Media Scandals Are Political Events

• DOI: 10.1177/1065912916684034
• Date: 01 March 2017
• Published date: 01 March 2017
• Subject Matter: Articles

684034PRQXXX10.1177/1065912916684034Political Research QuarterlyNyhan
research-article2017
Article
Political Research Quarterly
2017, Vol. 70(1) 223 –236
Media Scandals Are Political Events:
© 2017 University of Utah
Reprints and permissions:
sagepub.com/journalsPermissions.nav
How Contextual Factors Affect Public
https://doi.org/10.1177/1065912916684034
DOI: 10.1177/1065912916684034
journals.sagepub.com/home/prq
Controversies over Alleged Misconduct
by U.S. Governors
Brendan Nyhan1
Abstract
When political scandals erupt in the press, we usually blame misconduct by public officials, but these episodes are
political events whose occurrence and severity also depend in part on the political and media context. Using data
on U.S. governors, I show that several key factors affect the likelihood and intensity that alleged misconduct will be
politicized by the opposition and publicized by the press. First, lower approval ratings, which decrease the cost of
politicizing and publicizing an allegation, are generally associated with more frequent and intense media scandals. By
contrast, competing news events can crowd potential scandals off the news agenda. However, no evidence is found
that opposition control of state political institutions leads to more media scandal. These results suggest that the
occurrence of media scandal depends more on circumstance than we typically assume.
Keywords
scandal, misconduct, corruption, governor, media, press
Coverage of alleged scandals often dominates political
of media scandal as a political event, focusing on the
news in the United States, potentially increasing public
public recognition and subsequent coverage of alleged
cynicism about government and displacing policy issues
misconduct by a public official as a scandal in the main-
from public debate. While the most familiar examples are
stream press. Using this measurement strategy, I show
presidential scandals like Watergate and the Lewinsky
that the process that generates media scandals is system-
affair, we see a similar pattern at the state level, where
atically influenced by the political context and news envi-
scandals have dominated the headlines in coverage of
ronment. When conditions are favorable for the opposition
governors like Rod Blagojevich and Chris Christie.
party to politicize a potential allegation and for the press
Despite the significance of these episodes, social scien-
to publicize it, chief executives are more likely to fall vic-
tists know very little about when or why scandals occur.
tim to media scandals. By contrast, even well-supported
A greater understanding of this topic can thus help us bet-
allegations may be suppressed or ignored under less
ter understand contemporary American politics.
favorable conditions.
One reason little is known about scandal is that our
Based on this theory, which has previously been tested
understanding of the term is too simplistic. Scandal is
only at the presidential level (Nyhan 2015), I identify
often conceptualized as the public exposure of corrupt,
three political and contextual factors that are likely to
illegal, or unethical behavior by public officials and
influence the vulnerability of chief executives to media
largely (implicitly) treated as an exogenous event. It is
scandal: the chief executive’s approval rating, which
true, of course, that many scandals are the result of the
should affect the costs of pursuing potential scandals for
disclosure of unambiguous misconduct or criminal acts.
the opposition and the press; media congestion, which
However, the conventional definition fails to explain the
should increase the opportunity costs of covering
patterns we observe in the timing of these revelations;
possible scandals and thus make it less likely that they
why some allegations of misbehavior turn into media
scandals when the evidence is ambiguous or contested
1Dartmouth College, Hanover, NH, USA
and others do not; or why media scandals are more preva-
lent in some contexts than others.
Corresponding Author:
Brendan Nyhan, Department of Government, Dartmouth College,
Because underlying patterns of executive misconduct
HB 6108, Hanover, NH 03755, USA.
are largely unobservable, I instead study the occurrence
Email: nyhan@dartmouth.edu

---

224
Political Research Quarterly 70(1)
receive press coverage; and opposition party control of
approach is similar to how scholars study the legal sys-
state political institutions, which should enable investiga-
tem. Although we by definition lack data on unobserved
tions that reduce the cost of identifying potential scan-
violations of the law, careful research designs allow
dals. Each factor is likely to affect the incentives for the
scholars to estimate how political and institutional factors
opposition party to politicize a potential scandal and/or
influence arrests and prosecutions (e.g., Gordon 2010).
for the press to publicize it and thereby change the
More generally, research suggests political factors—
expected likelihood and intensity of media scandal.
not just evidence and facts—may influence the realiza-
I test these hypotheses using a comprehensive new
tion of many important types of events like scandal. For
dataset on U.S. governors for the 1977–2010 period and a
instance, Reeves (2011) considers how political incen-
series of novel research designs that allow me to conduct
tives can even influence disaster declarations. Similarly,
new and more rigorous tests of the effects of political and
Shaw (1999, 414–16) finds that significant campaign
news context on the incidence of media scandal. First, I
events are more likely when presidential candidates devi-
show that lower approval ratings are associated with
ate from expected levels of public support. In other
more frequent and intense media scandals in the future. In
words, the likelihood that an “event” will occur and per-
addition, I show that news congestion reduces the proba-
ceived as meaningful is shaped by the underlying politi-
bility of media scandal. However, opposition control of
cal context, not just the facts of what took place—the
the legislature or attorney general’s office has no effect
same pattern that I describe below for scandal.
on media scandal, contradicting many prior claims. As
This article focuses specifically on the event I call
with presidents, the occurrence of gubernatorial media
“media scandal,” which I use to refer to the identification
scandal seems to depend in part on circumstance and con-
of a controversy involving a chief executive as a scandal
text, not just objective evidence of misconduct.
in mainstream press coverage. This measurement
approach reflects the centrality of news coverage to the
Why We Should Study Media
existence of scandal in contemporary politics (Waisbord
Scandal as a Political Event
2004, 1079). Most importantly, defining the outcome in
this way allows me to distinguish politically important
Until recently, quantitative political science has neglected
cases of alleged misconduct from fringe allegations while
the study of scandal (C. M. Cameron 2002, 655). With the
remaining agnostic about the existence of misconduct in
exception of Nyhan (2015), previous studies that have
any particular case. The result is not only improved theo-
been conducted largely focus on the effects of scandal or
retical clarity and measurement precision but new insight
ethics charges on presidents, members of Congress, or
into how the prevalence and severity of media scandals
other politicians rather than their causes (see, for example,
are influenced by political and other contextual factors.
Basinger 2013; Meinke and Anderson 2001; Rottinghaus
In this study, I focus specifically on media scandal
2014a, 2014b, 2015; Welch and Hibbing 1997).
among U.S. governors, an important topic that has been
A key obstacle to research in this area is the difficulty
largely neglected. Previous studies have been largely
of defining and measuring scandal. The most common
qualitative and tend to rely on post hoc judgments about
approach is to try to study the incidence of official mis-
the existence of scandal from chronologies and other his-
conduct, which is often equated with scandal. torical sources. By contrast, I draw on contemporaneous
Unfortunately, we not only lack an objective definition of
news coverage to construct novel longitudinal measures
misconduct (which is often seen to extend beyond viola-
of the actual incidence of gubernatorial media scandal
tions of the law to various perceived improprieties), but
and to estimate the effects of political and other contex-
we cannot observe the existence or absence of miscon-
tual factors on its likelihood and intensity. This study rep-
duct—the private behavior of public officials is largely
resents the first systematic quantitative analysis of the
unobservable. Others prefer instead to equate scandal
onset and magnitude of gubernatorial media scandal,
with the disclosure of a moral or ethical transgression by
which has previously only been studied among presidents
a public figure (e.g., Markovits and Silverstein 1988).
(Nyhan 2015).1
However, this definition is also inadequate. First, disclo-
The need for this research is great. As the chief execu-
sures of genuine transgressions do not always generate
tives of their states, governors frequently suffer from
scandals (in part because the...