there are many of those types reading this blog but I hope the number of scholars threatened by such research are diminishing.

[Sean Wilson]

... at least that's the version they tell to the graduate students. Although it is good to see Jeff (and Harold?) admit that they were working within a deficient philosophic paradigm when they constructed their models of judging -- they are apparently no longer doubting a Dworkin-effect that causes "law" itself to structure judicial attitudes -- it is nonetheless difficult to see certain empirical falsehoods continue to exist.

I really don't know how political science will ever become relevant if we treat models as we do rhetoric. What I mean by that is the bivariate model that Jeff refers to (the one with the high R-squared) loses about two-thirds of its explanatory value if you don't put the voting data into percentages before conducting regression analysis. As I have quite clearly shown, aggregation is a modeling technique that causes exaggerated and unreliable estimates of goodness of fit. Moreover, there is no sound empirical modeling protocol that suggests one should IGNORE individual level data in favor of a regression of percentages. (Ecological regression is performed only when you do not have individual-level data or when multi-level analysis accompanies it).

To see the verification of this, visit my website:

http://ludwig.squarespace.com/latest-works/

The truth is that Jeff and Harold never had a model of attitudes that explained anything other than about 24% of the civil liberties votes. Still a good model, of course, but not one that could support the large language that came with the packaging. Indeed, one wonders whether Jeff's model isn't more compatible with a Dworkin-like view of judging, and whether it was simply stuffed into a hard-core realists (or "critical") framework because of the motivated reasoning of the scientists.

Also, keep in mind that Jeff's model only explains about 9% of the votes on today's Court because the extreme justices that used to anchor Jeff's model -- Rehnquist and the liberal members of the Warren Court -- are now gone.

I don't have time to get into the search and seizure model that doesn't understand the difference between "facts" and "doctrine." But take a look at Barry Friedman's article; he understands the point.