A wideneing divide: an examination of constitutional polarization in supreme court nomination hearings

At its core, this paper seeks to explore polarization and partisanship within Supreme Court nomination hearings. The role partisanship plays in contemporary nomination hearings has appeared to increase greatly, leaving many wondering if the Supreme Court has lost its sanctity as a neutral arbiter of the law. Is increased polarization just a perception, or is it a concrete reality? This research paper utilized Computer Aided Textual Analysis to parse the transcripts of 16 previous nomination hearings for phrases which could be considered polarized. In total, this accumulated to nearly 2.2 million words spoken between 127 unique speakers. Because polarization and partisanship are complex concepts to probe in nature, it is difficult to assemble a robust dictionary capturing target "polarized" words or phrases without human coding. Thus, the dictionaries used in this study are limited to constitutional polarization and speaker affect. The conclusions drawn show there has been a net decrease in constitutional polarization over the last 50 years, and neither gender nor party allegiance has significant effect on the use of constitutionally polarized words. However, Senate control does play a role in polarization levels. Of the five most polarized hearings, four were hearings in which the Senate was controlled by a different party than the nominating president. Ultimately, it seems other factors beyond the constitution are what is driving an apparent increase in tension and vitriol during the nomination hearings.

ABSTRACT At its core, this paper seeks to explore polarization and partisanship within Supreme Court nomination hearings. The role partisanship plays in contemporary nomination hearings has appeared to increase greatly, leaving many wondering if the Supreme Court has lost its sanctity as a neutral arbiter of the law. Is increased polarization just a perception, or is it a concrete reality? This research paper utilized Computer Aided Textual Analysis to parse the transcripts of 16 previous nomination hearings for phrases which could be considered polarized. In total, this accumulated to nearly 2.2 million words spoken between 127 unique speakers. Because polarization and partisanship are complex concepts to probe in nature, it is difficult to assemble a robust dictionary capturing target “polarized” words or phrases without human coding. Thus, the dictionaries used in this study are limited to constitutional polarization and speaker affect. The conclusions drawn show there has been a net decrease in constitutional polarization over the last 50 years, and neither gender nor party allegiance has significant effect on the use of constitutionally polarized words. However, Senate control does play a role in polarization levels. Of the five most polarized hearings, four were hearings in which the Senate was controlled by a different party than the nominating president. Ultimately, it seems other factors beyond the constitution are what is driving an apparent increase in tension and vitriol during the nomination hearings. ii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 LITERATURE REVIEW 3 METHODOLGY 19 RESULTS 25 DISCUSSION 33 CONCLUSION 38 REFERENCES 40 iii 1 Introduction One of the grievances outlined against King George III in the Declaration of Independence regarded the state of the British judiciary. The revolutionaries charged that the King had made “judges dependent on his will alone, for the tenure of their offices” (“Declaration of Independence,” 2019). In the minds of the founding fathers, the lack of an independent judiciary in the colonies was a crucial cog missing from the wheel of independence. Still today, countless states and government entities recognize the importance of an independent judiciary, untainted by politics. But in modern times, this notion of independence has been tested. The mantra of “packing the courts” and blatant ideological scheming has left some skeptical to how much the judiciary can be trusted. Undoubtedly, no judicial appointment is more important and influential than that of the Supreme Court. The stakes are the highest here, and partisan demagoguery can run rampant. This has turned nomination hearings into high-stakes political ventures. Though nomination hearings are often hours-long and full of complex legal questioning, much of the lay-public has tuned into these spectacles because of the widespread coverage in news media, and a perceived political battle. The recent nomination of Justice Brett Kavanaugh is perhaps the most infamous example of a nomination hearing turned explosive political war. During the initial fourday hearing and subsequent witness testimony, nearly 6 in 10 Americans said they were following the hearing closely, hoping to see if their preferred side “won” the nomination, and what the fate of Brett Kavanaugh would be (Montanaro, 2018). But for some, this troubling hearing sparked questions about the fate of the independent judiciary. 2 Prominent political leaders began to call for significant changes to the Supreme Court. Many questioned the efficacy of the nomination process itself. Despite all the tumult, the Supreme Court still stands and continues to handle some of the most difficult and pressing legal questions in modern society. But the state of the nomination process now poses a bevy of important questions. Has this process become polarized and hyper-political? Do nominees use language and phrases that indicate a political bias during their hearings? Is the Supreme Court becoming a political tool to be wielded by Congress and the executive branch? These questions are all integral to understanding the Supreme Court as an institution. If Supreme Court justices have become closeted partisans, the impact of their decisions become all the more tenuous. It would be easy to dismiss rulings or positions of justices who supposed ideological slant differs from your own. This paper explores the use of polarized language in the 16 most recent Supreme Court nomination hearings in an effort to draw conclusions about their political nature. As language is the foundation of these hearings, the qualitative and quantitative study of the words and phrases used can provide insights into their underlying meaning. Through the use of Computer Aided Textual Analysis (CATA), polarized words are identified and their tone can be illuminated. Importantly, the main aspect of polarization studied in this paper revolves around the constitution. Polarization can be defined in numerous contexts and pertain to many different subjects, so no all-encompassing measure exists. However, constitutional polarization is easier to capture in its entirety. As this is a particularly crucial aspect explore, as the constitution is the document the Supreme Court is charged with interpreting. 3 Surprisingly, when performing CATA on the nomination hearings, the results indicate that while constitutional polarization has waxed and waned over time, there has been a slight net decrease overall. The greatest determining factor on the use of constitutional polarization appears to be whether or not the Senate is controlled by the same party as the Presidency. Additionally, though both parties tend to mirror each other’s use of polarization, Republicans have typically increased their use of negative language to attack Democratic nominees over time. Democrats however, do not tend to use significantly greater polarizing language when speaking with Republican nominees. These findings suggest that other aspects outside of the constitution are what is causing the high levels of conflict seen in modern confirmation hearings. Literature Review This literature review follows a course which aims to outline the state of political polarization in America, as well as in Congress and the Supreme Court specifically. First, the development, role, and structure of the Supreme Court is briefly reviewed. Then, an outline of the confirmation process is given. Next, the state of polarization in the general public is reviewed, where I argue the literature shows that politically involved lay-public have become more polarized in recent years. After, the presence of polarization in congress, or “elite polarization” is discussed. Finally, the idea of elite polarization is transitioned the Supreme Court itself, before the main research questions are outlined. The Development and Role of the Supreme Court The Supreme Court of the United States website expresses its role within our democracy in terms of “equal justice under the law.” As inscribed over the main entrance to the high court building in Washington, D.C., the court is the “final arbiter of the law… 4 charged with ensuring the American people the promise of equal justice under the law” (“The Court,” 2019). However, the creation of a court with such powers was not a straightforward process. When the United States was still a fledgling democracy, the struggle to create a unifying constitution persisted. One aspect the failed Articles of Confederation lacked was a national entity capable of exercising judicial review to resolve disputes. As a result, many states took it upon themselves to establish their own judiciary to interpret their infant laws. Yet, separate state courts caused some discord. How to best unify the separate judiciaries was duly debated, and Alexander Hamilton penned Federalist 78 in an effort to assuage fears of creating a federal court system. Acknowledging that the necessity of such an institution had already been illuminated, Hamilton crafted an argument for how the court should function. He began by arguing a topic becoming hotly contested during modern times—court term limits. Hamilton believed appointed justices and judges should remain in office “during good behavior” (Hamilton, 2008). This, he said, was the most effective way “to secure a steady, upright, and impartial administration of the laws,” using the qualifier of good behavior to stop potential “despotism” observed in other political systems (Hamilton). But term limits were not the only concern. The use of courts as political weapons was also a valid fear. This is why Hamilton also strongly argued in favor of maintaining the courts as independent from other branches of government. The judicial branch does not “hold the sword of the community” like the executive, or “command the purse” as the legislature does (Hamilton). It exists to “exercise neither force nor will, but merely judgement” (Hamilton). Of the three branches, the judicial was meant to be the weakest, 5 and thus could not endanger liberty when independent. Hamilton argued we can only lose liberty, “if the power of judging be not separated from the legislative and executive powers” (Hamilton). Remaining staunchly independent and a lack of term limits are staples of the system we see today. With many major pitfalls addressed and myriad disagreements momentarily suspended, the framers were able to ratify the Constitution in 1788. Undoubtedly Hamilton’s writings were a major influence setting the groundwork for a federal judiciary. The final product is present within Article III of the Constitution: “the judicial power of the United States shall be vested in one Supreme Court, and in such inferior courts as the Congress may from time to time ordain and establish” (US Const. art. III). The authority of the courts was further specified in section two of Article III, setting forth the jurisdiction of the high court in “all cases, in law and equity, arising under this Constitution, the laws of the United States, and treaties made, or which shall be made, under their authority” (US Const. art. III). These powers drew the Supreme Court closer Hamilton’s vision. Politicization of the Confirmation Process The selection process of Supreme Court Justices is also laid out within Article II. Commonly referred to as the “appointments clause,” Article II vests authority in the president to nominate judges, with the Senate responsible for confirming these decisions. The president was given power “by and with the advice and consent of the Senate… to appoint judges of the Supreme Court” (US Const. art. II). 6 Of course, with Hamilton’s desire for lifetime appointments to the court, it could be argued that the involvement of political figures in the appointment process via Article II causes an ideological bias to be planted in those appointed. Indeed, introducing political actors of any aspect could be seen to taint this process. Though the justices are set out to be neutral arbiters of the law, the selection process is full of anything but neutrality. As Calvin Massey (1991) notes, “history tells us that politics are at the heart of the confirmation process. There is no way to drain the politics from this swamp” (Massey, p. 16). This can be seen throughout the history of the nomination process. Over time, one out of five nominations have failed, with 12 being rejected outright by the Senate (Massey, p. 1). Both failed and controversial nominations happen for a variety of reasons. The most blatantly ideological occurs when the Senate wishes to express their general opposition to the president at a time when the president is particularly “politically vulnerable for reasons unrelated to the nomination process” (Massey, p. 2). The Senate’s stonewalling of Merrick Garland in 2016 is a contemporary example of how the Senate can wield power over a lame-duck president they disapprove of solely along ideological lines. Nominees can also be rejected due to their own political beliefs. In 1987, Robert Bork saw his nomination fail in part because “many of his prior positions on constitutional law were [seen as] sufficiently ‘out of the mainstream’ of American constitutional law,” making his confirmation a politically unwise move for the Senate (Massey, p. 6). 7 A failed confirmation can also take place due to doubts of the nominee’s integrity and character. In modern times, it is rare to see nominees be questioned solely on this basis, and oftentimes it can be applied with other factors. One famous example come from the nomination of Brett Kavanaugh. Kavanaugh was already facing an uphill battle given his nomination by an extremely polarizing president. Many Senators also questioned his loyalty to precedent and the doctrine of stare decisis. Furthermore, this opposition became supplemented by the explosive accusations of sexual assault that came forward during the process. Ultimately, not only was Kavanaugh nominated by a politically unpopular president, and his judicial philosophy was deemed too conservative, but his integrity was also called into question. One could argue the combination of all three factors in the instance of Justice Kavanaugh’s confirmation created the perfect firestorm necessary to ignite his hearing to new levels of contention and partisanship. Quickly, both Congress and the public became divided. According to a 2018 Gallup poll, toward the end of the hearing, 46% of Americans supported the nomination, while 45% did not, leaving only 9% with no opinion (Jones, 2018). Additionally, the 71-point gap observed between Republicans and Democrats on their position of the nomination, was “the largest Gallup has measured for a Supreme Court nominee to date, by nearly 20 points” (Jones, 2018). Partisanship within the Public Though the nomination process has seen its levels of conflict ebb and flow over the years, not all directly agree how partisan attitudes in different sectors have influenced the court itself. The public is seen as one area of influence. General data regarding the 8 partisan and polarized attitudes of the American public shows many Americans are becoming more partisan and polarized when it comes to politics. According to the Pew Research Center, median Republicans are currently more conservative than 97% of Democrats, and median Democrats are more liberal than 95% of Republicans (Kiley, 2017). To put it more bluntly, 45% of Republicans and 41% of Democrats believe the opposing party to be dangerous enough to pose a threat to the health of the nation (Doherty, Kiley, & Jameson, 2016). Further recent research on this topic has shown “evidence that partisanship influences economic behavior, even when it is costly,” indicating these numbers from Pew are having tangible effects on behavior (Doherty et al.). Much of the academic and political world has taken notice on these recent developments. However, the idea of polarization in the public is not completely contemporary. First seen “in the 1990’s, media and interpreters of American politics began to promulgate a polarization narrative” (Fiorina and Abrams, 2008, p. 564). And this belief has begun to take hold across politics. Not only was this narrative established by the media, but “it also seems to have influenced what actually happened in politics” (Fiorina and Abrams, p. 565). For example, traditional electoral strategies of the past have emphasized the need to persuade moderate swing voters. By contrast, a more polarized strategy of maximizing base voter turnout has become more commonplace (Fiorina and Abrams, p. 565). Despite this data, some political scientists studying the phenomenon of polarization say it is not increasing. Fiorina and Abrams argue that while “elite polarization” amongst Congressional members has increased, the overall rates of 9 polarization amongst the American public has not. While there is concern in the field that increasing elite polarization may lead to more polarization within the general public, they argue their research demonstrates otherwise. Fiorina and Abrams note this elite polarization has made an impact on increasing recognition of party differences, and led to a heightened sense that political outcomes matter (Fiorina and Abrams, p. 582). Yet, they believe the actual political positions of Americans have not shifted substantially. They posit that scholars often wrongly focus their attention on elite polarization, assuming it fuels further divides the general population. Although the proportion of politically active Americans believing “there are important differences between the two parties has risen 30 percentage points in the past half century… we do not see analogous trends when we look at distributions of normal people— typical Americans who are not deeply involved in politics” (Fiorina, 2016, p. 6). In other words, it is “partisans, not independents, who have lost ground: independents are now the largest single ‘partisan’ category” (Fiorina, p. 8). In an argument that will resonate well with civics pessimists, Fiorina believes the data shows “Americans instinctively seek the center” because they “are not very well informed about politics, do not hold many of their views strongly, and are not ideological” (Fiorina, p. 12). Certainly, party activists and donors have begun to shift further away from the center, but they generally only make up about 10 percent of a national sample (Fiorina, p. 4). Additionally, much of the work done by Fiorina seems to suggest political scientists have been examining the notion of polarity and partisanship incorrectly, using 10 voting decisions, which are limited indicators of polarity at best. He differentiates between polarized choices that happen when voters are offered highly partisan candidates, and actual polarized values, which are more difficult to ascertain. Though Fiorina tends to focus on what the actual political values of the American public are, there are further ways to consider polarization and partisanship. McGhee and Krimm (2009) believe viewing American’s party registration preferences is an effective method. Traditional surveys on public opinion can shed some light on this question, but “there is no guarantee that the dimensions of difference scholars have explored are politically relevant in the sense that they can be mobilized by the parties to form strongly opposing coalitions” (McGhee and Krimm, 2009, p. 349). Voting data is also an imperfect measure, since voters may grudgingly select a party candidate—like many claimed to do in the 2016 election—while remaining ideologically centrist and dissatisfied with the choices presented (Chozick & Thee-Brenan, 2016). In their analysis, McGhee and Krimm use party registration choices to measure the ideological distance between the two parties. Importantly, since many states give voters the opportunity to register as independent, it allows for a clear and concise measurement. Though many individuals registered as independent may seem like “closeted partisans,” there is a difference between leaning toward a party in voting, and complete “party zealotry” (McGhee and Krimm, p. 349). These independent registrants who exhibit a specific party lean tend to be only weak partisans. When measuring partisanship on these lines, McGhee and Krimm’s results echo many of Fiorina’s findings. In the states where party registration data is tracked, there has been a sharp rise of people identifying as independent, or declining to state their party allegiance entirely 11 (McGee and Krimm, p. 348). This is despite restrictions registrants might face in voting for candidates within closed primaries. With the exception of slight growth in regional partisanship in the southern United States, neither major party has made gains in the rest of the country. Looking beyond this slight regional change, the only major difference in party registration is the increase in independent registration (McGee and Krimm, p. 366). While McGee and Krimm’s examination of voter registration and polarization appears to align with Fiorina’s findings, others still contend the opposite. In a response to Fiorina’s research, Alan Abramowitz and Kyle Saunders argue data does in fact indicate a pattern of polarization within the American public. Abramowitz and Saunders (2009) call upon ideology and party identification survey data from the American National Election Studies (ANES) to make their case. When viewing this data over the last several decades, they contend a pattern emerges of increasing party polarization. When reviewing data between 1972 and 2004, “the difference between the mean score of Democratic identifier and the mean score of Republican identifiers on the 7-point liberal conservative scale [has] doubled” (Abramowitz and Saunders, 2009, p. 547). Additionally, when considering positions held on topics from abortion to reparations, the divide along party lines has increased significantly (Abramowitz and Saunders, p. 547). Even when factoring in geographical polarization, the authors argue states are becoming more saturated with their dominant party. Though on a national scale, elections are becoming more competitive, the results within individual states are not. When compared to presidential elections of the 60s and 70s, which had close elections on both the national and state level, the average margin of 12 victory in states “has increased dramatically over time” (Abramowitz and Saunders, p. 548). Another indicator of this trend is seen with presidential approval ratings. In a 2004 ANES survey regarding President George W. Bush’s performance, 90% of Republicans approved and 66% approved strongly; by contrast 81% of Democrats disapproved and 64% disapproved strongly (Abramowitz and Saunders, p. 547). These deep divisions along party lines regarding George W. Bush’s presidential performance were the highest division for any president since ANES began asking a presidential approval question in 1972. And this example of high polarity was not an outlier. Since Richard Nixon’s presidency in 1972, the approval gap between Republicans and Democrats has only grown wider. Another key part of Abramowitz and Saunders’s findings regards the type of person they see polarization affects the most. As mentioned above, Fiorina contends that Americans “instinctively seek the center” and are generally apathetic about politics. Abramowitz and Saunders agree that in general, the least informed and politically active individuals are clustered near the center. But the most interested, informed, and active citizens are much more polarized in their political views (Abramowitz and Saunders, p. 543). However, the divide amongst Americans on politics does not only exist along political activity lines. Abramowitz and Saunders also believe the level of elite ideological polarization in Congress reflects the real divides seen within the American public. Gender also has also generated inquests on its role in polarization. According to Baldassarri and Bearman (2007), “most intergroup polarization is mild,” and with respect 13 to gender, “we observe stability or even instances of depolarization in intergroup differences” (Baldassarri and Bearman, 2007, p. 786). Some have pointed to the increase in support that Republicans see from men as evidence of a partisan cleavage along gender. When Barack Obama defeated Mitt Romney in the 2012 presidential election, he “defeated Obama by seven percentage points among male voters” (Gillion, Ladd, Meredith, 2014, p.1). This occurrence marked “the fourth time in the last five presidential elections that a plurality of men supported the Republican nominee, while a plurality of women supported the Democrat” (Gillion et al., p. 1). While this is an interesting point to note regarding gender and political affiliation, it does not translate to greater polarization levels for one gender over another. Though females may anecdotally tend to support Democratic candidates over Republican, their propensity to become polarized is no different than men. Growth has been seen with polarization for “those who are politically active” as they “tend to have more extreme positions than the rest of the population,” but polarization “along the lines of gender” has not been observed (Baldassarri and Bearman, p. 787). Ironic as it is, it seems we are unable to fundamentally agree on whether or not polarization and partisanship exist within the public in the first place. Despite the internal debate, it still seems clear that at minimum, the engaged public who are voting in elections, donating to campaigns, and following the issues, are becoming less moderate in their views. This can have a real effect on the nomination process. After all, these politically involved individuals are more likely to closely follow the nomination hearings, let alone know they are happening. Their involvement can put pressure on their Senators to behave in a more polarized manner to maintain support amongst their constituents. 14 Despite these disagreements, public influence is certainly a key factor to consider. Fortunately, a clearer picture comes from Congress regarding so-called “elite polarization.” Elite Polarization Using data from Poole and Rosenthal’s well-known DW-Nominate scores, most academics in the field agree that Congress is the most partisan it has been since the late 19th century (Hasen, 2013, p. 234). A frequent estimate as to when elite polarization began to develop in earnest was following the “ideological realignment in the South following the Civil Rights era, as conservative Southern Democrats moved to the Republican party” (Hasen, p. 207). With this Congressional trend beginning during the Civil Rights era, the gap is only growing wider today. Richard Pildes notes how politically engaged members of the electorate are far more polarized than unengaged, similar to Abramowitz. But he extends his findings a step further by arguing this fact indicates that the engaged public is “mirroring the polarization” seen in Congress (Pildes, 2011, p. 279). The phenomenon of increased polarization amongst the engaged public has followed the trend set by Congress. Similar to the ANES data presented by Abramowitz, the proportion of loyal partisan members of Congress has increased since the late seventies. In 1976, moderates constituted 41% of the Senate. Yet today, that proportion is just 5% (Theriault, 2006, p. 493). A near-identical pattern can be observed in the House. This is in part due to more extreme members of Congress replacing moderate outgoing members, as well as member adaptation. Adaptation is the process of a legislator becoming markedly more conservative or liberal over the course of their time in 15 Congress in response to a shift in their constituency or allies in Congress. These adaptions, along with member replacements, have “fundamentally separated the political parties in Congress” even further, and thereby “decimated the ideological center” (Theriault, p. 488). Importantly, both of these mechanisms for increasing polarization within Congress are not the only contributing factor. Scholars have also examined the impact the president has on Congress. It appears obvious that members of Congress do not enjoy seeing “‘their’ President, and thus ‘their’ party embarrassed” (Lee, 2008, p. 914). This is because presidents are often noted to be leaders of their respective party, despite how powerful they truly are in the eyes of Congress (Lee, p. 914). When examining how presidential leadership influences polarization within the Senate, a measurable impact can be seen. Even with “the most ideologically divisive political issues on which individual Senators are likely to have relatively well-defined preferences,” how the President interacts with these issues can change partisan outcomes. When grouping all votes together involving ideological issues, Lee finds presidential leadership “systemically widens the gap between parties” (Lee, p. 924). This finding gives credence to the theory that presidential leadership “exerts a contrary pull on the parties in Congress, diving them farther apart on the issues that presidents champion” (Lee, p. 915). When the president publicly takes a side on an issue, it widens the distance between the parties on said issue in Congress. This means when a president makes a point to support their Supreme Court nominee, members of their party in the senate will likely fall in step with this staunch support. 16 Given this increasing gap between the two parties, an analysis of Congressional speech can also be used to show the real-world effects polarization has on our legislators. When viewing “hyperbolic criticism” of members outside of a Congressman’s party, it can be seen these comments are being driven by those on the ideological extremes (Blum, Goodman, Grimmer, Parker, 2018, p. 1). Through an examination of constituent facing communications—press releases and newsletters—the reality of polarized speech within Congress becomes more apparent. As many of the aforementioned studies find, polarization began to significantly permeate into Congress during the post-civil-rights era. It is undeniable that the two dominant parties in U.S. politics became stronger ideological entities around this time (Blum et al., p. 6). This has also been shown by a shift in political speech from this point onward. While legislators have always engaged in “credit claiming,” or relaying to their constituents that they were responsible for a recent positive outcome for their district, attacks are becoming far more prevalent. In the modern era, both Republicans and Democrats afforded similar time engaging in credit claiming up “until President Obama’s election, after which Republicans began attacking the Democratic party at seismically higher rates” (Blum et al., p. 16). Following their loss of the presidency in 2008, Republicans began to shift their strategy away from claiming credit in their districts, to putting out attack-focused rhetoric. This change for Republicans is strongest among legislators who are noted for being much more ideologically conservative. In light of Congress becoming more polarized in all facets, the increasing number of highly partisan members has seen a marked rise in polarizing speech. 17 There is little denying the increase in polarization the country has seen amongst elected officials in Congress. After ideological realignment during the civil rights era, the difference between parties has become more important. This is seen in a variety of ways. The number of true moderates has dropped significantly, attacks of the opposing party are on the rise, and presidential leadership can only further widen the divide. Each of these findings suggests that Congress itself could be a perfect storm for polarization when it comes to Supreme Court nominations. Polarization within the Supreme Court In addition to the public and Congress, the Supreme Court must also be considered as a source of polarization. In this era of polarized politics, some scholars have suggested even the court is starting to exhibit a partisan swing. However, this was not always the case. In the past, “partisan affiliation was not always a reliable predictor of judicial ideology” (Clayton and McMillan, 2013, p. 134). In fact, after appointing Justice Earl Warren, President Eisenhower famously quipped how the appointment was “the biggest damnfool mistake I ever made” (Clayton and McMillan, p. 134). Warren, a Republican appointed to the Supreme Court by Eisenhower, delivered the court’s unanimous opinion on Brown v. Board which reversed the now-maligned “separate but equal” doctrine. During the oral arguments, Eisenhower is reported to have asked Warren” to consider the perspective of white parents in the Deep South” (O’Donnell, 2018). Eisenhower opined that all these parents were “concerned about is to see that their sweet little girls are not required to sit in school alongside some black (O’Donnell, 2018). 18 Yet, in light of all the aforementioned data, the continued ideological polarization of both parties, in addition to the Supreme Court’s position settling contentious issues, the justices being appointed seem to exhibit increasingly ideological behavior. Data which estimates the ideological position of each judge illuminates changes in the Court’s median ideological position. Since the Roosevelt presidency, justices appointed by Republican presidents have carried conservative mean ideological scores, and the opposite is true for justices appointed by Democrats, exhibiting a mean liberal score (Clayton and McMillan, p. 135). Ideological scores can be calculated in a variety of ways. Some scholars have used methods which estimate the “ideology of each justice at the time of their appointment based on public perceptions in pre-confirmation newspaper editorials” (Clayton and McMillan, p. 135). Other methods involve estimating the “ideological position of each justice based on their voting behavior in each term” by placing “a justice’s ideal preference point on a common ideological continuum” (Clayton and McMillan, p. 135). Regardless of methodology, these scores have begun to diverge considerably from the center since Clinton’s presidency (Clayton and McMillan, p. 135). The mean ideological position of the court has also begun to trend more conservative in recent years, prior to fluctuating around the mean. This is consistent with the prior observation, since of the 16 justices appointed to the Supreme Court since 1968, 12 were nominated by Republicans. This may seem like an obvious observation—should it not be logical that conservative presidents tend to nominate conservative justices, and vice versa? But when tangible data is put behind this thinking, the reality of how politicized the court truly is 19 becomes starker, and less palatable. As mentioned earlier, Hamilton noted the dangers of the court becoming more partisan. Questions About Nomination Hearings and Polarization Upon further review of partisanship and polarity within American politics, several things become clearer. While debates regarding the overall polarity of the general public may not yet be resolved, the level of partisan behavior and polarity amongst the engaged public has increased (Abramowitz and Saunders, 2008). This is in step with the changes in partisan loyalty seen from Congress (Theriault, 2006). As fewer moderates and those willing to break from party ranks remain in Congress, fewer engaged voters exhibit party neutrality. While the polarization of both Congressmembers and their engaged constituents have begun to play off one another, partisan thinking has been reflected in the Supreme Court as well (Clayton and McMillian, 2013). Keeping these findings in mind, I have arrived at several research questions centered around whether the trend of polarization and partisanship within Congress and the public extended to Supreme Court hearings: RQ1: Has a trend in constitutional polarization in the hearings grown over the years, or only recently developed? RQ2: Does one party exhibit greater constitutional polarization than the other? RQ3: Are nominees becoming more partisan? RQ4: Does gender influence the level of constitutional polarization? RQ5: Is the level of polarization different when a Senator is defending their party’s nominee, or attacking the opposing nominee? 20 In order to investigate these questions, this paper undertakes a content analysis of Supreme Court hearings. This a particularly effective method of study, as the hearings are very public forms of communication with transcript proceedings. As was noted earlier, there appears to be a relationship between the rising level of “elite polarization” in Congress, and the decline in centrism amongst the engaged public. And the amount of engaged public on this issue is no small amount. More than 20 million Americans tuned in to view the explosive hearings of then Judge Brett Kavanaugh that were televised across six major networks (Bauder, 2018). Therefore, a look at the confirmation hearings could prove to be a valuable piece of evidence to include in this debate, and improve our understanding of the political state of America. If the hearings for the Supreme Court, our most independent and apolitical branch of government, have become highly partisan and polarized, the future of our judicial system could be in peril. Method This study utilized a quantitative approach to content analysis in order to dissect remarks by United States Senators and successful Supreme Court nominees during Supreme Court nomination hearings from 1971 until 2017. During each nomination hearing, transcripts are created as official records of what is said by both Senators and nominees participating in the hearing. Typically, the hearings proceed in a fairly organized format as Senators have limited time to ask questions of the nominee. This makes the transcripts relatively easy to parse as there are clear exchanges taking place. These transcripts served as the main corpora this study. Though nominations have been occurring since 1789 with the confirmation of Justice John Jay, the time period of 1971 to 2017 was selected as a result of the 21 availability of text suited for analysis ("Supreme Court," 2018). Transcripts of nomination hearings going back further do exist within Congressional records, but these documents grow increasingly difficult to analyze with time, as they are more difficult to translate into a suitable digital format. To make the process of organizing these transcripts easier, the R Street Institute generously compiled all of the hearings from 1971 to 2017 into an aggregate Microsoft Excel sheet1. This spreadsheet contained the nomination hearings for fifteen of the most recent successful Supreme Court nominees (excluding Brett Kavanaugh), as well as William Rehnquist’s hearing to become chief justice, for a total of sixteen datasets. This amounted to a total of over 2.18 million words spoken between 127 different unique speakers. Importantly, the hearing transcripts from R Street were culled to only include the opening statements of nominees and back-and-forth exchanges between speakers. This provided a cleaner dataset as the transcripts verbatim contain procedural speech and dialogue that would have created unnecessary noise in the data. Transcripts were also broken up by speaking turn by R Street, so that each row of data constituted one speaking turn for an individual. The words used within these speaking turns were considered the units of analysis for this project. Regardless of length, each turn was examined for different kinds of speech. To begin the analysis, aggregate data across 46 years was broken up by hearing, which could be filtered in a variety of ways. One column was added to notate the speaker’s party affiliation, or in the case of nominees, the party of the President they were nominated by for each speaking turn. Additionally, codes indicating gender were added 1 https://www.rstreet.org/2019/04/04/supreme-court-confirmation-hearing-transcripts-as-data 22 to the data (1 for male, 2 for female). Every speaker during the hearings was also given a unique identifying number which carried over between datasets. For example, Senator Charles Grassley participated in several hearings, and retained the speaker code of 21 throughout. After being sorted into spreadsheets by nomination – with each speaking turn coded for party, speaker, and gender – the raw data was cleaned to only contain the speech from each turn and the assigned codes. Constitutionally Polarized Language One of the cornerstones of this research was establishing a way to operationalize polarization for measurement through Computer Aided Textual Analysis. Unfortunately, creating a custom dictionary of polarizing terms for each hearing is an intensive process beyond the scope of this study. However, robust research already exists on this topic. To this end, data was pulled from a recent research paper on constitutional polarization from Pozen, Talley, and Nyarko (2019). In their paper, Pozen et al. took a computational approach to understanding constitutionally polarized language within Congress. Using text from the Congressional Record, they created a machine-learning algorithm that would attempt to predict a speaker’s ideological identification based solely upon their remarks. Then, the authors would measure polarization by “evaluating how easy or hard it was for [the]… algorithm to predict political ideology or party affiliation” (Pozen et al., 2019, p. 24). Using this algorithmic method, the authors searched through previous Congressional Record documents by year to determine what phrases from each year could be considered the most party distinctive or party “owned.” In order to ensure the party owned phrases being captured were accurately linked to the constitution, the 23 authors created various corpus dictionaries specifically for constitutional discourse. Their expansive dictionary includes hundreds of phrases specific to constitutional discourse, as well as “phrases that appear in the text of the Constitution and lack a common extraconstitutional usage” like “advice and consent” (Pozen et al., p. 19). The expansive dictionary was also applied to the lists generated on party-owned terms. Crucially, sorting by year allows for more accurate measurement throughout time as phrases like “voting rights” could have been owned by Congressional Democrats in some years, and more owned by Republicans in later years. As a result, their inquest produced several lists of party owned terms from 1873 up until 2016. These lists were adapted to this paper to search for terms within the nomination transcripts that would indicate partisanship in correspondence with the year of the nomination. For example, for Justice Lewis F Powell’s nomination, which occurred in 1971, party-owned terms were pulled from the Pozen et al. list based upon 1971. It is important to note that the use of this dictionary bank meant only a specific type of polarization was sought after. Because the dictionaries were specific to constitutional terms, only constitutional polarization could be measured. Therefore, other types of polarization present in the hearings were not necessarily captured. It is crucial to distinguish between different categories of polarization, as this carries importance for the implication of results. The method used in this paper only related to constitutional polarization, and thus only conclusions about polarization along constitutional lines can be drawn. Constitutional Polarization Computational Analysis 24 Once the party “owned” terms from Pozen et al. were organized into separate Excel spreadsheets by year corresponding with hearing, they were uploaded into LIWC2015, a computerized linguistic analysis software. The nomination transcripts were also uploaded in their cleaned format. Analysis was then run one hearing at a time. To conduct its search, LIWC2015 parses the uploaded text, searching each speaking turn for terms contained in a specified dictionary. Regarding the constitutionally polarized language dictionaries, each speaker was measured to determine their usage of both Republican “owned” or polarized terms, as well as Democratic terms. These two categories were also combined to reflect a speaker’s total use of polarized language, regardless of party affiliation. Following its analysis, LIWC2015 then produces percentages for the frequency of dictionary terms contained within a speaking turn. For example, during the Lewis F Powell hearings, for one speaking turn, Senator Philip Hart of Michigan returned a total polarization frequency of 1.52%. The figure of 1.52% essentially indicates that during a specific speaking turn, 1.52% of Senator Hart’s words were considered to be polarized. This also could also be seen as a Democratic polarization frequency of 1.52%, as the only constitutionally polarized terms used by Senator Hart during that speaking turn were Democratic. For each hearing, similar numbers were generated for every speaking turn. Affect As was discussed in the literature review, “credit claiming” and positive language, as well as attacks on opponents, are becoming far more common (Blum, et al., p. 6). The tone or emotion of words used can also be seen as a reflection of partisanship during the 25 hearings. Harsh attacks on hearing participants, or extreme praise can showcase polarization not captured by the constitutional polarization dictionaries. LIWC2015 contains several internal dictionaries which can measure various “basic emotional and cognitive dimensions often studied in social, health, and personality psychology” (Pennebaker et al., 5). One of these internal dictionaries concerns affect, and it was also used as a dimension in this study. Affect is a concept most frequently referred to in psychology describing the “experience or intensity” of urgency associated with the feeling of an emotion (Shouse, 5). In order to produce an affect dictionary, the creators of LIWC first collected words pertinent to this category often associated with psychological emotion rating scales. These collections of words were then rated by a group of four to eight judges who “qualitatively rated [the phrases] in terms of ‘goodness of fit’ (Pennebaker et al., p. 5). Once assessed by the judges, the dictionary was then tested and refined on various bodies of text. These scales were also tested to ensure internal reliability and external validity. For the purposes of this study, the affect dictionaries used pertained to negative and positive emotion. The negative emotion affect dictionary contained words such as “enemy, ugly, and nasty,” while the positive emotion dictionary encompassed words like “love, nice, and sweet” (Pennebaker et al., p. 3). Affect Computational Analysis In tandem with analyzing the hearing texts for Congressionally polarized language, each hearing also had its affect measured in terms of positive and negative emotion. Similar to the constitutional polarization angle, this measure was taken for each individual speaking turn for both nominees and Senators. If a speaking turn contained any 26 of the key phrases listed in the affect dictionary, LIWC2015 identified its presence and counted it as a percentage of the entire text for the speaking turn, just as it did for the constitutional polarization dictionaries. Additionally, a composite positive and negative emotion score for each entire hearing was calculated using an average from each speaking turn. The Senator’s and nominee’s emotionality was also cross compared. On one dimension, the “out party” or party whose members were not the same as the nominating president’s, were measured. In other words, Democratic Senator’s affect towards Republican nominated judges was measured for both positive and negative, and Republican Senators were measured on a similar level. Results The first research question (RQ1) concerned the change in polarization over time during Supreme Court nomination hearings. Based upon much of the research described in the literature review, it would certainly be unsurprising if speech used in the nomination hearings became more polarized in step with rising “elite polarization” from Congress (Hasen, 2013, p. 244). However, when examining the results, this is not the case. Over time, the hearings produced an average polarization score of 0.11, meaning on average, speaking turns contained 0.11% polarized words or phrases. This did see fluctuations from hearing to hearing, but no single nomination scored higher than 0.19 or lower than 0.05. As a result, the overall trend seen in Figure 1 shows a slight net decrease in polarization over time from 1971 to 2017. It is particularly interesting to note that the first hearing in this study 27 for Justice Powell yielded a score (0.09) not unlike the most recent hearing examined for Justice Gorsuch (0.10). Average % of Polarized Words per Speaking Turn Figure 1 - Average Constitutional Polarization 0.25 0.20 0.15 0.10 Aveage Const. Polarization Tr endl ine 0.05 0.00 w Po 1 ell 1 97 h Re t uis nq 7 6 17 86 81 10 05 91 75 93 09 06 94 90 98 98 19 19 20 20 19 19 20 20 20 19 19 19 )1 y1 o ts or er er as rg or ch lia an ns ef lit ed er m ut ay su ca ey ag hi ve bu A nn n b r o r S o C s o K e n m ( B S n C Ro Th St Go to Ke Gi O' is t So qu hn e R 7 19 1 Judge and Year of Hearing Given the results showing net polarization has subtly trended downwards, it was also of interest to see the results of RQ2, pertaining to difference between parties and their use of polarized language. The research used to create constitutionally polarized language dictionaries for this study from Pozen et al., found Republicans “have been responsible for the post-1970’s rise in constitutional polarization” (Pozen et al., p. 40). This is partly due to Pozen et al.’s belief that members of Congress have grown more internally unified, and therefore more externally divided (Pozen et al., p. 31). Already, the initial results refuted the notion of a “post-1970’s” rise in constitutional polarization during the hearings. But, would Republicans still account for a majority of the polarization? Figure 2 displays the average constitutional polarization score by hearing broken up by party identification. In general terms, it appears that hearings which are high or low-scoring are so for each party. In other words, if one party used particularly polarizing 29 nomination to rise to such high levels of polarization (McMahon, p. 929). Clearly, the Republicans resorted to more constitutionally polarized language to make their argument. The element of Senate control brings into perspective another interesting way of examining the polarization present in a hearing and related to RQ3. Figure 3 shows the average nominee polarization, but also accounts for who controlled the Senate. The color of the bar indicates the party who the Justice was nominated by (or the party of the president), and the purple colored bars reflect a difference in the party who controls the Senate. Here it becomes easier to see that of the five most polarized hearings, four were hearings in which the Senate was controlled by a different party than the nominating president. It also can be seen that of the five least polarizing hearings, four were when the Senate was controlled by the same party as the nominating President. Another key question for this study concerned the role gender played in polarization (RQ4). Generally, studies confirm that “those who are politically active… tend to have more extreme positions” on politics, but “polarization along the lines of gender” is far less common (Baldassarri and Bearman, p. 787). Thus, questions arose 30 surrounding gender in the hearings, both for the Senators asking questions and the nominees responding. On the broadest level, an overall comparison can be made when looking at constitutional polarization levels by gender. Unfortunately, only half of the hearings had women participating, and even in those hearings contributions made by women were scant. Additionally, many of the female senators on the judiciary committee did not change from year to year. In total, only nine women—including both nominees and Senators—participated in Supreme Court Nomination hearings from 1971 to 2017. If anything, this fact serves to hammer home the stark reality that men have played an outsized role in these monumental decisions. Nevertheless, trends in polarization for gender were monitored as a total average throughout all the hearings. Women averaged a polarization level of 0.11, while men averaged a total of 0.12. Clearly, this overall average trend fits with the observation that polarization along the lines of gender is not a common occurrence. Even amongst political elites, it would appear that gender does not play a significant role. Seeing consistency in gender polarization for the hearings as a whole, individual hearings were also examined for their trends. Figure 4 shows the average constitutional polarization by gender for each hearing. Even when broken down by individual hearing, the trend of men and women showing the same amount of polarization across time is maintained with one notable exception. Justice Breyer’s hearing presents itself as an outlier, with the females participants in that hearing showing significant polarization levels around 0.26, while the male participants returned a drastically lower score of 0.05. 32 One of the most immediate observations to be made is that the frequency of positive remarks is greater than the frequency of negative remarks for every hearing. That is to say, generally speaking, speech used in hearings is more positive than negative, regardless of political affiliation, or who is speaking to whom. However, looking at party trends more closely gives a different perspective. Surprisingly, it does not appear that Republican Senators tend to use greater positive speech than Democrats when Republican nominees are involved. Looking at Figure 5, it is clear when Republican nominees are involved, the positive language used is fairly identical between parties. As the rate of positive language use goes up or down for Democrats toward a Republican nominee, Republicans follow suit. Figure 5 - Democrat vs Republican Positive Affect toward Republican Nominees Avg % of Polarized Words Per Speaking Turn 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Powell 1971 Rehnquist 1971 Stevens 1975 O'Connor 1981 Scalia 1986 Rehnquist (Chief) 1986 D + Affect to R Nominee Kennedy 1987 Souter 1990 Thomas 1991 Roberts 2005 Alito 2006 Gorsuch 2017 R + Affect to R Nominee Negative affect was not quite the same story. Though its frequency was lower than positive affect, Republican and Democratic use of it was not closely linked. Generally, Republican negative affect towards Republican nominees was lower. Interesting notable exceptions were for John Paul Stevens’s hearing in 1975, where Democratic negative affect was significantly higher, and David Souter’s hearing in 1990, where Republicans were higher. This can be seen in figure 6. 33 Figure 6 - Democrat vs Republican Negative Affect toward Republican Nominees Avg % of Polarized Words per Speaking Turn 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Powell 1971 Rehnquist 1971 Stevens 1975 O'Connor 1981 Scalia 1986 Rehnquist (Chief) 1986 R - to R Nominee Kennedy 1987 Souter 1990 Thomas 1991 Roberts 2005 Alito 2006 Gor such 2017 D - to R Nominee When looking at affect used toward Democratic nominees, a clearer picture can be seen. Democrats and Republican’s use of positive affect towards Democratic nominees was on average higher than the positive affect towards Republican nominees. Additionally, the trend seen in positive affect between Democrats and Republicans mirrors each other. Though Republicans typically used fewer positive remarks than Democrats in this instance, they never use significantly less positive affect than their Democratic counterparts. Figure 7 - Democrat vs Republican Positive Affect toward Democratic Nominees Avg % of Polarized Words per Speaking Turn 6 5 4 3 2 1 0 Ginsbur g 1993 Breyer 1994 D+ Affect to D Nominee Sotomayor 2009 R+ Affect to D Nominee Kagan 2010 34 Negative affect towards Democratic nominees painted a separate picture. Democratic negative affect on this scale showed a remarkable consistency. For the four hearing in question, Democrat’s use of negative language wavered between 0.63, and 0.66. The result is a trend with almost no variation. On the other side, Republicans use of negative language has only increased over time. Starting at 0.65, and nearly equal to Democrats for Justice Bader Ginsburg in 1993, Republican’s negative affect frequency has only climbed, to finish at 1.00 for Justice Kagan in 2010, as seen by Figure 8, creating a growing gap. Figure 8 - Democrat vs Republican Negative Affect toward Democratic Nominees Avg % of Polarized Words per Speaking Turn 1.2 1 0.8 0.6 0.4 0.2 0 Ginsburg 1993 Breyer 1994 D - Affect to D Nominee Sotomayor 2009 Kagan 2010 R - Affect to D Nominee Discussion Implications: Overall, the results of this study showed a variety of polarization levels across several different measures. Though constitutional polarization has always been present in the time period studied, it has decreased slightly over time. This is surprising for several reasons. 35 First, there is evidence to believe that members of the public, or at least involved members of the public, have begun to shift further away from the center (Fiorina, p. 4). While members of the public do not directly play a role in the nomination process, they still can influence the proceedings. Senators facing a divided and polarized constituency, when paired with an election year, can feel greater pressure to bend to the will of their potential voters. Additionally, public interest groups have begun to play a greater role in the nomination process. In an era where “presidents place near-exclusive focus on ideological compatibility and reliability,” when considering nominees, it grows increasingly important to accurately pick judges who reflect a presidential administration’s values (Hasen, 2019, p. 5). Interest groups can play a crucial role in aiding this pick. Nowhere is this more evident than the influence of the Federalist Society, a private organization of conservative jurists. As Hasen (2019) notes, during President Donald Trump’s term, nearly all of the federal judges he has appointed have some connection to the conservative group (Hasen, 2019, p. 5). Despite the presence of competitive elections and the increasing influence of groups like the Federalist Society, constitutional polarization across time has dropped slightly. This result becomes more shocking when considering the state of “elite polarization,” or the increased partisanship of those in Congress. Since the early 70’s, when moderates constituted nearly 41% of the Senate, this number has steadily dropped, where it now hovers at around 5% (Theriault, p. 493). This “decimation of the ideological center” would logically suggest greater partisanship seen during the nomination hearings over (Theriault, p. 488). Furthermore, after the election of 36 President Obama in 2008, Republicans began attacking the Democratic party at seismically higher rates” (Blum et al, p. 16). What better stage to attack your political opposition on than vitally important judicial nomination hearing which has captivated the attention the nation? In light of the findings of increased public polarization and greater influence on the hearings, combined with increasing “elite” polarization, it is surprising to see that constitutional polarization has decreased. Given this reality, it appears that members of the senate judiciary committee take their role in the nomination process rather seriously. This is counter to previous beliefs that “politics are at the heart of the confirmation process” and there is “no way to drain the politics” from it (Massey, p. 16). However, it is important to recall this study measured constitutional polarization. This means there was likely indications of polarization in the transcript that were not picked up by the dictionary used. Other ways of operationalizing polarization may return less surprising trends. Another important finding from this research, is that Republican’s use of negative language, or affect, has only increased over time when they are facing a Democratic nominee. Perhaps this indicates an increased sense of importance of the judiciary for the Republican party. Members of the Republican party seem to be more willing to fight tooth and nail—using negative language—to attack Democratic nominees. On a broader scale, a recognition that the confirmation process has not been thoroughly tainted by politics is of high value to the court. According to a recent poll by Marquette University’s law school, the Supreme Court has the highest confidence rating out of several different government instructions, including Congress, the presidency, and 37 local court systems (Conway). Similarly, when respondents were asked which branch of the federal government they trusted the most, the Supreme Court and judicial branch secured the top spot, with 57% indicated they trusted it the most (Conway). Additionally, 50% of respondents indicated they believed the court exhibited a moderate position, rather than liberal or conservative, and 64% believed the Justices’ decisions mainly follow the law, rather than politics (Conway, 2019). While these results could help increase confidence in Congress’ ability to perform its job, they are even more important for the Supreme Court. The entire foundation of our independent judiciary hinges on an element of trust. As Alexander Hamilton noted in Federalist 78, the Supreme Court “exercise neither force nor will, but merely judgement.” (Hamilton, 2008). The court frequently hands down decisions that large portions of the country may find untenable. As Ruth Bader Ginsburg noted in an interview with NPR, "the court has no troops at its command, doesn't have the power of the purse, and yet time and again, when the courts say something, people accept it.” (Totenberg, 2019). Limitations and Future Study: Though this study did produce an intriguing dataset to consider for polarization and the judiciary, there were several limitations. One of the biggest limitations exists in the methodology. Two separate dictionaries were used to parse the text for polarized phrases and affect. Undoubtedly, the polarization dictionary was able to capture words which were considered to be “party owned” for the time, and the affect dictionary captured different emotional concepts, these were not all-encompassing. For future study, a greater number of dictionaries with larger word banks and could be utilized. Though it would be intensive work to ensure all these dictionaries were 38 valid, a more exhaustive and expansive use of dictionaries would better ensure more aspects of polarization were captured, thus providing even more accurate data. Another limitation concerned the use of a computer-aided software analysis like LIWC2015 and the dictionaries themselves. Though the number of social science and humanities scholars recommending these kinds of software for large-scale textual analysis is growing, its application is still imperfect. This particularly can be an issue for political texts. Frequently, “when dictionaries are created in one substantive area and then applied to another, serious errors can occur” with validity (Grimmer and Stewart, 2013, p. 8). Political texts are often very context specific, and when dictionaries are taken from one context and applied to another, the fit will not always be perfect. Even something as broad as affect analysis, existing dictionaries can “establish granular scales” for a specific sentiment, it is “essentially impossible to derive gold standard evaluations” for scales for emotions like fear or happiness (Grimmer and Stewart, p. 9). For this study, the political polarization dictionary used was taken from a study which examined “party owned” phrases spoken during Congress for specific timeframes. While this is much more applicable than a generic dictionary, greater specificity to the nomination hearings could prove to be beneficial as much of what is said is very context dependent. Given the existing skepticism of dictionary based measures, future studies may choose to incorporate a combination of computer coding as well as supervised human coding. Though the inclusion of human coders can be cost and time prohibitive, it allows for greater flexibility and adaptability to the context of each hearing. The dictionaries 39 used in this study, for example, did not account for phrases like “sexual assault” which would have surfaced in the Thomas and Kavanaugh hearings, because they were not pertaining to the constitution. Hearing-specific phrases like this can be better captured by human-centric coding. Another limitation for this study was the lack of the Kavanaugh hearing transcripts. It was this explosive nomination that sparked much of my interest in this research. However, full transcripts from this hearing have yet to be made available. Another possible area for future study, would be the inclusion of more “Article III judges”, or federal judges who must go through the nomination and confirmation process, like circuit and district court judges. Supplemental studies should be sure to include these aspects to paint an even clearer picture of polarized language within our federal judiciary. Conclusion This study was initially born out of the chaos of Brett Kavanaugh’s Supreme Court nomination hearing, which left much of the country reeling from the ensuing partisan firestorm. It seemed that the sanctity of the Supreme Court had been compromised, and trust was eroding. A basic question was then proposed: How partisan were the nomination hearings? Quantitatively speaking, the words used in the hearings could provide some foundation of understanding. While this paper mostly probed the presence of constitutional polarization, the constitution is at the heart of the hearings, and thus could be considered the most important subset of polarization to examine. Surprisingly, it does not appear that Senators or nominated justices speak in polarizing terms, at least when it comes to the constitution. These data suggest that it might be other political issues that can cause contention at 40 hearings. Though this implication does not lead to a total exoneration of those participating in the nomination hearings, it does bring a level of confidence to the process. Supreme Court Justices are charged with being the final interpreters of the constitution. To know that Senators and nominees are using constitutionally polarized phrases less frequently as time has gone on means their commitment to unbiased constitutional language has improved—albeit only slightly. The country places a great deal of trust in the Supreme Court. They are charged with handing down decisions on some of the most fundamentally difficult questions our society faces. While some may disagree with their interpretations, those decisions still become the law of the land. An independent and trustworthy judiciary is critical to the health of our democracy. It is imperative that we place a genuine trust in and promote an apolitical nature to the nomination process. As Chief Justice Charles Evans Hughes remarked about the Supreme Court upon the completion of the Supreme Court building’s construction in 1932, “the republic endures, and this is the symbol of its faith.” 41 References Abramowitz, A. I., & Saunders, K. L. (2008). Is polarization a myth? The Journal of Politics, 70(2), 542-555. Retrieved from The University of Chicago Press database. Baldassarri, D., & Bearman, P. (2007). Dynamics of political polarization. American Sociological Review, 72(5), 784-811. https://doi.org/10.1177%2F000312240707200507 Bauder, D. (2018, September 28). More than 20 million people watched the brett kavanaugh hearing. The Chicago Tribune. Retrieved from https://www.chicagotribune.com/nation-world/ct-brett-kavanaugh-hearing-tvratings-20180928-story.html Blum, R., Goodman, C., Grimmer, J., Parker, D. C.W., & Zlotnick, F. (2018, July 17). Disunity in the ranks: Ideological extremism and out-party status in congressional rhetoric. Retrieved from SemanticScholar database. Chozick, A., & Thee-Brenan, M. (2016, July 14). Poll Finds Voters in Both Parties Unhappy With Their Candidates. Retrieved March 31, 2020, from https://www.nytimes.com/2016/07/15/us/politics/hillary-clinton-donald-trumppoll.html Clayton, C. W., & McMillan, L. K. (2013). The roberts court in an era of polarized politics. The Forum, 10(4), 132-146. Retrieved from Research Gate database. Conway, K. (2019, October 21). New nationwide marquette law school poll finds confidence in U.S. supreme court overall, though more pronounced among conservatives. Retrieved March 4, 2020, from https://www.marquette.edu/news- 42 center/2019/new-nationwide-mu-law-school-poll-finds-confidence-in-ussupreme-court-overall.php The court and constitutional interpretation. (2019). Retrieved September 9, 2019, from https://www.supremecourt.gov/about/constitutional.aspx Declaration of independence. (2019, November 19). Retrieved March 19, 2020, from https://www.archives.gov/founding-docs/declaration-transcript Doherty, C., Kiley, J., & Jameson, B. (2016, June 22). Partisanship and political animosity in 2016. Retrieved September 24, 2019, from http://assets.pewresearch.org/wp-content/uploads/sites/5/2016/06/06-22-16Partisanship-and-animosity-release.pdf Fiorina, M. P. (2016, September 14). Has the american public polarized. Retrieved September 28, 2019, from https://www.hoover.org/research/has-american-publicpolarized Fiorina, M. P., & Abrams, S. J. (2008). Political polarization in the american public. Annual Review of Political Science, 11, 563-588. Retrieved from Annual Reviews database. Gillion, D. Q., Ladd, J. M., & Meredith, M. (n.d.). Education, Party Polarization and the Origins of the Gender Gap in U.S. Party Identification. University of Pennsylvania, 1-61. Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.639.814&rep=rep1&ty pe=pdf 43 Grimmer, J., & Stewart, B. M. (2013). Text as data: The promise and pitfalls of automatic content analysis methods for political texts. Political Analysis, 21(3), 267-297. https://doi.org/10.1093/pan/mps028 Hamilton, A. (2008). The federalist papers: No. 78. Retrieved September 9, 2019, from https://avalon.law.yale.edu/18th_century/fed78.asp Hasen, Richard L. "Polarization and the Judiciary." Annual Review of Political Science, May 2019, pp. 1-30. SSRN, doi:10.1146/annurev-polisci-051317-125141. Hasen, R. L. (2013). End of the dialogue? Political polarization, the supreme court, and congress. Southern California Law Review, 86(2), 205-262. Retrieved from HeinOnline database. Jones, J. M. (2018, October 3). Americans Still Closely Divided on Kavanaugh Confirmation. Retrieved March 31, 2020, from https://news.gallup.com/poll/243377/americans-closely-divided-kavanaughconfirmation.aspx Kiley, J. (2017, October 23). In polarized era, fewer americans hold a mix of conservative and liberal views. Retrieved September 24, 2019, from http://www.pewresearch.org/fact-tank/2017/10/23/in-polarized-era-feweramericans-hold-a-mix-of-conservative-and-liberal-views/ Lee, F. E. (2008). Dividers, not uniters: Presidential leadership and senate partisanship, 1981-2004. The Journal of Politics, 70(4), 914-928. Retrieved from JSTOR database. 44 Massey, C. R. (1991). Getting there: A brief history of the politics of supreme court appointments. Hastings Constitutional Law Quarterly, 19(1), 1-16. Retrieved from HeinOnline database. McGhee, E., & Krimm, D. (2009). Party registration and the geography of party polarization. Polity, 41(3), 345-367. Retrieved from JSTOR database. McMahon, K. J. (2007). Presidents, political regimes, and contentious supreme court nominations: A historical institutional mode. Law and Social Inquiry, 32(4), 919954. https://doi.org/10.1111/j.1747-4469.2007.00083.x Montanaro, D. (2018, September 26). New Poll Finds Americans Deeply Divided By Gender On Kavanaugh Nomination. Retrieved March 31, 2020, from https://www.npr.org/2018/09/26/651647131/poll-nearly-6-in-10-to-closely-watchkavanaugh-ford-hearing-many-undecided-on-tr O'Donnell, M. (2018, April). Commander v. Chief: The lessons of Eisenhower's civilrights struggle with his chief justice Earl Warren. Retrieved March 31, 2020, from https://www.theatlantic.com/magazine/archive/2018/04/commander-vchief/554045/ Pennebaker, J. W., Boyd, R. L., Jordan, K., & Blackburn, K. (2015). The development and psychometric properties of liwc2015. Retrieved January 10, 2020, from https://repositories.lib.utexas.edu/bitstream/handle/2152/31333/LIWC2015_Lang uageManual.pdf Pildes, R. H. (2011). Why the center does not hold: The causes of hyperpolarized democracy in america. California Law Review, 99(2). Retrieved from 45 https://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?article=1039&context=c alifornialawreview Pozen, D., Talley, E., & Nyarko, J. (2019, September 19). A computational analysis of constitutional polarization. Retrieved January 30, 2020, from https://scholarship.law.columbia.edu/faculty_scholarship/2271/ Shouse, E. (2005). Feeling, emotion, affect. M/C Journal, 8(6). Retrieved from http://www.journal.media-culture.org.au/0512/03-shouse.php Supreme court nominations (Present-1789). (2018). Retrieved January 31, 2020, from https://www.senate.gov/legislative/nominations/SupremeCourtNominations1789p resent.htm Theriault, S. M. (2006). Party polarization in the US congress: Member replacement and member adaptation. Party Politics, 12(4), 483-503. Retrieved from SAGE Journals Online database. Totenberg, N. (2019, July 24). Justice ginsburg: 'I am very much alive'. Retrieved March 8, 2020, from https://www.npr.org/2019/07/24/744633713/justice-ginsburg-i-amvery-much-alive U.S. Const. art. III. Retrieved from https://www.law.cornell.edu/constitution/articleiii Weissmann, S., & Marcum, A. (2019, October 7). Supreme Court Confirmation Hearing Transcripts, As Data [Microsoft Excel]. Retrieved from R Street Institute website: https://www.rstreet.org/2019/04/04/supreme-court-confirmation-hearingtranscripts-as-data/