The Subconscious Effect of Subtle Media Bias on Perceptions of Terrorism

• Published date: 01 May 2021
• Author: Katherine T. McCabe,Karsten Donnay,Lukas Feick
• Date: 01 May 2021
• DOI: 10.1177/1532673X20972105
• Subject Matter: Articles

https://doi.org/10.1177/1532673X20972105

American Politics Research

2021, Vol. 49(3) 313 –318

© The Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1532673X20972105

journals.sagepub.com/home/apr

Article

Introduction

The use of specific words or expressions in news media cov-

erage affects the degree to which their audiences perceive

certain violent events as threatening. As a case in point, con-

sider the 2018 surge in homicides in London: While The

Guardian (2019) reported “London killings in 2018: how

homicides in the capital rose to a decade high,” other head-

lines on the topic read “London BLOODBATH” (Express,

2019), “Carnage in the capital” (Daily Mail, 2019), and

“London Bloodsoaked After Five Stabbings in 24 Hours”

(NewsWars, 2019). Correspondingly, the extent of threat and

urgency we associate with the same event varies with the

way it is reported on, or framed, and can play a substantial

role in triggering or sustaining dynamics of insecurity felt by

the public (Baele et al., 2019, p. 520).

This paper uses a survey experiment on a U.S. sample of

respondents to quantify how variation in media framing

influences perceptions of personal threat of violent events.

Specifically, we embed a set of labeling and wording choices

in a multivariate survey experiment alongside a range of

contextual factors—location, death toll, and type of a

reported attack. This design is chosen to capture the full

complexity of actual media reporting where individuals are

exposed to coverage that can vary along all of these dimen-

sions. Our strategy allows us to isolate which specific

aspects of the coverage affect how personally threatening a

person views an incident, on average, while allowing for the

fact that different article cues may also vary, as they would

across actual media coverage.

We find that the probability that an incident is perceived

as more personally threatening increases, on average, if arti-

cle tone is more negative and sensational, the event is geo-

graphically proximate and large in size. Even though article

tone transports no factual information, in our experiment

these wording choices carried a greater impact on threat per-

ceptions than many types of factual information and explicit

“labeling” choices included in news stories. The results have

several implications for the potential impact of media cover-

age on public opinion, as well as normative implications for

the role of media in society.

Existing Evidence and Theoretical

Framework

There exists a rich literature examining the relationship

between incidents of violence and their coverage in the

media. Studies have considered the effects of the labels used

to describe the perpetrators of an incident (Baele et al., 2019;

Montiel & Shah, 2008), the location where it took place

(Finseraas & Listhaug, 2013; Fischhoff et al., 2003), the

972105APRXXX10.1177/1532673X20972105American Politics ResearchFeick et al.

research-article2020

1University of Konstanz, Germany

2University of Zurich, Switzerland

3Rutgers, The State University of New Jersey, New Brunswick, USA

Corresponding Author:

Karsten Donnay, Department of Political Science, University of Zurich,

Affolternstrasse 56, Zurich ZH 8050, Switzerland.

Email: donnay@ipz.uzh.ch

The Subconscious Effect of Subtle Media

Bias on Perceptions of Terrorism

Lukas Feick1, Karsten Donnay2,

and Katherine T. McCabe3

Abstract

Media outlets strategically frame news about violent events using sensationalist labels such as “terrorist” or “Islamist” but

also more subtle wording choices that affect the overall article tone. We argue theoretically and show empirically using a

conjoint experiment that, contrary to existing studies, the effect of these two framing devices on readers’ perceptions of

terrorist events should be carefully separated. Even though article tone transports no factual information, in our experiment

negative and sensational wording choices carried a greater impact on threat perceptions than the explicit “terrorist” and

“Islamist” labels. In a realistic news article setting, which varied other salient context cues such as proximity or event size,

subtle shifts in article tone still subconsciously influenced threat perceptions. This highlights the potential dangers of media

coverage fueling otherwise unjustified fears by injecting unnecessary editorial tone.

Keywords

media bias, framing, labeling, public opinion, conjoint analysis