UHCM #561918, VOL 0, ISS 0
Testing the Effect of Framing and Sourcing
in Health News Stories
RENITA COLEMAN, ESTHER THORSON, AND
This page lists questions we have about your paper. The numbers displayed at left can be
found in the text of the paper for reference. In addition, please review your paper as a whole
Q1: Au: Please provide a complete reference entry for ‘‘Sherman, Mann, & Updegraff, 2006’’
or remove its in-text citation.
Q2: Au: Please note that anonymous reviewers should not be acknowledged, so the second
footnote was deleted.
Q3: Au: Please cite ‘‘Campos-Outcalt, D. (2004)’’ in the text or remove its reference entry.
Q4: Au: Please spell out ‘‘AEJMC.’’
Q5: Au: Please cite ‘‘Rest, J. R. (1993)’’ in the text or remove its reference entry.
Q6: Au: Please cite ‘‘Signorielli, N. (1993)’’ in the text or remove its reference entry.
TABLE OF CONTENTS LISTING
The table of contents for the journal will list your paper exactly as it appears below:
Testing the Effect of Framing and Sourcing in Health News Stories
Renita Coleman, Esther Thorson, and Lee Wilkins
Testing the Effect of Framing and Sourcing
in Health News Stories
School of Journalism, University of Texas–Austin, Austin, Texas, USA
ESTHER THORSON AND LEE WILKINS
School of Journalism, University of Missouri–Columbia, Columbia,
This study examines whether changing the way news stories report on health can
induce shifts in readers’ perceptions of problems of obesity, diabetes, immigrant
health, and smoking. The authors manipulated two variables in a controlled experi-
ment: the quality of sourcing—the number of sources and their expertise—and the
framing—changing from an episodic, traditional frame to a thematic frame that
incorporated information on context, risk factors, prevention strategies, and social
attributions of responsibility. The authors found that a thematic frame made readers
more supportive of public policy changes and encouraged them to improve their own
health behaviors. However, it did not alter their attributions of responsibility for
health problems from one of blaming individuals to seeing the larger social factors.
Adding richer sourcing to the thematic frame did not increase these effects, nor did
readers find the thematic stories to be more interesting, relevant, believable, impor-
tant, and informative. In addition, there were differential results because of story
topics that represent uncontrolled effects. The implications for improving health
reporting to encourage positive change in society are discussed.
For most people, the news media are their most important and consistent health
information source (Schwitzer et al., 2005). When it comes to increasing awareness
and knowledge of health issues, news media are possibly even more important than
interpersonal communication (Fishman, 2006).
Public health experts are not always satisfied with the way the media report
health news. The focus on individuals and anecdotes at the expense of context and
societal contributions to disease gives people a distorted view of the problem, they
say. Campaigns to change the way the media report health stories to reflect a ‘‘public
health model of reporting’’1are under way (Dorfman, Wallack, & Woodruff, 2005;
Higgins, Naylor, Berry, O’Connor, & McLean, 2006; McManus & Dorfman, 2005),
1We use the term public health model of reporting in accordance with the definition of this
approach used by its framers, Lori Dorfman and Jane Stevens. This may not be the most pre-
cise term for a public health audience, but it is the terminology recognized by journalists as
referring to inclusion of societally focused factors in health reporting.
The authors thank Ph.D. student Liz Gardner for her expert assistance creating the stor-
ies. This research was funded by a grant from the California Endowment.
Address correspondence to Renita Coleman, School of Journalism, University of Texas–
Austin, 1 University Station A1000, Austin, TX 78712, USA. E-mail: firstname.lastname@example.org
Journal of Health Communication, 0:1–14, 2011
Copyright # Taylor & Francis Group, LLC
ISSN: 1081-0730 print=1087-0415 online
3b2 Version Number
Date and Time
7.51c/W (Jun 11 2001)
03/05/11 and 09:50
and is being taught by schools and fellowships such as the University of Southern
California Health Journalism Fellowships Program. However, no research has been
done to determine whether these new reporting methods affect audiences the way
public health experts hope. No empirical evidence exists to support the desired
effects of this model of reporting health stories. The purpose of this research is to
determine whether rich sourcing and thematic framing—two hallmarks of the public
health model of reporting—affect audiences’ perceptions of health issues in the way
that public health experts intend.
The public health model is defined as an approach that sees the causes of death and
injury as preventable rather than inevitable. By studying the interaction among the
victims, the agent, and the environment, this approach seeks to define risk factors,
then develop and evaluate methods to prevent problems that threaten public health.
The goal of the model is to alter the basic conditions in society that give rise to
and sustain such problems (Mercy, Rosenberg, Powell, Broone, & Roper, 1993).
Although developed for use with stories of crime and violence, the public health
model of reporting also has been adopted for use with stories of traditional health
problems such as cancer and obesity (Hatley-Major, 2009), heart disease (Kim,
Kumanyika, Shive, Igweatu, & Kim, 2010), and diabetes (Gollust & Lantz, 2009).
One prominent theoretical basis for the public health model of reporting is framing
theory. Events are ‘‘framed’’ or given a field of meaning within which they can be
understood (Severin & Tankard, 1992). This theory implies that cues learned from
the media also can be used to make sense of our experiences and social situation
(Baran & Davis, 1995). The frames that the media use in stories help define problems
and call attention to some things while obscuring others (Entman, 1993). According
to Entman, frames have at least four functions: to define problems, diagnose causes,
make moral judgments, and suggest remedies. The public health model of reporting
seeks to alter the frame of a health story by focusing on causes of disease, risk fac-
tors, and prevention strategies. Using Entman’s four functions, public health-framed
stories would define problems as caused by social factors in addition to individual
ones; change moral judgments to focus on society and individuals, and offer
additional remedies such as changes to public policies in addition to actions by
Thematic vs. Episodic Framing
The framing work of Iyengar (1991) bears the most resemblance to the framing goals
of the public health model of reporting in that both focus on shifting attention to
who is responsible for causing and treating health problems. Iyengar studied how
framing news influenced whom people blamed for problems and found that tele-
vision news stories on crime framed a certain way that he termed ‘‘episodic’’ led
people to attribute responsibility to individuals, but that crime stories framed differ-
ently, which he called ‘‘thematic’’ led people to attribute responsibility more to
societal causes. His definition of thematic coverage has many commonalities with
2 R. Coleman et al.
the type of reporting advocated by the public health model. Thematic coverage is
associated with increased societal attributions while episodic coverage—the kind
employed by most news stories—is related to increased attributions of individualistic
causal responsibility as well as punitive treatment. Iyengar’s work has been repli-
cated with newspaper health stories, with thematically framed stories leading to
more societal attributions and episodically frames stories leading to individual attri-
butions (Hatley-Major, 2009). Whether health coverage is thematic or episodic mat-
ters because if people attribute some blame for health problems to society, they are
more likely to support changes in public policies, laws, regulations, prices, product
standards and institutional practices (Chapman, 2001), which is necessary to fully
address a health problem. However intuitive this sounds, no research has been done
to determine whether this shift in responsibility actually occurs in audiences who
read health stories written in this style. Nor has research been done to determine
whether thematic or episodic message framing affects people’s intentions to change
their own health behaviors. This has been studied in the context of message framing,
but under the rubric of gain=loss framing rather than Iyengar’s episodic=thematic
framing (Sherman, Mann, & Updegraff, 2006). This study explores whether beha-
vioral intentions also are affected by this type of message framing.
An example of health reporting that includes contextual, thematic information
would be stories on obesity and obesity-related diseases that include information
on research that shows less affluent neighborhoods where obesity rates are higher
have fewer options for healthy foods. Using the public health model of reporting,
a journalist would explain that grocery stores in these areas are less likely to sell fresh
fruits and vegetables compared to stores in more affluent neighborhoods (California
Center for Public Health Advocacy, 2008). It would also explain that poorer neigh-
borhoods are more likely to have fast food restaurants as opposed to wealthier areas
(California Center for Public Health Advocacy). In addition, the story might explain
that poorer neighborhoods have fewer places to exercise and that safety is a concern
for those who seek outdoor exercise for themselves and their children. Explained this
way, it is clear that the problem involves the underlying conditions of limited access
to exercise, availability of healthier foods, and socioeconomics rather than simply
individual responsibility and choice. Incorporating this contextual information
would help readers see that obesity also has underlying social causes and that
individuals struggling with obesity are not just lazy or unwilling to eat less or
Evidence shows that public health issues are rarely described thematically in
news stories (Chavez & Dorfman, 1996; Dorfman & Schiraldi, 2001; Dorfman &
Wallack, 1998; Dorfman et al., 2005; Lawrence, 2004; McManus & Dorfman,
2005; Woodruff, Dorfman, Berends, & Agron, 2003), although some change has
begun to appear on the topic of obesity (Kim & Willis, 2007). Primarily, the episodic
and individual-centered reporting on health issues that currently exist influence
people to think about health problems as something that only individuals are respon-
sible for and capable of changing; they do not emphasize the social factors that
contribute to disease, or the policy changes that could help lower the incidence of
There is evidence that people’s perceptions of risk are subject to large and
systematic biases. These misconceptions undoubtedly influence the way
that people think about and respond to hazards in their personal lives.
Testing Framing and Sourcing3
Such biases may misdirect the actions of public interest groups and
government agencies, resulting in less than optimal control of risk.
(Combs & Slovic, 1978, pp. 837–838).
Thus it is important to learn whether changing the way the media report on
health issues to a public health model can actually change people’s perceptions. If
it is the case that a public health approach to news stories can bring about positive
attitude change, then the media may play an important role in altering the conditions
in society that give rise to disease. If it is not, then efforts should be focused
On the basis of the foregoing, this study makes three predictions regarding
stories framed according to the public health model:
Hypothesis 1: Because the public health model of reporting gives readers
more societal information, readers who see public health
stories will attribute more responsibility for health pro-
blems to society than readers of traditional stories.
Hypothesis 2: Because the public health model of reporting gives readers
more societal information, readers who see public health
stories will endorse public policy changes more than read-
ers of traditional stories.
Hypothesis 3: Because the public health stories frame health problems as
being preventable and offer suggestions for prevention,
readers of public health stories will report greater inten-
tions to change their own health behaviors than readers
of traditional stories.
We also ask one research question regarding how much readers like the public
Research Question 1: Will readers find public health stories more
interesting, relevant, believable, important, and
informative than traditional stories?
Examination of sourcing patterns goes hand-in-hand with framing analysis (Kim &
Weaver, 2003). It has long been acknowledged that those who provide information
are a major influence on how people view issues (Cohen, 1963). Journalists rely on
sources for interpretation, and many argue that journalists frame their coverage via
the sources they use (Kim & Weaver). Research on sourcing patterns has mostly
focused on elite versus ordinary people as sources. However, some work in science
reporting has shown that a greater quantity of sources and experts such as scientists,
research reports, and health professionals (Ramsey, 1999) lend greater credibility to
news stories than do fewer sources and sources that included celebrities or non–
health-related sources (Hatley-Major & Coleman, 2006). There is also evidence that
competent or expert sources without a vested interest in the information are more
credible (Salwen, 1992). Outside of experts as sources, there is some evidence that
4R. Coleman et al.
people who have personal experience with the disease also make credible sources that
can lead people to change their own health behaviors (Frisby, 2006). Furthermore,
‘‘ordinary people’’ are frequently sourced in traditional health stories as are medical
personnel and other authorities (Brannstrom & Lindblad, 1994; Mercado-Martinez,
Robles-Silva, Moreno-Leal, & Franco-Almazan, 2001). This research uses the term
rich sourcing to describe stories with more sources and higher quality sources, and
poor sourcing to describe the opposite.
These ideas are encouraging for public health advocates who wish to change the
way journalists routinely report on health issues. However, most research has shown
that the media have not routinely included greater numbers of expert sources in their
reports (Dorfman, Woodruff, Chavez, & Wallack, 1997; Stevens, 1998), and almost
none has looked at whether these changes have the kinds of effects on audiences that
the researchers and experts hope for.
Main effects of sourcing were predicted for all dependent variables and were
tested along with main effects of framing in Hypotheses 1 through 3. In addition, this
research predicts an interaction effect between the public health=rich sourcing com-
bination of stories, controlling for demographics and other variables:
Hypothesis 4: The combination of public health framing and rich sour-
cing will be more significant than the other combinations
on societal attribution of responsibility, public policy
endorsement, and behavioral intentions.
A between-subjects experiment using a 2?2 factorial design was used to test whether
public health framing and rich sourcing had a causal effect on peoples’ attitudes
toward health problems. The first factor was framing (public health=traditional);
the second factor was sourcing (rich=poor). The repetition factor was four health
topics (obesity, diabetes, immigrant health, smoking). Participants received all four
stories in one condition (public health=rich, public health=poor, traditional=rich,
We first performed a manipulation check to establish that readers could in fact dis-
tinguish between the rich and poor sourcing stories and between the traditional and
public health framed stories. The sourcing questions asked readers to use a 5-point
Likert-type scale ranging from 1 (strongly agree) to 5 (strongly disagree) to determine
whether there were more than three sources in the stories, and if the sources were
experts, credible, believable, and trustworthy. The mean scores of 15 participants
who saw the rich source versions were higher for all questions than the mean scores
of 15 participants who saw the poor source versions. The total number of 30 was too
small to assess statistical significance, but the means were all higher for the rich
source versions, signaling that readers did indeed agree that there were more and bet-
ter sources in those stories. Those 30 manipulation check participants also answered
questions on the same 5-point Likert-type scale about whether the stories empha-
sized social causes, provided statistics and factual information, included context, risk
Testing Framing and Sourcing5
factors and consequences, and offered prevention strategies. One question was
reverse coded that asked whether the stories highlighted individual responsibility
more than social factors. Mean scores were all in the direction predicted, indicating
that readers did indeed detect more public health information in the public health
A total of 136 participants were randomly assigned to the treatment or control
group. Approximately 60% of the participants were students from a large
Southwestern university and the other 40% were adults drawn from the community
by contacting various social groups and organizations whose meetings were
announced online and in newspapers. The average age was 24 years. Sixty-eight
percent were female. Individual group sizes ranged from 31 to 37. Each group
was exposed to one of four versions of a mock news story containing stories about
four health issues.
Four health news stories on the topics of diabetes, smoking, obesity, and immigrant
health were selected from the LexisNexis database. Using real news stories contrib-
uted to the validity and credibility of the stimulus. The stories were re-written and
edited by a former newspaper journalist.
Four versions of each story were constructed; the versions that represented tra-
ditional reporting of newspaper health stories were written with episodic framing,
and emphasis on individuals. These stories were changed only slightly, mainly to
conform to length requirements, from the versions of actual health news stories
obtained from the LexisNexis database. The public health versions incorporated the-
matic framing and societal context. For example, the obesity story included infor-
mation about social factors such as ‘‘More hours at work means more eating
outside the home where larger portions are the norm,’’ and ‘‘increased television
viewing also leads to a fatter America.’’ Base-rate information about the percentage
of Americans with the particular health problem and the rate of increase in that
health issue also were included. Prevention strategies were also offered; for example,
the smoking story included a statement that ‘‘States with strong tobacco control laws
reap strong benefits from the bans. They have markedly lower smoking rates and
fewer people dying or suffering from lung cancer. For example, in the first 18 months
after the town of Pueblo, Colorado, enacted a smoking ban in 2003, hospital admis-
sions for heart attacks dropped 27%. Admissions in neighboring towns without
smoking bans showed no change.’’ The poor sourcing versions were operationalized
with two, low-quality sources—one was a person affected by the health problem, the
other was either a friend or family member of that person, a celebrity spokesperson,
or priest or religious official. The rich sourcing versions used five sources that were
better quality expert sources—again, a person affected by the health problem, but in
addition four sources who were health professionals such as doctors, nurses, county
health office workers, and nutritionists and also researchers or sources from expert
organizations such as the Centers for Disease Control and Prevention or the
National Institutes of Health. This resulted in sixteen stories, one for each of the four
health issues that represented traditional=poor sourcing, traditional=rich sourcing,
6 R. Coleman et al.
public health=poor sourcing, public health=rich sourcing (contact authors for exact
stories). The design allowed the researchers to determine whether there was an effect
for framing alone, for sourcing alone, and for the combination of the different
frames and sources.
The stories were written in the typical length for print news stories, 15 to 20
inches each. The stimulus stories were all approximately the same length and con-
tained the same information except for the public health material; in the traditional
or control condition, additional, unrelated information about the individuals, their
family, and quotations or comments from family, friends, or officials were included
to make the stories the same length. This information was designed to have minimal
or no effect on readers’ perceptions of story liking, public policy, attributions of
responsibility, and behavioral intentions—the dependent variables of interest in this
study. Except for the immigrant health story, all references to ethnicity were elimi-
nated including ethnic-sounding names.
Immediately after reading each story, participants answered a questionnaire measur-
ing the four dependent variables. All were measured on 7-point scales from 1
(strongly agree) to 7 (strongly disagree). Participants completed the same question-
naire with slight wording changes appropriate to the story immediately after reading
each story until they read all four stories. The order in which participants received
the stories was rotated so that each story topic appeared first, second, third, and
fourth an equal number of times. The dependent variables were operationalized as
We measured story liking with an index of five questions that asked how inter-
esting the story was, how relevant to their lives, how believable, important, and
informative it was, and how much they liked it (Cronbach’s a¼.78).
We measured attribution of responsibility (Cronbach’s a¼.75) with an index of
11 questions including ‘‘Diabetes is a societal problem,’’ ‘‘Diabetics have only them-
selves to blame’’ (reverse coded), ‘‘If people work hard they can almost always man-
age their diabetes,’’ ‘‘Fatty food and ‘super-sized’ portions in restaurants are partly
to blame for diabetes,’’ ‘‘Communities can help prevent diabetes by offering safe,
free places to exercise,’’ and ‘‘People with diabetes are innocent victims.’’ Question
wording for all dependent variables was varied slightly to be appropriate to the
health issue of the story. For example, one immigrant health care question was
reworded to say ‘‘If people work hard they can almost always find a way to obtain
health care’’ (reverse coded). (Contact authors for exact question wording and
The public policy endorsement index (Cronbach’s a¼.77) asked to what extend
participants favored or opposed five policy changes including requiring health care
companies to pay for immigrants’ health care treatment and prevention, giving dis-
counts on health insurance premiums to immigrants, and providing tax breaks to
employers who provide health services to immigrants.
The intention to change behavior index (Cronbach’s a¼.724) included four
questions about the participants’ likelihood of doing things in the next 30 days, such
as these for the obesity story: ‘‘buy more fruits, vegetables, and salads,’’ ‘‘park far-
ther away and walk more,’’ ‘‘take the stairs instead of the elevator,’’ and ‘‘go to the
gym or exercise at home more often.’’
Testing Framing and Sourcing7
After completing all four stories, another questionnaire asked about parti-
cipants’ media use, political party identification, liberal-to-conservative ideology,
and the usual demographics.
We used analysis of covariance to determine whether significant differences exist
between the different types of stories, which will determine whether the public health
framing, rich sourcing, or a combination had a causal effect on participants liking
for the stories, attribution of societal responsibility, public policy endorsement,
and intentions to change their behavior. The covariates statistically controlled for
included age, race, education, gender, having an interest in health issues, and paying
attention to health news. Because this is a controlled experiment, we were able to
determine whether the story framing and sourcing had a direct causal effect on par-
ticipants’ attitudes regardless of individual differences.
Hypothesis 1 predicted that because the public health model of reporting gives
readers more societal information, readers of public health=rich sources stories will
attribute more responsibility for health problems to society than readers of tra-
ditional stories. This hypothesis was not supported. Neither public health framing
(F¼.915, p¼.340) nor rich sourcing (F¼.145, p¼.740) caused readers to attribute
responsibility to society more than traditionally framed and sourced stories. The
covariates of gender (F¼4.14, p<.05) and education (F¼4.2, p<.05) were signifi-
cant, however (see Table 1). Women (M¼4.5, SD¼0.57) were more likely than men
(M¼4.23, SD¼0.82) to blame society, as were those with higher education levels
(some graduate school: M¼4.8, SD¼0.77; high school: M¼4.3, SD¼0.67).
Hypothesis 2 predicted that readers who saw public health stories would endorse
public policy changes more than readers of traditional stories because the public
health model of reporting gives readers more societal information. This hypothesis
was supported. Public health framing was significantly more likely to cause people
to support public policy changes than traditional framing (F¼5.6, p<.05) after con-
trolling for demographics. Significant covariates included gender (F¼9.89, p<.01),
age (F¼5.79, p<.05), and education (F¼7.13, p<.01). Again, women (M¼5.0,
(M¼4.44, SD¼0.89); as were those with more education (some graduate school:
M¼5.7, SD¼0.88; high school: M¼4.6, SD¼1.08). There was no significant main
effect for sourcing (see Table 1).
Hypothesis 3 predicted that because public health stories frame health problems
as being preventable and offer suggestions for prevention, readers of public health
stories will report greater intentions to change their own health behaviors than read-
ers of traditional stories. This hypothesis was supported. Public health framing
(M¼4.0, SD¼1.1) was significantly more likely to cause people to say they intended
to change their own behavior than traditionally (M¼3.6, SD¼0.94) framed stories
(F¼4.5, p<.05) after controlling for demographic covariates. Significant covariates
included gender (F¼20.4, p<.001) and race (F¼8.58, p<.05). Women (M¼4.1,
SD¼0.86; men M¼3.2, SD¼1.10) and Hispanics (M¼4.5, SD¼1.2) the most
likely to change their behavioral intentions. Again, there was no main effect of sour-
cing (see Table 1).
weremore likely to supportpolicychanges thanmen
8R. Coleman et al.
Table 1. F value, eta squared, means, and standard deviations of framing, sourcing, the interaction of framing and sourcing, and
covariates on societal attribution, policy support, behavior change, and liking of stories
F¼.915, ? g g¼.007
F¼5.6?, ? g g¼.042
F¼4.5?, ? g g¼.034
F¼2.24, ? g g¼.017
F¼.145, ? g g¼.001
F¼3.32, ? g g¼.026
F¼.696, ? g g¼.005
F¼2.43, ? g g¼.019
F¼.007, ? g g¼.001
F¼1.31, ? g g¼.01
F¼1.5, ? g g¼.012
F¼.03, ? g g¼.001
F¼2.68, ? g g¼.021
F¼5.79?, ? g g¼.044
F¼.056, ? g g¼.001
F¼.058, ? g g¼.001
F¼4.2?, ? g g¼.032
F¼7.13??, ? g g¼.053
F¼.009, ? g g¼.001
F¼1.09, ? g g¼.009
F¼4.14?, ? g g¼.032
F¼9.89??, ? g g¼.072
F¼20.4???, ? g g¼.138
F¼1.13, ? g g¼.009
F¼1.26, ? g g¼.01
F¼9.89??, ? g g¼.072
F¼6.5?, ? g g¼.049
F¼1.37, ? g g¼.001
Research Question 1 asked whether readers would find public health stories
more interesting, relevant, believable, important, and informative than traditional
stories; we found no differences. The public health stories did not cause readers to
find them significantly more likable than the traditionally written stories (F¼2.24,
p¼.137) (see Table 1).
Hypothesis 4 predicted that the combination of public health framing and rich
sourcing would be more significant than the other combinations on the dependent
variables of societal attribution of responsibility, public policy endorsement, and
behavioral intentions. This was not supported as there were no significant interac-
tions of framing and sourcing on any of the dependent variables (see Table 1).
The Effects of Individual Stories
This study involved replications across four story topics—immigrant health, smok-
ing, obesity, and diabetes. It was expected that the pattern of results would be the
same for each of these stories, given the careful controlling for length, presentation
style, and so forth. It is interesting to note that some results that were not hypothe-
sized varied according to these story topics. For example, rich sourcing in the dia-
betes story had a significant effect on policy support, and a nearly significant
effect in the obesity story; people who read the rich sourcing versions of those two
stories were more likely to support policy changes, but not readers of the rich-
sourced smoking or immigrant health stories. The immigrant health story did show
a significant effect of rich sourcing on people’s intention to change their own health
behaviors, and also a significant interaction effect of sourcing and framing on beha-
vioral change. The smoking story had no significant effects of either sourcing or
framing on any of the dependent measures. Furthermore, the smoking story per-
formed differently from the other three stories on the attribution of responsibility
index. The Cronbach’s alpha measure of internal consistency was too low for the
smoking story to support creating an index of questions that measured how much
people blamed society or individuals for smoking-related problems; that was not
the case in the other three stories. These inconsistent findings support the idea that
health story topics matter as much or more than sourcing and framing and represent
The results of this study were a mixed bag. The most central hypothesis was that
richer sourcing and public health framing would cause readers to attribute health
problems to societal problems, rather than blaming the individuals who were fea-
tured in the stories. However, neither sourcing nor framing affected attribution to
society. This may be because a single story—or even four—cannot have a great effect
on people who are so accustomed to blaming individuals for their health problems
and are not used to thinking, as public health professionals do, that the greatest
effect on health comes from people’s environments. Although framing effects have
been found with one or only a few stories for other topics, it was not the case for
these health news stories. It is possible that these few stories could not adequately
replicate the real world where audiences experience repeated exposure to health
news. The change in the context of political discourse about responsibility generally
may be another explanation; the political discourse about responsibility has changed
10 R. Coleman et al.
so dramatically since the early 1990s when Iyengar conducted his research that it
now takes more stimuli to move the responsibility needle.
More encouraging was the fact that public health framing did cause people to
support public policy changes more than traditional framing. This suggests that
the more reporters come to understand public health framing and use it, the more
people will understand and support the environmental changes that can positively
influence people’s health and be willing to support policies that would change those
There were no effects of the better and more authoritative sourcing on support
for public policy changes. Because ordinary people who have experience with the
health issue, medical personnel and other authorities are most often the sources of
traditional health stories (Brannstrom & Lindblad, 1994; Mercado-Martinez et al.,
2001), it is possible that our manipulation was not strong enough. News conventions
usually call for a minimum of three sources per story; our traditional stories had two
and our public health stories had five – even though readers could detect the differ-
ence in manipulation checks, it may not have been a big enough difference to affect
attitudes and behavior. Further, people with health problems are traditional sources
of health stories, and have been found to be highly credible, so it would have been
unrealistic for us not to include these credible, real people in our traditional stories.
It appears that the traditional way journalists source stories is already quite good,
and that adding more expert sources would not reap enough rewards to be worth
the extra effort. Again, women and those with more education were more likely
to support public policy changes, regardless of sourcing or the presence of a
public health frame, and this was also somewhat expected from previous research
(Brannstrom & Lindblad, 1994).
Public health framing also caused people to be more likely to intend to change
their own health-related behaviors. There was no effect of the quality of sourcing
on willingness to change self behaviors. It is interesting to note that women and
Hispanics were more likely to be willing to change their own health-related beha-
viors, regardless of condition.
Last, we expected that the very ‘‘best’’ stories, those with a combination of rich
sourcing and a public health frame, would show enhanced effects over either of the
variables by themselves. This did not occur.
Perhaps the most problematic aspect of the results was that there were not con-
sistent effects across the four story topics. In some of the stories, the missing effect of
rich sourcing did actually appear. For example, rich sourcing in the diabetes story
had a significant effect on policy support, and a nearly significant effect in the obes-
ity story; people who read the rich sourcing versions of those two stories were more
likely to support policy changes, but not readers of the rich-sourced smoking or
immigrant health stories. Again, why rich sourcing would have an effect in some
stories and not others suggests sources of uncontrolled variation in story topics.
We note that even Iyengar’s (1991) results on attributions of responsibility for crime
showed strong interactions with the subject matter.
While we continue to be puzzled about this result, it is possible that, at this point
in history, smoking represents a special case. Even though smoking is a powerful
addiction, public health messages through the media, through many other interper-
sonal channels, and through public policies such as the establishment of nonsmoking
sections in restaurants have been consistent for decades about the potentially deadly
effects of smoking (Dorfman & Wallack, 2007). The fact of the matter is that people
Testing Framing and Sourcing 11
still smoke despite all this negative information. It is perhaps unreasonable to sug-
gest, in this environment with multiple messages from multiple sources, that a single
news story—no matter how it is crafted—would have an effect on a belief and beha-
vior system that has developed over many years of messages about this particular
health habit regardless of whether people themselves smoke.
The results regarding the immigrant health stories obviously do not fit this
explanation. We note that we tested this on participants who reside in a Southwest-
ern state where immigrants represent significant portions of the population,
assuming that background knowledge of the issue in our participants would make
it more likely for us to find effects of our manipulations. Although we did not test
for this, it is possible that our subjects may be cognitively overwhelmed by the poli-
tics of the illegal immigration political debate. While there is some empirical evidence
(plus a lengthy history) to suggest plausible alternative explanations on the results of
the smoking portion of the experiment, all we can say at this point is that communi-
cation about public health questions as they are applied to immigrants is certainly
deserving of additional research.
Because so much of health reporting training focuses on improving the level and
quality of sourcing, in order to understand its effect on people, it is important to
identify what it is about that sourcing that creates the desired effect on readers or
whether having journalists concentrate on obtaining more hard-to-reach sources is
The strength of this study is that it succeeds in showing just how important pub-
lic health framing is, while ruling out some of the approaches that cost journalists
time and effort but may not produce desired results. Although public health framing
does not affect attribution of health problems to society rather than to individuals, it
leads to more endorsement of changing public policy in ways that encourage better
health. It stimulates people to think about changing their own health behaviors. It
therefore provides important support for the basic concept underlying training
reporters to use a public health frame.
Baran, S. J., & Davis, D. K. (1995). Mass communication theory: Foundations, ferment and
future. Belmont, CA: Wadsworth.
Brannstrom, I., & Lindblad, I.-B. (1994). Mass communication and health promotion: The
power of the media and public opinion. Health Communication, 6(1), 21–36.
California Center for Public Health Advocacy. (2008). Designed for disease: The link between
local food environments and obesity and diabetes. Retrieved November 1, 2008, from
Campos-Outcalt, D. (2004). HIV prevention enters a new era. The Journal of Family Practice,
Chapman, S. (2001). Advocacy in public health: Roles and challenges. International Journal of
Epidemiology, 30, 1226–1232.
Chavez, V., & Dorfman, L. (1996). Youth and violence on local Spanish language television
news. International Quarterly of Community Health Education, 7, 121–138.
Cohen, B. (1963). The press and foreign policy. Princeton, NJ: Princeton University Press.
Combs, B., & Slovic, P. (1979). Newspaper coverage of causes of death. Journalism Quarterly,
56, 837–843, 849.
Dorfman, L., & Schiraldi, V. (2001). Off balance: Media coverage of youth crime. Guild
Practitioner, 58, 75–78.
12 R. Coleman et al.
Dorfman, L., & Wallack, L. (1998). Alcohol in the news: The role for researchers. Contempo-
rary Drug Problems, 25(1), 65–84.
Dorfman, L., & Wallack, L. (2007). Moving nutrition upstream: The case for reframing obes-
ity. Journal of Nutrition Education and Behavior, 39, S45–S50.
Dorfman, L., Wallack, L., & Woodruff, K. (2005). More than a message: Framing
public health advocacy to change corporate practices. Health Education & Behavior, 32,
Dorfman, L., Woodruff, K., Chavez, V., & Wallack, L. (1997). Youth and violence on local
television news in California. American Journal of Public Health, 87, 1311–1316.
Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of
Communication, 43(4), 51–58.
Fishman, J. M. (2006). Mass media and medicine: When the most trusted media mislead.
Mayo Clinic Proceedings, 81, 291–293.
Frisby, C. (2006). A matter of life and death: Effects of emotional message strategies on
African American women’s attitudes about preventative breast cancer screenings. Journal
of Black Studies, 37(1), 103–126.
Gollust, S. E., & Lantz, P. M. (2009). Communicating population health: Print news media
coverage of type 2 diabetes. Social Science & Medicine, 69, 1091–1098.
Hatley-Major, L. (2009). Break it to me harshly: The effects of intersecting news frames in
lung cancer and obesity coverage. Journal of Health Communication, 14, 174–188.
Hatley-Major, L., & Coleman, R. (2006, August). From the doctor’s office to the pulpit:
Creating more effective HIV=AIDS messages for African Americans by examining source
credibility and evidence format. Paper presented to the Minorities & Communication
Division of AEJMC.
Higgins, J. W., Naylor, P. J., Berry, T., O’Connor, B., & McLean, D. (2006). The health buck
stops where? Thematic framing of health discourse to understand the context for CVD
prevention. Journal of Health Communication, 11, 343–358.
Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago:
University of Chicago Press.
Kim, A. E., Kumanika, S., Shive, D., Igweatu, U., & Kim, S.-H. (2010). Coverage and
framing of racial and ethnic health disparities in U.S. newspapers, 1996–2005. American
Journal of Public Health, 100, S224–S231.
Kim, S.-H., & Willis, L. A. (2007). Talking about obesity: News framing of who is responsible
for causing and fixing the problem. Journal of Health Communication, 12, 359–376.
Kim, S. T., & Weaver, D. H. (1993). Reporting on globalization: A comparative analysis of
sourcing patterns in five countries’ newspapers. Gazette: The International Journal for
Communication Studies, 65, 121–144.
Lawrence, R. G. (2004). Framing obesity: The evolution of news discourse on a public health
issue. Press and Politics, 9(3), 56–75.
McManus, J., & Dorfman, L. (2005). Functional truth or sexist distortion? Assessing a
feminist critique of intimate violence reporting. Journalism, 6, 43–65.
Mercado-Martinez, F. J., Robles-Silva, L., Moreno-Leal, N., & Franco-Almazan, C. (2001).
Inconsistent journalism: The coverage of chronic diseases in the Mexican Press. Journal of
Health Communication, 6, 235–247.
Mercy, J. A., Rosenberg, M. L., Powell, K. E., Broone, C. V., & Roper, W. L. (1993). Public
health policy for preventing violence. Health Affairs, 12, 7–26.
Ramsey, S. (1999). A benchmark study of elaboration and sourcing in science stories for eight
American newspapers. Journalism & Mass Communication Quarterly, 76(1), 87–98.
Rest, J. R. (1993). Research on moral judgment in college students. In A. Garrod (Ed.),
Approaches to moral development: New research and emerging themes. New York:
Teacher’s College Press.
Salwen, M. (1992). The influence of source intent: Credibility of a news media health story.
World Communication, 21, 63–68.
Testing Framing and Sourcing13
Schwitzer, G., Mudur, G., Henry, D., Wilson, A., Goozner, M., Simbra, M., et al. (2005, Download full-text
July). What are the roles and responsibilities of the media in disseminating health infor-
mation? The PLoS Medicine Debate, 2, 215–222.
Severin, W. J., & Tankard, J. W. (1992). Communication theories: Origins, methods, and uses in
the mass media. New York: Longman.
Signorielli, N. (1993). Mass media images and impact on health. Westport, CT: Greenwood
Stevens, J. E. (1998). Integrating the public health perspective into reporting on violence.
Nieman Reports, Winter, 38–40.
Woodruff, K., Dorfman, L., Berends, V., & Agron, P. (2003). Coverage of childhood nutrition
policies in California newspapers. Journal of Public Health Policy, 24, 150–158.
14 R. Coleman et al.