The importance of empirical analysis (with forecasts of Bucklew & Madison)

The two Eighth Amendment cases show hard cases make for hard predictions

Bad facts make bad law. So too, hard cases make for hard predictions. The Court’s two current death penalty cases, illustrate both phenomena. But these cases also illustrate the added value of empirical analysis of oral argument over purely qualitative or impressionistic readings.

The Supreme Court’s two current pending death penalty cases both have very peculiar facts. In Bucklew v. Precythe, argued last week, the prisoner argues that execution by lethal injection would be cruel and unusual given his particular medical history. Madison v. Alabama, argued on October 2nd, presents the question of whether a prisoner can be executed for a crime he cannot recall.

The Supreme Court had the option of taking a case raising a question with potentially much broader significance for Eighth Amendment jurisprudence. Hidalgo v. Arizona challenged whether Arizona’s death penalty legislation insufficiently narrowed application of the ultimate punishment. True to form, the Supreme Court avoided that important question of the frequency of the ultimate punishment, and focused instead on two cases raising as applied challenges, with limited applications for the 2,743 people currently on death row.

Forecast of Bucklew v. Precythe

Argument in Bucklew v. Precythe involved delving into the grisly facts of Petitioner’s medical condition and the painful possibilities involved in his execution. Along the way, the argument saw some unusual judicial behavior. Immediately, Justice Sotomayor began grilling Petitioner’s advocate for the lack of a clear record on those facts, and whether the horrific nature of his death could simply be avoided by the trach already installed in his throat.

ScotusOA prediction based on the oral argument in Bucklew v Precythe

Commentators on the First Mondays podcast suggested that this interplay may indicate that Petitioner’s advocate may have lost Justice Sotomayor, which would make it “tough” to win, given her general reliable pro-defendant vote. This much-qualified prediction misunderstands what is going on. Although Sotomayor was very active in questioning the Petitioner, and that level of activity often signals disagreement, she was even more active during Respondent’s time. Sotomayor spoke 17 times to Petitioner and 18 times to Respondent; but when she spoke to Respondent, she had much more to say—921 words versus 497 words. In spite of the opening salvo, by the end of the argument Sotomayor registered a significant disagreement gap in favor of Petitioner. Thus, even though she described herself as “upset” with Petitioner’s advocate, our analysis, represented in the figure below, predicts she is 90% likely to vote for Petitioner, and is the Justice most likely to do so.

Sotomayor normally reserves her toughest questions for the prosecution side in criminal cases—perhaps reflecting, as well as her liberal ideology, that as a former prosecutor herself she has particularly high standards for the profession. That fact made her sharp critique of the capital defendant seem more important than it really was. Our empirical approach to reviewing oral argument helps put episodes like this in a more balanced context and avoid salience bias (a recognized common behavioral irrationality that causes people to focus on prominent information and ignore potentially more significant but less noticeable contra-indicators).

A far clearer signal of a Justice breaking with expectations was Justice Kavanaugh. He spoke only to Respondent’s advocate and asked tough questions as to whether there was any limit on the potentially “gruesome and brutal pain” the state is permitted to impose in executing Bucklew, and demanded a yes or no answer from the state’s advocate.

In contrast, the numbers on Justice Gorsuch are likely misleading. He spoke only twice to Respondent’s advocate and once to Petitioner’s advocate, making his signals very weak. Equally importantly, both questions he put to Respondent’s advocate were open-ended questions asking him to get to a point he had said he would make—that is, a seemingly friendly inquiry. Given Gorsuch voted to deny the stay of execution, we expect he will join the other conservatives in voting with Respondent. (Although Justice Thomas was his normal reticent self, he too had joined the dissent from the stay of execution).

With Justice Ginsburg unusually silent (even before breaking her ribs), we can only go on her prior voting record, which is generally pro-defendant. All of this would lead to a prediction of a 5:4 vote for Petitioner, however the prediction depends on the untested vote of Justice Kavanaugh—a tough prediction, indeed.

Forecast of Madison v. Alabama

Oral argument in Madison v. Alabama, on the question of whether someone who cannot remember the crime he committed due to multiple strokes, also involved descriptions of peculiar facts. According to Petitioner’s advocate, inmate Madison regularly soils himself because he cannot remember that he has a toilet in his cell. The law on the issue was as messy as the facts. The following was a typical confused interaction:

Sonia Sotomayor: Mr. Stevenson, part of the problem is the use of the word “loss of memory.” And I — in your briefs, you seem to go back and forth on this. Are you conceding that amnesia about the incident alone, where you can function in every other way in society, would you be incompetent then?

Bryan A. Stevenson: No.

Sonia Sotomayor: To be executed?

Bryan A. Stevenson: Yes, that’s right.

On another point Justice Alito complained: “No, I don’t understand — I don’t understand your answer.”

And Chief Justice Roberts questioned whether there was an issue at all in the case, saying to Petitioner’s advocate: “There are two questions. You concede on one, and the state concedes on the other.” Petitioner had conceded that simply not remembering the crime is not enough to avoid execution, and the state had admitted that if the person is incompetent, they cannot be executed.

The figure below shows our predictions for the case.

ScotusOA prediction based on the oral argument in Madison v Alabama

Thomas was silent as always and Gorsuch was silent in this case also. Furthermore, both Breyer and Roberts presented ambiguous signals. Again, Thomas, Alito, and Gorsuch had opposed the imposition of the stay, suggesting again support for Respondent. This time, however, Roberts had not joined that order. His far more mixed signal in this argument supports Amy Howe’s prediction that as Chief Justice, he “might have a greater incentive than his colleagues to avoid deadlocking on Madison’s case” (Kavanaugh was not yet on the Court when the case was heard) and so could provide the fifth vote for a very narrow victory for Madison by remanding the case back to state court to consider the specific question of whether Madison is incompetent because of his dementia.

We think this analysis makes sense of Roberts’ tentative signal. But we would go further and say that Roberts will only vote in favor of Petitioner for a very narrow ruling, otherwise he will vote for Respondent. We believe Roberts’ ambiguous signal indicates his concern over the potential for this second very specific medical circumstance to create an enormous slippery slope to a flood of future death penalty challenges. Whereas Bucklew’s case is limited to an n of approximately one, making it hard to justify the Court using one of its approximate 75 spots to effectively act as a court of last instance, in contrast, Madison’s case could drastically change death penalty jurisprudence. With death penalty appeals dragging the ordinary execution process out over decades, the chances of other inmates developing memory-related medical issues are very high. A broader ruling for Petitioner could spawn a tide of challenges that would make Atkins IQ challenges seem a narrow set. As such, we expect Roberts to either support a very narrow ruling for Petitioner or rule in favor of Respondent. The empirics indicate he is right on the borderline.

Thus, we predict a 5:3 win for Petitioner on narrow grounds, or a 4:4 default win for Respondent.

 

Prediction for Bucklew: 5:4 vote for Petitioner (Bucklew)

For Petitioner: Sotomayor, Kagan, Breyer, Kavanaugh, Ginsburg

For Respondent: Roberts, Alito, Thomas, Gorsuch

Most likely the switch: Kavanaugh

 

Prediction for Madison:

5:3 vote for Petitioner (Madison) and remand

or

4:4 vote for Respondent (Alabama)

For Petitioner: Sotomayor, Ginsburg, Kagan, Breyer, (Roberts)

For Respondent: Alito, Thomas, Gorsuch, (Roberts)

Most likely to switch: Roberts

Interruptions among the justices as a predictor of voting disagreement

In a just published article in the Boston College Law Review, Kyle Rozema and I show that interruptions between two justices at oral argument are significantly associated with voting disagreement between pair in the eventual case outcome. We show that on average, a given judicial pair is 7% less likely to vote together in a case where an interruption occurs between them at oral argument. This adds to the mounting evidence that interruptions constitute good predictors of voting outcomes in cases.

Not all justices are alike, but the pattern is clear

Kyle and I show that interruptions at oral argument, as in the many other areas where interruptions have been studied, capture a form of conflict between the justices. Whether interruptions cause ill will or reflect existing conflict, either way they provide a glimpse into otherwise hidden forms of disagreement between members of the Court. Of course, scholars have long recognized regular dividing lines among the justices, most notably ideological divisions, but those factors apply across all cases—we are able to show that interruptions in a given argument reflect an increased probability of disagreement in the case at hand, even accounting for more general ideological division.

The correlation between interruption and disagreement is not the same for all justices. In general a justice is more likely to disagree with someone who has interrupted them or whom they interrupted, but there is variation. The figure below shows the extent to which each justice is more likely to vote contrary to someone with whom they were involved in interruption in a given case. Points to the left of the zero line in the center indicate a greater tendency to agree, points to the right of the zero line indicate a greater tendency to disagree. All but four justices serving on the Court between 1960 and 2015 lie to the right of the zero line, being more likely to disagree with someone with whom they interrupt or are interrupted by.

Differences in Disagreement for Cases with and without Interruptions (Jacobi & Rozema 2018)

Interruptions as conflict, and other theories

We show that a substantially and significantly significant relationship exists between justice to justice interruptions and ultimate disagreement between those justices in the case at hand: a greater probability of voting disagreement between justices of 7% based on one interruption is a very large effect. The question, then, is what is the mechanism of that disagreement?

We explore several explanations for why justices who interrupt each other in a case are less likely to vote together. Our hypothesis is a “conflict” theory: interruptions constitute a type of observable conflict that is systematically associated with disagreement. The conflict theory is consistent with psychology research, which defines interruptions as “deviations from the turn-taking rule that specifies that only one party should talk at the time.

But we also explore other theories. A second “exposure” theory would suggest that justices who speak more in a case might be more exposed to interrupting or being interrupted simply by virtue of taking up more airtime. A third “dissatisfaction” theory is that a justice is interrupting because he or she is at odds with the rest of the Court about the direction of the oral argument and the anticipated outcome of the case. For example, this may occur if the case is very salient to the individual justice but not necessarily salient to the rest of the Court. A fourth “difficult case” theory would suggest that interruptions are simply reflections of something about the case generally that is common to all justices, and cases with more disagreement are more prone to interruptions. In difficult cases, an interruption may not be specific to either of the justices involved in the interruption but is simply a response to the nature of the case itself

We find that the conflict explanation accounts for over half of the effect: the probability of two justices agreeing when one has interrupted the other in a case is reduced by over 4%, even after we have accounted for the other explanations. Exposure is the next most significant factor: a 10% increase in the time a pair of justices spend speaking decreases the likelihood of their agreement by 5.3%. We find no evidence for the dissatisfaction theory. We also find evidence for the difficult case theory: agreement between the justice-pairs is 1.3% lower when there is this kind of interruption. The effect of exposure and difficult cases, however, is less than the effect of conflict.

 

 

Zombies in the graveyard! Knick v. Township of Scott rises from the dead

Reargument ordered in Knick v. Township of Scott

I wrote a post last Wednesday (on Halloween) all about the  spooky metaphysics of deciding when a graveyard taking comes into existence and predicting a 4:4 split in Knick v. Township of Scott. The post was scheduled to go live at our usual time, Monday morning at 9am. However, on Friday the Court restored Knick to the calendar for reargument and directed the parties to file additional briefing. This order basically confirms our prediction that the case was a 4:4 split, although it is possible that some of the justices were genuinely undecided and thought additional argument and briefing could clarify matters.

The reargument order calls for briefing on the Petitioner’s argument in footnote 14 of their brief that there is a key distinction between actions of the responsible government entity at the time of the taking and a compensation order (or denial thereof) by a state court under an inverse condemnation action. Petitioner argues that her takings claim is ripe as soon as the Township required she dedicate a public easement without offering compensation.

I strikes me that the call for 10 additional pages of briefing is just papering over the current 4:4 split that presumably Justice Kavanaugh will resolve when the case is reargued.

The original post follows:

Just in time for Halloween

In Knick v. Township of Scott, the Supreme Court waded into the somewhat metaphysical question of when exactly a takings claim against a state comes into being and the important practical question of where such cases can and should be litigated. This case was argued on October 3rd, but we thought it was perfect for a Halloween themed post.

The Williamson County Catch-22

As the Knick case itself illustrates, plaintiffs who would prefer to take their takings cases to federal court can easily fall into in a Catch-22 situation. Petitioner Rose Mary Knick argues that a local cemetery ordinance compelling her to allow the public daytime access to a grave site on her property amounts to a Fifth Amendment taking unless compensation is forthcoming. The Catch-22 arises because under the 1985 precedent of Williamson County Regional Planning Commission v. Hamilton Bank, Knick cannot pursue her federal Section 1983 claim until she has gone through the Pennsylvania procedure for seeking just compensation and been denied. However, because courts since Williamson County have applied issue preclusion to prevent unsuccessful state plaintiffs re-litigating the same question in federal court, once she loses in a Pennsylvania state court, she will no longer have access to federal court.

Ghost in the machine?

Our predictive models suggest that Chief Justice Roberts (75%), Justice Gorsuch (94%) and Justice Alito (67%) are solidly in the petitioner’s camp.

Knick v. Township of Scott Prediction Based on Supreme Court Oral Argument

It seemed clear at oral argument that the Chief could see the merits of both sides and was concerned about  “the letters that we’re going to get from district court judges around the country who are not going to be very happy learning that they now have to adjudicate state inverse condemnation actions, which can be fairly elaborate.”

The model is understated in the case of Alito. True, Alito engaged with both sides of the argument, but his exchanges with Petitioner were almost entirely in the form of “That’s a good argument, but wouldn’t it be even better if you said it like this?” For example, about 10 minutes into the argument, Alito cut off Justice Kagan to say the following:

Samuel A. Alito, Jr.:

Let me see if I understand your claim, because a little — some of the questions and the discussion up to this point is a bit confusing to me. I thought your claim was that there is a violation of the takings clause and you can, therefore, bring a suit under 1983 when the state does something that constitutes a taking but at the same time says we’re not paying you anything for this. Now it’s not a question of when they would have to pay once they’ve admitted that there’s a taking, but when they do something that constitutes a taking, and they say, no, this isn’t a taking at all, and, therefore, you’re getting zero, which I understand to be your claim here, then you can go directly to federal court and bring an action under 1983. And to require you to go to state court before you do that is essentially to require you to exhaust state remedies before you can bring a 1983 claim, which is never required under 1983.

I thought that that was your argument.

J. David Breemer for the Petitioner could only agree.

Our predictive model counts only two clear votes for the Respondent, Justice Sotomayor (85%) and Justice Kagan (64%). However, even though our model suggests Justices Ginsburg and Breyer are leaning slightly in favor of the Petitioner, a vote for the Respondent seems just as likely. Based on his comments at oral argument, Breyer seems likely to vote for the Respondent and will leave Williamson County intact, but he will soften the effect of issue preclusion that creates the Catch-22.

The other federalism

Conservative justices sometimes care deeply about federalism and state sovereignty, but not so much when it comes to the authority of state courts to decide issues of state law. One of the striking things about the oral argument in Knick v. Township of Scott was that only Justice Sotomayor addressed the Petitioner’s most likely motivation for trying to keep her case in federal court.

Knick denies that there is a grave site located on her property; but she argues that even if there is, she loses an important property right if the state can compel her to grant public access during daylight hours. Assuming that the grave site exists (why take the case to the Supreme Court if it doesn’t?), one of the key questions on the merits of this takings claim will be to what extent Pennsylvania property law had always allowed for limited public access in a case like this. Sotomayor raised the question of whether the restriction was inherent in the property prior to the ordinance, describing it as the “whole issue.” She may have also been thinking that the real point of framing a case like this as a Section 1983 action is not to just to vindicate a federal constitutional right, it is to make the federal courts the final arbiters of Pennsylvania state property law.

The looming specter of a rehearing

By unwritten tradition, justices do not vote in cases for which they were not present at oral argument, and so Justice Kavanaugh should not break the 4:4 tie we predict. However, there is precedent for cases to be re-argued, particularly when otherwise the case is likely to split 4:4. But other cases in the past have simply been handed down with an even split by the original makeup of the Court. If Knick is reargued, and Kavanaugh has the deciding vote, judging by the apparent ideological split of the case in the first hearing, and the similarity between Kavanaugh’s and the other conservatives’ behavior at oral argument in the few cases he has heard so far, that could turn the outcome into a 5:4 decision for Petitioner Knick.

Knick v. Township of Scott prediction: 4:4

For Petitioner Knick: Gorsuch, Roberts, Alito, Thomas

For Respondent Township of Scott: Sotomayor, Kagan, Ginsburg, Breyer

Most likely to switch: Breyer and Roberts

Conservative Principles v. Conservative Outcomes: The Clash of the Two Trump Appointees on the ACCA

Gorsuch versus Kavanaugh: what is a conservative?

The Supreme Court heard two cases about the reach of the Armed Career Criminal Act (ACCA) in October that highlight stark differences between the two Trump appointees. Justice Gorsuch appears to be, as promised, a conservative in the mold of the late Justice Scalia: sitting far to the right on the Court but willing to side with the liberals when issues of methodology or fairness in criminal cases demand. In contrast, Justice Kavanaugh was uncritical of the potential harshness of the government’s position in each case, seemingly focused on promoting the law and order outcome rather than refining the means of analysis.

The two cases involve complex distinctions about the applicability of the much-litigated ACCA, which imposes a 15 year sentence enhancement for persons convicted of three crimes within certain specified categories, at least one of which involves weaponry. Burglary is one recognized category, and U.S. v. Stitt queries whether burglary of a residence can include burglary of a vehicle—which previous cases have said are otherwise not covered—if those vehicles can be adapted for residency or are otherwise being used as residences. Robbery is not recognized as a category under the ACCA, but Stokeling v. U.S. raises the question of whether robbery can nevertheless be covered by the sentence enhancement as a crime that has “an element of use, attempted use, or threatened use of physical force” against another, even if it only involves using slight force to overcome minimal victim resistance.

Early in the argument in Stitt, Gorsuch started listing problems with the government’s position, drawing on a variety of approaches, from textualism—including drawing significance from the state legislature’s use of a disjunctive and the government’s interpretation raising problems of surplusage—to whether congressional intent could be inferred if the legislation covered so few states that defined burglary in such a way. He made similar points in Stokeling, along with raising concerns about the meaning of the legislation at the time of enactment, the traditional use of the term ‘robbery’ at common law, as well as more general critiques in Stitt about widely held dissatisfaction with the Court’s jurisprudence on the ACCA.

In contrast, Kavanaugh focused on carefully maneuvering around prior precedent to make a stronger case for the government. For instance, in Stitt, he appeared quite well-prepared, tailoring a path for getting around what had appeared to be a bright line rule against the inclusion of vehicles in burglaries. For a number of justices, ACCA cases raise questions of fairness, notice, due process, and proportionality, as well as difficult questions of statutory interpretation. Kavanaugh seemed unfazed by these worries, saying for instance of the notice question: “if you’re convicted three times of burglary for burglarizing an RV, you’re on notice, presumably. . . I don’t understand the notice point.

These differences are likely to be determinative, and we predict that the two Trump appointees will diverge in their votes on these two cases, as the following two figures show.

Stokeling v US Predictions

Figure 2

US v. Stitt Predictions

The figures above show the results of a new predictive model we have developed. The model transforms statistical observations about oral argument directly into predicted probabilities based on the prior behavior and voting patterns of each justice. We blended an average of Gorsuch, Roberts, and Alito for the 2017 Term to estimate baseline parameters for the newly appointed Kavanaugh.

In Stokeling (the robbery case), the model predicts a winning coalition for the Petitioner (the criminal defendant) of Justices Sotomayor, Kagan, Gorsuch, and Ginsburg, with probabilities of 88%, 87%, 84%, and 68% respectively. Justice Breyer was uncharacteristically silent but we are somewhat confident he will make up the fifth vote (Breyer is only silent in about 7% of arguments). The model predicts at least a three justice minority consisting of Justice Alito, Chief Justice Roberts, and Kavanaugh, with 91%, 66%, and 61%, respectively. These numbers struck us as understatements after we listened to the argument.

In Stitt (the burglary case), we believe that the Chief and Justice Thomas will vote for the Petitioner (the state), although Thomas said nothing and Roberts said nothing of substance during the oral argument. Assuming Roberts will vote consistently in the two cases and that Thomas will vote against the defendant in both, the model predicts a winning coalition of Kavanaugh, Kagan, Alito, Roberts, and Thomas, at rates of 88%, 73%, and 58% for Kavanaugh, Kagan, and Alito, respectively. The model also predicts at least a three justice minority consisting of Ginsburg, Gorsuch, and Sotomayor (all north of 85%), but it is equivocal about Breyer, as indicated by the fact that he shows up on both sides of the 50/50 dividing line.

One of the striking aspects of the two arguments is that Justice Kagan seems likely to vote a split ticket, holding for the criminal defendant in one case and for the government in the other. For Kagan, the central question in both cases appeared to be the feasibility of line-drawing. The difference between the two cases is that Kagan appeared comfortable with the distinction between mobile homes and other vehicles in Stitt, but not with the different shades of robbery implemented by the various states in Stokeling. It is also possible that Breyer will split the same way for largely the same reasons, but this is far from clear from his numbers. We expect the rest of the justices to vote consistently pro- or anti-criminal defendant in both cases.

 

Stokeling prediction: 5:4 for Petitioner (Stokeling)

For Stokeling: Sotomayor, Kagan, Gorsuch, Ginsburg, and Breyer

For the government: Alito, Chief Roberts, Kavanaugh, and Thomas

Most likely to switch: Breyer

 

Stitt prediction: : 5:4 for Petitioner (the state)

For the government: Kavanaugh, Kagan, Alito, Roberts, and Thomas

For Stitt: Ginsburg, Gorsuch, Sotomayor, and Breyer

Most likely to switch: Breyer

Using interruptions to predict Supreme Court cases

Reading Justice Gorsuch

If you wanted to know which way Justice Gorsuch was going to vote in the 2017 Term, you could have placed your bets with 86% accuracy by observing just one statistic from oral argument—how many times Gorsuch interrupted each side.

In the 2017 Term, Gorsuch heard 60 arguments and voted 40 times for the Petitioner and 19 times for the Respondent (Washington v. United States was affirmed by an equally divided Court, and by convention no individual justice votes are recorded). On average, he initiated 1.6 more interruptions during the Respondent’s time than during the Petitioner’s (a few of these interruptions were of other justices but most were of the advocate). But like Chicago’s annual average temperature of 49°, that number masks all the interesting variation. In those cases where Gorsuch voted for the Petitioner, he interrupted the Respondent an average 3.8 more times than the Petitioner. Conversely, when voting for the Respondent, he interrupted the Petitioner 3.1 more times on average.

One way to visualize this kind of data is with a histogram.

Gorsuch Interruption Gap, Pro-Petitioner and Pro-Respondent Histograms 2017 Term

In the figure above, we have overlaid two different histograms of the interruption gap (interruptions of Petitioner minus those Respondent) for Gorsuch in the 2017 Term. The orange bars indicate the interruption gap in cases where the Justice voted for the Petitioner: it is clear that in the vast majority of these cases, Gorsuch interrupted less during the Petitioner’s time than Respondent’s. The blue bars show the interruption gap for the pro-Respondent cases. Here, in contrast, most of the weight of the figure is above zero—i.e. Gorsuch interrupted the Petitioner more often, though the shape of the blue distribution is flatter. We should note that because Gorsuch voted for the Petitioner twice as often as he voted for the Respondent (as the Court in general does), an orange bar represents twice as many cases as would a blue bar of the same height.

In the next figure, we further explore the distribution of Gorsuch’s interruption voting patterns through a boxplot.

Gorsuch Interruption Gap Boxplots 2017 Term

Boxplots are cool (trust us!) but they require a little explaining. The “box” part of a boxplot indicates the range separating the top 25% of the data from the bottom 25% (the interquartile range); think of this as the space between the 25 and 75 yard lines on a football field. The vertical line segmenting this box is the median of the data. So, in the figure above, a quick comparison of the light blue pro-Respondent box with the orange pro-Petitioner box shows that there is no overlap between the middle 50% of each of these categories. What is more, the orange box is entirely to the left of zero on the interruption gap scale and the blue box is entirely to the right. The white space between the boxes tells you at a glance that 75% of cases in the orange category have a lower interruption gap than 75% of cases in the blue category. And thus, when Gorsuch interrupts one side more than the other, he is usually showing his hand and telegraphing which way he will vote.

The whiskers extending away from the boxes extend close to the extremes of the data: they indicate about how far the data extends before you get to true outliers (in a normal distribution, 0.7% of observations lie beyond the upper and lower limits). The fact that the whiskers cross the zero point confirms that although the interruption gap is a very strong predictor for Gorsuch, it is not infallible. In fact, in the 2017 Term, relying on the interruption gap to predict Gorsuch’s votes would have been misleading in just 7 cases.

Some broader implications

The Supreme Court is under close scrutiny like never before. In this era of intense political polarization, the Court merits attention for reasons that go beyond the outcomes of individual cases or issues of methodology and jurisprudence. The justices themselves have become icons—or demons—representing the moral future of the country. Ruth Bader Ginsburg’s visage appears on cups, T-shirts, and various other paraphernalia, and the late Antonin Scalia is fetishized equally by the right. The two most recently appointed associate justices were subject to extraordinary confirmation processes. Gorsuch holds a “stolen” seat in the eyes of some, and Kavanaugh is defective in character and temperament, according to others. In a divided country where justices are feted and demonized with equal fervor, the public is now paying close attention to every aspect of the Court’s decision-making process, down to the meanings attributed to every word uttered. Oral argument deserves attention because it is the one public part of the Court’s process, and because it is public it gives us an opportunity to more rigorously assess aspects of judicial character that would otherwise be matters of supposition. One way to explore those more amorphous judicial traits is to look at the language justices use in the relatively unguarded arena of oral argument, as compared to tightly scripted case opinions.

Websites such as SCOTUSBlog provide excellent coverage of current Supreme Court cases using traditional modes of legal analysis: close reading and expert opinion. In addition, reporters such as Marcia Coyle, Dahlia Lithwick, Adam Liptak, and Nina Totenberg provide qualitative analysis of Supreme Court oral arguments, and there also are podcasts such as Amicus, First Mondays, and even an army of teenagers who write about the Supreme Court at the website High School Scotus. There is also excellent analysis by other empirically minded scholars, such as Tim Johnson, who has shown that oral argument can be important in judicial decision-making, and some have used transcript data in their research.

But ScotusOA offers something new. Unlike other commentators, we make voting predictions for each justice as the cases are heard. Predicting outcomes is always treacherous, as even a largely accurate model will have variation and noise: even an attempt to model prior data, where the results are known, will lead to imperfect “predictions.” Consequently, we buttress our empirical analysis by listening to the oral arguments, in order to follow the nuances of each case and to understand what the justices think they are communicating.

We also base our predictions on justice-specific models of features of oral argument that our early research has told us are important in the aggregate (Jacobi & Schweers, “Justice, Interrupted”; Jacobi & Rozema, “Judicial Conflict and Voting Agreement”; Jacobi & Sag, “Justices as Advocates”). The challenge is to develop a model that is a good predictive tool in individual cases. To do that, we have developed multiple models for each justice to predict their votes based on oral argument. What is striking about Justice Gorsuch is that he is most easily predicted using just one variable, the interruption gap.

Nielsen v. Preap forecast: Kavanaugh shows his colors and they are bright red

On a major issue for the Trump administration, Kavanaugh signals he will be loyal

At oral argument in Nielsen v. Preap last week, the Supreme Court considered whether non-citizens released from criminal custody should be subject to mandatory detention when the Department of Homeland Security fails to take them into custody immediately. Coming on only his second day on the Bench, the natural question is: how did Justice Kavanaugh behave, and what does it tell us of his likely vote in the case? Given the Trump administration’s crackdown on illegal immigrants, Nielsen v. Preap presented Kavanaugh with his first opportunity to repay those who stood by him in the face of allegations of sexual misconduct, lack of candor, and an injudicious temperament.

After his confirmation hearings, many expressed concern that Kavanaugh could not be impartial, given the arguably unjudicial anger he expressed, with even the nonpartisan ABA reconsidering its review of him based on concerns about his temperament. In response, Kavanaugh said that he could be impartial. It is difficult to infer how the nomination hearings have affected Kavanaugh because he was always expected to be both conservative and partisan. But whether the cause is ideological alignment, inherent partisanship, or indebtedness and resentment in response to the nomination process, our empirical analysis of the Preap oral argument suggests that Kavanaugh will be loyal to the Republicans. The following figure shows the disagreement gap—the differential between a justice’s treatment of the side he or she ultimately votes with and against—in terms of both justice words and speech episodes in the Preap argument.

United States v. Preap, Disagreement Gap in Speech Events and Words

The disagreement gap is a better predictor for some justices than others. As we noted in a previous post, Justice Gorsuch and Justice Alito have particularly obvious tells in this regard: the more each talks to an advocate, the more likely each is to vote against him or her. The early numbers on Kavanaugh suggest that he will leave Gorsuch and Alito in the dust in this regard. At first in Preap, Kavanaugh seemed to be still finding his feet, with his initial attempts to ask a question thwarted by his colleagues. But once he finally took the floor, he signaled his favor for the Government. He had as much to say to the respondents as anyone—Gorsuch and Kavanaugh both had 13 speech events directed at the Respondents—and he had the highest disagreement gap, with nothing at all to say to appellants. The extreme tilt in his participation on this issue and the proportion of his questions and non-questions (a 5:11 ratio, if you were curious), as well as their content, suggest he is a safe vote for the government and for the Trump administration’s immigration policies.

The rest of the Court

As for the rest of the Court in Preap, the split was fairly predictable, with the liberals strongly supporting the Respondent immigrants and the conservatives showing equal support for the government. On a highly polarized Court, the biggest surprise is the ambiguous signals from Justice Ginsburg.

The issue presented is one of statutory interpretation, since the statute refers to detention “when” the person is released; however, with some of the non-citizens represented in this case having been free for years, one for 11 years, the liberal side of the Court could not help querying whether a constitutional claim also arose (much to the annoyance of Justice Alito). As the figures above show, however, this strategy is unlikely to be successful. We predict a 5:4 win for the government, falling along standard ideological lines (as one of us predicted on CBS on Tuesday, prior to the argument).

The Advocates

The performance of the advocates in Preap is worth noting. Assistant Solicitor General, Zachary D. Tripp, was unusually weak for that office. He struggled to answer basic questions, such as where the government draws authority to detain non-arrestees and whether the same rule would apply to somebody 50 years past detention rather than 11 years. He also interrupted and spoke over the justices an extraordinary 23 times. In contrast, advocate for Respondents, Cecillia D. Wang, began exceptionally strongly, speaking for 4 minutes and 15 seconds before any Justice asked her a question. She was interrupted only 28 times, compared to Tripp’s 52 times, another sign of strength, and she interrupted only 8 times (these were mostly cross-talk).

However, Wang struggled a little later in her argument, repeatedly insisting that the statute’s use of the term “when” a person is detained necessarily requires that detention be initiated on the same day as release from arrest. That strict timetable clearly alarmed the pragmatists on the Court, both left and right. Instead, she should have responded that 24 hours would be the ideal rule, but no matter what the rule, 11 years is clearly unreasonable. Perhaps having had such a perfect beginning, she was shooting for the moon, but in the process she may have lost any chance she had to bring on board the conservative pragmatists, Chief Justice Roberts and Justice Alito.

Justice Gorsuch attempted to flex his formalist muscles, debating Wang on questions of grammar, but it is foolish to think that formalism can decide this question. The term “when” is inevitably context dependent: the time constraint implicit in the command “make the tea when the kettle boils” raises a very different expectation than the instruction “come visit when you are in town.” The government would have the Court believe that tea can be made with water boiled 11 years previously.

Beyond Preap: Kavanaugh’s first week on the Court

Considering all four cases the Court heard in Kavanaugh’s first week on the Bench, the newest justice may not be ideologically moderate, but so far he has been moderate in his behavior, in some senses. As seen on the figure below, on average this week, Kavanaugh talked slightly less than the median level among the justices and interrupted at the median. We saw the same trend in the Preap case.

Kavanaugh’s first week, by the numbers

However, these numbers might conflict with the subjective experience of listening to the oral argument because Kavanaugh barely spoke at all in Air And Liquid Systems v Devries and spoke only to one side in both Preap and in United States v. Stitt. It was only in Stokeling v. United States that Kavanaugh had a significant amount to say to both sides of the argument. It seems likely that when Kavanaugh settles in to the Court he will become even more vocal, and it would not be surprising to see him continue to strongly favor one side in any given case. We also expect that Kavanaugh will interrupt even more in the weeks and years to come.

Prediction: 5:4 for the government

For the government (Petitioner): Kavanaugh, Gorsuch, Thomas, Roberts, Alito

For Preap (Respondent): Kagan, Sotomayor, Breyer, Ginsburg

Likely to switch: none

Gundy v. U.S. Forecast: A highly unusual meeting of minds

Gorsuch and Sotomayor find uncommon common ground

Last week’s oral argument in Gundy v. U.S. leads us to predict a unique alliance: a dissenting coalition of the liberal Justice Sotomayor and the conservative Justice Gorsuch. Gundy addressed whether the Sex Offender Registration and Notification Act (SORNA) unconstitutionally delegated unconstrained authority to the Attorney General to determine its pre-authorization application. Petitioner’s theory of the case has the potential to radically reshape the Court’s approach to the modern administrative state.

The following two figures show the disagreement gap—the differential between a justice’s treatment of the side he or she ultimately votes with and against—in terms of both justice words and speech episodes in Gundy’s oral argument. Our qualitative and quantitative analysis leads us to predict a 5:3 ruling in favor of the government, with Sotomayor and Gorsuch joined by the silent Justice Thomas, based on his previous positions. This alignment has never occurred before. For context, Sotomayor and Thomas have only once dissented in isolation from the other justices and have formed two of three dissenters in just three cases previously, but never with Gorsuch as the third vote. Sotomayor has never been part of a two or three justice dissent with Gorsuch. If we are correct, this will be a truly unique alignment among the justices.

Gundy v. US, Disagreement Gap (Words)

Gundy v. US, Disagreement Gap (Speech Events)

Listening to the oral argument, the reasons for this unusual alignment become clear: Sotomayor and Gorsuch have found common ground even though each is clearly guided by their contrasting ideologies. Sotomayor’s questions indicate a liberal distrust of the criminal justice process: she considered that who is covered by a statute is “at the core of what a law is” and here defines who is a criminal. In contrast, Gorsuch’s questions and comments show he is motivated by conservative distrust of government: he worried about the “blank check” Congress gave to the Attorney General. Neither justice seemed too bogged down in concerns about how limiting delegations like the one in SORNA would imperil the broader edifice of the administrative state.

We argued recently that you should prefer to play poker with Gorsuch than with any other justice on the Court, as the early data suggests his vote on case outcomes is the easiest to predict by the bias in his questions and comments at oral argument. The comments of two of his fellow justices and an advocate he was trying to help, as well as the laughter of the courtroom, implied agreement with that conclusion, at least in the case at hand. Gorsuch gave two very detailed leading speech episodes suggesting he agreed with Petitioner’s argument, ending with the rhetorical question “What’s vaguer than a blank check to the Attorney General of the United States to determine who he’s going to prosecute?” In response, Petitioner’s advocate Baumgartel repeatedly answered simply “yes.” Justice Ginsburg then pointedly leave quipped “That’s your argument stated very, very concisely,” to which Baumgartel piled on, saying “I’ll cede my time.” The courtroom laughed, and Justice Kagan, who had ceded the floor to Gorsuch after he interrupted her, continued the joke, saying “Well, then I’ll take back my time.” Laughter followed once more.

That Gorsuch was so clearly willing to rule against the government in Gundy is little surprise: along with Thomas, Gorsuch was predicted to be most likely to take that position based on his prior statements, with some suggesting  Ginsburg could possibly also agree with those two conservative justices. But much more surprising is our prediction here that it is Sotomayor, not Ginsburg, who is the most likely liberal justice to vote with Gorsuch, as the two figures above show.

Ideology was not all that mattered in Gundy’s oral argument. Justice Kagan is almost as liberal as Sotomayor, yet she measured very clearly in the other direction. On the numbers, she was even more predictable than Gorsuch, but her more subtle style makes her views less obvious to a casual observer. The other liberal justice, Justice Breyer, appeared on the figures above to be sending mixed or unclear signals. In fact, we are confident that he will join the strongest advocates for the government’s position: in order, Justices Kagan, Alito, and Ginsburg. Early in the argument, Breyer got his own laugh by advising Baumgartel that she was wise not to try to make an argument to strike down all 300,000-odd legislative delegations to agencies. A highly pragmatist justice, we predict Breyer will rule for the government for this very practical reason, but he is torn in this case between his liberal ideology and his pragmatism. He described the concern “gnawing” at him: the “danger when you combine [in one person] prosecuting a person with the writing of the law under which you prosecute.” With his ideology and his legal methodology pulling him in opposite directions, Breyer spoke similar amounts to each side. Yet, ultimately, he showed his hand, beseeching the government advocate Wall to help him write the rule to avoid such a danger.

The advocates seem aware of the dissonance between ideology and methodology that the case presents for some of the justices. Wall finished his argument with a long description of the many harms that would flow from ruling for the Petitioner, an attempt to ensure that Breyer, Alito, and the Chief stayed with him for pragmatic reasons. In rebuttal, Baumgartel tried to pick off Alito and Roberts by saying that the state legislative equivalent can avoid those harms. At least Breyer and Alito seem unpersuaded by this. The hardest vote to read is Chief Justice Roberts, who did question both sides closely, measuring as mixed on both measures, one in each direction.

Forecast: 5:3 for the U.S.

For the U.S.: Kagan, Alito, Ginsburg, Breyer, Roberts

For Gundy: Gorsuch, Thomas, Sotomayor

Most likely to concur or switch: Roberts

by Tonja Jacobi and Matthew Sag

Empirical evidence of Roberts’ switch in the ACA case

The metrics suggest Roberts was not the only one to switch

With a Supreme Court nominee declaring there has been a left-wing conspiracy against him, naturally people are wondering whether a Justice Kavanaugh, if confirmed, could eventually act in a more judicious fashion. If not, might the new median of Court, Chief Justice Roberts, rise to the challenge of putting the institutional interests of the Court above his own political preferences? Looking back at the Chief’s role in deciding the fate of the Affordable Care Act (aka Obamacare), we are optimistic that he can.

When the fate of the ACA was decided in NFIB v. Sebelius, many were surprised to see the conservative Chief side with the liberal justices to uphold the signature legislation of the Democrat President. Others anticipated it, expecting Roberts to subvert his political preferences to avoid a 5:4 conservative-liberal divide on such a salient case in the lead-up to the 2012 election. On reading the opinions, many were struck by the fact that Roberts’ majority opinion read in large part as a manifesto against the legislation he was unhappily upholding; in addition, the dissent at times referred to the majority as “the dissent.” This suggested that Roberts may have initially sided with the conservatives – a view furthered by rumors to that effect. Until now, this has all been speculation. Here, we assess the empirical evidence from oral argument as to whether Roberts switched, and find some interesting results: not only did Roberts switch, but he was not the only one who changed his vote on some of the issues.

The four separate parts to the Obamacare oral argument

There were four questions addressed in four separate oral arguments over three days: (1) whether the constitutional validity of the ACA should even be heard by the Court; (2) whether Congress has power to require Americans to purchase healthcare, via the individual mandate, either under the Commerce Clause or the Taxing and Spending Clause; (3) whether, if the individual mandate was unconstitutional, it was severable, or alternatively, if the rest of the ACA must fall as a result; (4) and whether Congress exceeded its powers under the Spending Clause by threatening to withhold Medicaid spending to non-cooperating states. The following figures show how the justices behaved at each of the four oral arguments in order, measured by the disagreement gap. As we explained in more detail in a previous post, the justices tend to have more to say to the side they ultimately vote against. This can be measured in terms of speech events or just the difference in the number of words. If the gap is very small, we treat it as a weak signal, colored gray; occasionally, the size of the disagreement gap on one measure will be large but other factors, such as the balance of statements versus questions, or the disagreement gap on another measure, indicates that the signal should be treated as weak.

ACA Issue 1 Disagreement Gap (Words) [Click on the image to expand]
 

ACA Issue 1 Disagreement Gap (Speech Events) [Click on the image to expand]
On the first issue, of whether the Court should address the substantive issues at all in light of the Anti-Injunction Act, the above figure shows that at least four of the justices strongly indicated a desire to decide the case on its merits, and the others did not give a clear signal. Only Justice Scalia seemed consistently against a decision on the merits, but only with a weak signal from a justice who seldom held his cards close to his chest. The first issue, then, looked cross-partisan, with Justices Sotomayor and Kagan wanting to hear the case, presumably to uphold the legislation, as clearly as Justice Alito, who presumably wanted to strike down the legislation. Roberts, too, made clear his desire to decide the case on the merits. Consistent with this, the Court ultimately unanimously held that the Anti-Injunction Act did not apply and that the case should be decided on the merits. This suggests that our metrics are working reasonably well.

ACA Issue 2 Disagreement Gap (Words) [Click on the image to expand]
ACA Issue 2 Disagreement Gap (Speech Events) [Click on the image to expand]
The second issue was the key question in the case and here our analysis in the figure above strongly suggests that Roberts did indeed switch his vote from striking down the legislation to upholding it under the Taxing Power. On both measures, the disagreement gap in words and speech episodes, Roberts consistently measures as the second most vehement against the ACA, behind Alito on one measure and Scalia on another. At oral argument, Roberts seemed to be the most consistently opposed to the ACA and yet he ultimately provided the swing vote in favor.

However, it is important to note that this oral argument addressed the government’s power to pass the ACA under both the Taxing Clause and the Commerce Clause. In authoring the majority opinion, although Roberts found the ACA constitutional under the Taxing Clause, he went out of his way to make clear that the Commerce Clause did not support the passage of the ACA. If Roberts had gone into the oral argument thinking that the ACA was valid as a tax but not under the Commerce Clause, we would have expected to see at least mixed signals from him during this section of oral argument rather than a signal strongly against the ACA. Furthermore, given that if the law was valid as a tax, its status under any other head of constitutional power should have been irrelevant, if Roberts held his tax position from the outset, he should have been less interested in the Commerce Clause issue.

ACA Issue 3 Disagreement Gap (Words) [Click on the image to expand]
ACA Issue 3 Disagreement Gap (Speech Events) [Click on the image to expand]
So it looks like Roberts switched his vote on the all-important second issue, but he was not alone in making a switch on an issue in this case. On the third issue, as seen in the above figure, Justices Kagan, Breyer, and Sotomayor all come out very strongly on both measures in favor of severability, pointing to a way that even if the mandate was struck down, the rest of the Act would stand. Yet, eventually all three of those justices joined the 5:4 majority that did not answer this question, even though the dissent addressed the issue in detail. This is not surprising: compromise is a standard part of the political and diplomatic process of building a coalition on a multi-judge court. What is noteworthy is that it shows that Roberts was not the only one who appeared to have switched – our empirical analysis of the oral argument suggest that other Justices were also acting strategically in NFIB v. Sebelius.

ACA Issue 4 Disagreement Gap (Words) [Click on the image to expand]
ACA Issue 4 Disagreement Gap (Speech Events) [Click on the image to expand]
Given Justice Thomas’ habitual silence, oral argument on the fourth issue indicated a very clear 4:4 liberal-conservative divide among the speaking justices would translate into a conservative 5:4 ruling. Such an ideological division would have fit with previous Spending Clause decisions (e.g. New York v. United States, 1992). Yet, the ultimate outcome was 7:2 against the ACA. At oral argument, Justices Kagan and Breyer led the pro-ACA charge when it came to the disagreement gap in speech events and both were close behind Sotomayor in terms of the disagreement gap in the number of words, and ahead of Justice Ginsburg. And yet Kagan and Breyer joined the opinion striking down this part of the ACA. The majority decision looks to be a product of politics: perhaps Kagan and Breyer were making a strategic concession – compromising on the Spending issue to ensure the constitutionality of the legislation under the Taxing power – or perhaps they were following the Chief’s lead and moderating their positions to preserve the institutional legitimacy of the Court.

* * *

Our empirical analysis of the oral argument in NFIB v. Sebelius indicates that four justices ultimately voted against at least one of their priors on display at oral argument. It is, of course, possible that our metrics are inaccurate, but they align with many commentators’ and experts’ expectations at the time. Possibly, these justices changed their fundamental views some time after oral argument, but another explanation is more likely. We see NFIB v. Sebelius as a case study in judicial strategy, the politics of compromise and coalition building, and how the justices occasionally sacrifice their individual policy preferences to protect the legitimacy of the Court.

By Tonja Jacobi and Matthew Sag

Notes: (1) Thanks to Barry Weingast for suggesting this inquiry. (2) For a more detailed account, see Tonja Jacobi, Obamacare As A Window on Judicial Strategy, 80 Tenn. L. Rev. 763 (2013) (ssrn link)

Favoring friend versus foe in Supreme Court oral arguments

As we discussed in last week’s post on predicting Supreme Court votes based on oral argument metrics, on average the justices have much more to say during oral argument to the party they ultimately vote against. Last week we focused on the micro-picture and how this “disagreement gap” could be used to predict the outcomes of cases based on oral argument. It is also worth looking at the longterm trend, and what the disagreement gap tells us about the Court as an institution.

The Disagreement Gap Over Time

The Disagreement Gap Over Time (1955-2017 Terms) [CLICK IMAGE TO EXPAND]
The above figure shows the average number of words spoken by each justice per oral argument, differentiated on the basis of whether the justice ultimately voted for or against the position of the advocate. That is, whether the justice ultimately agreed or disagreed with the side that the advocate was arguing. The average number of words in agreement per case in any given year is indicated by the hollow black circles. Likewise, words in disagreement are represented by the solid red circles. The gray area shows the confidence interval for words in agreement and the light blue area does the same for words in disagreement.

We divided the data into two time periods: from the 1955 to 1994, and from 1995 to 2017. The shaded regions shows the 95% confidence interval for the trendline in each period. The division of the data between before and after the 1995 Term comports with the theory and findings in our forthcoming article, The New Oral Argument, that much of the increase in judicial activity during oral argument can be attributed to broader trends in political polarization, which escalated dramatically after the Republican Revolution in the 1994 Congressional election.

The figure above clearly shows that although a small disagreement gap has been a feature of Supreme Court oral argument for some time, the size of the gap drastically increased in the mid-1990s. Since the mid-1990s, the justices are now significantly favoring their “friends” over their “foes.”

The Disagreement Gap in the 2017 Term

The New Oral Argument examined 60 years of data, from 1955 to 2015. Now we are going to examine whether the disagreement gap applies evenly to all the justices. We focus on the recent 2017 Term, the first full term with Justice Gorsuch and without Justice Scalia.

The first two columns in the table below indicate the disagreement gap for each justice in the 2017 Term, in terms of words and speech events. The third and fourth columns show words and speech events in disagreement as a percentage of the total for each justice.

Justice

Average Disagreement Gap

Percent of Disagreement

(Words)
(Speech Events)
(Words)
(Speech Events)
Thomas 0 0 0 0
Kennedy 10 0.7 52% 54%
Breyer 151 2.0 59% 55%
Sotomayor 147 3.2 61% 56%
Ginsburg 79 1.6 63% 58%
Kagan 158 1.1 63% 53%
Roberts 126 1.3 64% 53%
Alito 162 3.7 69% 64%
Gorsuch 147 4.1 70% 64%

Every single justice who spoke in the 2017 Term follows the pattern that we identified, of speaking less to the side that he or she ultimately agreed with. Only Justice Thomas is truly evenhanded, by speaking not a word to either side.

Yet there is significant variation even among the eight speaking justices. If the justices were as easy to read at the poker table as they are in oral argument, then you would rather be on a table with Justices Gorsuch and Alito than with Justices Thomas and Kennedy. Gorsuch does not just follow the trend that we identified prior to his arrival on the Court: he personifies it. Justice Gorsuch shows the largest difference in number of speech episodes directed at each side; in the average case, he directed about 5.5 questions or comments to the side he voted with and 9.5 to the side he voted against. Another way to look at this is that Gorsuch and Kennedy each spoke an average of slightly over 100 words to the side they voted for, but whereas Kennedy spoke almost as many words to his foes, Gorsuch had more than twice as much to say to the side he voted against.

It is not surprising that Justice Gorsuch fits the disagreement gap pattern so well. We identified this effect as a product of partisan polarization, and Gorsuch was appointed in a climate of extreme political polarization in Congress, in an era in which the public is more polarized than ever.

Predicting Supreme Court votes based on oral argument metrics

Applying forecast techniques to patent cases

Court watchers and interested parties pay close attention to the ebb and flow of oral argument in individual cases. Usually the best way to understand oral argument is to read the briefs and listen to the argument, but there is also mounting evidence that a more quantitative approach is useful in predicting the votes of individual justices and, ultimately, case outcomes. A big data approach allows for analysis that is not possible in a case by case analysis: for instance, it is not feasible to listen to every case in order to discern what historical trends are emerging in judicial behavior since 1955.

The disagreement gap

In our forthcoming article, The New Oral Argument: Justices As Advocates, we show that the justices overwhelmingly tend to have more to say to the party they ultimately vote against. Sarah Shullman made this suggestion in 2004 based on a study of cases from the 2002 Term; John Roberts came to the same conclusion in a 2005 article before he was appointed to the Bench; and Johnson, et al, among others revisited the issue in 2009.

In The New Oral Argument we show that this disagreement gap has been a feature of Supreme Court oral argument since at least the 1960s but that the size of the gap ballooned in the mid-1990s and has been increasing ever since. We also have a number of other metrics, taking into account more granular data including word counts, interruptions, and the difference between questions and comments, all of which we show follow patterns that help predict case outcomes.

Reduction to practice

Our disagreement gap analysis in the figures below shows the difference between the number of times each justice spoke to counsel for petitioner and counsel for respondent. Because being spoken to more often is actually a very bad sign for the advocate, we then invert those numbers so that a positive score reflects a gap favoring the petitioner (dark navy bars) and negative scores favor the respondent (red bars). Those bars reflect not only speech episodes to each side, but the ratio of questions to comments, patterns of interruptions, and other factors. We also indicate uncertainty or a “weak signal” from the data where appropriate (gray bars). The actual votes of the justices in these cases are indicated on the right-hand side of the figure.

In preparation for this year’s Supreme Court IP Review at the Chicago Kent Law School, Professor Ed Lee asked us whether there was anything interesting to report from the oral argument data for last term’s patent cases. We thought it was a great opportunity to test out some predictive models we have been working on, by applying our metrics to last Term’s intellectual property cases, and seeing if the outcomes could have been predicted based on their oral arguments.

WesternGeco LLC v. ION Geophysical Corp

Prediction based on oral argument in WesternGeco v. ION Geo [CLICK TO ENLARGE]
In WesternGeco v. ION the Court held 7-2 that lost profits in overseas markets attributable to patent-infringing exports are recoverable in patent litigation. Our model correctly predicted that Justices Gorsuch and Breyer would vote in favor of the respondent and that Justices Kennedy, Sotomayor, Alito and Kagan would vote in favor of the petitioner. The WesternGeco analysis highlights how useful the model can be in predicting outcomes that don’t fall along traditional liberal-conservative fault-lines, such as the Breyer-Gorsuch coalition.

SAS Institute Inc. v. Iancu

Prediction based on oral argument in SAS Institute Inc. v. Iancu [CLICK TO ENLARGE]
In SAS Institute v. Iancu, the Court held 5-4 that when the Patent Trial and Appeal Board institutes inter partes review of a party’s challenges to the validity of an issued patent, it must make a decision on all of the patent claims contested by that party. The majority arrived at this conclusion by holding that that the word “any” meant “every” in the relevant statute. In SAS Institute, our model correctly predicted the votes of all of the justices except for the habitually silent Justice Thomas. Given the liberal-conservative divide evident in the eight speaking justices, we would have predicted that Justice Thomas would vote with the conservative majority in favor of the petitioner.

Oil States Energy Services, LLC v. Greene’s Energy Group

Prediction based on oral argument in Oil States v. Greene’s Energy [CLICK TO ENLARGE]
In Oil States, the Court rejected a constitutional challenge to the system of inter partes review introduced in the 2011 patent reform legislation, the America Invents Act. Oil States is the most intriguing of the three cases because so much rested on the outcome of the case and because, as seen below, our predictive model misread Justice Breyer’s eventual vote. Earlier this year, we were engaged as paid consultants and asked to predict the outcome of Oil States.

After reading the briefs, listening to the argument, and crunching the numbers, we predicted a 7-2 vote in favor of the respondent, with Justice Breyer concurring. Happily, this proved to be the exact outcome. Understanding the issues in the case and the substance of Justice Breyer’s comments and questions, we were confident that our model was misleading in this particular instance. However, the model was extremely useful in helping us to read the intentions of Justices Alito and Kennedy.

Patterns are not rules, and so even a very accurate metric will not accurately predict every judicial vote in every case. Hence in Oil States, we adjusted the empirical prediction in accord with what we heard of Justice Breyer’s tone. But by analyzing trends and patterns, we are able to go beyond impressionistic accounts in predicting case outcomes. The proof will be in the pudding, so check in here for our forecasts, and check back to see how they line up with the ultimate case outcomes.