The New Median: Ideology in the Post-Kennedy Court

Guest post by Maveric Searle

This week we have a guest blogger, a Northwestern Pritzker law student, Maveric Searle, who developed a new measure of judicial ideology for a final paper in Tonja Jacobi’s Supreme Court cases seminar. He also applied it to the current Term, based on the limited decisions released so far, to answer the very pressing question of who we can expect to be the new median of the Court. Maveric has turned his paper into a ScotusOA post, providing the first take on the new ideological spectrum of the Roberts Court.

A new median

The retirement of Justice Anthony Kennedy in 2018 represented a potentially pivotal moment in the modern history of the Supreme Court. Kennedy had long held the spot of the median justice on the Court. The median justice, whose vote is often the deciding vote in contentious cases, is typically thought of as being the most powerful or the most influential justice. Kennedy’s retirement represented not only the appointment of a new justice, but also the opportunity for a new justice to fill the position at the Court’s center. Chief Justice John Roberts is commonly seen as the favorite to become the new median justice. This post analyzes the Court’s early rulings from the October 2018 Term to see if Roberts has lived up to those expectations, or if the Court’s newest member, Justice Kavanaugh, could be its new median.

Conceiving the Median: Agreeability and Majoritarianism

One simple way to identify the median justice is to look at who has voted in the majority most often this Term. The median justice should have a high rate of voting in the majority because historically the median is in the majority in closely decided cases more often than any other justice.

            The numbers above suggest that the median is likely Kavanaugh or Roberts, each of whom has been in the majority at a higher rate than their colleagues. Kavanaugh has a higher ranking than Roberts, but importantly Kavanaugh did not participate in several cases this term, and a dissenting vote in any one of those cases would have moved him below Roberts.

Another straightforward approach is to look at how often each Justice agreed with the other Justices. The median justice should have a high agreement rate because, by virtue of being centrist, he or she will generally agree with colleagues on either end of the spectrum more often than will the opposite extreme justices.

Once again, Kavanaugh and Roberts are duking it out for top spot, tied for who looks most  likely to be the median in terms of average agreeability rate. Given the similarities in both their rates of voting in the majority and their average agreeability rating, a more sophisticated analysis is needed to determine which of the two is the new median justice.

Ranking the Justices on a Left-Right Ideological Spectrum

It is now well accepted who is a liberal justice and who is a conservative justice, not only because the justices are nominated and confirmed by increasingly polarized political actors, but because each of the currently serving Justices has prior affiliations with either the Federalist Society or the American Constitution Society. The key question is: what are the differences among the liberals and among the conservatives? And in particular, with a new five justice conservative majority, who will be the most moderate conservative justice, and thus the median of the Court?

To identify the median justice by ideology along a left-right continuum, the next measure examines how often each justice votes with the conservative justices as compared to the liberal justices. Justices are classified as liberal or conservative based on the political party of the president who appointed them. It then counts how many liberal and conservative justices each justice agreed with in each case (using the percentage of the ideological bloc that the justice agreed with, to account for the fact that there are more conservative than liberal justices).

In a unanimous decision, every justice would agree with 100% of both the conservative and liberal justices. In a 5-4 ideologically ordered decision, a conservative justice would agree with 100% of the conservative justices and 0% of the liberal justices. In a case like Garza v. Idaho—a 6-3 decision with Thomas, Alito, and Gorsuch dissenting—each justice in the majority agreed with 40% of the conservatives (Roberts and Kavanaugh) and 100% of the liberals. Each justice in the minority agreed with 60% of the conservatives and 0% of the liberals. Justices are given weight for agreeing with themselves to ensure that decisions with a lone dissenter have an impact on the dissenting justice’s final score.

These percentages are then turned into an aggregated 0–1 “liberal score” and an aggregated 0–1 “conservative score,” with e.g. 60% agreement being a score of .6. A net ideology score is then determined by subtracting the liberal score from the conservative score for each justice. Using this method, unanimous decisions do not impact the final score because they result in an equal increase to both the aggregated liberal and the aggregated conservative score.

Finally, the net ideology scores are standardized by taking the final score and dividing it by the number of non-unanimous cases that each justice participated in. This yields a number between -1 and 1 for each justice, who can now be ranked them from left to right.

The resulting ranking lines up well with common intuitions, with the possible exception of Justices Gorsuch and Alito: Gorsuch lying slightly to the left of Alito may surprise some, particularly given Gorsuch’s emphasis on originalism and Alito’s rejection of that methodological approach most commonly associated with conservatism.

However, the scores may be capturing some nuance that would be missed by overly focusing on the two Justices’ rhetorical approaches, as there are some areas where Gorsuch has shown himself to be more liberal than Alito. For instance, in Biestek v. Berryhill, Gorsuch wrote a dissent that was joined by Justice Ginsburg. This dissent likely represents the two formalists agreeing based on factors that were methodological rather than ideological. Given the limited data available from the cases decided so far this Term, an unusual case like Biestek had a significant impact on Gorsuch’s final score: this decision alone moves Gorsuch from slightly to the right of Alito to slightly to the left. But Biestek may be quite representative: when it comes to matters of fairness—a key concept in the standard “underdog”–sympathetic model of liberalism—there is a strong argument to be made that Gorsuch lies to the left of Alito in some areas.

Unlike the statistics on justice agreeability and rate of voting in majority, the standardized ideology scores of Roberts and Kavanaugh demonstrate a significant difference between the two Justices. Although Kavanaugh has been in the majority at the highest rate, and he is tied with Roberts in terms of average agreeability, Roberts’ ideology score puts him distinctly to the left of Kavanaugh.

While it is still early in the post-Kennedy era of the Court, and a number of the Court’s most controversial cases have yet to be decided this term, this method suggests that Chief Justice Roberts will in fact be the Justice to look to as the Court’s new median. The early numbers suggest that we really are seeing “the Roberts Court.”

October 2018 Term in review — Part IV (Interruptions, continued)

Gender versus seniority

In this post, we continue to examine the role of interruption at oral argument in the 2018 Term. In our last post, we showed that the overall rate of justice-to-justice interruptions went down in the 2018 Term, as did the gender imbalance of those interruptions. In this post, we scrutinize who is interrupting and who is being interrupted at the individual level. The data suggests that, in 2018 at least, gender is not the only factor seemingly at play: seniority also seems to be an important factor in this Term’s judicial interactions.

Interrupting versus being interrupted: who is disproportionate?

The figure below looks at justice-to-justice interruptions. It shows the relationship between the rate of being interrupted and the rate of interrupting for each justice in the 2018 Term. We published a similar figure for the 1998-2016 and 2017 Terms in a previous post. This is a good way to assess whether someone is being interrupted because they simply talk a lot: if that were true for an individual justice, high rates of being interrupted would correlate with high levels of interrupting. The dashed 45° line represents parity between being interrupted as often as one interrupts. Thus, even though the data is presented in terms of raw number of interruptions, the 45° line effectively accounts for variation in the rate of each individual justice’s speech episodes. For instance, Justice Thomas sits at the zero point on both axes, since he rarely speaks. Justices who are interrupted more than they interrupt sit above and to the left of the 45° line (represented by solid red circles), justices who interrupt more than they are interrupted sit below the 45° line (represented by hollow black circles).

Consistent with the finding that female justices are interrupted at significantly higher rates than male justices, once again in the 2018 Term, Justices Sotomayor and Kagan were the two justices most frequently interrupted by other justices. The figure shows the raw number of interruptions for each axis: Sotomayor and Kagan were each interrupted approximately 49% and 35% more often than the most interrupted male justice, Justice Alito.

However, there is an obvious difference between Sotomayor and Kagan. Sotomayor was once again the most interrupted justice in the 2018 Term, and she was interrupted by her fellow justices many more times than she interrupted them. In contrast, Kagan interrupted her fellow justices even more than she was interrupted. In previous years, this was not the case. Previously, we found that Kagan was learning to reduce her polite language—which seems often to simply make her easier to interrupt—but she had lagged behind Sotomayor and other female justices in making this adaptation. Kagan may be learning to more effectively stand up for herself at oral argument.

The other striking result from the first figure is that, unusually, there are two male justices who sit on the upper left side of the 45° line. Both Justices Kavanaugh and Gorsuch were interrupted more than they interrupted, which is atypical. Since these are the two most junior justices, and Kagan and Sotomayor are the next two most junior justices, it seems likely that there is a seniority effect in justice-to-justice interruptions at play in the 2018 Term. In prior research, one of us found that seniority was a statistically significant factor in interruption rates—more senior justices interrupt more junior justices at significantly higher rates than vice versa—however the effect in substance was overwhelmed by the impact of gender (as it also was by ideology).

Although we have not yet conducted regression analysis for the new Term, the figure suggests that seniority may have been more significant this Term. But this does not mean gender is no longer significant. We note that Kavanaugh and Gorsuch were interrupted only slightly more than half as many times as Sotomayor and Kagan. Gender still seems to be heavily at play in terms of who is interrupted, even though Kagan is breaking the pattern somewhat in terms of who does the interrupting. (We do not want to make too much of Justice Ginsburg’s lower rates of both interrupting and being interrupted, due to her absence from some arguments. Previously, she was lower on both fronts.)

Advocate interruptions of justices: improvement?

We see more or less the same pattern when we examine how often the justices were interrupted by the advocates. The figure below shows how often each justice was interrupted by one of the advocates, per thousand words spoken by the justice. In contrast to previous findings, there is no self-evident gender effect here. Seniority seems to be at play as much as gender.

When looking at which advocates do the interrupting, we also see less of gender effect than was found in previous Terms. The figure below indicates which advocates interrupted the justices the most, normalized per thousand words.

Of significant interest, the number one interrupter on this measure was a woman, Barbara Underwood. Underwood appeared in a particularly contentious, and potentially significant, case, Department of Commerce v. New York, on the question of whether federal government may include a citizenship question in the 2020 Census. Having a woman in the number one spot is not the only shift: there does not appear to be the previous gender pattern at all in the other advocates’ interruptions of the justices in the 2018 Term, as the two genders are distributed quite evenly. We think that the fact that women can now interrupt as much as men is a great step forward, although technically advocates should never interrupt the justices, according to the Supreme Court rules. This advance is particularly important given our recent finding that female advocates were given less of an opportunity to speak in 2018 oral argument than male advocates.

Technical notes:

Note that some advocates do not appear in the third figure, as the transcript indicates that they did not interrupt at all during their appearance/s.

When we talk about interruptions in this post, we include any time one speaker interrupts another speaker as indicated in the transcript by the “–” notation at the end of the line. Previously, we have differentiated between quickly occurring interruptions, which may be accidental (called “crossovers” to capture two people beginning to speak at almost the same time), and more distinct interruptions, occurring when a speaker clearly has the floor, capturing more explicitly impolite or deliberate interruptions. We do not yet have the timestamps for this Term to make this distinction. Consequently, there may well be instances where the transcript reflects an interruption which some of our readers may feel was not “really” an interruption. That may be so, however, readers should also consider that perceptions of whether one speaker has interrupted another are affected by the biases and expectations of the listener and we trust the court reporters to at least be consistent. We have used the same definition of interruption in all of the data in this post.

October 2018 Term in review — Part III (Interruptions)

A more civil Term?

As we have now well documented, there has long been a gender imbalance in the rate of interruptions of female justices versus male justices. Previously, we asked whether that imbalance had improved in response to the attention being given to the Virginia Law Review article that revealed this pattern, as both Justice Sotomayor and Justice Ginsburg had each suggested it had. As of OT2017, we found that it had not. But an analysis of justice to justice interruptions from the 2018 Term suggests there has been improvement at the Court.

Female Justices are still disproportionately interrupted, but less so

The next two figures are based on our calculation of the normalized rate of interruption for the male and female justices, per argument. Given that Justice Thomas is virtually silent at oral argument and in light of the fact that a number of arguments in the 2018 Term took place before an eight-person bench, it is important to normalize the rate of interruption by the gender ratio of the participating justices in each case. For example, where Justice Ginsburg was absent and Justice Thomas was silent, we treated the Court as comprised of two women and five men. We went back and applied the same methodology to the oral arguments transcripts for previous terms, going back to 1998.

The figure immediately below shows the difference in the normalized rate of interruption for the female justices compared to male justices (orange bars). All bars that appear above the zero line mean that women are interrupted more than men; bars below the line show the reverse. Adjusting for their relative numbers on the Bench, the female justices were interrupted at a higher rate than the male justices in every Term except for 2004 and 2005. What is more, that difference was growing on average from 2006 to 2017. But the 2018 Term shows an improvement.

The 2018 Term was not only less gender imbalanced than the extraordinary 2017 Term, the gender difference was also somewhat lower than in every Term since 2011. Perhaps the change that Justice Sotomayor perceived, of an improved and more civil environment at oral argument, is now manifesting more clearly?

A recent improvement, or a broader trend?

However, the difference in the normalized rate of interruption may be less informative than the same information looked at as a ratio of interruptions between men and women. The figure below shows the ratio of interruptions of female versus male justices (blue bars): a Term measuring at 1 would have an equal number of interruptions of female justices and male justices, normalized by their respective numbers on the Court.

Looked at in this way, not only is the 2018 Term a vast improvement on recent years, it has a lower gender imbalance than any Term since 2004, back when Justice Ginsburg was the sole woman on the Court. It also suggests that perhaps our pessimism last year was unwarranted: this shows that in terms of the ratio of interruptions, female to male interruptions also dropped significantly in 2017, after the issue became one of public comment.

A more civil Court?

Although the difference in the normalized rate of interruption between female and male justices (orange) went up in the 2017 Term and down in the 2018 Term, the ratio of those same numbers (blue) went down in each of those terms. What explains this difference between the orange bars and the blue bars? The answer is that the number of justice to justice interruptions is not constant, not by any means, as the next figure shows.

The 2018 Term was not only a less gender imbalanced term in terms of interruptions, it was also seemingly more civil (or at least less heated) than the 2017 Term. The rate of interruptions in 2017 was exceptionally high, and 2018 saw interruptions drop somewhat. We do not want to overstate the civility that represents: 2018 still saw the third highest level of interruptions in the last 20 years, but the increase over previous years is more in line with a general linear trend, rather than the exponential shift that could have been inferred from last Term’s interruption levels.

We also have to be wary of reading too much into the lower rate of justice to justice interruptions in the 2018 Term. The decrease in interruptions might simply be attributable to the large number of cases where at least one justice was absent—Justice Ginsburg through illness and Justice Kavanaugh by joining the Court partly through the Term. In the next post we will break down interruptions by justice.

One final point to note: remember that the 2004 and 2005 Terms were the only two terms in which men were interrupted more, proportionally, than women. We see from the above figure that those two Terms were also outliers in terms of the number of interruptions. This suggests that even with just one relatively senior woman on the Court, the unusual lack of gender imbalance in those two Terms was seemingly a result of more men being interrupted, rather than women being interrupted less.

Technical note:

When we talk about interruptions in this post, we include any time one speaker interrupts another speaker as indicated in the transcript by the “–” notation at the end of the line. Previously, we have differentiated between quickly occurring interruptions, which may be accidental (called “crossovers” to capture two people beginning to speak at almost the same time), and more distinct interruptions, occurring when a speaker clearly has the floor, capturing more explicitly impolite or deliberate interruptions. We do not yet have the timestamps for this Term to make this distinction. Consequently, there may well be instances where the transcript reflects an interruption which some of our readers may feel was not “really” an interruption. That may be so, however, readers should also consider that perceptions of whether one speaker has interrupted another are affected by the biases and expectations of the listener and we trust the court reporters to at least be consistent. We have used the same definition of interruption in all of the data in this post.

October 2018 Term in review — Part II (Advocates)

In our October Term 2018 in review part I, we identified patterns of behavior among the justices. In this post, we look to the other side of the Bench, the advocates. As we have shown elsewhere, the proportion of female advocates appearing before the nation’s highest court has increased over time, but the Supreme Court Bar is still largely a “boys club.” Our analysis suggests there has been little improvement in representativeness at Supreme Court oral argument in the 2018 Term.

Me too? Where are the women?

In the 2018 Term, there were 124 individual advocates who appeared before the Supreme Court: 26 were female and 98 were male. Over the 71 oral arguments, these advocates collectively made 185 appearances: 32 by women and 153 by men.

One might imagine that women would get better opportunities through the Solicitor General’s office than in the private sector. And it is marginally better, at least this Term. In OT2018, of the 124 advocates, 20 appeared for the Solicitor General’s office in some capacity: Noel J. Francisco as the SG, Jeffrey B. Wall as the Principal Deputy SG, and 18 other deputy SGs and assistants to the SG. Five of these 20 government lawyers were women; however, those women accounted for only 9 of the 53 appearances. So, in sum, women account for 20.97% of Supreme Court advocates in OT2018, and 25% of advocates appearing for the SG’s office.

Repeat players: is it any better for female advocates next time around?

The figure below shows that the odds only get worse when we look at repeat performances. Of the 34 repeat players (those appearing twice or more) in 2018, only 6 were female. Four were from the SG’s office and two others also appeared twice or more: Ann O’Connell Adams, Rachel P. Kovner, Erica L. Ross, and Morgan L. Ratner, all assistants to the SG, Lisa S. Blatt, Chair of Williams & Connolly’s Supreme Court and Appellate practice, and Teresa Ficken Sachs, Co-Chair, Appellate Advocacy and Post-Trial Practice Group at Marshall Dennehey Warner Coleman & Goggin.

Let me talk: are women treated differently?

So much for appearances: once before the Supreme Court, were women also treated differently? We are interested in particular in whether the justices allowed male and female advocates the same chance to speak. The next figure shows the number of words each advocate spoke on average, per appearance this Term.

Looking at the average number of words spoken per appearance, there does not seem to be a clear gender pattern: 14 women were in the top half of the graph, 12 in the bottom half. That suggests that though their numbers may be low, women are getting to speak their turn without any greater interruption. That impression, however, is wrong.

To see why, think about the role of the Solicitor General: many believe the SG’s office receives greater deference from the justices, but looking only at the number of words may not show this. Given that the SG’s office regularly appears as an intervening amicus, the SG’s representative often talks for a shorter period than the non-government advocates.

To see if men and women, and government and non-government lawyers, are given the same chance to get a word in between the back-and-forth between the justices, the next figure shows the number of words spoken by each advocate as a ratio of the number of words spoken to him or her by the justices.

Now we see that the Solicitor General’s office did in fact tend to receive more deference in OT2018 than appeared in the raw numbers. In terms of gender, of the 24 women to appear in OT2018, only 7 were in top half in the ratio of words spoken, one was exactly in the middle and the remaining 18 wallowed in the bottom half of the table.

Once accounting for the fact that there are differences in the times of various advocates appearing for the Court, we see that not only are women grossly underrepresented in the Supreme Court Bar, but when they do get a chance to appear, they are not given as much chance to actually speak. Gender differences persist at the Supreme Court in the 2018 Term.

* Revised to correct advocate misidentification in earlier versions at Thursday May 9, 2019 7:39PM.

October 2018 Term in review — Part I

The Supreme Court has finished hearing oral arguments for the October Term 2018; analysis involving success rates will need to wait for all the vote outcomes to be revealed, but we can summarize much of the behavior at oral argument already.

As we have described in detail inThe New Oral Argument: Justices as Advocates, one of the defining features of modern Supreme Court oral arguments is how active the justices are. In The New Oral Argument, we argue that much of this activity is directed at judicial advocacy, and we found that “losers on the Court” tend to be more active, fighting back through their words. This includes justices who were in the minority when the case was decided. It also includes losers in a big picture sense: we found that when the Court has been dominated by conservatives, liberals have been more active—in contrast, in the 1950s and 1960s, when the liberals dominated, the conservatives were relatively more active. With the Court becoming even more conservative in the 2018 Term—with the retirement of the moderately conservative Justice Kennedy being replaced by now Justice Kavanaugh, anticipated to be considerably more conservative—we expected to see the liberals being more active than the conservatives. As the figure below shows, that is what occurred.

The figure above shows both words spoken and speech events by each bloc of justices, for the average justice within each category. On average, liberal justices each spoke approximately 37,000 words to the conservatives’ approximately 26,000 words per justice. Liberals engaged in more than 1,100 speech episodes on average, the conservatives only 950. Note that for both these calculations, we treat the conservative bloc as consisting of four speaking justices: Justice Thomas speaks so infrequently that he is in a class of his own (the “monastic conservative” bloc?).

Looking at the justices in more detail, the next figure shows the differences for the 2018 Term.

True to form, Justice Breyer was the most verbose justice in the Term, but he was not as far ahead of the other justices as he had previously been. Justice Sotomayor was not far behind Breyer, and in fact spoke more often. But as we know, Sotomayor is the most interrupted justice on the Court, as she was again last Term, so it makes sense that, as shown above, even though Breyer talked more than Sotomayor, Sotomayor needed more speech episodes to say less than Breyer.

For once, Thomas appears as a blip instead of a vacancy, due to those 41 words—all spoken in one case, all directed to one side (disagreement gap, anyone?), all during the advocate’s rebuttal, and all challenging the side making a Batson challenge in a criminal case. The facts of the case, Flowers v. Mississippi, were so outrageous that even the ordinarily prosecution-friendly Justice Alito was shocked by the astonishing behavior of the local prosecutor. Thomas’ intervention in Flowers was not inspiring: his three questions were premised on a misunderstanding of the law or the facts. We were also struck by the fact that the official court transcript contains extra words Thomas did not say but which were required/implied by traditional rules of grammar.

Focus on Kavanaugh and Ginsburg

In light of all the controversy over his nomination and his remarkable conduct during the nomination process itself, for which he subsequently apologized, Justice Kavanaugh’s performance on the Bench is obviously worthy of scrutiny.  When Kavanaugh first joined the Court, he seemed unconfident; after a few halting attempts to question the advocates, he soon became more practiced. However, throughout the Term, we noticed that Kavanaugh continued to cede the floor more often than is typical for a male justice.  

The other justice on everyone’s mind this year was Justice Ginsburg. Attention was drawn to Ginsburg due to her recent illness and ostensible frailty, but also because she was the star of an Oscar nominated biopic. The top three justices on both sides of the figure above are liberals, but Ginsburg was the exception, second bottom only to Thomas. Her activity cannot be assessed in raw totals due to her first absence from oral arguments due to illness. The next figure is also revealing on that question.

The figure above explores how Ginsburg’s and Kavanaugh’s performance at oral argument varied over the course of the Term. The figure shows the moving average in words spoken by our two key justices of interest, Ginsburg and Kavanaugh, as compared to the average for the other three liberals and other three conservatives (we have excluded Thomas who, despite breaking his silence, still barely registers). It shows the 71 cases of the Term, including Knick v. Township of Scott appearing twice, once for argument and once for re-argument. A moving average (here, of the last 7 cases) makes it easier to read any time trend that emerges, but it obscures the extremes—for instance, it will not show silence in a given case, but we can note that where it is interesting.

Ginsburg broke her ribs between cases 22 and 23, but her activity levels were already dropping from case 11 onwards, and she was silent in the two cases prior to breaking her ribs. We can only speculate as to whether she was already feeling ill from the cancer that was discovered after the break. She then missed 11 cases as she recovered from surgery. But the good news is that she is back up to full RBG activity levels—in fact, in the last dozen cases, Ginsburg has been more active than she was prior to the rib break or the cancer diagnosis. We hope this is a sign she is now in very good health. She is still not as active as the other liberals, all of whom are younger than her, but she is up to the level of the average conservative justice (who are also all younger than her).

Speaking of conservatives, our prediction that Kavanaugh became more active over time is borne out in the figure. He began as the second quietest justice on the Court until Ginsburg’s temporary decline; Kavanaugh is now on average very similar to the other conservatives. It is interesting to note that, comparing the conservatives and the liberals overall over the course of the Term, even though the conservatives were consistently less active than the liberals, the two groups moved largely in lockstep (except for a bump in the activity of the liberals during the 40s cases). This suggests that, with the exception of that group of about 10 cases, while there are systematic differences between the two groups, the justices as a whole agree on which cases are more worthy of their input.