The confirmation hearings of then judge Brett Kavanaugh reminded many of the hearings for Justice Clarence Thomas. In both situations, there were allegations of sexual misconduct after the initial hearings concluded. In both cases, the Senate Judiciary Committee held an additional hearing to investigate the claims. The ideological balance of the Court was also at stake in both confirmations. Clarence Thomas replaced the liberal icon Thurgood Marshall, swinging the Court further to the right; Brett Kavanaugh replaced the moderately conservative swing vote Anthony Kennedy, solidifying the Court’s conservative majority. Both men were confirmed by incredibly small margins (50–48 and 52–48, respectively). And finally, both confirmation battles were followed by large wins by women in the next national election.

Although these similarities are certainly striking, it is the hearings’ differences that are perhaps more noteworthy and consequential for examining senatorial behavior. Kavanaugh was accused of sexual assault as a teen, whereas Thomas allegedly sexually harassed colleagues as an adult and supervisor at the Equal Employment Opportunity Commission. Anita Hill, Clarence Thomas’s accuser, was questioned directly by the senators in 1991; Christine Blasey Ford, Brett Kavanaugh’s accuser, was questioned directly by the Democratic senators, but the Republican senators used a proxy—a female prosecutor with experience in the area of sexual assault.^[1] The Thomas hearing spurred awareness of sexual harassment (Black and Allen 2001). The Kavanaugh hearings, on the other hand, occurred in the wake of the Me Too movement and after several high-profile men had lost significant positions of power and privilege. In 1991, there were only two women in the Senate and none on the Judiciary Committee; by 2018, there were twenty-three women sitting as senators and four serving on the Judiciary Committee.

These differences resulted in the Kavanaugh hearing presenting a context especially conducive to senators making rhetorical overtures both supporting women and recognizing the issues of sexual harassment and assault. In this short research note, we leverage the different contexts of the Thomas and Kavanaugh hearings to explore whether institutional and broader social contexts affect the representational and strategic behavior of senators. To do so, we systematically examine the frames senators used in the floor speeches that were made during both confirmations. To put it more simply, we investigate whether there is an association between a change in the institutional and social context and a change in use of rhetorical representation.

A relative constant of senatorial behavior is the use of floor speeches to declare and explain publicly an intended vote (Mayhew 2004; Jacobson and Carson 2019). To be sure, current members of Congress have many options for communicating with their constituents and signaling colleagues and interest groups (see Lublin 1999; Tate 2003; Gamble 2011; Minta 2011; Stout 2019). The senatorial speech justifying a vote remains part of this arsenal. Senator Collins’s forty-minute floor speech explaining her support for Kavanaugh is a clear example of the continued viability of the floor speech as a signaling method.^[2] Since at least some senators engaged in this behavior in the two time periods described, these speeches are excellent tools to investigate whether the different contexts of the hearings resulted in senators responding differently to the issue of sexual misconduct by a Supreme Court nominee.

Scholarly investigations into congressional behavior are legion. Mayhew (1974) provides the seminal explanation for why senators explain their votes. Hill and Hurley (2003) show that senatorial speeches are clear examples of representational and strategic behavior, and Osborn and Mendez (2010) find that gender affects the content of these speeches. The highly salient and controversial nature of both the Thomas and Kavanaugh confirmations would spur significant senatorial talk for representational and/or strategic ends (Haines, Mendelberg, and Butler 2019).^[3] Given this research, the emergence of the Me Too movement, and the increase in female senatorial representation, we expect to see different frames employed in the Senate floor speeches prior to the Thomas and Kavanaugh confirmation votes (Thomas 1991; Lawless 2015; Wade and Brittan-Powell 2001). Specifically, we expect there would be an increase in recognition of the issues of sexual harassment and assault or women’s issues more obliquely in the speeches from the Kavanaugh hearing regardless of a senator’s party.

Although we expect the speeches during the Kavanaugh confirmation overall to contain more references in support of women and women’s issues, we also anticipate that there will be differences in the speeches’ recognition of sexual assault, harassment, and support for women in general at the level of the individual senator for two reasons. First, the GOP in the current period suffers from a significant deficit when it comes to support from women (Cassese and Barnes 2018; Barnes and Cassese 2017). Republican senators, then, signaling their support for women or the Me Too movement in their justification for Kavanaugh would provide some rhetorical representation that could benefit their reelection goals. While some research suggests that politicians might avoid the topic altogether, fearing the potential backlash any statement might incur (Pietryka 2012; Milita et al. 2014), it is also clear that if there is an electoral reward or the political context simply requires action, politicians will speak out (Sides 2006; Pietryka 2012). The high-profile nature of both of these confirmation events places the situation in the latter category. Thus we expect that the increased rhetorical support for women and women’s issues during the Kavanaugh confirmation should be greater for Republican senators than for Democratic senators.

Second, signaling support for women (an increasingly mobilized and potent voting bloc) should yield electoral dividends. Thus for those senators facing reelection soon after the hearings, the benefit of speaking out is clear. Accordingly, we expect higher levels of senatorial speech among senators facing reelection directly after the Kavanaugh confirmation than for the Thomas vote.

To state each of these hypotheses more plainly, if the Thomas hearing is considered our first time period and the Kavanaugh hearing the second period, our expectations are as follows:

Research Design

To study the senators’ speeches, we identify the underlying messages (i.e., frames) of each speech. Individuals “actively classify and interpret their life experiences to make sense of them” using “schemes of interpretation called ‘primary frameworks’” (Scheufele 2009; Goffman 1974). Within a political speech, such as a Senate floor speech, “there are ways of…depicting events…that depend on the framework employed” by the speaker (Scheufele 2009). By framing the issues in each speech and in giving a speech, senators hope to connect to their constituents, and in doing so, take the appropriate position and explain their vote (Mayhew 2004; Jacobson and Carson 2019), thereby ensuring their key supporters that they are voting as expected.

To begin, all floor speeches made by the senators after the second set of confirmation hearings for either Justice Thomas or Kavanaugh were read and analyzed for frameworks (N = 98; N = 63 for Thomas; N = 35 for Kavanaugh). Initially, we content-analyzed the speeches at a granular level, identifying seventeen frames. We then collapsed these frames into ten categories for analysis (see table 1).^[4] In addition, we gathered data on every senator including age at the time of the speech, race/ethnicity, gender, date of their upcoming reelection, and political party. We also recorded the vote of each senator at the end of the hearing: yea or nay. Our analyses of these data employ both contingency tables and bivariate logit analysis with interaction effects.^[5]

Findings

The test of our first hypothesis, an increased emphasis or mention of issues of sexual harassment or assault, revealed that the frame most closely associated with these salient topics, the I BELIEVE HER frame, was not a commonly employed frame. Only 16 percent (N = 16) of the speeches overall contained this frame, and almost half (N = 7) of these occurred during the Thomas confirmation, well before the Me Too movement. The LASTING IMPACT frame, which is more obliquely related to these issues and includes references to the weight of the hearings on women or the status of women, was more popular, occurring in 68 percent (N = 67) of the speeches overall, but again the bulk of its use occurred during the Thomas confirmation (44 versus 23). We calculated the bivariate relationship between the confirmation and these two frames and found a significant difference between the nominations in the use of the LASTING IMPACT frame (.002), though the direction is not as we anticipated. At first blush, it would seem that we do not find support for our first hypothesis.

Another way we can test our first hypothesis is to examine the relative ranking of the frames. The I BELIEVE HER frame ranks last among all the frames during the Thomas hearing but rises to seventh place during the Kavanaugh period. The LASTING IMPACT frame ranks second during the Thomas hearing and ties for first during Kavanaugh hearings. In other words, while the absolute number of times these frames were employed was higher during the earlier hearing, the I BELIEVE HER and LASTING IMPACT frames were actually employed more frequently during the Kavanaugh hearings relative to the other frames (see figure 1). The change in ranking shows that the language used to discuss issues of sexual harassment or assault in these types of speeches may not have changed all that dramatically; however, the emergence of the Me Too movement and the addition of several women to the Senate and the Senate Judiciary Committee seems to have inspired at least increased reference to the status of women.^[6] Thus we claim partial support for our first hypothesis.

Across the confirmations, the most frequently used frames, based on raw numbers, were QUALIFICATIONS BOOST for the nominee or testifier (70 percent), QUALIFICATIONS ATTACK (57 percent), PROCEDURAL FAILURE and PROCEDURAL FAIRNESS (47 percent, 33 percent, respectively), JUDICIAL TEMPERAMENT (47 percent), and PARTISIAN STRATEGY (31 percent). These results, we believe, would likely match up with other confirmation hearings and suggest that the unusual aspects of these two hearings did not much alter Senate behavior (see Watson and Stookey 1995; Solberg and Waltenburg 2014).

Our second hypothesis relates to partisan difference in behavior across our two cases. We anticipated that Republicans would use the I BELIEVE HER and/or the LASTING IMPACT frames more in the later time period. We employed cross-tabulations and a logit model to test our hypothesis. Again, our expectations were not borne out by our data (see table 3). During the Thomas confirmation, there were no significant differences between the parties in the employment of the I BELIEVE HER frame. Although the difference is significant (.006), during the Kavanaugh confirmation it is the Democrats who utilized this frame appreciably more often. The same pattern repeats for the LASTING IMPACT frame. Significant differences only appear between the parties during the Kavanaugh hearing (.006), with Democrats again setting the pace for using this rhetorically representative frame. For their part, GOP senators used the frame at about the same rate as their Democrat counterparts during the Thomas confirmation, but the inclusion of the frame in their speeches fell off dramatically during the Kavanaugh confirmation.

To further substantiate this finding, we combined the two frames associated with the Me Too era—LASTING IMPACT and I BELIEVE HER—and employed a logit model with the senator’s party, the nomination, and an interaction term of those variables as the independent variables (see online appendix for the logit estimation results). Figure 2 clearly shows that the likelihood of employing these frames decreased for senators in both parties, and especially so for the GOP. Indeed, the probability of a Democratic senator making a speech with one of these frames drops meagerly from .49 to .41 while the probability of a Republican senator using this frame decreases dramatically, from .44 to .12.

Therefore, we are again forced to recognize that our hypothesis—that the Republicans would increase their use of frames related to sexual assault and harassment, or at least maintain their recognition of these issues in their senatorial speeches—is incorrect. Despite the changing context and the Republican Party’s recognition of the gender gap, the use of these frames actually decreased among those few Republican senators that gave a speech justifying their vote on Kavanaugh. In other words, the Republicans were significantly less likely to use rhetorical representation or to reach out to women in their speeches.

Thus far we have found only partial support for one of our hypotheses; however, in doing so we have also uncovered a most intriguing finding: the rather dramatic decline in the number of speeches made from 1991 to 2018. During the confirmation hearing for Clarence Thomas, 63 senators spoke, while only 35 chose to make a speech to explain their vote before casting their vote for or against Brett Kavanaugh. Furthermore, 47.6 percent of the speeches after the Thomas hearing were made by Republicans and 52.4 percent were made by Democrats—nearing parity. After the Kavanaugh hearing, 60 percent of the speeches made were by Democrats, and only 40 percent were made by Republicans.^[7] This within-party difference for the GOP is highly significant (<.0001). As with the initial comparison, this unexpected finding begs for further investigation. Therefore, we examined the within-party differences in the use of several of the frames.^[8]

We calculated the bivariate relationship between the confirmation and the frames under examination within each party. As mentioned earlier, the LASTING IMPACT frame, as anticipated from the earlier results, is highly significant when we examine only GOP speeches (.001), and the I BELIEVE HER frame approaches conventional levels of statistical significance (.055) for the Republican senators. Similarly, the use of two other frequently employed frames—QUALIFICATIONS BOOST, QUALIFICATIONS ATTACK—are significant among the GOP but not the Democrats. The Republicans spent more time in their speeches boosting the nominee or, in the rare case, the testifier,^[9] during the Thomas hearing (63 percent versus 22 percent). The discrepancy is not as large for QUALIFICATIONS ATTACK but runs in the same direction (37 percent versus 18 percent) and is still notable. The only frame that showed a difference in behavior for the Democrats is PROCEDURE FAIR, and there was a significant difference in the GOP use of this frame as well. For both parties, the use of this frame decreased significantly between the two confirmations. Senators from either party were less likely to tout the attempts at due process or characterize the proceedings as fair during the Kavanaugh hearing than during the Thomas hearing. In fact, in all frames discussed here, the trend is decreasing employment by the GOP. In other words, the GOP spoke less after the Kavanaugh hearings and used fewer frames in those scant speeches.

These findings all suggest that there has been some shift in the behavior of Republican senators, even if it is not the adaptation we initially hypothesized. It is possible that the differences we find are still a reflection of typical Congressional behavior, which leads to our last hypothesis. Our third hypothesis examines whether senatorial behavior, in terms of making these types of speeches, reflects the reelection goal (Mayhew 2004; Jacobson and Carson 2019). Examining the speeches overall, we find that about 64 percent (N = 23) of the senators facing reelection immediately after the Thomas confirmation and 33 percent (N = 12) of the senators on the ballot directly after the Kavanaugh hearings gave speeches justifying their votes. While the high percentage after the Thomas hearings supports our hypothesis, the significantly smaller percentage before the vote on Kavanaugh does not.^[10] Moreover, senators in their penultimate year are no more likely to give a speech following either hearing than their more electorally secure counterparts, an observation clearly displayed in figure 3. Therefore, we cannot accept our third hypothesis—that the likelihood that a senator gives a speech is spurred by their concerns about backlash or waning support in the election immediately after the confirmation vote.

Figure 3 also confirms the dramatic drop-off in the number of senators giving speeches following the confirmation hearing reported earlier. Clearly something changed between the two hearings and that change deserves at least a modicum of further investigation. To that end, we further examined the participation by party. Of the senators whose reelection year was immediately following the Thomas confirmation vote, 10 Republican senators and 13 Democratic senators gave a speech (70 percent of the Republicans up for reelection and 68 percent of the Democrats). After the Kavanaugh hearing, only two Republican senators but 10 Democratic senators who were up for reelection chose to justify their vote (only 25 percent of the Republicans versus 43 percent of the Democrats).^[11] While the difference between the two parties is not significant, the behavioral change here is stronger for Republicans than Democrats. We are not suggesting that these findings are pathbreaking; we recognize that we are only looking at two time points and that the changes are in the same direction for both parties. However, it is possible that since Republicans are polarizing more quickly than Democrats (Theriault 2006), they feel less need to justify or qualify their vote to an increasingly homogenous base. While this is not strong evidence, it is revealing and suggests that further investigation into this change in behavior is necessary.^[12]

Conclusion

Overall, these results paint a picture of a change of behavior for Senators, and especially the Republicans in terms of how they approach a confirmation vote complicated by issues of sexual harassment or assault. Our main finding is that there is no greater discussion or employment of frames related to sexual assault or harassment after the Kavanaugh hearings despite the advances made by women and the burgeoning Me Too movement. In fact, the use of the main frames related to these issues was less, in absolute terms, during the Kavanaugh than the Thomas confirmation. Despite the prevalence of the “I believe her” hashtag in the Twittersphere, senators, particularly Republican senators, were not using this or similar phraseology as they explained their vote. Similarly, fewer senators highlighted the historical significance of the Kavanaugh hearing and what it means for women and the country moving forward though the use of such language was more frequent than almost any other frame. While both parties saw a large reduction in the number of speeches, the decrease in the use of this tactic overall and of the frames of interest here was considerably one-sided. There was a significant drop in the use of these frames among the GOP during the Kavanaugh hearing when compared to the earlier Thomas hearing. The Democrats, on the other hand, took the opportunity to employ rhetorical representation in their speeches and were the reason these two frames ranked a bit higher in usage in 2018 than in 1991. While it is possible that Republican senators simply did not see the Kavanaugh hearing as setting any precedent or impacting history, we find that explanation implausible. It is more likely that GOP senators simply thought the less said the better for their position and party. Indeed, the largest behavior change we find is an appreciable difference in the number of senators making floor speeches; specifically, Republican senators were much less likely to give a speech in 2018 than in 1991 (see figure 4). To put it concretely, the probability of a Republican giving a speech declines from .7 during the Thomas hearings to .27 during the Kavanaugh hearings, and this pattern is not affected by proximity to reelection as seen in figure 3.

This decline in the use of explanatory or position-taking speech could be related to the change in the political climate. Pew Research finds that both voters and their representatives are more ideologically polarized than in previous time periods (Desilver 2014). Political ideology accounts for almost all congressional choices so that the members’ votes have become less about the issue at hand and more about falling in line with the party (Theriault 2006). There are likely many behavioral consequences of these changes, and we posit that the decrease in senatorial speeches—in position taking—is one of them.

It is also possible that in our more polarized era, Republican senators simply do not need to justify or qualify their vote. Their constituency is driven more by ideology and such rhetorical representation is not as critical today as it was in the 1990s. Additionally, it may be that the emergence of social media (i.e., Twitter) means that senators no longer need to justify their vote on the floor of the chamber. We intend to investigate if the senators were more prolific via Twitter or if this behavioral pattern extends to social media as well. While the evidence presented here is certainly not definitive, it is revealing and suggests that further investigation into this change in behavior is warranted.

Finally, the findings reveal that while the Me Too movement resulted in unprecedented numbers of women being elected to Congress, the effect of the influx of these members may not result in a definitive change in the overall conversation. The purported LASTING IMPACT discussed by members during the Thomas confirmation battle did not yield concrete changes in speeches after the Kavanaugh hearings. Senators still used their speeches to attack the credibility of the nominee and the testifier (see figure 1), and while the LASTING IMPACT frame moved up in the rankings, it was used mostly by one side of the partisan aisle. In other words, the impact of the Thomas hearings and indeed the Me Too movement in general did not seem to affect business as usual in the Senate.