Cameras in the High Court: An Empirical Examination of Support for Supreme Court Justices – Black, Owens, Johnson, and Wedeking

Posted by on Apr 18, 2023 in Per Curiam

Cameras in the High Court: An Empirical Examination of Support for Supreme Court Justices – Black, Owens, Johnson, and Wedeking
Download PDF

Cameras in the High Court: An Empirical Examination of Support for Supreme Court Justices

Ryan C. Black,* Ryan J. Owens, Timothy R. Johnson, and Justin Wedeking§

On March 16, 2023, Senators Dick Durbin (D-IL) and Chuck Grassley (R-IA) introduced legislation that would require the United States Supreme Court to allow cameras in its Courtroom.[1] The bill instantly made news.  It provided a rare splash of bipartisan color in an otherwise black and white polarized Washington, D.C. That the bill also might trigger a direct confrontation with the High Court made it even more newsworthy. But while the bill made news, it hardly represented a new effort. For years, politicians, the media, and interest groups have lobbied and rattled sabers to dragoon the Court into allowing cameras in its Courtroom. Throughout, the Court has remained unshakably opposed.

In a study we published in Political Behavior,[2] we discovered that Justices have legitimate reasons to be cautious about placing cameras in their courtroom. There, we examined how people’s exposure to oral argument video footage changed their attitudes about judicial legitimacy. We found that, under certain conditions, cameras can help enhance the Court’s legitimacy. Yet, under other conditions, cameras can damage the Court’s legitimacy. From our perspective, the fact that exposure to a single snippet of footage could immediately impact a person’s general attitudes towards the judiciary was noteworthy. Those results underscore the consequential nature of the proposed legislation.

But we assert here that it is not just the Court as an institution that cameras will affect. Cameras also could influence what people think about individual Justices. While Justices are not above public scrutiny, it is important to know what influences people’s views of them and how those views might change. That is, how might exposure to video change what people think about the judges they observe? What is more, with recent threats on Justices’ lives, it seems particularly critical to know how the public’s views on Justices may change as a function of cameras. We do so here.

As with our earlier findings, the results here are mixed. On the one hand, there is some good news for people who want cameras in the Supreme Court. Respondents who watched a neutral exchange between a judge and attorney wound up supporting the judge more than those who listened to that neutral exchange. Additionally, respondents thought no less (nor more) of a judge when they watched him in a contentious exchange with an attorney versus listening to that exact same contentious exchange.

On the other hand, there is also some worrisome news. Respondents supported the judge significantly less when they watched him engaged in a contentious exchange versus watching him in a neutral exchange. And this effect existed in two different styles of video presentation: a dynamic “Zoom-like” format that focused on the individual speaker and a static, wide-angle, shot of the full bench and attorney’s lectern. We say that these results are worrisome because the media may be more likely to disseminate contentious exchanges than neutral exchanges. Viewed on the whole, then, these findings counsel caution and more investigation before the Court adopts cameras.

We begin by briefly presenting the arguments for and against cameras. We then explain our experiment, how we measured our variables, and then present our results. We conclude with a discussion of what may come to pass if the Court adopts cameras.

Cameras at the Supreme Court

On Thursday March 21, 2013, spectators began forming lines to reserve seats at oral arguments for two highly salient cases: US v. Windsor and Hollingsworth v. Perry.[3] Interestingly, neither of those cases were on the Court’s hearing list for the day. Nor were they on the list for the next day or even the following Monday. The Court heard Hollingsworth on Tuesday, March 26th—a full five days later—and Windsor on Wednesday, March 27th —a full six days later.

It is probably a sign of the Court’s institutional strength that so many people waited in line to observe oral argument; but spectator space is limited. There are typically fewer than 100 seats available for those who stand in line to watch the Court.[4] These space constraints might be irrelevant if people could watch the Court’s arguments via television or livestream. But they cannot.  The Court repeatedly has denied requests by various entities to place cameras in the courtroom. As a result, many people continue to push for cameras and have lobbied Congress in such efforts.

Those who support cameras at the High Court believe cameras will make the Court more transparent and accountable to the public. They argue that those features would increase the Court’s legitimacy. Similarly, cameras might educate and inform the public about the Court. People would learn about the Court; and, after all, scholarship shows that to know the Court is to love the Court.[5] In this vein, Justice Elena Kagan once said that cameras might “allow the public to see an institution working thoughtfully and deliberately and very much trying to get the right answers, all of us together.”[6]

But others worry about the effects of cameras. Some Justices opine that the media will reduce a complex oral argument to a brief, unrepresentative, video clip that make the Justices look bad. Justice Antonin Scalia expressed this apprehension:

I am against it [cameras] because I do not believe, as the proponents of television in the Court assert, that the purpose of televising our hearings would be to educate the American people. That’s not what it would end up doing. If I really thought it would end up educating the American people I would be all for it . . . but they wouldn’t see all of [what we do] . . . .[W]hat most of the American people would see would be 30-second, 15-second takeouts from our argument, and those takeouts would not be characteristic of what we do.[7]

Other Justices worry that televising oral argument will change the way they and their colleagues behave and thereby lessen the value of oral argument. For example, Chief Justice John Roberts once declared: “I do think that cameras in the Courtroom would impede [the] process . . . . I think that if there were cameras, the lawyers would act differently. I think, frankly, some of my colleagues would act differently and that would affect what we think is a very important and well-functioning part of the decision process.”[8] Other Justices echo this concern. Justice Elena Kagan once stated that cameras could cause Justices to “filter ourselves in ways that would be unfortunate.”[9] These Justices are concerned that televising oral argument not only will influence the Court’s legitimacy but also will force Justices to change their behavior so as to avoid public blowback.

The question remains whether people would in fact change their views of the Justices if they watched them at oral argument. If so, it stands to reason that the Justices themselves might change in response. And so, we sought to investigate whether people’s views of the Justices might change. Given that we already have published a study on how exposure to video might affect the Court’s legitimacy, it seemed that the logical next step would be to examine how the public’s views of the Justices they watch may change. It is to this task that we now turn.

The Survey Experiment

We implemented a survey experiment that exposed respondents to real video clips from two state supreme courts. The experiment relied on 1,475 respondents and used a 2 (video v. audio) x 2 (contentious v. neutral) x 2 (dynamic v. static) plus control design. The treatment exposed individuals to a single 50–60 second clip of one of two state supreme court oral argument exchanges between an attorney and a Justice. Some of our respondents listened to an oral argument while others watched it. Some respondents observed a Justice aggressively question an attorney while others observed a polite and deliberate exchange between the two. And some respondents saw a dynamic camera angle that showed the judge and the attorney in a Zoom-like format while others observed a static, wide-angle shot of the full bench from a distance. After observing these oral argument clips, we asked respondents a series of questions about their views of the judges they observed. This included questions about judicial legitimacy, which we have discussed elsewhere,[10] as well as questions about the specific judge whose behavior our survey respondents observed.

Respondents watched or listened to clips from the Minnesota and Indiana supreme courts. We selected these two courts for a handful of reasons. First, we sought to mitigate possible confounders like respondents’ preferences for or against certain accents. The two accents in our sample were quite similar. Therefore, respondent differences will not be a function of accent preferences. Second, we needed videos that were relatively high quality. Few states provided video footage, let alone consistently good footage that we could use. These states were among the best. The states also provided important variation for one of our treatments (camera angle), with Indiana using the dynamic Zoom-style format and Minnesota employing a static, wide-angle approach.[11]

We utilized Lucid Theorem to execute an experiment that focused on the effects of watching and listening to portions of an oral argument.[12] Though a convenience sample, Lucid samples provide demographic and experimental results that track well with U.S. national benchmarks[13]  and are very common in experimental studies like ours.[14]

Because we seek to examine how the public’s views on Justices will change because of watching oral argument, we calculate a Judge Support Score for each respondent. The score is a composite of answers to three separate questions: (1) Whether the respondent believed the judge behaved professionally, (2) whether the respondent had confidence in the judge’s ability to do his job, and (3) whether the respondent believed the judge would decide future cases fairly. For each of these statements, our participants could say they strongly agreed, somewhat agreed, neither agreed nor disagreed, somewhat disagreed, or strongly disagreed. As an empirical matter, we have good evidence that responses to all three of these questions is driven by a single “latent force,” which allows us to condense them into the single Judge Support variable.[15] Larger values indicate more support for the judge (e.g., agreement that he behaved professionally) and smaller values indicate less support (e.g., disagreement that he behaved professionally).


As was the case with our original study, the results here are mixed. But taken as a whole, we believe they too counsel caution on this important decision.

We begin with an analysis of respondent views toward judges after exposure to a neutral exchange between the judge and an attorney. Again, these are clips where a Justice and attorney engaged in a polite exchange over a legal point. Figure 1 displays the effect that the medium (i.e., audio or video) had on average levels of Judge Support. These results show that people who watched a neutral exchange from the dynamic angle held more positive views toward the judge than people who listened to that exchange. In the left half of the figure, we see that respondents who listened to the neutral exchange in the Indiana court held a Judge Support score of approximately 0.77 but those who watched that neutral exchange held a Judge Support score around 0.81. The difference between these two values is statistically significant. Somewhat surprisingly,[16] we observe no differences between listeners and watchers in the neutral Minnesota case. On balance, this is a positive finding for those who want cameras in the Court.

Interestingly, Figure 2 shows that when we examine the difference between audio and visual in the contentious context, there are no differences in Judge Support among our respondents. Those who listened to the contentious exchange in the Indiana context had a support score of 0.72 while those who watched the contentious exchange had a score of around 0.74—a difference that does not approach standard levels of significance. The results are similar in the Minnesota case. Those who listened to the contentious exchange had a support score of 0.66 while those who watched the contentious exchange had a score of 0.68. These results would seem to support those who want cameras in the court, at least to the extent that they satisfy the “do not harm” standard.

Our last analysis is perhaps the most revealing and, from the Justices’ perspective, the most concerning. Here, we examine how respondents differentially react when exposed to a video clip of contentious versus neutral exchanges between a Justice and an attorney. We contend this comparison is revealing because it partially accounts for the likely filtering role that the media would play in a world where cameras are allowed. Editors would showcase contentiousness over neutral content.[17]

The question here is whether respondents who are exposed to the contentious video think any differently of the Justices than those exposed to the neutral video. As Figure 3 shows, respondents supported judges significantly less after watching a contentious exchange versus watching a neutral exchange. What is more, this deleterious effect exists in both styles of video presentation that we examined.

Consider respondents exposed to the Indiana clip. Those who watched the neutral exchange had a support score of 0.81 while those who watched the contentious exchange had a support score of 0.73, a statistically significant difference. While this 10 percent relative decrease is not overwhelming in terms of its substantive magnitude, recall that it obtains after exposure to just a single video clip. Iterative exposures could yield larger impacts.

Despite a less dramatic and engaging format, the Minnesota format produces similar effects. Respondents who watched a neutral exchange had a support score of 0.81 while those who watched a contentious exchange had a support score of 0.68, a relative decrease of 16 percent that is also statistically significant.

Additional study is needed to address whether this effect might magnify after seeing multiple clips, but at the very least it suggests that there are potentially harmful consequences that policymakers must consider.[18]


These findings show that scholars need to conduct more research—and communicate their findings to policymakers—before Congress and the Court introduce cameras into the Courtroom.[19] The results in Political Behavior showed that, under likely conditions, people will support the Court less after watching oral argument. Our results here are similar. There are some conditions under which people think more favorably of judges after watching oral argument. But when we examine the clips that people are likely to see if the Court moves to cameras, we find some immediate and significant decreases in support for judges. And it is important to note that this drop occurs as the result of fairly typical judicial behavior.

Going beyond the data for a moment, we are wary of the unintended consequences of introducing cameras. As we stated above, one reason to oppose cameras is that Justices might change their behavior at oral argument to avoid providing fodder for clips that might be used to attack them or the Court. Justice Souter told the House of Representatives that he in fact changed his behavior as a state supreme court Justice because of cameras:

“I can testify from personal experience that the cameras certainly affected my behavior because I knew that there were some questions that I might ask just within the context of a case, which if I asked, would be the excerpt, the soundbite totally out of context on the six o’clock news . . . my fifteen second question would be there . . . and quoted that way it would create a misimpression either about what was going on in the Courtroom, or about me, or about my impartiality, or about the appellate process.”[20]

Souter changed his behavior to avoid precisely what our results show may happen: people will think less of the judge when he or she asks hard questions. The judge, in anticipation, will ask fewer questions and pull punches at oral argument. Cameras could shut down or reduce Justice questioning. Reduced Justice questioning will lead to poorer information for the Justices. And poorer information will diminish the quality of the Court’s decisions.

It is also likely that the Court might decrease the amount of time it devotes to oral argument. Today, oral arguments can last roughly two hours. The petitioner attorney receives two minutes of uninterrupted time to summarize her argument to the Court. She then spends roughly 28 minutes answering any Justice’s questions. After that time concludes, each Justice (in order of seniority) can ask the petitioner follow up questions. When all the Justices have finished asking their follow up questions, the petitioner concludes, and the respondent begins. The respondent also has two minutes of uninterrupted time, 28 minutes to make her case, and then must answer each Justice’s (in order of seniority) follow up questions.

With all this time, Justices and counsel are bound to say things that others could easily take out of context. To minimize the possibility that Justices would say something the media would report out of context, it is easy to imagine the Justices significantly reducing the amount of time for oral argument or perhaps even deciding more cases on the written briefs alone.

These two possible outcomes could diminish the quality of the Court’s opinions. Oral argument is an important part of the Court’s decision-making process. As it stands today, Justices use oral argument to persuade each other. They play devil’s advocate to gauge their colleagues’ positions, to begin coalition building, and to begin formulating opinions. They use it to probe for each party’s weaknesses and limiting principles. By engaging the attorneys, Justices can obtain more information about the consequences of possible likely decisions. In fact, 80 percent of the issues raised at oral argument appear for the first time there (i.e., do not appear in the parties’ briefs); and 33 percent of the issues raised uniquely at oral argument make it into the Court’s final opinions.[21] All this is to say that oral argument is useful for the Court. But if the Justices believe that usefulness has faded—or that the costs of oral argument outweigh the benefits—they may retreat in a way that reduces the overall quality of their opinions.


When Senators Durbin and Grassley introduced their legislation to require the Court to allow cameras, they highlighted the transparency and accountability cameras would bring to the Court. They might be correct. But what seems clear from our analyses here and elsewhere is that they have overlooked the potential costs associated with cameras. One of those costs concerns the Court’s legitimacy. Another cost concerns how people view the Justices themselves.

It seems quite likely that the polarization in people’s views towards particular Justices would increase with cameras. Seeing the Justices could exacerbate underlying positive or negative feelings, especially where those Justices engage in contentious exchanges with counsel. There is simply not yet enough data to know whether the putative benefits of cameras would outweigh the costs. Our suggestion, therefore, is to engage in further study of this important topic before undertaking a major change. With so much at stake, caution is the best option.

* Professor of Political Science, Michigan State University; Faculty affiliate, MSU College of Law.

George C. and Carmella P. Edwards Professor of American Politics, Leon Epstein Faculty Fellow, University of Wisconsin-Madison.

Horace T. Morse Distinguished Professor of Political Science and Law, University of Minnesota.

  • Professor of Political Science, University of Kentucky.

[1] Cameras in the Courtroom Act, S. 807, 117th Cong. (2021).

[2] See Ryan C. Black, Timothy R. Johnson, Ryan J. Owens, and Justin Wedeking, Televised Oral Arguments and Judicial Legitimacy: An Initial Assessment, Pol. Behav. (forthcoming 2023),

[3] See Jeremy W. Peters, Cold, Wet Wait for Tickets to Supreme Court’s Same-Sex Marriage Cases, N.Y. Times, Mar. 25, 2013.

[4]  Id.

[5] See James L. Gibson and Gregory A. Caldeira, Knowing the Supreme Court? A Reconsideration of Public Ignorance of the High Court, 71 J.  Pol. 429 (2009).

[6] See Richard Wolf, Cameras in the Supreme Court? Not Anytime Soon, USA Today, Mar. 7, 2019.

[7] See C-SPAN, Justice Scalia on Cameras in the Supreme Court, YouTube (Jul. 26, 2012),

[8] See C-SPAN, Chief Justice Roberts on Cameras in the Court, YouTube (Oct. 17, 2018),

[9] See Wolf, supra note 6.

[10] See Black et al., supra note 2.

[11] Our Minnesota clip came from a 2009 case, which was argued at a time when the court utilized a single, static camera. As of this writing in April 2023, Minnesota now appears to show oral argument using three camera angles: the attorney at the podium and two medium-angle views that portray the left and right side of the bench. See, e.g., Oral Argument, Johnson v. Freborg, Case No. A21-1531 (2023),  [].

[12] Lucid is a nationally reflective sample that improves upon earlier platforms like Amazon’s Mechanical Turk by using respondent quotas to achieve a census-balanced sample. While Lucid samples are convenience samples, they provide demographic and experimental results that track well with U.S. national benchmarks (Coppock & McClellan, 2019) and are increasingly common in experimental studies like ours. See, e.g., Alexander A. Coppock and Oliver A. McClellan, Validating the Demographic, Political, Psychological, and Experimental Results Obtained from a New Source of Online Survey Respondents, 6 Rsch. and Pol. 1 (2019); Albert H. Fang and Greogary A. Huber, Perceptions of Deservingness and the Politicization of Social Insurance: Evidence From Disability Insurance in the United States, 48 Am. Pol. Rsch. 543 (2020).

[13] See Alexander A. Coppock and Oliver A. McClellan, Validating the Demographic, Political, Psychological, and Experimental Results Obtained from a New Source of Online Survey Respondents, 6 Rsch. and Pol. 1 (2019).

[14] See, e.g., Albert H. Fang and Greogary A. Huber, Perceptions of Deservingness and the Politicization of Social Insurance: Evidence From Disability Insurance in the United States, 48 Am. Pol. Rsch. 543 (2020).

[15] To make this determination we calculated a reliability statistic called Cronbach’s alpha, which allows us to assess the reliability between those three items and the underlying scale (latent force) that we believe generated them. The value for our three measures is 0.84, which corresponds to “good” reliability (see Joseph A. Gliem and Rosemary R. Gliem, Calculating, Interpreting, and Reporting Cronbach’s Alpha Reliability Coefficient for Likert-Type Scales, Midwest Research-to-Practice Conference in Adult, Continuing, and Community Education (2003) (available online at [last accessed 4/16/23]). Relatedly, if we estimate an exploratory factor analysis using the three questions, we find evidence that a single factor explains 98% of the variance.

[16] In a perfect world, we would have been able to portray the exact same content through three different approaches: audio only, dynamic video, and static video.  Since we could not do this, however, we are unable to offer a definitive explanation for why this difference exists.

[17] We admittedly take an “old world” approach to thinking about the media to the extent we envision what snippets are potentially featured on the evening news after oral argument has taken place. Our arguments, however, do easily generalize to more contemporary conceptualizations of the news media, as well. Digital providers are driven by clicks and page views and so it takes very little imagination at all to see potential click-bait headlines like “Justice Sotomayor Schools State Attorney.”

[18] To be clear, our results do not support the conclusion that the medium of video is uniquely harmful as compared to audio. That is, we observe analogous content effects among the audio-only conditions where attitudes towards the judge are significantly lower in the contentious versus neutral content conditions. Our belief, however, is video will be uniquely featured in a way that audio content – contentious or otherwise – currently is not.

[19] Whether Congress has the constitutional authority to require the Court to allow cameras is a different question and one into which we do not wade here.

[20]  U.S. Supreme Court Appropriations: Hearing Before House Appropriations Subcommittee, 104th Cong. (1997), [].

[21] See Timothy R. Johnson, Oral Arguments and Decision Making on the United States Supreme Court (2004).

Post a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.