Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 1 Ebola Joke? You Probably Won’t Get It—Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media Morgan Cullen Conrad Foreman Liad Lehavy
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 1
Ebola Joke? You Probably Won’t Get It—Interreality Comparison of Ebola Threat and
Perceived Threat Expressed in U.S. Media
Morgan Cullen
Conrad Foreman
Liad Lehavy
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 2
Abstract: This study examined American news coverage of Ebola, focusing on episodic framing,
dramatization, and geographical bias in coding articles. A content analysis of 44 randomly
selected articles from The Washington Post, The New York Times, and The L.A. Times from June
1, 2014 to February 28, 2015 demonstrated dramatized and episodically framed news coverage
of Ebola. Gallup survey data indicates American’s believe Ebola to be a much greater concern
in the United States than data from the Centers for Disease Control and Prevention claims. The
news coverage of Ebola also demonstrated geographical bias through an ethnocentric emphasis
on Ebola in the United States while minimizing or omitting the epidemic in Africa. Although we
cannot prove the dramatized, ethnocentric, episodically framed coverage of Ebola caused
American’s skewed perception of the prevalence of the disease in the United States our data
suggests it may have affected it.
Introduction
News coverage of crime can alter the public’s perception of its prevalence in society
(Gans, 1979). The more frequently crime stories are presented in the news, the more prevalent
the viewer may perceive crime to be in his or her environment. Repeated depictions of crime,
especially violent crime, may cultivate perceptions of a violent world (Romer, 2003). Our
research seeks to ascertain if the same understanding is true of news coverage about diseases and
if news coverage of diseases shapes the public’s perception of their prevalence. If so, the more a
specific disease is presented in the news, the more prevalent viewers may believe that disease is
in their country.
This study analyzes coverage of Ebola in the U.S. media. Ebola has been plaguing
Africa’s population for years and corresponding American news coverage is anomalous. Long
after the Ebola crisis began in Africa, the U.S. news media began presenting Ebola as a pressing,
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 3
newsworthy topic and cover it extensively in the second half of 2014. All of a sudden, it seemed
the U.S. news environment became completely saturated with Ebola coverage. This study is
interested in examining what may have led to the spark of Ebola coverage in U.S. media. We
desire to assess if the immense coverage of Ebola in the U.S. cultivated an excessive fear of
contracting Ebola among U.S. residents. If dramatized news coverage of Ebola skews public
perception of its prevalence in the U.S. and/or the likelihood of contracting the disease, the news
may be skewing the perception of the significance of other issues as well. To examine this idea,
we conducted a content analysis of Ebola coverage in U.S. newspaper articles and an analysis of
survey data from the Gallup organization.
Literature Review
In constructing sampling procedures, and methods for examining the sensationalization of
health risks in media, past studies were referred to, and utilized in acquiring methods. Some of
these studies, specifically a 2011 study by Villar and Zamith, broadly examine disease coverage,
and are excellent precursors for studies discerning Ebola sensationalization. In today’s
technology-driven industry controlling news distribution, mass media becomes a main channel of
health communication to the general public. In light of this immense public dependence, media
will influence what audiences will perceive to be important regarding these diseases, coinciding
with agenda setting theories (Villar, Zamith, 2011). Further enabling this agenda setting is how
prevalent a source news is becoming, with people gaining more health knowledge from informal
sources (press reports, internet) than personal experience (Villar, Zamith, 2011). Because of how
epidemics become framed through journalists’ experiences, which shape what aspect consumers
focus on, this study hypothesized that: 1) The frequency of media coverage correlates with
public perceptions of personal and societal risk for these certain diseases 2) Frequency of media
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 4
coverage is not correlated with actual mortality of disease. 3) The public’s perception of personal
and societal risk of diseases does not correlate with actual mortality of diseases/injuries
While mortality data was acquired to assess news environments, a series of article coding
was also completed using data bases similar to those used in our Ebola study. Using key-word
searches in analyzing articles from prominent newspapers, the study measured the public’s
perceived risk of diseases in an online survey, asking respondents to rate their levels of concern
over each inquired-about disease. After mortality data identified the three most, and least-
common causes of death, and content analysis reported that they were accurately represented,
perceived risk surveys showed that respondents believed the causes of death to be greater threats
to society than to themselves (Villar, Zamith, 2011). While survey data reflected accuracy in H1,
the other two hypotheses were not supported. Possible limitations of this study, though, stemmed
from the higher socioeconomic and educational status of the survey sample, perhaps implying
higher media literacy. Additionally the article keyword search approach to coding could have
skewed the way in which media coverage framed health issues. We considered the limitations of
various studies when designing our study’s methodology.
To further discern effective methods of content analysis, we referenced a studied done in
2005 about representations of Mad Cow disease in British newspapers, and how they evolved
during the ten-year period of 1986-1996. Specifically, it looks at risk as a variable in assessing
portrayal-accuracy of diseases, maintaining that risk is often over-reported (Washer, 2005). This
study also examined social representations of how new threats are shaped, and introduced to
news circulation, specifically through shaping of historical events and other symbolism.
Similarly to past and current studies, Washer chose prestige British newspapers,
searching for mentions, and presented contexts of Mad Cow disease within articles published in
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 5
the ten year span. While some trends were found, a theme that emerged in Mad Cow disease
coverage was that of panic (Washer, 2005). This was done mostly by reporting the measures that
schools and other public institutions took to avoid beef. While multiple health experts vocalized
the excessiveness of these measures, panic-inducing words kept appearing in articles. What was
found was that disease’s representations were often determined by authors, but that it was
government, media, and the popularity of health-interest which made Mad Cow more salient
(Washer, 2005).
In studies examining health news reporting, “panic” is a recurring concept. The media
used a dramatized, panicked tone in articles to convey significance of risk to audiences. This is
something we expected to find upon examining recent Ebola coverage. One article published by
the London School of Economics and Political Science in 1995 does this with the concept of
“moral panic.” This is something often incited by politicians, and media in general (McRobbie,
Thornton, 1995). What is so effective in inciting this moral panic, according to the article, is its
ability to inspire the necessary emotional involvement to keep audiences interested in an issue
and partially guaranteeing its relevancy. Due to this persuasive effect, moral panics have now
become a way to bring attention to certain issues media deem important, and that act “on behalf
of the dominant social order” (McRobbie, Thronton, 1995).
A study which builds on this notion of moral panic was done by Sheldon Ungar in 2008,
and examined the dramatization of Bird Flu which, at the time, was not a new phenomenon
(Ungar, 2008). To do this, the study chose “Daily Topics” from news articles, which were
selected if they corresponded to predetermined indicators of fear or reassurance, with indicators
being: 1) Predicted number of deaths from bird flu 2) Coverage of actual deaths due to the bird
flu 3) Spread of the bird flu, especially when threat to Europe was mentioned 4) Medical plans
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 6
and preparations to deal with bird flu announced by EU countries. After analyzing news content
throughout the crisis, results showed that three stages of communication discourse developed,
referred to as “sounding the alarm”, dominated by fearful claims, “mixed messages”, where
continuation of threats was combined with reassurance, and “hot crisis and containment”,
involving efforts to undo the frightfulness of the threat stage (Ungar, 2008).
A similar, earlier study essentially studied the “alarm” stage of Unger’s study, and looked
at whether British media constructed Ebola as a threat in the mid-1990’s. The study, by Helene
Joffe and Georgina Haarhoff in 2002, viewed Ebola as way to study responses to diseases that
have no local implications on the people consuming news about them. Emphasizing the
insignificance that Ebola should have had on British citizen’s perceived-risk at the time, this
study most importantly sought to understand whether Ebola was represented as threat or as
contained and if there were group differences in social representations of Ebola in Britain. To
study this, British newspaper articles covering Ebola during the 1990’s were collected, spanning
48 articles over eight newspapers. A coding framework was then constructed to discern the
different themes related to Ebola, and the images that accompanied them. Researchers then
compiled 50 respondents to undergo interviews asking them of their knowledge of Ebola, how it
is spread, whether they believe they are at risk, and other questions. Analyses of the articles
essentially showed that while newspapers attempted to “make Ebola ‘real’ and relevant” to their
audiences, many consumers treated it as if it were essentially fictional. Readers tended to label
Ebola as an African epidemic, and not something that could realistically infringe on Western
cultural, and medical practices.
This study was conveniently done after Sheldon Unger, the researcher who explored Bird
Flu coverage, also analyzed Ebola coverage before its recent media breakout. Unger’s study
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 7
(1998) argued that media shifted from alarming to reassuring coverage in the face of potential
grass root panic. To infer this, researchers analyzed coverage of Ebola compared to coverage of
other “emerging diseases”, and found that in presenting Ebola, certain coverage packages were
used to frame it at certain points in its development. It was found that initially, after “following
their standard practices,” media framed it in terms of a “mutation-contagion” package--a highly
dramatized, crisis-ridden frame. After presenting Ebola as an “impending epidemic of fear” they
were able to switch to the more reassuring “containment package.” This was ultimately able to
“other” the situation--meaning it located Ebola as a problem unrelated, geographically and
culturally to the Western consumer base. The analysis further suggested that the strategy of
“othering” is not based on any existing cultural boundaries, but is used as a way to curve
audience concerns, and act as reassurance to them (Unger, 1998).
Data/Procedure
The theoretical population in this content analysis is all news media stories about Ebola
in the United States from June 1, 2014 to February 28, 2015. To set the scope of the news articles
to correspond with the scope of American population we desired to examine (the entire
American population on a national scale), we chose three national news organizations: The
Washington Post, The New York Times, and The Los Angeles Times. Through a library database
at the University of Michigan, we searched “Ebola” to retrieve articles from each news
organization. Each resulting article was assigned a number in numerical order from most recent
to least recent—this list of articles was the sampling frame. Each of the three news organizations
had an individual sampling frame from which we randomly selected 15 articles. To do so, we
utilized a random number generator to randomly choose 15 unique numbers. The numbers that
correspond to the articles in the sampling frame were selected to be a part of the sample. This
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 8
was repeated for each news organization, resulting in 15 articles each from The Washington Post
and The New York Times and 14 articles from The Los Angeles Times — a total of 44 randomly
selected articles in our sample (one was removed because it was not relevant to Ebola).
Measurement
To establish inter-coder reliability when coding items from our sample, the 3 coders met
to code 2 articles independently. After each article, we compared our results. We found that we
attained over 80% inter-coder reliability, as 1 coder differed from the others by 1 variable of the
6 variables coded (3 for each article). Since we had attained high inter-coder reliability, we felt
comfortable dividing the articles to have each coder code 15 articles independently, as we would
likely assign the same values regardless of the coder. The articles were assigned by distributing
each coder one article in alphabetical order of first names—article 1 was assigned to Conrad,
article 2 was assigned to Liad, article 3 was assigned to Morgan, article 4 was assigned to
Conrad, and so on.
For the content analysis of Ebola national news coverage, Hypothesis 1 and Hypothesis 3
are relevant. Hypothesis 1 broke down into two variables. The first H1 variable, EPF, denotes the
presence or absence of episodic framing in an article. If the issue in the article is presented as an
event-oriented report that focused on individual instances instead of the broader impact of the
virus, then we coded it as using episodic framing. This variable is set on a 0 to 2 value system. A
0 score means that the article does not contain elements of episodic framing and focuses on the
issue in its broader context. A score of 2 exemplifies extreme episodic framing, where the
majority of the article is focused on one instance while ignoring the broader context of the issue.
A strong prevalence of episodic framing in an article affects blame attribution and may skew the
perception of the prevalence of Ebola by presenting isolated incidences without broader context.
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 9
The second H1 variable is DRA (dramatization). It is characterized by emphasizing emotional
impact of Ebola and exaggerating its burden. This variable is also set on the same 0 to 2 value
system. A 0 score indicates the absence of dramatization, relying on facts and statistics rather
than emotional burden. A score of 2 means the article focuses on the emotional burden of Ebola
and exaggerates the threat it poses.
In the content analysis, Hypothesis 3 is analyzed through Variable 3—GEO. GEO
concerns geographical bias and ethnocentrism in the United States’ news coverage of Ebola. This
variable emphasizes the extent to which an article focuses on the United States compared to
Africa. These articles were coded on a scale from 0 to 3. If the article discussed the Ebola crisis
in Africa without discussing the United States, the article was coded as a 0. A 0 means that the
article had no ethnocentric geographical bias. If Ebola in Africa was not discussed in the article
at all, it was coded as a 3 for the GEO variable. A 3 in the GEO variable shows a high level of
ethnocentric geographical bias for the United States according to the codebook.
To analyze Hypothesis 2, we utilized survey data from Gallup about the public’s
perception of Ebola. We chose to use their data to assure the scope of the news coverage we
were analyzing corresponded to the survey data (both were national). If we distributed a survey
we created, we would not reach a representative group of respondents on the national level. We
used 3 Gallup surveys to look at the public’s perceived threat of Ebola for H2.
One survey — conducted between October 4-5, 2014 — through the Gallup Daily
tracking survey system examined a random sample of 1,016 adults (18 and older) living in any of
the 50 U.S. states and the District of Columbia. The first survey question in this survey reads,
“Now thinking about the Ebola virus that has been in the news, did you, personally, worry
yesterday about getting the Ebola virus, or not?” The second question asks, “How likely do you
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 10
think it is that you or someone in your family will get the Ebola virus– very likely, somewhat
likely, not too likely, or not likely at all?” Question 3 asks, “How confident are you that the
federal government will be able to handle an outbreak of the Ebola virus in this country — very
confident, somewhat confident, not too confident, or not confident at all?” The fourth and final
question from this survey reads, “Which comes closest to your view about the Ebola virus —
[ROTATED: it will not strike the United States at all, there will be a minor outbreak in the
United States, there will be a major outbreak in the United States, but it will not create a crisis,
(or) it will strike the United States and create a crisis]?”
The second Gallup survey conducted telephone interviews (cellular and landline) from
November 6-9, 2014 to a random sample of 828 adult residents (18 years and older) across all 50
U.S. States and the District of Columbia. The sample included 430 men and 398 women. The
random sample was selected using random digit dialing programs. Gallup weighed the samples
to correct for non-response bias, unequal selection probability, and double coverage of telephone
users in the two sampling frames (men and women). This survey asked one open-ended question:
“What would you say is the most urgent health problem facing this country at the present time?"
The third Gallup survey is based on telephone interviews conducted on October 12-15,
2014 with a random sample of 1,01 adults (18 years or older) from all 50 states in the U.S. and
the District of Columbia. This Gallup survey does not disclose the survey question(s) to the
public, but it asked about which issues Americans find most problematic. Together, these 3
Gallup surveys provide insight into the American public’s perceived threat of Ebola and
connects to H2.
Findings
Interreality Comparison of Ebola Threat and Perceived Threat Expressed in U.S. Media 11
The average score of all 44 articles for episodic framing was 0.86, which is just under the
threshold for “significant episodic framing”; the average score for dramatization was exactly 1,
landing right on the mark for “significant dramatization” (the reference line in Table 1). Overall,
24 of the coded articles were found to have some episodic framing (coded as 1 or 2), with 20
showing no episodic framing (coded as 0). 29 of the coded articles were found to have some
degree of dramatization (coded as 1 or 2), with the other 15 articles showing no dramatization
(coded as 0).
This data supports our first hypothesis, H1, that American news articles about Ebola will
be dramatized and use episodic framing in that the majority of articles demonstrated some degree
of episodic framing and dramatization. The average scores for dramatization (1) supports our
hypothesis as well. The average score for episodic framing (0.86) also supports our hypothesis,
though not as strongly.
Table 1
The Washington Post produced the highest number of articles that were coded as “0” for
both episodic framing (8/15) and dramatization (6/15), as can be seen in Table 2 and Table 3.
The Washington Post also had the lowest average score for episodic framing (0.67). However,