How Partisanship Influences What Congress Says Online and How They Say It

• Author: Patrick D. Tucker,Richard T. Wang
• DOI: 10.1177/1532673X20939498
• Published date: 01 January 2021
• Date: 01 January 2021
• Subject Matter: Articles

939498APRXXX10.1177/1532673X20939498American Politics ResearchWang and Tucker
research-article2020
Article
American Politics Research
2021, Vol. 49(1) 76 –90
How Partisanship Influences What
© The Author(s) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
Congress Says Online and How They
https://doi.org/10.1177/1532673X20939498
DOI: 10.1177/1532673X20939498
journals.sagepub.com/home/apr
Say It
Richard T. Wang1 and Patrick D. Tucker2
Abstract
We investigate the influence of partisanship on congressional communication by analyzing 180,000 press releases issued
by members of Congress (MCs) between 2005 and 2019. Specifically, we examine whether partisan factors such as party
control of the White House and/or Congress influence the tone used by MCs and whether MCs are more likely to focus on
issues that their respective party owns. Our analyses include the use of multiple OLS models, the machine learning approach
gradient boosting, and Grimmer’s topical modeling software “expAgenda.” We find that (1) partisanship influences the tone
MCs use when communicating online; and (2) MCs are unable to prioritize discussing issues that their respective party own
but devote slightly greater attention to their party’s issues than MCs from the opposite party. Our study ultimately finds
strong evidence of partisan influence in the way MCs design their press releases and has important implications for online
congressional communication.
Keywords
political communication, congress, partisan communication
Does partisanship influence what members of Congress
are not filtered through journalists (Lipinski & Neddenriep,
(MCs) say online and how they say it? To answer our research
2004, p. 7; Malecha & Reagan, 2012, p. 59). They encapsu-
question, we exploit a vast database of press releases pub-
late how MCs generally communicate to their constituents
lished by MCs on their official websites. With these press
(Adler et al., 1998; Grimmer, 2010; Lee, 2017, p. 117;
releases, we specifically examine whether party control of the
Mayhew, 1974) as well as how they behave in Congress
White House and/or Congress influences the tone used by
(Grimmer, 2013). In addition, MCs consider them to be
MCs when communicating and whether MCs prioritize their
“instrumental” in providing information about local services,
communication on issues owned by their respective parties.
“communicating their views[,] and explaining their actions
Our analyses include the use of several multiple regression
to their constituents” (Malecha & Reagan, 2012, p. 18).1
models regressing the tone of press releases on party control
While press releases may not be consumed often by the
of the White House and of Congress, while controlling for
average constituent directly, they remain an essential part of
other sources of influence such as legislators’ effectiveness.
the congressional communication process. Traditionally,
We also employ gradient boosting machine learning models
MCs’ communications staff understand that constituents
to assign sentiment scores to press releases. Finally, using the
consume a great deal of their information through local
“expAgenda” software, we assign press releases topic labels.
media and they take great care to develop relationships with
We observe several findings from our study, such as that
local press to build a receptive audience (Cook, 1989). In
partisanship influences the tone MCs use when communicat-
turn, local reporters report on press releases and hopefully
ing online, and that MCs are unable to prioritize discussing
present them in positive terms. As the resources for local out-
issues that their respective party owns, but they devote
lets continue to decline, the importance of the press release
slightly greater attention to their party’s issues than MCs
has not necessarily diminished. Grimmer demonstrates that
from the opposite party. Our findings have important impli-
local papers often engage in what he terms “ventriloquism”
cations for congressional communication on the Web.
1Cornell University Law School, Ithaca, NY, USA
Press Releases
2Yale University, New Haven, CT, USA
Press releases serve as an appropriate medium for the type of
Corresponding Author:
Richard T. Wang, Cornell University Law School, Myron Taylor Hall,
analysis to evaluate our question as they are low-cost (Adler
Ithaca, NY 14853-4901, USA.
et al., 1998), spatially limitless (Druckman et al., 2010), and
Email: rtw65@cornell.edu

---

Wang and Tucker
77
(Grimmer, 2013, p. 32). Rather than report on the MC’s press
ignores the strength of the direction. A continuous variable
release, the press will simply lift text from the press release
for tone may provide a more nuanced analysis of partisan-
and print it in the local paper as a news story. In this way, the
ship’s influence on communication since MCs’ behavior
constituent may not come into contact with the press release
may vary by electoral context.2 Finally, Lipinski only consid-
in its original form, but the likelihood that they will encoun-
ers majority control of Congress and does not include party
ter it in their daily newspaper, perhaps with minimal filter-
control of the presidency—another potential source of parti-
ing, is relatively high.
sanship’s influence.
Issue ownership theory suggests that partisanship also
Previous Studies of Legislator
influences the policies discussed by elites. The theory posits
Communication Using Press Releases
that the public “believe[s] that one of the major parties is bet-
ter suited to deal with particular issues” (Druckman et al.,
Congressional communication not only provides constituents
2010, p. 5) as a result of the party’s “history of attention,
a window into the legislative activities of their MCs, but may
initiative, and innovation” towards those issues (Petrocik,
also shape their attitudes (Ansolabehere et al., 1993; 1996, p. 826). Previous research on issue ownership has
Druckman, 2001; Grimmer, 2013; Jacobs & Shapiro, 2000,
mainly focused on congressional candidates’ strategies rather
pp. 50, 61; Krosnick & Kinder, 1990; Malecha & Reagan,
than incumbents (Ansolabehere & Iyengar, 1994; Simon,
2012; McCombs & Shaw, 1972; McGraw & Ling, 2003;
2002; Spiliotes & Vavreck, 2002; Xenos & Foot, 2005).
Miller & Wanta, 1996; Page et al., 1987). As citizens and MCs
Although Grimmer’s (2013) study of US senator press
devote more attention to Internet modes of communication
releases finds that “both parties engage in debates about the
and information gathering (Esterling et al., 2013; Fitch, 2016;
same topic[s]” (p. 103), Grimmer does not consider House
Fitch et al., 2005; Johnson, 2013, p. 84; Lupia & Philpot,
members’ press releases and thus potentially neglects the
2005), understanding how MCs behave online becomes more
effects of the institutional differences between the two cham-
important as such behavior may differ from more traditional
bers (Smith, 2007).
mediums of communication (Esterling et al., 2013).
Although several studies focus on the Internet’s influence
Framework
on the political process more broadly (Davis, 1999; Johnson,
2013; Malecha & Reagan, 2012; Oleszek, 2007), as well as
Partisanship Influence on Tone
the Internet’s impact on certain congressional behavior (Adler
We argue that party control of the White House leads MCs to
et al., 1998; Bimber, 1999; Carter, 1999; Druckman et al.,
employ a more positive tone. The president’s copartisans
2010; Owen et al., 1999; Simon, 2002; Xenos & Foot, 2005),
have a vested interest in ensuring that the president’s agenda
very few studies have attempted to “understand the content of
and his performance is viewed favorably by the public for
Congressional communication through the capturing and cat-
several interrelated reasons: First, the increase in political/
egorizing of individual messages” (Golbeck et al., 2010) and
ideological polarization among voters has largely aligned the
how such content is specifically influenced by partisanship.
president’s constituents with those of his copartisans
Furthermore, few studies consider the relationship (Jacobson, 2017, pp. 56–57). Consequently, the policy goals
between partisanship and tone of congressional communica-
of the president and his copartisans should largely be aligned
tions. Tone may reveal systematic patterns of partisanship
(Smith, 2007, p. 72). Second, even if the president’s policy
influence in congressional communication. Tone is also a
agenda diverges from his copartisans on several key issues,
significant component of framing (Brunken, 2006; Gunther,
copartisan MCs still possess an interest in supporting the
1998; Hester & Gibson, 2003), which in turn influences the
president’s general image since, following the increase in
public in various policy debates (Druckman, 2001; Jacobs &
political polarization, MCs are closely attached to their par-
Shapiro, 2000, pp. 50, 61; Krosnick & Kinder, 1990; Malecha
ty’s reputation which, in turn, is closely linked to the presi-
& Reagan, 2012; McCombs & Shaw, 1972; McGraw &
dent and his platform (Bond & Fleisher, 1990; Jacobson,
Ling, 2003; Miller & Wanta, 1996; and Page et al., 1987).
2017, p. 59; Malecha & Reagan, 2012, p. 21; Smith, 2007,
Consequently, understanding MCs’ use of tone is essential to
p. 28). Indeed, midterm elections are often perceived as a
understand MCs’ framing practices. For...