VU Research Portal

Making a drama out of a crisis?

Klemm, C.

2016

document versionPublisher's PDF, also known as Version of record

Link to publication in VU Research Portal

citation for published version (APA)Klemm, C. (2016). Making a drama out of a crisis? A multidisciplinary study of news media coverage of publichealth crises and the role of emotion. Vrije Universiteit.

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

E-mail address:vuresearchportal.ub@vu.nl

Download date: 22. Mar. 2021

Chapter 4

Swine Flu and Hype: A System-atic Review of Media Dramatiza-

tion of the H1N1 InfluenzaPandemic

This chapter is published as: Klemm, C., Das, E., & Hartmann, T. (2014). Swine Flu and Hype: A Systematic Review of Media Dramatization of the H1N1 Influenza Pandemic. Journal of Risk Research. doi:10.1080/13669877.2014.923029

86

AbstractHighly disconcerting at the time, in retrospective the 2009 H1N1 influenza pandemic looks like much ado about nothing. As a consequence, many accused the media of having created an artificial hype or hysteria around the new virus, thus contributing to unwarranted public fear. The current paper set out to examine the validity of such accusations. We integrated empirical findings on whether the media dramatized H1N1 on a global scale through systematically reviewing prior content-analytic studies. We developed a coding scheme specifying three indicators of dramatized media coverage that – together –inform about how mass media coverage about H1N1 may amplify risk perceptions in the public: (a) the volume of media coverage, (b) the media content presented, particularly an overemphasis of threat while neglecting measures of self-protection, and (c) the tone of coverage. Results show that media attention was immense, that news content stressed threat over precautionary measures, while the pattern of coverage tonality remained nebulous due to conflicting findings. The present review also revealed a critical gap in existing knowledge about the tone of media coverage on H1N1, and discusses implications for future research on dramatization of public health risks by the media.

Chapter 4

87Swine Flu and Hype

Spring 2009 brought the emergence of the first influenza pandemic since 1968, and with it the world verged on a swine flu panic. First discovered in Mexico, the new flu virus was sprawling around the globe, with two alarm-ing characteristics: it was affecting the young and healthy, rather than the usual risk groups of the elderly and chronically ill, and it resembled the strain that had caused the fatal Spanish flu in 1918. Scary at the time, the 2009 pandemic retrospectively looks like much ado about nothing. The toll was far from matching early apprehensions about the scale it could reach. According to the World Health Organization (WHO), H1N1 influenza – as swine flu was later termed – caused approximately 18,000 deaths, com-pared to around 250,000-500,000 deaths seasonal influenza causes each year, and 40-100 million people that died of the Spanish flu (Taubenberger & Morens, 2006; World Health Organization [WHO], 2009, 2010, 2012).

When the initially announced risk of a fatal pandemic did not ma-terialize, many began to blame the media of having created an artificial H1N1 hype or hysteria, blowing up the threat to bait for a bigger audience (e.g., Keil, Schönhöfer, & Spelsberg, 2011). “Now I and my colleagues in the media are the ones accused of crying wolf ”, Independent’s health editor Jeremy Laurance (2009) moaned. Epidemiologists and public health schol-ars Bonneux and van Damme (2010, p. 1308) called the H1N1 pandemic an “iatrogenic pandemic of panic”, an artefact caused by the treatment of the problem, that is created by the way health officials and the media han-dled the health crisis. Wagner-Egger and colleagues (2011, p. 461) found that many laypeople viewed the media as villain, as “fear mongering or as a puppet serving powerful interests”. Also, WHO Director-General Marga-ret Chan, and her Assistant Director-General, Keiji Fukuda implicated that the media was a contributing cause of heightened risk perceptions among the public (Durodié, 2011).

Are such accusations valid? It is important to answer this question because the media serves as a key communication platform in a public health crisis (Glik, 2007; Yu, Frohlich, Fougner, & Ren, 2011). Whether they can fulfil this important role, however, depends on the media’s un-tainted reputation as a trusted information source just as much as on the actual information they broadcast. Like in Aesop’s classic fable of the boy who cried wolf, journalists - if perceived as scaremongers - might lose their credibility and their warnings may be overheard when a future risk emerges.

88 Chapter 4

Therefore the current paper set out to examine the validity of such accusations. Although previous studies have examined media coverage of H1N1, very few studies have explicitly examined whether the media dram-atized H1N1, suggesting either that they did or that they did not (Vaster-man & Ruigrok, 2013; Yu et al., 2011). These different conclusions may be due to different operationalization of what constitutes dramatization. Media coverage has been defined as dramatizing if it exaggerates existing risks, if it awards it with a disproportionate amount of attention considering the actual relevance of the threat (e.g., , a media hype; Vasterman, 2005), or if coverage portrays the (health) threat primarily based on arousing or emotional language as well as based on formal features rather than factual ones (Aust & Zillmann, 1996; Dudo, Dahlstrom, & Brossard, 2007; Yu et al., 2011; Zillmann & Gan, 1996; Zillmann, Gibson, & Sargent, 1999; Zill-mann, 2006). Although these definitions have clear merits, they also have a downside: they need an objective reference point as to what constitutes ‘too much’, or ‘exaggerated’, or ‘too emotional’ media coverage. As Kitzinger (1999) points out, answering the question of whether the media dramatize risks often also entails two underlying assumptions, which are rarely expli-cated. The first is normative. It proposes that there is something like ideal risk reporting, such as an objective representation of risks. The second is that the official providers of risk information, namely health officials and journalists, act on a purely ‘scientific’ basis. We ought to be aware of those assumptions when speaking about dramatization of mass media coverage. Both assumptions suggest that research on dramatization seeks to contrast media coverage, either with an explicated definition of ‘ideal journalism’ or objective quantifications of the ‘actual real risk’.

In the present research we introduce a non-normative, quantified approach to dramatization of risk based on scientific theories on risk per-ception, health communication, and media sensationalism. We developed a coding scheme specifying three indicators of dramatized media coverage that – together – may inform about how mass media coverage about H1N1 may amplify risk perceptions in the public: (a) the volume of media cover-age, (b) the media content presented, particularly an overemphasis of threat while neglecting measures of self-protection, and (c) the tone of coverage.

The goal of this study is thus to provide an answer to the question whether the media dramatized H1N1 by means of a systematic review of

89Swine Flu and Hype

prior content-analytical studies on news coverage of H1N1. We include all prior content analyses of media coverage of H1N1 - also those that did not explicitly examine the question of media dramatization - if they include valuable information that may help answer the question of dramatization. In the next section we elaborate on our quantitative operationalization of dramatization and the theoretical and empirical basis for our key constructs.

Quantitative Indicators of Dramatized Media Coverage of Health Risks

Communicating risk in an effective manner often entails “finding compre-hensible ways of presenting complex material that is clouded by uncertainty and is inherently difficult to understand” (Calman et al, 1999, as cited in Berry, 2004, p. 26). Risk communication scholars have provided a rich em-pirical fact base about the way individuals understand and evaluate risk and risk information and evidence abounds that perception of risk by laypeople and by ‘risk assessors’ barely overlap (e.g., Slovic & Weber, 2002; Slovic, 1987). Applied to the current context of media coverage on the H1N1 pan-demic, these findings from risk research indicate that the media can serve as ‘amplifier’ of risk (Kasperson, Kasperson, & Pidgeon, 2003), however, overly negative perceptions of health risks among the public or even public scares need not necessarily be triggered by gross exaggerations in the media. They can result from flawed or biased information processing, due to the way the audience assigns meaning to the information presented (Loewenstein, Hsee, Weber, & Welch, 2001; Slovic, Finucane, Peters, & MacGregor, 2007), and from media external factors such as the general ‘Zeitgeist’ of perceiving the world as a risky place, as well as from the interpretation of the new, emerg-ing risk against the backdrop of a collective memory of similar risk such as past pandemics (af Wåhlberg & Sjöberg, 2000; Durodié, 2011).

As news reporting on a public health crisis such as an emerging pan-demic, relates to various academic disciplines, ranging from health commu-nication to journalism studies, the current review chose an interdisciplinary approach, integrating findings from the various disciplines when develop-ing the following three indicators of dramatization.

90 Chapter 4

Volume of Media CoverageResearch in the field of risk communication found that “giving too much or too prominent space or time” to a certain risk may lead to sensational-ism and dramatize risks (Dunwoody & Peters, 1992). Studies on the social amplification of risk framework found that extensive coverage may serve as risk amplifier, regardless of whether the risk portrayal is actually accu-rate (Kasperson et al., 1988). Similarly, quantity of coverage theory found that increased media coverage often “turns public opinion in a negative di-rection” – regardless of the tone of coverage itself – and can thus produce heightened opposition and fear (Mazur & Lee, 1993, p. 683). Based on these findings, we define the sheer volume of news coverage on H1N1 as a first indicator for how much the mass media (implicitly) suggested H1N1 to be an alarming health risk, i.e., dramatization.

Media ContentStudies in the field of health communication and psychology have pro-vided ample evidence that specific types of information presented – or the lack thereof – can influence individuals’ perceptions and health behaviours (Witte & Allen, 2000; Witte, 1994). One of the most prominent models in this context is Witte’s Extended Parallel Process Model (EPPM). The model distinguishes between information on the severity of and individual susceptibility to a threat from information on preventive measures, their effectiveness (response efficacy) and the efficacy beliefs of the individual to perform these measures (self-efficacy). Witte’s EPPM model has been extensively researched, as well as been applied to health messages in news media (Love, 2011). While risk information and an understanding of the threat is an essential foundation for individuals to develop a motivation to protect themselves, empirical evidence demonstrates that it can also result in fear and maladaptive responses to risk if it comes without information on protective measures to avert the threat (efficacy information) (Floyd, Prenctice-Dunn, & Rogers, 2000; Rogers, 1983).

Furthermore, effects of health messages differ, depending on wheth-er they purely include a mention of protective measures or also explicit statements about the effectiveness of measures (response efficacy). In order to motivate individuals to take preventive measures, news reports should ideally include both, mentions of preventive measures, and statements

91Swine Flu and Hype

about their effectiveness (Witte & Allen, 2000). Based on these findings, we propose that media content - specifically, a one-sided focus on the risk of H1N1 with neglect of available intervention possibilities such as hand washing or vaccination, is a second indicator for how much the mass media (implicitly) suggested H1N1 to be an alarming health risk, i.e., dramatiza-tion.

Tone of CoverageScholars in the field of journalism or media studies, in particular those researching sensationalism, have demonstrated that coverage tonality im-pact audiences’ risk perceptions. Certain attention grabbing production features such as video manoeuvres (e.g., slow motion, eyewitness camera) and decorative effects (e.g., sound effects, fast-pace cutting) are commonly suspected to exaggerate factual content (Grabe, Kamhawi, & Yegiyan, 2009; Grabe, Zhou, & Barnett, 2001). Further, Zillmann and colleagues found that exemplars of victims in news, particularly when emotionally intense, can strongly and lastingly increase (health) risk perceptions (Aust & Zill-mann, 1996; Zillmann, 2006). Based on Zillmann’s findings, Vettehen and colleagues propose to define vivid storytelling through concrete personal stories or interviews with laypeople as a sensationalist or dramatic feature (Hendriks Vettehen, Nuijten, & Beentjes, 2005).

Studies from health communication demonstrate that these findings are likewise applicable to media coverage on health risks. Biener and col-leagues report that the emotional tone of messages about health risks im-pacts message perception, and that negative affect can increase risk percep-tions (Biener, Ji, Gilpin, & Albers, 2004; Loewenstein et al., 2001). Dudo and colleagues (2007) suggest that emotion-laden language and the use of worst-case scenarios can be defined as sensationalist or dramatic. Based on these findings, we propose the tone of H1N1-related media coverage (to-nality as well as formal features) as a third indicator of media dramatization that may have (implicitly) contributed to heightened public risk percep-tions.

92 Chapter 4

Methods

Search ProtocolFor the purpose of the present literature review, we identified relevant pri-or publications through a search in two computerized scientific databases (EBSCO Host, Web of Science) on Aug, 1st 2012. We combined the search term Media or Newspaper or Radio or TV or Online or Twitter or Facebook or YouTube or Blog* or “Content Analysis” or Framing with each of the search terms H1N1, “Swine Flu”, or “Pandemic Influenza”, including only articles published after the H1N1 outbreak. Since Web of Science is a multidis-ciplinary database, we filtered out articles from unrelated disciplines. Our search resulted in 235 articles (EBSCO Host: 42, Web of Science: 193).

Inclusion and Exclusion CriteriaThe 235 articles identified were sorted using a number of inclusion and exclusion criteria (Figure 1).

In the first step we excluded 15 duplicates. We then screened the remaining 220 articles and excluded 19 articles that were not peer-reviewed journal articles, and 12 articles not English-language. We further checked for relevance, considering two inclusion criteria: articles had to refer to the H1N1 pandemic (no other pandemic), and studies needed to analyse media content. We excluded 26 articles that addressed other influenza pandemics and 153 articles for not employing content analyses.

Ten relevant articles remained. We searched in the reference lists of the obtained articles and contacted several researchers to identify further relevant articles. A closer inspection revealed three articles (Pandey, Patni, Singh, Sood, & Singh, 2010; Vasterman & Ruigrok, 2013; Yu et al., 2011) that were not identified in the initial search, which we added to our sys-tematic review. In sum, 13 articles containing content analyses of the mass media coverage of H1N1 were retrieved for the present literature review (Table 1).

93Swine Flu and Hype

Figure 1. Selection process of sampled articles.

94 Chapter 4

Table 1. Summary of countries and media analysed in the reviewed studies

Country Media sampleCanada Globeandmail.com, Vancouversun.com,

Cbc.caRepublic of China (Tai-wan)

The Liberty Times, Apple Daily, United Daily News, China Times

The Netherlands NRC Handelsblad, de Volkskrant, De Telegraaf, Trouw, Algemeen Dagblad, De Pers, Metro, Spits; TV: NOS Journaal, RTL Nieuws, Hart van Nederland

UK Guardian, The Independent, The Daily Tel-egraph, The Daily Mail, The Express, The Sun, The Mirror, The News of the World (& Sunday issues)

US The New York Times, USA Today, Los An-geles Times, Washington Post, CNN.com, MSNBC.com

Main Findings

A summary of the identified 13 articles that analysed news coverage on the H1N1 influenza pandemic worldwide is presented in Tables 1-4. All stud-ies consisted of quantitative analyses. Eight articles investigated the content of H1N1-related news in traditional media outlets. Most studies analysed print newspapers like The New York Times, Washington Post, or China Times, some tabloids such as The Sun, while few examined TV news. Six articles examined new media such as Twitter, YouTube, Wikipedia, and blogs. Al-together, the studies cover the media reporting in 32 European countries, Canada, the U.S., Taiwan, and Australia. However, except for the UK (Hil-ton & Hunt, 2011) and The Netherlands (Vasterman & Ruigrok, 2013), news coverage in European countries was only analysed for the first week of the H1N1 pandemic (Duncan, 2009).

95Swine Flu and Hype

Dramatization Indicator 1: Volume of Media CoverageTen out of 13 studies (Table 2) analysed the volume of H1N1-related cov-erage over the course of the pandemic. All studies demonstrate that overall the H1N1 pandemic presented a major theme on the media agenda; the amount of coverage was immense. In the first four days of the pandemic (27 – 30 April 2009), the three top newspapers of each of 31 European countries published around 650 – 800 articles on H1N1 per day. In the first week (27 April – 3 May) this summed up to a total of 3,463 articles, an enormous number considering the same media together had only published 2,824 articles on all other health-topics in a period of one month (Duncan, 2009). Media attention was similarly huge in the online sphere. A great deal of information was available on the video platform YouTube. In the first two months of the pandemic (April/May) a total of 142 videos contained relevant information about H1N1 influenza. In the early pandemic days, Twitter had over 50,000 tweets per day that mentioned H1N1 or swine flu (Chew & Eysenbach, 2010).

Development of volume of media coverage over pandemic course. The fact that the reviewed studies all analysed different time frames (rang-ing between seven days and eight months) complicated finding overall pat-terns across studies. Still, when integrating findings, three peaks in news coverage on H1N1 emerged, all of which coincided with important real-world events.

1st peak end of April/start of May 09 – pandemic outbreak. All studies demonstrated that H1N1-related coverage was by far the highest at the very start of the pandemic, followed by a rapid fade-away within the first month. The outbreak of H1N1 in Mexico, the spread of the virus and the WHO’s issue of a first outbreak notice in late April 2009 together prompt-ed the first and largest peak (Duncan, 2009; Goodall, Sabo, Cline, & Eg-bert, 2012; Hilton & Hunt, 2011; Tausczik, Faasse, Pennebaker, & Petrie, 2012; Vasterman & Ruigrok, 2013; Yu et al., 2011). Online media such as Twitter, blogs and Wikipedia visits showed an equivalent pattern, yet the surge in media coverage at the start of the H1N1 outbreak was even more immediate (Chew & Eysenbach, 2010; Signorini, Segre, & Polgreen, 2011).

2nd peak July-Sept 09 – declaration of pandemic and first summer wave. A second peak in media attention was triggered by the WHO’s declaration of H1N1 as a “pandemic” on 11 June 2009 together with a first wave in

96 Chapter 4

H1N1 cases in June and July. Five studies provide information on news vol-ume in this period. All but one study (Yu et al., 2011) found that the June events resulted in a second peak in H1N1 news coverage in summer 2009. When comparing the amount of volume during the second peak to the earlier April/May peak, country differences emerged, particularly between the UK and the other countries (Goodall et al., 2012; Hilton & Hunt, 2011; Vasterman & Ruigrok, 2013; Yu et al., 2011). Such differences may be ex-plained by differences in epidemiology, as the H1N1 virus in the UK had a different epidemiology from the other countries. Again, the online sphere revealed similar patterns. Chew and Eysenbach (2010) found that Twitter activity increased during summer 2009, and the study clearly documents that the surge in volume was triggered by the WHO’s declaration.

3rd peak Oct/Nov 09 – mass vaccinations and autumn wave. Lastly, the start of mass vaccination programs paired with a second wave of H1N1 cases (Centers for Disease Control and Prevention [CDC], 2012; Euro-pean Centre for Disease Prevention and Control [ECDC], 2010) resulted in a third peak of media attention in October and November 2009. Six studies investigated news coverage in this period. They all indicate a last spark of media attention in autumn, yet the studies reveal country differ-ences in terms of the intensity of the peak compared to the previous spring and summer peak (Goodall et al., 2012; Hilton & Hunt, 2011; Rachul, Ries, & Caulfield, 2011; Vasterman & Ruigrok, 2013; Yu et al., 2011). In a com-parable fashion to traditional media, Twitter showed an increase in activity end of October, as well as immediate responses to the arrival of vaccines in the U.S. on 6 October (Chew & Eysenbach, 2010).

Dramatization Indicator 2: Media ContentEight studies that we identified explicitly analysed media content. Two out of the thirteen studies chose a deductive approach in the development of content categories, both using as underlying framework the extended par-allel process model (EPPM: Witte 1994; Witte & Allen 2000). The re-maining studies applied an inductive approach, however in doing so, found comparable content categories to those specified in the EPPM. Hence, it is worthwhile to use the EPPM distinction between threat and efficacy information as guideline for our analysis of the content-specific indicators of dramatization.

97Swine Flu and Hype

Threat information: severity of and susceptibility to H1N1. Across countries and studies information on the severity of and susceptibil-ity to H1N1, such as reference to the risk of hospitalization or death, the spread of the disease, and mortality tallies, was by far the leading theme in H1N1 news coverage. Goodall and colleagues (2012) reported that severity information was present in 86% of stories in U.S. news. Since the individual reviewed studies differed in terms of the countries and media analysed as well as in their specific research interests, noteworthy dissimilarities could be observed. Studies found differences in whether news focused on the se-verity of the threat versus individual susceptibility to it. For instance, U.S. news showed a much stronger focus on severity information (Goodall et al., 2012), while UK news showed a stronger focus on the individuals sus-ceptibility, particularly on local infections (Hilton & Hunt, 2011). Chang (2012), who made an interesting addition by not only analysing the pres-ence of threat information but also noting whether messages sounded alarm or stressed the need for alertness, found for Taiwanese news that the main message when tackling the threat of the H1N1 virus, was the need for en-hanced alertness rather than sounding an alarm.

Efficacy information: protective measures. Next to severity, the second most prominent theme in news coverage on H1N1 was personal protection against the disease, i.e., efficacy information (Chang, 2012; Fog-arty et al., 2011; Goodall et al., 2012; Yu et al., 2011). The majority of the reviewed studies looked exclusively at the mention of preventive measures, only few included response efficacy in their analysis. This complicates con-clusions on the prevalence of efficacy information because it is not entirely clear if a found lack of response efficacy information is due to the fact that the news indeed did not report on it, or whether they did, but researchers did not analyse and hence report it in their findings. From those studies that analysed the presence of response efficacy, we can conclude that although news reports frequently included information on protective measures, they only rarely addressed the effectiveness of recommended actions. Taken to-gether, all but one study (Hilton & Hunt, 2011) found that efficacy infor-mation was the second most dominant information in news coverage on H1N1. However, studies that analysed also the presence of response efficacy information in news, find that it was only seldom mentioned.

In sum, all but one of the reviewed studies found that threat infor-

98 Chapter 4

mation accounted for the largest share of information provided in H1N1-related news, while efficacy information was the second most prominent theme. The only study that found efficacy information to exceed threat in-formation was Yu and colleagues’ (2011) study, possibly due to the fact that they measured only severity and no susceptibility information. Results in-dicating whether severity or susceptibility information was presented more frequently were inconsistent; half the studies found severity, the other half susceptibility to be the predominant information.

Dramatization Indicator 3: Tone of Media CoverageMost studies investigating tone of news coverage found little evidence of the media ‘over-hyping’ the HINI pandemic. Studies varied greatly in their operationalization of tone of coverage. While some studies coded an overall tone (e.g., Duncan, 2009, Hilton & Hunt, 2011), or tone in form of frames (e.g., alarming, reassuring; Vasterman & Ruigrok, 2013) others coded refer-ences to emotions, such as fear, (Goodall et al., 2012; Henrich & Holmes, 2011), relief, or anxiety (Tausczik et al., 2012, Chew & Eysenbach, 2010), and one study explicitly analysed sensationalism (Yu et al., 2011). Findings regarding the overall tone are mixed. Duncan (2009) reported that in the first week of the pandemic 70% of articles in European media portrayed factual information, and Hilton and Hunt (2011) found the same for the further course of the pandemic (83% of articles were factual or neutral). However, Vasterman and Ruigrok (2013) find the opposite when analys-ing tonality of frames; 74% of messages contain alarming frames. Like-wise, Chang (2012) finds that most reports adopt high alarm frames. The study differentiates these frames further into alarming and alerting frames though, and finds that the majority of alarming frames have the tone of high alert, rather than sounding alarm. Regarding media referring to cer-tain emotions, Goodall and colleagues (2012) report that around one third of news reports referenced to fear. Half of these suggested too much fear, while the other half did not comment on whether fear response was justi-fied. Tausczik and colleagues (2012) report more negative emotion words in swine flu blogs than in other blogs, yet they find twice as many references to positive emotion words (2%) than to negative emotion words (death-relat-ed, and anxiety-related; 1%). Tausczik’s study reveals interesting intermedia effects also in terms of tonality. For instance, anxiety measured in blogs

99Swine Flu and Hype

preceded information seeking on Wikipedia; both followed similar trajec-tories, peaking shortly after the announcement of H1Nl and then declining rapidly. Further, there was a strong correlation between the number of posi-tive and negative emotion words in blogs and newspapers on the same day. Yu and colleagues (2011), who explicitly analysed sensationalism, reported that around 5% of articles used loaded words such as ‘fatal’ or ‘lethal’, and al-most 20% described worst-case scenarios. All studies investigating changes in sentiment over time found a decrease of negative sentiments over time (Goodall et al., 2012; Tausczik et al., 2012; Vasterman & Ruigrok, 2013).

100 Chapter 4

Tabl

e 2. S

umm

ary

of R

evie

wed

Stud

ies –

Vol

ume o

f Med

ia C

over

age (

Dra

mat

izat

ion

Indi

cato

r 1)

Cou

ntry

Med

iaM

ain

Find

ings

Stud

y31

Eur

opea

n co

untri

es3

top

news

pa-

pers

of e

ach

coun

try1

»Im

men

se vo

lum

e dur

ing

first

pand

emic

week

(397

9 ar

ticles

). »U

K b

y fa

r the

hig

hest

no. o

f arti

cles,

while

Eas

tern

E

urop

ean

coun

tries

hav

e in

gene

ral f

ar lo

wer v

olum

e of

cove

rage

.

Dun

can

(200

9)

The N

ethe

r-lan

ds8

news

pape

rs

(bro

adsh

eet &

ta

bloi

d), 3

mai

n T

V n

ews

»Im

men

se vo

lum

e (22

70 ar

ticles

), av

erag

e of 1

.4 ar

ticles

pe

r new

spap

er/d

ay.

»H

ighe

st vo

lum

e of n

ews r

epor

ts: fi

rst t

wo p

ande

mic

week

s. »M

edia

atte

ntio

n co

mes

in 6

wav

es, 3

mor

e int

ense

pea

ks.

Vaste

rman

&

Rui

grok

(201

3)

UK

8 ne

wspa

pers

(b

road

shee

t &

tabl

oid)

»H

ighe

st vo

lum

e: Ju

ly/A

ug (5

50 ar

ticles

), 2

small

er w

aves

in

Apr

il/M

ay, &

Oct

. »M

edia

atte

ntio

n ac

ross

the p

ande

mic

in th

e UK

diff

ered

fr

om o

ther

coun

tries

.

Hilt

on &

Hun

t (2

011)

US

4 ne

wspa

pers

(b

road

shee

t),

2 on

line n

ews

sour

ces

»H

ighe

st vo

lum

e: lat

e Apr

il/M

ay, w

ith a

sudd

en d

rop

in

volu

me b

y Ju

ne.

Goo

dall

et al

(2

012)

, Yu

et al

. (2

011)

Aus

tralia

5 Sy

dney

TV

sta

tions

1 »H

1N1

was l

eadi

ng h

ealth

stor

y fo

r 8 w

eeks

, rem

aine

d in

th

e top

5 m

ost f

requ

ently

repo

rted

healt

h sto

ries f

orFo

garty

et al

. (2

011)

101Swine Flu and Hype

20 w

eeks

.

Wor

ldwi

deBl

ogos

pher

e »M

edia

atte

ntio

n sh

ows a

rapi

d in

crea

se, p

eak,

and

de-

cline

Apr

il 24

– M

ay 7

, 200

9.Ta

uscz

ik et

al.

(201

2W

orld

wide

Wik

iped

ia »A

t the

pan

dem

ic sta

rt th

e num

ber o

f visi

ts to

Wik

iped

ia

page

s inc

reas

ed ra

pidl

y bu

t ret

urne

d to

bas

eline

with

in a

few

week

s.

Taus

czik

et al

. (2

012

Wor

ldwi

deTw

itter

»Tr

end

in tw

eet v

olum

e is c

ompa

rabl

e to

prin

t med

ia.

»H

ighe

st vo

lum

e of H

1N1-

relat

ed tw

eets:

late

Apr

il/

May

(>1%

of s

ampl

e twe

et vo

lum

e), t

hen

rapi

dly

decli

ne

with

in th

e 1st

week

of M

ay to

≈0.

3%.

Che

w &

Ey-

senb

ach

(201

0),

Sign

orin

i et a

l. (2

011)

Wor

ldwi

deYo

uTub

e »14

2 H

1N1-

relat

ed vi

deos

with

a to

tal d

urat

ion

of 5

69

min

utes

.Pa

ndey

et al

. (2

010)

Not

e. 1 A

naly

sed

med

ia n

ot fu

rther

spec

ified

.

102 Chapter 4

Tabl

e 3. S

umm

ary

of R

evie

wed

Stud

ies –

Med

ia C

onte

nt (D

ram

atiz

atio

n In

dica

tor 2

)

Cou

ntry

Med

iaM

ain

findi

ngs

Stud

yA

ustra

lia5

Sydn

ey T

V

statio

ns1

Thre

at in

form

atio

n »Se

verit

y: 6

3% o

f sta

tem

ents

»Su

scep

tibili

ty: n

.a.E

ffica

cy in

form

atio

n »Pr

even

tive m

easu

res:

37%

of s

tate

men

ts »R

espo

nse e

ffica

cy: n

.a.Bo

th (t

hrea

t & effi

cacy

) in

one n

ews i

tem

: n.a.

Foga

rty et

al.

(201

1)

US

4 ne

wspa

pers

(b

road

shee

t),

2 on

line n

ews

sour

ces

Thre

at in

form

atio

n: 6

7.1

% o

f new

s rep

orts

(Stu

dy 2

2 ) »Se

verit

y: 8

6% o

f new

s rep

orts

(Stu

dy 1

) »Su

scep

tibili

ty: 3

0% o

f new

s rep

orts

(Stu

dy 1

) »E

ffica

cy in

form

atio

nPr

even

tive m

easu

res:

»St

udy

1: 5

6% |

Stud

y 2:

74%

of n

ews r

epor

ts »R

espo

nse e

ffica

cy: 1

5% o

f rep

orts

that

addr

esse

d pr

even

-tiv

e mea

sure

s clea

rly su

gges

ted

effec

tiven

ess,

8% q

ues-

tione

d eff

ectiv

enes

s »Bo

th (t

hrea

t & effi

cacy

) in

one n

ews i

tem

: 47%

of n

ews

repo

rts

Goo

dall

et al

. (2

012)

, Yu

et al

. (2

011)

UK

4 ne

wspa

pers

(b

road

shee

t),2

onlin

e new

s

Stud

y lo

oked

at th

reat

and

effica

cy in

form

atio

n in

det

ail,

belo

w a m

ain

sum

mar

y of

the t

rend

in th

emes

(plea

se se

e or

igin

al ar

ticle)

.

Hilt

on &

Hun

t (2

011)

103Swine Flu and Hype

sour

ces

Thre

at in

form

atio

n: »Se

verit

y: p

redo

min

ant t

hem

e »Su

scep

tibili

ty: 2

nd m

ost p

rom

inen

tE

ffica

cy in

form

atio

n »Pr

even

tive m

easu

res:

3rd

mos

t pro

min

ent

»R

espo

nse e

ffica

cy: n

.a.Bo

th (t

hrea

t & effi

cacy

) in

one n

ews i

tem

: n.a

Chi

na (T

ai-

wan)

All

4 m

ajor

news

pape

rs in

Ta

iwan

»Th

reat

info

rmat

ion:

38%

of h

eadl

ines

, 35%

of a

rticle

s »E

ffica

cy in

form

atio

n: 1

8% o

f hea

dlin

es, 1

5% o

f arti

cles

»Bo

th (t

hrea

t & effi

cacy

) in

one n

ews i

tem

: 20%

of s

torie

s

Cha

ng (2

012)

Can

ada

3 on

line n

ews

sour

ces

»Th

reat

info

rmat

ion:

n.a.

»E

ffica

cy in

form

atio

n: 6

% o

f com

men

ts »Bo

th (t

hrea

t & effi

cacy

): n.

a.

Hen

rich

&

Hol

mes

(201

1)

Wor

ldwi

deTw

itter

»Th

reat

info

rmat

ion:

0.2

%-1

.3%

of t

weet

s »E

ffica

cy in

form

atio

n: tw

eets

relat

ing

to h

and-

hygi

ene

(2-3

%),

facia

l mas

k (<

7%),

vacc

inat

ion

(1%

-3%

) of a

ll H

1N1

relat

ed tw

eets

»Bo

th (t

hrea

t & effi

cacy

) in

one n

ews i

tem

: n.a.

Sign

orin

i, Se

gre,

& P

olgr

een,

(2

011)

Wor

ldwi

deYo

uTub

e »61

% (n

=87)

of v

ideo

s had

thre

at an

d/or

effica

cy in

form

a-tio

n (n

ot an

alys

ed se

para

tely

)Pa

ndey

et al

. (2

010)

Not

es. 1

Ana

lyse

d m

edia

not

furth

er sp

ecifi

ed. 2

Stud

y 1

(Goo

dall,

Sab

o, C

line,

and

Egb

ert,

2012

) loo

ked

at se

verit

y an

d su

scep

tibili

ty in

form

atio

n se

para

tely,

whi

le stu

dy 2

(Yu,

Fro

hlich

, Fou

gner

, and

Ren

, 201

1) lo

oked

at th

em to

geth

er (i

.e., t

hey

look

ed at

risk

info

rmat

ion

in g

ener

al).

104 Chapter 4

Tabl

e 4. S

umm

ary

of R

evie

wed

Stud

ies –

Ton

e of C

over

age (

Dra

mat

izat

ion

Indi

cato

r 3)

Cou

ntry

Med

iaM

ain

findi

ngs

Stud

y31

Eur

opea

n co

untri

es3

top

news

pa-

pers

of e

ach

coun

try1

»In

the fi

rst p

ande

mic

week

, the

vast

majo

rity

of ar

ticles

we

re fa

ctua

l (70

%),

the s

econ

d lar

gest

bulk

was

supp

ort-

ive.

Dun

can

(200

9)

The N

ethe

r-lan

ds8

news

pape

rs

(bro

adsh

eet &

ta

bloi

d), 3

mai

n T

V n

ews

»N

ews c

over

age w

as m

ainl

y ala

rmin

g (7

4% o

f mes

sage

s). »To

ne ch

ange

d be

twee

n di

ffere

nt st

ages

(alar

m st

age,

prep

arat

ory

stage

, cris

is sta

ge).

As t

he p

ande

mic

pro-

gres

sed,

alar

min

g m

essa

ges d

ecre

ased

, whi

le re

assu

ring

and

neut

ral m

essa

ges i

ncre

ased

.

Vaste

rman

&

Rui

grok

, (20

13)

US

4 ne

wspa

pers

(b

road

shee

t),

2 on

line n

ews

sour

ces

»33

% o

f new

s rep

orts

refe

renc

ed to

fear

, half

sugg

est-

ing

too

muc

h fe

ar; t

he o

ther

half

not

com

men

ting

on

whet

her f

ear w

as ju

stifie

d. »Th

e am

ount

of r

efer

ence

s to

fear

dec

reas

ed o

ver t

he

pand

emic

cour

se.

»Se

nsat

iona

lism

: 5%

of a

rticle

s use

d em

otio

n-lad

en la

n-gu

age,

11%

por

traye

d H

1N1

as a

glob

al pa

ndem

ic, 8

%

used

wor

st-ca

se sc

enar

ios.

Goo

dall

et al

. (2

012)

; Yu

et al

. (20

11

Wor

ldwi

deBl

ogos

pher

eW

ikip

edia

»Bl

ogs i

nclu

de tw

ice as

muc

h po

sitiv

e (2%

) tha

n ne

gativ

e em

otio

n wo

rds (

1%).

»O

ver t

ime e

xpre

ssio

ns o

f neg

ativ

e em

otio

ns d

ecre

ased

sig

nific

antly

, whi

le po

sitiv

e em

otio

ns in

crea

sed.

Taus

czik

et al

. (2

012)

105Swine Flu and Hype

Can

ada

3 on

line n

ews

sour

ces

»To

nalit

y no

t dire

ctly

mea

sure

d, b

ut re

fere

nce t

o fe

ar:

»Fe

ar o

f H1N

1 (1

9% o

f com

men

ts po

sted)

, of v

accin

e (8

%).

Hen

rich

&

Hol

mes

(201

1)

UK

4 ne

wspa

pers

(b

road

shee

t), 2

on

line

»Th

e majo

rity

of ar

ticles

are n

eutra

l (83

%),

13%

are

alarm

ing,

4% ar

e rea

ssur

ing.

Hilt

on &

Hun

t (2

011)

Chi

na (T

ai-

wan

All

4 m

ajor

news

pape

rs in

Ta

iwan

»M

ost r

epor

ts ad

opt h

igh

alarm

fram

es (3

8% o

f hea

d-lin

es, 3

5% ar

ticles

).C

hang

(201

2)

Wor

ldwi

deYo

uTub

e »16

% (2

3 vi

deos

) wer

e misl

eadi

ng m

ostly

por

trayi

ng

H1N

1 as

ove

rhyp

ed, c

allin

g it

a man

mad

e con

spira

cy o

r go

vern

men

t pro

paga

nda.

Pand

ey et

al.

(201

0)

Wor

ldwi

deTw

itter

»Tw

eets

com

mon

ly ex

pres

sed

hum

our (

13%

), co

ncer

n (1

2%),

ques

tions

(10%

).C

hew

& E

ysen

-ba

ch (2

010)

Not

e. 1 A

naly

sed

med

ia n

ot fu

rther

spec

ified

. 2

106 Chapter 4

Discussion

The present literature review set out to examine if the mass media drama-tized H1N1, as suggested by criticism of the media during and after the 2009 pandemic (Bonneux & Damme, 2010; Keil et al., 2011; Laurance, 2009; Wagner-Egger et al., 2011). We systematically reviewed existing content-analytical studies of H1N1 coverage including also those that did not explicitly examine the question of media dramatization. Drawing on findings from risk research, health communication, journalism, and media studies, we developed a coding scheme specifying three features of H1N1-related media coverage that indicate dramatization: volume of coverage, media content (specifically, focus on threat versus efficacy information), and emotional tone of coverage.

Was the Volume of Coverage Expedient – or Excessive?The 2009 H1N1 pandemic received immense media attention, as all of the 13 studies we reviewed found. Our review revealed that attention was not parallel to the trajectory of the epidemic, i.e., reflecting the number of in-fections, but was instead rather event-oriented. Precisely, we identified three peaks in media attention triggered by real-world events. Judging whether the amount of attention the media gives to certain risks is justified or exces-sive, is a difficult, generally value-laden issue (Kitzinger, 1999). Yet, some of the reviewed studies conclude that the enduring media attention, govern-ment vigilance and the daily fuelling of news with reports on new infections and deaths might have made recipients believe that the threat is bigger than it was (Fogarty et al., 2011; Goodall et al., 2012). Their conclusions are in line with prior research that found that the pure volume of information may lead to risk overestimation and increase feelings of fear (Kasperson et al., 1988; Mazur & Lee, 1993). Based on the present findings, the extensive coverage on H1N1-related risk provides an indication that mass media cov-erage was - perhaps inadvertently - dramatized.

Did Content Emphasize Threat While Neglecting Means of Protection?As regards content-related causes of dramatization, our review showed that

107Swine Flu and Hype

most stories focused on H1N1 as a threat and featured the seriousness of and susceptibility to the new H1N1 virus. We further found that efficacy information, despite being the second most prevalent information in news on H1N1, was far less prevalent than threat information. Interestingly, the reviewed studies draw divergent conclusions from these findings. Goodall et al. (2012, p. 14) concluded that stories “overreported H1N1-related death and hospitalization”. They argued that death and hospitalization were men-tioned too frequently considering the likelihood of these consequences according to the Center for Disease Control and Prevention (CDC). Al-though Fogarty et al. (2011, p. 5) also found that the majority of newspaper statements referred to the seriousness of and susceptibility to H1N1, they concluded that “statements [...] were generally non-alarmist and reassur-ing”. Their conclusion suggests that while media content may fuel risk per-ceptions by stressing the seriousness of health threats, the tone of cover-age may still remain sober and reassuring. These conclusions illustrate that dramatization cannot be defined by a single indicator but emerges in an interplay of several factors, as proposed in the current paper. In sum, we found that content of H1N1-related news coverage may have contributed to dramatization in two ways: First through a disproportionate mention of risk compared to the actual risk as specified by health institutions, which may heighten public perceptions of risk (Goodall et al., 2012). Second, through an emphasis on H1N1-related threat while neglecting efficacy in-formation, which may increase public fear and maladaptive responses in the public (Rogers, 1983; Singer & Endreny, 1987, as cited by Kitzinger, 1999).

Was the Tone of Coverage Dramatic?Concerning the tone of news coverage, the majority of studies found no evidence for media dramatization but rather concluded that news report-ing was factual and non-alarmist (Duncan, 2009; Hilton & Hunt, 2011). However, in stark contrast the two studies analysing frames (Chang, 2012; Vasterman & Ruigrok, 2013) found alarming messages to be predominant; Vasterman and Ruigrok found an immense high number of alarming frames (74% messages). Such sharply conflicting findings could be routed in coun-try differences, such as differences in the actual risk H1N1 posed (epide-miology), media systems or local news cultures. They could also be rooted in the complexity and subjectivity of what constitutes ‘alarm’, though, and

108 Chapter 4

consequences in differences in operationalization. Vasterman and Ruigrok (2013) define alarmist frames when an event is described in terms of risk, including updates on the number of infections, hospitalisation or deaths, as well as a description of the threat as “extremely contagious” or “deadly”. Similarly, Chang defines alarming frames when the article emphasizes the severity of an issue, people’s vulnerability to its threats, and the need for enhanced alertness. Whereas these studies on framing include informa-tion, namely severity and susceptibility, in their definition of alarmist, the other studies distinguish more strictly between information and the tone in which information is presented (Duncan, 2009; Hilton & Hunt, 2011). In part, these studies are also narrower in their definition. For instance, Hilton and Hunt (2011, p. 942) use the same three categories (alarmist, neutral, reassuring) as the other studies, however, employ a more restricted defini-tion of ‘alarmist’ headlines, namely as those that were judged as potentially able to cause the reader anxiety. Yu and colleagues (2011) draw on sensa-tionalism literature and define media coverage dramatic or sensationalist when swine flu is depicted as a global pandemic (i.e., the extreme negative outcomes, worst-case scenario), or a deadly disease (i.e., emotionally-loaded words such as ‘deadly’, ‘lethal’). The described variations in the conceptual-izations combined with the apparent disparities in findings, illustrates the importance to find shared definitions of dramatized, alarmist, or exagger-ated media portrayal of risk.

The present review also revealed a critical gap in existing knowledge about the tone of media coverage on H1N1. Formal features have been identified as a crucial component of dramatization (Grabe, Zhou, & Bar-nett, 2001; Hendriks Vettehen et al., 2005) and have also been researched in the context of other epidemics (e.g., Dudo, Dahlstrom, & Brossard, 2007). However, almost none of the reviewed studies examined formal features commonly linked to the tone of coverage such as images, language, person-alized narratives, or camera effects. An exception was the study by Yu and colleagues (2011) that analysed (emotion-laden language and worst-case scenarios, and found that 5.3% of articles used emotional-laden words such as ‘deadly’, ‘lethal’, ’fatal’, and ‘huge death toll’, and around 11% portrayed H1N1 as a global pandemic (worst-case scenario). Such a distinction be-tween textual information and formal features is paramount, as these two factors have been found to impact risk perceptions in different manners,

109Swine Flu and Hype

both in terms of information processing and in the eventual evaluation of risk (Visschers, Meertens, Passchier, & de Vries, 2008; Zillmann, 2006). Therefore, in the current review, conclusions on the tone of media coverage of H1N1 can only be limited as the sampled studies have only very limit-edly investigated formal features that signal dramatization.

Limitations and Suggestions for Future ResearchAs with any study, the present findings have to be interpreted consider-ing several limitations. First, as only one study explicitly but marginally analysed dramatization of news coverage on H1N1, we were limited to the material at hand. Consequently, we cannot report empirical evidence for all factors that we consider relevant for dramatization but were limited to an analysis of the information on the three indicators that could be distilled from the existing content analytic studies. Particularly, evidence on formal features is lacking. We suggest that future studies explicitly investigate news dramatization in the context of H1N1.

Second, we propose that dramatization emerges in a complex in-teraction of various factors, as has been suggested by risk scholars (Kitzinger, 1999). Yet, none of the reviewed studies illuminates this interplay. As our review looked at three factors and whether they were simultaneously pre-sent, we may conclude that the combination of an immense amount of news on H1N1 with an emphasis on H1N1-related threat possibly contributed to dramatization. Yet, our conclusion is limited by the fact that none of the sampled studies jointly examined all three indicators of dramatization. As to avoid making too simplistic judgments of whether the media dramatized risks, future studies should examine the co-occurrence of various character-istics of dramatization.

Lastly, as stated earlier, the question of whether the media drama-tize risks entails underlying normative assumptions, which are rarely expli-cated. These suggest that research on dramatization often aims to contrast media coverage, either with an explicated definition of ‘ideal journalism’ or with an objective quantification of the ‘actual real risk’ (Kitzinger, 1999). However, risk research has illustrated that rather than being quantified ob-jectively, risk perceptions are highly subjective. Durodié (2011) stresses that risk perceptions are socially mediated. They are a “function of the times” (Durodié , 2001, p. 512), i.e., impacted by our general perception of the

110 Chapter 4

world as a dangerous, risky place, as well as interpreted in the context of our experiences with similar past risks. Further, often the presentation of risk in the media is the outcome of a power struggle behind the scenes, a struggle between several claims-makers who seek to impose their preferred vision of the objective risk (Allan, Anderson, & Petersen, 2010). The fact that some of studies examined in the present literature review reached different con-clusions about the level of dramatization, even if they obtained comparable empirical results, illustrates that the explication of benchmarks should help to advance research on the dramatization of mass media coverage.

Conclusion

Was mass media coverage of H1N1 overly dramatic? Dramatization builds on various factors, three of which we developed based on theory and em-pirical evidence in the realm of risk research, health communication, and journalism studies: excessive media attention, content-related features such as an overemphasis on threat, and the tone of coverage. For the case of H1N1-related news coverage we found that two out of three indicators were prevalent, yet the role of coverage tonality remained nebulous, as find-ings were sharply conflicting. From this, we can conclude that media may have – inadvertently – contributed to heightened risk perceptions through a high volume of coverage and an unbalanced emphasis on the threat of H1N1, however, tonality or formal features as a key factor in determining whether media portrayal was dramatized, needs further empirical investi-gation. To provide deeper insight into what constitutes “drama-laden” vs. “drama-free” news reporting on health risks, future studies should jointly analyse different indicators of dramatization and explicate “benchmarks” of an ideal or non-dramatic coverage.