Article Summary
Introduction
The U.S. public generally favors politicians who engage in bipartisan lawmaking. Moderate policy and compromise with the other party are popular. Despite this broad public preference for bipartisanship, politicians in the U.S. are highly polarized. Why does this intense partisan polarization not then alienate the public? This paper suggests that what politicians do matters less than what they say. Constituents know very little about politicians’ positions on issues, the laws they create, and how much responsibility politicians have for passing those laws. Politicians take advantage of this by branding themselves in speeches, media, and campaigns in ways that make them appeal to a broad set of voters. Because ‘bipartisanship’ is so popular in the public, politicians often try to gain public support by signaling that they engage in bipartisan lawmaking, even when they really do not. If this argument is correct, the way the public thinks about bipartisanship should be very different to the way politicians do. Moreover, politicians branding themselves and their policies as bipartisan should increase their appeal.
Analytical Approach
The paper tests these arguments in three ways. First, it examines what ordinary people in the U.S. public make of bipartisanship. To do so, the author surveyed 1055 people in the United States.. Participants were asked about a scenario in the House of Representatives and asked whether it describes bipartisanship or not. Participants were randomly assigned to read one of two scenarios. In the first, three Republicans joined the Democrats to pass a bill; in the second, three Democrats joined the Republicans to pass a bill. After reading the scenario and indicating whether they felt it described bipartisanship, participants were asked to define bipartisanship in an open-text box, then finally to indicate whether they thought bipartisanship was good or bad.
Second, it examines the ways that politicians in Congress use the term ‘bipartisanship.’ The paper collects all 434,266 speeches made on the floor of Congress between 1992 and 2018. These speeches are an opportunity for politicians to both state their opinions and connect with voters, so politicians craft their speeches carefully to gain attention from the media and the public. To capture politicians’ discussions of bipartisanship, the author searched all of the speeches for the term ‘bipartisan,’ then extracted all of the sentences the term was used in. The author considers who uses the term, whether references to bipartisanship are related to actual bipartisan behavior and bipartisan bills, and how the use of the term changes over time.
Third and finally, the paper tests how voters respond when politicians use the word ‘bipartisan’ to describe legislation. The author does so with two experiments. In the first experiment, 1206 people in the U.S. were randomly assigned to read a short, artificial speech from a leader of the opposing party. The speech detailed a plan for a bill that was either overtly partisan or not overtly partisan, and the speaker either described that plan as bipartisan or not. Then, participants were asked how much they support or oppose the policy and how moderate or ideologically extreme it was.
| Experimental Conditions in Main Study | Elite Issue Polarization | |||
|---|---|---|---|---|
| High Polarization | Low Polarization | No information | ||
| Elite Civility | Civil | 1 | 2 | - |
| Uncivil | 3 | 4 | - | |
| No information | - | - | 5 | |
In the second experiment, the author explores how people respond to bills that are described as bipartisan compared to bills that are actually bipartisan (i.e., ones with a substantial number of opposition members supporting them). 1324 people in the U.S. were randomly assigned to one of three versions of a news article. All three versions describe a piece of Republican legislation: they vary in how Republicans describe the bill and how many Democrats are reported to vote for it. In the first condition, Republicans call the legislation bipartisan; in the second, they call it important. (This second condition helps determine whether effects are driven by the term ‘bipartisan’ in particular or just by generically positive terms.) Within these first two conditions, the number of Democrats reported to vote for the bill was randomized from 1 to 100 (i.e., from a very small number of party defectors to a very large coalition). Finally, in the last version, Republicans describe the legislation as explicitly partisan and no Democrats vote for it. After the manipulation, participants are asked how much they support or oppose the legislation and how moderate or ideologically extreme it was.
| Table of Conditions for Experiment 1 | Partisanship of Bill | ||
|---|---|---|---|
| Overtly partisan | Not overly partisan | ||
| Description of Bill | Bipartisan | 1 | 2 |
| Not bipartisan | 3 | 4 | |
Main Findings
In the first study, the author asks participants about three representatives joining the other party to vote for a bill. In reality, this is a tiny amount of support from the opposition, not nearly enough to reflect meaningful bipartisanship. Nonetheless, most participants felt that this situation described bipartisanship, regardless of whether the representatives were co-partisans joining the opposition or opposition members joining co-partisans. Participants generally felt that bipartisanship was good, but most of them were unable to clearly define it. In other words, people know that they like bipartisanship in general, but they don’t know exactly what it is.
In the second study, the author finds that use of bipartisan language in Congress appears to be strategic. Politicians refer to bipartisanship not because they are actually bipartisan, but because it appeals to voters. The use of bipartisan talk has increased over time. During presidential election years, when parties feel the need to appeal to broad groups of voters, bipartisan language increases; during midterm election years, when parties feel the need to appeal to their base, bipartisan language decreases. There is also no relationship between bipartisan intent and bipartisan language. Extreme politicians refer to bipartisanship just as much as moderate politicians, and bills described as bipartisan are no more likely to have meaningful bipartisan sponsorship. However, there are some contexts that encourage politicians to use bipartisan appeals: when politicians are campaigning in general elections, they use more bipartisan appeals, as do those who barely win elections (and thus can’t rely on just their own party’s voters to stay in office in the future).
In the two experimental studies, the author finds that describing legislation as bipartisan makes people support it more, even if the legislation itself is overtly partisan. It is the description that matters—as long as the Republican bill is described as bipartisan, it barely matters whether 1 or 100 Democrats vote for it. Describing the bill as important increases support by a comparable amount, but only when at least 10 Democrats vote for the bill; for smaller numbers, ‘bipartisan’ is more effective.
Implications
Bipartisan talk is strategic: politicians describe themselves and their policies as bipartisan not because they actually are bipartisan, but instead because they understand it will appeal to voters. And this appeal is substantial. Many contested bills in Congress pass with <10 minority votes. The experimental results show that simply describing these as bipartisan—even though they are clearly not—is likely to substantially increase public support for them. This undermines any motivation politicians might have to compromise with the other party, and makes it much more difficult for voters to assess what their representatives are actually doing.
Questions left unanswered
There is more research that can be done to understand who uses bipartisan language and when. The author particularly highlights the value of interviews and case studies to understand politicians’ motivations and strategies related to using bipartisan language.
The author also notes that, if appeals to bipartisanship are so effective, what happens when members of one party describe the other party as partisan? And what happens when both parties engage in this discourse, describing themselves as bipartisan and criticizing the other size as uniquely partisan? If voters hear this message from both parties simultaneously, what do they make of it?
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: No
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: Yes
- Largest p-value presented as significant: 0.05
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
The author’s focus was on the term ‘bipartisan.’ However, in the experimental studies, the effects of describing a bill as bipartisan were generally quite similar to the effects of describing a bill as important. The author might have spent more time discussing the (lack of) differences between these two manipulations and their implications for congressional language. How much is this dynamic attributable to the unique appeal of bipartisanship rather than voters being receptive to politicians’ endorsements and positive self-branding generally?
Open Data & Analyses
Does the article make the replication data publicly available?: Yes
Does the article make the replication analysis scripts publicly available?: Yes
Article Citation
Westwood, S. J. (2022). The partisanship of bipartisanship: How representatives use bipartisan assertions to cultivate support. Political Behavior, 44(3), 1411–1435. https://doi.org/10.1007/s11109-020-09659-6
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
@article{westwood_partisanship_2022,
title = {The partisanship of bipartisanship: {How} representatives use bipartisan assertions to cultivate support},
volume = {44},
url = {https://doi.org/10.1007/s11109-020-09659-6},
doi = {10.1007/s11109-020-09659-6},
number = {3},
journal = {Political Behavior},
author = {Westwood, Sean J.},
month = sep,
year = {2022},
pages = {1411--1435},
}