Representation by the Minute: The Influence of Ethnicity, Partisanship and District Fit on Legislator One-Minute Floor Speeches

Abstract

Much of the research on minority representation in the U.S. House has focused on how group preferences are reflected in recorded votes, yet most opportunities for position-taking exist outside the roll call arena. This research explores the conditions under which descriptive representation enhances substantive representation by examining an area where members have a great deal of freedom to register their preferences - speeches on the House floor. Using a novel database of immigration-related one-minute speeches from the 109th-113th Congress, this research explores both the decision to deliver and the position taken in speech on the floor. It finds that not only do Latino members devote a greater share of attention to the issue of immigration than their colleagues, but they take distinct positions that are more reflective of Latino interests, and they are less vulnerable to district pressures than their non-Latino colleagues. These results suggest a clear value in descriptive representation and more diverse legislatures.

Notes

Acknowledgments

The author would like to thank Kristin Tennessen, Lauren Ernat, and three anonymous reviewers for helpful comments and suggestions.

Notes

1 Rouse (2013), looking at state legislatures, finds the effect of Latino descriptive representatives on the substantive representation of Latinos to vary throughout the legislative process. She finds that Latino legislators introduce a greater number of Latino interest bills when their numbers in the legislative body are relatively low. However, Rouse (2013) also finds party to be a better predictor of vote choice on both general interest and Latino interest legislation, suggesting descriptive representation is more visible and necessary for substantive representation in some conditions over others.

2 One-minute speeches are traditionally delivered at the prerogative of the Speaker, with the Speaker determining the amount of time available for speeches or the number of speeches to be given on a particular day. The majority and minority leadership typically receive advance notice of any limitation. Unlike many practices in the House, one-minute speeches are freely available to all members of both parties, without systematic bias, such as is present in floor debate. These policies stem from a 1984 announcement and have been continued into more recent sessions. Specifically, the policy states, “The chair will alternate recognition for one-minute speeches between majority and minority Members, in the order in which they seek recognition in the well under present practice from the Chair's right to left, with possible exceptions for Members of the leadership and Members having business requests” (Congressional Record, daily edition, vol. 161, January 6, 2015 H34).

3 Members are also given the opportunity to add additional information, longer speeches, or speeches in absentia into the Congressional Record. While these words are not delivered on the House floor, they are offered into the record, reducing the time constraints found in one-minute speeches delivered on the floor.

4 While one-minute speeches may appear symbolic in terms of the long range goal of passing legislation, many speeches are decisively substantive in advocating support or opposition for a piece of legislation. As previously indicated by Chabot (1997, March 5), most members know they will receive limited time or opportunities to weigh in or shape legislation that makes it to a final vote. Additionally, votes do not always allow members to register a so-called “good” position, or even one of their choosing, limiting members to a vague “yea” or “nay” on legislation they had little input in crafting without the opportunity for further explanation. Therefore, for some members, their only opportunity to make their voice heard on the floor and in the congressional record comes in the form of a one-minute speech. One example of using one-minute speeches to shape the agenda and with the focus of national media and public opinion comes from July 2013, when one-minute speeches gained national attention for their use to condemn comments made by Representative Steve King (R-IA). King stated that for each immigrant success story that comes with the DREAM (Development, Relief, and Education for Alien Minors) Act, “there’s another 100 out there who, they weigh 130 pounds and they’ve got calves the size of cantaloupes because they’re hauling 75 pounds of marijuana across the desert. Those people would be legalized with the same act.” Following this comment, many Latino members of Congress used one-minute speeches not only to denounce Representative King’s remarks, but also to argue for their preferred position: a clear path to earned citizenship and full integration for the undocumented (Llamas 2013). The attention this received in Congress and the coverage in the national media and Latino media demonstrates the importance of the speeches but also their ability to register with the public.

5 Future research could explore in more detail how issues that appear in one-minute speeches are literally translated to the legislative agenda, if at all, and under what conditions.

6 This effect should hold regardless of the race/ethnicity of the legislator. Non-Latino Democrats, regardless of race, will not have the same shared experiences or pressures to represent Latinos nationally, and are therefore expected to respond in a similar way to other Democrats. Appendix D expands on this idea, and offers further support for the distinct behavior of Latino members of Congress on this issue.

7 I include both speeches delivered on the floor and the material inserted into the Congressional Record in the House section, not the Extension of Remarks section. This approach is appropriate for a number of reasons. First, while permission is almost always granted for a member to address the chamber, including the entire speech avoids potential bias or a selection effect that would occur from omitting the end of the speech. Second, while members are permitted to extend their complete speech in writing, when they cannot finish their remarks, the printing requirements are not always correctly observed. As Schneider (2015) points out, the Joint Committee on Printing regulations for publication of the Congressional Record provide for extraneous matter to be printed in the “Extensions of Remarks” section, but noted in the Members’ remarks. However, she found this requirement was not always observed, with material often being printed in the House section along with the one-minute speech, making identifying this material accurately challenging, and potentially introducing bias into the data. Further, often the full speech is inserted into the record, and while the pdf of the record distinguishes the addition or remainder of the speech as distinct, the text file of the speech does not. It is worth acknowledging a counterargument, which suggests that inserting speeches into the record and yielding back time indicates they are intended for a different audience, namely that inserted speeches are intended for readers of the Record, while delivered speeches reach a larger audience through C-SPAN (see Maltzman and Sigel-man 1996; Morris 2001). However, this argument does not account for a number of factors, including the small C-SPAN viewership relative to the size of the electorate, and the varying reasons members may have to extend their remarks, including lengthy speeches, supportive materials, or slower delivery. Additionally, as many members choose to disseminate their one-minute speeches by mail, press releases, or websites, members can reach any individual or organization who may be interested in receiving their speech, regardless of the whether all, some, or none of the speech was verbally delivered on the House floor (see Schneider 2015). To account for this, I include speeches that appear in some way in the House section of the record, rather than solely the Extension of Remarks, so at least part of each speech was delivered on the floor.

8 To verify the accuracy of the LIWC program, a graduate student assistant and the author checked a sample of the speeches and content against hand-coding that had been done ahead of time.

9 Computer-based content analysis has been used to extract political positions from party manifestos (Laver, Benoit, and Garry 2003; Proksch and Slapin 2009), legislative speeches (Giannetti and Laver 2005; Kalaf-Hughes 2013; Kluver 2009; Laver and Benoit 2002; Proksch and Slapin 2009; Schonhardt-Bailey 2008), campaign speeches (Schonhardt-Bailey 2005), and constitutional negotiations (Benoit et al. 2005).

10 Details on the Wordfish procedure and associated robustness checks are provided in Appendix A.

11 I use Will Lowe’s Jfreq program, available at http://www.williamlowe.net/software/.

12 See Slapin and Proksch (2008) for a detailed discussion of implementing Wordfish.

13 This measure is based on self-identification and data from the Congressional Hispanic Caucus and Congressional Hispanic Conference. While identity is flexible, self-selecting into one of these Congressional organizations suggests a commitment to the representative's Hispanic or Latino identity.

14 As previously stated, immigration policy is personal for many Latinos, with a strong shift in the liberal direction in recent years (see Sanchez 2011).

15 Comparison results for a model of all speeches in the time period is presented in Appendix C.

16 Predicted values are calculated using the Margins post-estimation command in Stata 14.

17 Summary statistics for the raw Wordfish estimates and a discussion of the methodology appear in Appendix A.

18 Interactions between fit and district demographics were not statistically significant, and therefore are not included in the model.

19 This finding runs counter to the work of Wallace (2014) who finds evidence that both African American and Latino legislators can provide substantive representation to Latinos, as evidenced by their cosponsorship behavior. To offer an explicit test of this idea as it relates to one-minute speeches, Appendix D re-estimates the models presented in Tables 1 and 2, and finds that African American members do not provide substantive representation to Latinos in speech.

20 As previously discussed, there is evidence that one-minute speeches are distinct from the rest of the legislative process, including hearings and bills, because speech allows members to focus on the issues salient to their constituents to a greater extent than is observed in the traditional legislative process (see Hughes 2018). Future research could explore this link in greater detail by looking at the literal translation of words into policy.

21 As Lo, Proksch and Slapin. (2016) acknowledge, Wordfish, and other approaches treating text as data, assumes words are conditionally independent of one another - an assumption that does not hold in reality as grammar matters. However, previous literature, robustness checks, and manual spot checking demonstrate that the approach works in practice. Table 2 and Figures 3 and 4 in this appendix provide additional details and discussion.

Figure 3. Wordfish position estimates.

Figure 4. Word weights vs. word-fixed effects 109th-113th of Congress.

22 These results are consistent with previous literature (see Hall 1996; Rocca 2007).

23 These results hold if the model is re-estimated to exclude all immigration-related speeches.