Agenda Setting and Framing Theory

week you should:

1) Watch Max MCCombs Agenda-Setting Theory

2) Read: the article by McCombs, Shaw, and Weaver (2014) “New Directions in Agenda-Setting Theory and Research,” the article by Entman: (2007) “Framing Bias  Media in the Distribution of Power” and the article by Entman and Usher (2018) “Framing in a Fractured Democracy: Impacts of Digital Technology on Ideology, Power, and Cascading Network Activation.”

a) Ritter, Matt (2020) “Intraday Intermedia Agenda-Setting in the Manic World of Online News Reporting.”

b) Wallington, SherrieFlynt, Kelly Blake, Kelly, Kalahn Taylor-Clark, and K.  Viswanath (2010) “Antecedents of Agenda Setting and Framing in Health News: An Examination of Priority, Angle, Source, and Resource Usage from a National Survey of U.S. Health Reporters and Editors”

5) Post two critical questions for each of the articles assigned for presentation.

6) You may have been assigned an article presentation this week, in which case you only need to post questions to the articles you are not presenting.

– Reaction Paper IV:

Compare and contrast Agenda Setting Theory and Media Framing.  Draw your arguments from the article by McCombs, Shaw, and Weaver (2014)  “New Directions in Agenda-Setting Theory and Research,”  Herman’s article “The Propaganda Model: a Retrospective, and the article by Entman and Usher (2018) “Framing in a Fractured Democracy: Impacts of Digital Technology on Ideology, Power, and Cascading Network Activation.”

The paper should reflect a critical reading of the texts and your reaction.  Think of the paper as having two parts:

1) Make sure to identify the key arguments/relevant concepts of each theory based on the readings. Summarize what the authors meant, or at least what you think the authors meant, and try to see the world with that theoretical lens.

2) Critique the  theories from an outside perspective, finding their strengths and flaws. Try to include an original example that illustrates the theories and connects them to the media. Mass Communication and Society, 17:781–802, 2014
Copyright # Mass Communication & Society Division
of the Association for Education in Journalism and Mass Communication
ISSN: 1520-5436 print=1532-7825 online
DOI: 10.1080/15205436.2014.964871
DEUTSCHMANN SCHOLARS ESSAY
New Directions in Agenda-Setting
Theory and Research
Maxwell E. McCombs
School of Journalism
University of Texas at Austin
Donald L. Shaw
School of Journalism
University of North Carolina at Chapel Hill
David H. Weaver
School of Journalism
Indiana University – Bloomington
As agenda-setting theory moves toward its 50th anniversary, its productivity
in the past and at present augurs a highly promising future. In this essay, the
original theorists trace the development of agenda setting and identify seven
Maxwell E. McCombs (Ph.D., Stanford University, 1966) is the Jesse H. Jones Centennial
Chair in Communication Emeritus at the University of Texas at Austin. His current research
is focused on a new theoretical area, network agenda setting, the connections among the objects
or attributes defining each agenda.
Donald L. Shaw (Ph.D., University of Wisconsin, 1966) is Kenan Professor of Journalism
Emeritus at the University of North Carolina at Chapel Hill. His research interests include
American journalism history and communication theory.
David H. Weaver (Ph.D., University of North Carolina, 1974) is Roy W. Howard and
Distinguished Professor Emeritus at Indiana University. His research interests include studies
of journalists, and news media uses and effects, especially agenda-setting effects.
Correspondence should be addressed to Donald L. Shaw, School of Journalism, University
of North Carolina at Chapel Hill, Carroll Hall, CB 3365, Chapel Hill, NC 27599. E-mail:
cardinal@email.unc.edu
781
782
MCCOMBS, SHAW, WEAVER
distinct facets. They explore three of the seven facets—need for orientation, network agenda setting, and agendamelding—in greater detail because those are
particularly active arenas of contemporary research. Grounded in more than 40
years of productive collaboration among the authors, this inaugural Deutschmann Scholars Essay offers numerous new ideas about recent trends in and future
directions for agenda-setting theory and research. The three authors are all recipients of AEJMC’s Paul J. Deutschmann Award for Excellence in Research recognizing a career of scholarly achievement. The Deutschmann scholars observed
that this may well be the most original article they have ever written together.
As agenda-setting theory moves toward its 50th anniversary, its productivity
in the past and at present augurs a highly promising future. Beginning with a
tightly focused study in Chapel Hill of media effects on the salience of issues
among the public, agenda setting has evolved into a broad theory with seven
distinct facets:
Basic agenda setting, the impact of the media agenda on the public
agenda regarding the salience of issues, political figures and other objects
of attention (the first level of agenda setting).
. Attribute agenda setting, the impact of the media agenda on the public
agenda regarding the salience of the attributes of these objects (the second
level of agenda setting).
. Network agenda setting, the impact of the networked media agenda of
objects or attributes on the networked public agenda of object or attribute
salience (the third level of agenda setting).
. Central to understanding the strength of agenda-setting effects is the
concept of need for orientation, which details the psychology of each
individual’s encounter with the media. More recently, dual psychological
paths linking media exposure and agenda-setting effects have been detailed.
. Consequences of agenda-setting effects at all three levels for attitudes,
opinions, and behavior.
. Origins of the media agenda, which range from the prevailing cultural
and ideological environment to news sources, the influence of the media
on each other, the norms and routines of journalism, and the individual
characteristics of journalists.
1
. Agendamelding , the way we merge the civic agendas of the media and
our valued reference communities with our personal views and experience
to create a satisfying picture of the world.
.
1
Agendamelding is the unhypenated construction we coined to describe the intimate, often
unconscious process by which we borrow from a variety of agendas to find, or create, the
personal communities in which we choose to live.
DEUTSCHMANN SCHOLARS ESSAY
783
All of these facets are appropriate venues for research guided by agenda-setting
theory, both now and in the future. The core concepts of agenda-setting theory
are an object agenda, attribute agenda, and the transfer of salience between
pairs of agendas. In the now vast research literature on agenda setting,
there are many different operational definitions of these core concepts. The
issue agendas of the news media and the public are the most dominant
operational definitions, but far from the only ones. Contemporary research,
which continues to build on these theoretical foundations, is characterized by
two trends:
a centrifugal trend expanding to domains beyond the original focus on
public affairs;
. a centripetal trend of research, further explicating agenda-setting theory’s
core concepts.
.
Both of these trends will introduce new and diverse operational definitions
to the research literature.
Three of the seven facets listed are discussed here in greater detail because
they are particularly active theoretical arenas of contemporary research.
These three facets are need for orientation, network agenda setting, and
agendamelding.
NEED FOR ORIENTATION AND THE PSYCHOLOGY
OF AGENDA SETTING
The concept of need for orientation (NFO) was created more than 40 years ago
at the University of North Carolina-Chapel Hill in the early 1970s by Maxwell
McCombs and David Weaver, and was first introduced to a wider academic
audience in a paper presented at the April 1973 annual meeting of the
International Communication Association (ICA) in Montreal, Canada
(McCombs & Weaver, 1973). In that paper, the authors asserted on pages 2
and 3:
Explication of an agenda-setting function of the press must include the relevant
social and personal characteristics that mediate such an effect. The traditional
‘effects orientation’ of communication research must be combined with the
‘information-seeking’ or ‘uses and gratifications’ approach . . . . the basic
explanation of those effects lies at the psychological level, within the individual
citizen. At the psychological level, our major theoretical assertion is that every
individual has a need for orientation. Each individual feels some need to be
familiar with his surroundings, both his physical and cognitive environment.
In terms of Tolman’s (1932) concept of cognitive mapping, each individual
784
MCCOMBS, SHAW, WEAVER
will strive to ‘map’ his world, to fill in enough detail to orient himself, to
intellectually find his way around.
McCombs and Weaver (1973) defined need for orientation as a combination
of relevance and uncertainty, such that low levels of relevance led directly to
low levels of NFO, a high level of relevance coupled with low uncertainty led
to a moderate level of NFO, and a high level of relevance coupled with high
uncertainty led to a high level of NFO. They operationally defined political
relevance in terms of interest in, and discussion of, the 1972 presidential campaign and political uncertainty in terms of the consistency of voting record,
strength of political party identification, and degree of certainty about choice
of presidential candidate. They found strong support for their hypotheses
that use of newspapers and television for political information would increase
monotonically with levels of NFO and that the agenda-setting effect of newspapers would also increase in a linear pattern with levels of NFO. In a later
book chapter, Weaver (1977) found support for NFO’s impact on television
agenda-setting effects as well as on newspaper agenda-setting effects in the
1972 U.S. presidential election.
Since that early 1972 election study, there have been a number of tests of
NFO in a variety of settings, with somewhat different models, and mostly
consistent support for it as a useful predictor of both frequency of news
media use for political information and different levels of agenda-setting
effects, especially the first level of issues or general topics.
In 1980, Weaver tested whether NFO was a stronger predictor of media
exposure-effects relationships than more specific measures of gratifications. In this article, NFO was measured in terms of four groups based
on a 2 by 2 typology of relevance and uncertainty for the first time (low
relevance-low uncertainty, low relevance-high uncertainty, high relevancelow uncertainty, and high relevance-high uncertainty). The two middle
groups were both considered to have ‘‘moderate’’ levels of NFO and thus
equal status as predictors of news media use and agenda-setting effects, an
assumption that was not fully supported in later studies. Weaver (1980)
found that the more general measure of NFO had more influence as a
contingent condition on relationships between media use and media effect
than did specific gratifications measures, supporting Blumler’s (1979)
proposal that we need to use basic audience orientations to predict and
explain media effects.
A careful explication and testing of the NFO concept by Matthes (2006)
found that the original assumption by McCombs & Weaver (1973) that
relevance precedes uncertainty appears to be valid – that is, when there is
low relevance, NFO will be low, regardless of uncertainty level, but when
there is high relevance, uncertainty does matter. The Matthes study also
DEUTSCHMANN SCHOLARS ESSAY
785
found that another related concept, need for cognition, was a predictor of
NFO along with relevance, and it suggested that there are at least three
separate types of NFO – toward issues, facts, and journalistic evaluations.
In an experimental study, Chernov, Valenzuela, and McCombs (2011)
compared the measures of NFO introduced by Weaver (1977) with those
used by Matthes (2006), and found that both measures were correlated
and equally strong predictors of first-level agenda-setting effects (the relative
salience of issues), but the new measures of Matthes did not predict orientations toward facts or evaluations (second-level agenda-setting effects).
Chernov et al. (2009, p. 14) conclude, ‘‘It might be that the first exposure
to certain information leads to heightened awareness and subsequent
perceived issue importance while only after repetitious exposure do the
second-level agenda-setting effects start to build up.’’
A more recent study by Camaj (2014), using survey data from Kosovo,
distinguished between the two ‘‘moderate’’ NFO groups identified by
Weaver (1980) and found that those who were high in relevance and low
in uncertainty (the ‘‘moderate-active’’ group) were more likely than the other
groups to use partisan news, in keeping with their tendency to be partisans
(highly interested and low in uncertainty). This ‘‘moderate-active’’ group
was also as likely to use independent media as the ‘‘high’’ NFO group, and
both groups used independent media more frequently than the other
‘‘moderate’’ group (those she calls ‘‘moderate-passive’’ who were low on
relevance and high on uncertainty) and the low NFO group that was low
on both relevance and uncertainty. Camaj’s evidence suggests that the two
‘‘moderate’’ NFO groups are not equal as predictors of media use.
Another study by Camaj and Weaver (2013) using survey data from the
2008 U.S. presidential election found that NFO was a stronger predictor of
attention to political news than sheer frequency of news media exposure,
and that media attention was a better predictor of second-level agendasetting effects (the salience of attributes of presidential candidates) than
media exposure. Similar to the Chernov et al. 2009 study, this study found
that NFO did not predict opinions regarding the salience of candidate attributes (a second-level agenda-setting effect). This finding is also consistent
with results of a second study by Matthes (2008, p. 450), who concluded
that, ‘‘NFO does predict that individuals will turn to news media in order
to gather information, but it fails to predict which specific issue attributes
will be chosen as orienting cues.’’
And in another study of the 2008 U.S. presidential election relying on
interviews with voters in Chapel Hill, NC, the site of the original 1968
agenda-setting study, Weaver et al. (2010) found that those voters with a
higher level of NFO did tend to use vertical (traditional) news media,
especially newspapers and news web sites, to deepen their knowledge of
786
MCCOMBS, SHAW, WEAVER
issues. They also found that traditional media use was not strongly related to
which attributes of the candidates were emphasized by voters, whereas more
specialized and partisan (horizontal) media use was more strongly associated
with the attributes that voters associated with candidates. Again, these
findings seem consistent with those of other studies mentioned above that
did not find NFO predicting second-level agenda-setting effects.
Why Agenda Setting Occurs
In a comprehensive analysis of the psychology of agenda-setting effects,
McCombs and Stroud (2014) conclude that NFO is only one part of the
answer to the question of why agenda setting occurs. They review recent
research that outlines a second way to answer the why question by describing
the psychological processes that lead to agenda-setting effects, including
accessibility and applicability. They conclude that those who use media more
passively have lower levels of NFO than those actively using media, and
those with ‘‘moderate-active’’ NFO (high relevance and low uncertainty)
use partisan media more than those with a high level of NFO (high relevance
and high uncertainty). Because of this pattern, they speculate that those with
a moderate-active level of NFO may be motivated by directional goals and
those with high NFO by accuracy goals. Those with low and moderatepassive (low relevance and high uncertainty) NFO tend to process mediated
information passively and use the news media relatively infrequently, resulting in limited agenda-setting effects. Those with moderate-active NFO and
high NFO engage in more active information seeking and are less susceptible
to an accessibility (top of the mind) bias. Those with moderate-active NFO
seek more partisan media outlets, resulting in high first- and second-level
agenda-setting effects, whereas those with high NFO seek mainstream, less
partisan, media, resulting in strong first-level (issue) agenda-setting effects,
but only moderate second-level (attribute) effects.
To conclude, NFO and agenda-setting research have been elaborated and
expanded since the studies of the early 1970s. It’s now clear that there are
conceptual and practical benefits to expanding NFO to at least four different
categories instead of the original three, and that the effects of NFO vary
depending on which type of media are analyzed (mainstream=vertical vs.
niche=horizontal) and which level (first or second) of agenda-setting effects
is being predicted. It’s also becoming increasingly clear that agenda setting
can occur from casual or passive exposure to media mainly through the
accessibility process and also from more active information seeking and
reasoning through the applicability process, and that the role of NFO in predicting these psychological processes differs depending on the type of media
use and the kind of agenda-setting effects being predicted.
DEUTSCHMANN SCHOLARS ESSAY
787
FIGURE 1 David H. Weaver’s need for orientation model (March 2013).
Another way of summarizing the more active information-seeking relationships between NFO, media attention, agenda-setting effects and agendamelding is illustrated in a model of Need for Orientation proposed by
Weaver. Displayed in Figure 1, this model assumes that high NFO (Need
for Orientation) is more likely to lead to more attention to vertical rather than
horizontal media, and that moderate-active NFO (high relevance and low
uncertainty) is more likely to lead to more attention to horizontal media than
to vertical (because uncertainty is lower than with the high NFO group). It
also assumes that the most likely outcome from attention to vertical media
is first-level (object) agenda setting, and the most likely outcome from attention to horizontal media is second-level (attribute) agenda setting, although
attention to vertical media can result in some second-level agenda setting,
and attention to horizontal media can result in some first-level agenda setting.
Also, attention to vertical and to horizontal media are sometimes correlated,
as are first-level and second-level agenda setting. Finally, this model assumes
that agendamelding is more likely the result of second-level agenda setting
than first-level agenda setting, and that agendamelding has more influence
on attention to horizontal media than on attention to vertical media.
A WORLD OF AGENDAS
Although there are seven distinct facets of agenda-setting theory, many
regard agenda setting as the transfer of issue salience from the news media
to the public agenda. Although this facet of the theory has received the dominant attention from researchers over the years (Kim, Kim & Zhou, 2014),
788
MCCOMBS, SHAW, WEAVER
even within this arena agenda setting has expanded to include many other
channels of communication–political advertising, conversations, and social
media–as well as content other than public issues that ranges from major
news events to cultural topics (Kliger-Vilenchik, 2011; Bantimaroudis,
Zyglidopoulos & Symeou, 2010). Among the hundreds of agenda-setting
studies, there are many different operational definitions of agendas. The issue
agendas of the news media and the public are the most prominent
operational definitions, but far from the only ones.
With the widespread diffusion of social media, agenda-setting theory can
be applied to a much wider array of channels and more easily to an array of
content extending far beyond the traditional focus on public affairs. Scholars
have the opportunity to examine the transfer of salience between many different kinds of agendas. Even within the dominant news media agenda=
public agenda dyad, numerous operational definitions of these agendas are
emerging. And as these emerging agendas are defined by wider ranges of content and communication channels, agenda setting as we have known it–the
flow of the salience of the top issues of the moment from the news media
to the public agenda–will be only one of numerous agenda-setting processes.
These new versions of the familiar media agenda=public agenda comparison are very different from most previous studies of agenda setting, research
in which the most frequent starting point was the Gallup Poll’s MIP question: ‘‘What is the most important problem facing this country today?’’ With
its tight focus on the public issues that respondents nominated as ‘‘most
important,’’ this operational definition can appropriately be called the
prioritized agenda. In other words, this agenda is a highly filtered version
of what people found in the news and evaluated as most important. Although
at any moment dozens of issues compete for attention, only a very small
number garner a constituency of 10% or more among the public to rise to
the top of public issue agenda. Ten per cent has been identified as the threshold for significant public attention (Neuman, 1990). For example, a review of
60 monthly Gallup Polls asking the MIP question from 2009 to 2013 found a
range of 2 to 5 issues in these polls that met the criterion of 10%. The modal
number of issues meeting this criterion in a poll was three.
The Social Media Civic Conversation
The prioritized agenda is a stratified sample of the broader civic conversation
on public affairs. A much larger sample of this civic conversation is the social
media conversation taking place on Twitter, Facebook, blogs and numerous
other channels. Of course, much of social media conversation is focused on
personal interests and activities and has little to do with public affairs (Kelly,
2009). Nevertheless, a substantial record of both media and public discussion
DEUTSCHMANN SCHOLARS ESSAY
789
of public affairs is now available for observation and analysis. In the 2012
presidential election, Vargo et al. (2014) compared the Twitter issue agendas
of Romney and Obama supporters with the Twitter issue agendas of mainstream, conservative and liberal journalists. Across the entire year of 2012,
Neuman et al. (2014) compared the discussion of 29 public issues in both traditional news media and social media–Twitter, blogs, and forums=discussion
boards.
The social media issue agenda is far more expansive than the priority issue
agendas defined by responses to the Gallup MIP question. And the vast
quantity of the available data enables continuous monitoring of the civic
conversation about public issues. However, despite its huge size, this stratum is still a limited sample of the civic conversation. In regard to the central
hypothesis over the years in research on the priority agenda–the transfer of
issue salience from the news media to the public agenda–the social media
issue agenda is simultaneously too inclusive and too limited. Both of these
limitations will be discussed.
The agenda-setting processes involving the social media issue agenda
extend beyond the relationship between the media and the public that has
been at the center of agenda-setting research over the years. This expansion
can be detailed in terms of the origins for three distinct subsets of the social
media issue agenda.
Some of the messages that define the social media issue agenda originate
in citizens’ longstanding–and often passionate–interest in particular issues.
Hot button issues such as abortion, gay marriage, and regulation of guns
immediately come to mind. From time to time news events may trigger a
spike in social media messages on such topics, but by and large the news
media agenda play little part in stimulating citizens to converse about these
topics on social media, In rare instances citizens directly observe events and
subsequently comment about them on social media. This is mostly commentary about sports events or political events such as debates, but some aspects
of citizen journalism also fit here.
Both of these first two subsets are a small portion of the messages defining
the social media issue agenda, but they do introduce noise into our observations of the media=public relationship when we rely upon large data sets collected from the social media. In other words, the social media issue agenda is
too inclusive, a mélange of messages with highly diverse origins.
A primary source of the messages that make up the public issue conversation on social media are the news events of the day, which call attention
to a wide variety of topics and issues. Citizens then use the social media to
comment, distribute information or seek additional information on these
topics and issues (Meraz, 2013; Vargo et al., 2014). Increasingly, news organizations monitor this conversation with the result that high levels of social
790
MCCOMBS, SHAW, WEAVER
media activity sometimes prompts continuing news coverage (Messner &
Garrison, 2011; Wallsten, 2007).
The first two of the social message subsets just described are largely original contributions by the public to the social media issue agenda. The third is
a broadening and redefinition of the traditional agenda-setting role of the
news media. This latter social media issue agenda encompasses a much larger
set of issues than the priority issue agenda. Sometimes this social media issue
agenda is part of a two-stage agenda-setting process in which the news media
initially stimulate the pubic agenda and subsequently the public agenda
stimulates the news media agenda. In other words, there is a flow in both
directions (Meraz, 2011a, 2011b).
The Acapulco typology (McCombs, 2014) identifies three research designs
that have been used to study the agenda-setting process, designs that frequently have documented strong agenda-setting effects of the media priority
issue agenda on the public priority issue agenda. Type I studies measure the
correspondence between the salience of a set of issues on the media and
public agendas. In the Chapel Hill study, this was five issues on the media
and public priority issue agendas. The size of the social media issue agenda
is much greater and a high degree of correspondence between two very large
sets of issues is very unlikely. Type III and IV studies focus on the salience of
individual issues. Because of the much greater number of individual issues
that can be observed, there will be considerable variance in the salience of
these issues in the news media. There also will be a wide range of variance
in the salience of these issues on the social media agenda because of the
numerous variables other than media coverage that result in internet activity
by the public. In terms of the traditional agenda-setting hypothesis that the
media agenda sets the public agenda, this wide variance in the salience of
issues will attenuate this relationship.
The agenda-setting effects of the media on the public’s perception of
the priority issues of the day result from the high degree of redundancy in
the media messages received by the public about the priority issues of the
day. However, as our operational definitions of the media agenda expand
beyond the handful of priority issues of the day, this redundancy is greatly
reduced. As we consider larger and more inclusive issue agendas, the flow of
information experienced by the public has very low levels of redundancy for
many issues.
Atkinson et al.’s (2014) longitudinal study of 90 issues on the agendas of
12 major news organizations found a vast range in the salience of these issues.
The most frequently covered topic on the media agenda, international organizations, averaged 687 articles per month. The issue that ranked 5 out of the
90 averaged 438 articles per month. At rank 15 the monthly average is 192;
at rank 50, the average is 42; and at rank 90 the average is less than 2. The
DEUTSCHMANN SCHOLARS ESSAY
791
statistical description of the salience of these 90 issues is very much a
long-tailed distribution.
A factor analysis of the data for each of these 90 issues found that issue
salience, the frequency with which an issue appears on the media agenda, is
a key determinant of the emergence of a strong national agenda, that is, an
agenda characterized by homogeneity across news media. Across these 90
analyses, the proportion of variance explained by the first factor–the extent
of homogeneity across the news media–ranges from 0.12 to 0.93. This considerable variance in the redundancy of issues on the media agenda makes
it clear that the topic matters considerably. Earlier research on priority agendas introduced a number of concepts that identify topics more susceptible to
the agenda-setting influence of the news media and those that are less susceptible. These concepts include obtrusive and unobtrusive issues (Zucker, 1978)
and abstract versus concrete issues (Yagade & Dozier, 1990). Recent
attention to expanded agendas calls for renewed theoretical attention to the
characteristics of issues that impact the public’s response to media coverage
of the topics.
Finally, a caveat and research question. A recent report by the American
Association for Public Opinion Research’s task force on the use of social
media as a data source noted that the unit of analysis in social media data
commonly is the message, not an individual (AAPOR, 2014). Do the measures of an issue’s standing on the internet agenda based on the number of
messages correspond to the frequencies in the Gallup MIP question about
which issues are perceived by the public as the most important ones? This
difference in the units of analysis needs careful scrutiny.
Even if there is little difference in the two approaches to measurement, the
internet issue agenda remains an incomplete portrait of public opinion.
Among the caveats raised by the AAPOR task force about the use of social
media to determine public opinion is the continued existence of a digital divide that is especially pronounced among older citizens. According to the Pew
Internet and American Life Project, as of 2013, 81% of the US adult population had internet access, and among those with access, only 73% used social
media (Duggan & Smith, 2013). That leaves more than a third of all Americans out of the internet conversation, particularly older Americans. While
nine in 10 of 18-29 year-olds use social media, fewer than half of the
65 þ population do. For Twitter, a popular source of data because of the
large volume of publicly available messages and relatively simple process
of obtaining them (O’Connor et al., 2010), considerably less than a quarter
of the US population uses this social media channel (Duggan & Smith, 2013).
Big data analyses of the social media issue agenda afford a continuous
look at a major portion of public opinion on the issues of the day, but nevertheless are a limited outcropping of public opinion. Survey research will
792
MCCOMBS, SHAW, WEAVER
continue to play a role in determining the issue agenda of interpersonal
communication and even the agenda of those who remain silent until asked
by an interviewer. Neuman et al. (2014) stated:
Big data methodologies do not represent a panacea or a substitute for carefully
designed surveys, experiments, and content analyses. Instead they represent a
complement, an additional resource for better understanding a fast-changing
public sphere.
A New Vantage Point – A Third Level of Agenda Setting
All of the agendas considered to this point have one thing in common.
Theoretically and analytically, they treat objects and their attributes as
separate and distinct disaggregated elements. Of course, in reality objects
and their attributes are bundled together in media messages and in public
thought and conversation. Lippmann’s concept of the ‘pictures in our heads’
raises the question: To what extent are the media able to transfer the salience
of an integrated picture?
Some psychologists and philosophers hold that people’s mental representations operate pictorially, diagrammatically or cartographically. In other
words, audiences map out objects and attributes as network-like pictures
according to the interrelationships among these elements. From this perspective, the news media transfer the salience of relationships among a set of
elements to the public. These sets could be the objects on the media or public
agendas, the attributes on the media or public agendas, or a combination of
objects and attributes. These sets of relationships among elements of the
media and public agendas are the third level of agenda-setting (Guo, 2014).
The initial exploration of the extent to which the news media can transfer
the salience of relationships among a set of elements to the public focused
on the transfer of the salience of the relationships among a set of attributes
in the media to the public. To afford a comparison with traditional attribute
agenda setting, this pilot study conducted network analyses on datasets
initially collected by Kim and McCombs (2007). Studying candidates for
Texas governor and US Senator, Kim and McCombs found strong attribute
agenda-setting effects in their analyses for each candidate separately and for
all four candidates combined. Their analysis compared the frequency with
which various attributes appeared in news stories and citizens’ descriptions
of the four candidates. Reanalysis from a network perspective examined the
co-occurrence of attributes in the news stories and citizens’ descriptions of
the four candidates and found significant network agenda-setting effects
consistent with the strength of the effects in the original study. For example,
the overall correlation between the media and public attribute agendas in
DEUTSCHMANN SCHOLARS ESSAY
793
Kim and McCombs (0.65) corresponds with the correlation (0.67) between
the media and public network agendas (Guo & McCombs, 2011).
Another study examining the transfer of object salience compared the networked issue agenda of the U.S. news media, using the Project for Excellence
in Journalism’s extensive national samples, with the public’s networked issue
agenda measured by monthly national polls for 2007 through 2011. The correlations for each of these five years ranged from 0.65 to 0.87. Similar to the
analysis of attribute agendas just discussed, the statistical results for network
analysis and traditional correlation analysis are very similar (Vu, Guo, &
McCombs, 2014).
While it seems likely that the redundancy in the presentation of these
relationships by the media is again key, the level of redundancy necessary
to create these effects among the public is among the new research questions
presented to scholars. Conceptually and methodologically distinct, this new,
broader perspective on the bundling of agenda elements–the third level of
agenda-setting–tests a comprehensive agenda-setting hypothesis that the salience of relationships on the media network issue agenda can be transferred
to the public network issue agenda. The operational definitions of agendas
continue to expand, affording scholars a rich terrain to explore.
AGENDAMELDING AND CIVIC COMMUNITY BALANCE
If there is a media message and no audience, there is no agenda setting. Since
the rediscovery of a powerful media in the 1960s and 1970s (Gerbner, Gross,
Morgan, Signorielli & Shanahan, 1994; Noelle-Newmann, 1993; McCombs
& Shaw, 1972), we have become increasingly aware of the role audiences play
in selecting among media agendas. The available media in the marketplace
have exploded. Some of us read newspapers on the Web, find or create
groups of similar interest on Facebook or Twitter, and monitor news channels of many types throughout our day. Some of us, especially if we are older,
still read daily newspapers and watch the television evening news. We have
choices and we are using them to mix agenda messages into agenda communities that satisfy our individuality. We sample among media, such as
network broadcasting, that reach for everyone. We draw from media, such
as sports magazines, Websites, or blogs that fit our personal interests or
the interests of others whom we value.
Network and local broadcast stations reach for the diversity of a large
community as if shouting from the top of an Egyptian pyramid to a vast
audience below. These vertical media usually cover and reflect the basic institutions of our society. By contrast magazines, cable television, Twitter and
Facebook, among many others, usually connect us via valued special interest
794
MCCOMBS, SHAW, WEAVER
and personal interest communities, as if we lived, horizontally, on one level of
the pyramid. Of course, we do live in a mix of vertical, institutional
community and valued personal community. We are not passive. We find
or create our own communities by mixing information about vertical and
horizontal communities to match our own experiences and preferences.
Agendamelding is the social process by which we meld agendas from various
sources, including other people, to create pictures of the world that fit our
experiences and preferences. In this section, we use the word ‘‘agendamelding,’’ rather than ‘‘agenda melding’’ because the process is so intimate and
personal that we are not aware we are doing it.
Agenda-setting correlations measure the level of agreement between a
medium and audience. If there is no agreement on the basic issues, there cannot be a stable civic community unless police or military forces maintain
order. If there is a stable society there is a modicum of agreement among leaders, institutions, and citizens. The civic agenda is an evolving set of priorities
around which a social or political system exists. Walter Lippmann’s Public
Opinion argues that media connect us in complex modern society. From this
point of view, media do not just cover community. An agenda community, in
a way, is community. The media cover the communities where we live, work,
and play. Agenda setting represents a sharing of saliences, and agendamelding describes the process by which we borrow from a variety of agendas to
find, or create, the personal communities in which we choose to live.
Elements of Agendamelding
What level of agenda agreement is needed for a functioning community? One
might imagine that a perfect correlation between media and public of 1.00
would represent a rigid, controlled community. In the late 1930s and early
1940s, Adolf Hitler and his Nazi party controlled the German national press
agenda and imposed penalties on those caught attending to any alternative
political agenda, such as the BBC. Today, one imagines that the agreement
between media and public in North Korea, if it could be measured, would
be very high. The 2011 Arab Spring in Egypt, enabled in part by social media,
seems to have become stalled as significant segments of Egyptians cannot
seem to agree on national priorities. The correlation between media and
public needs to have some reasonable level of agreement.
In the third century B.C., the Greek mathematician Eratosthenes used the
known distance between Alexandria and Cyrene to determine the angle of
shadows made by the moving sun to compute the size of the earth. Drawing
from Eratosthenes who triangulated distances and angles to compute physical measurements, we can use triangulation to compute social distances. And
we can devise a formula to measure agendamelding.
DEUTSCHMANN SCHOLARS ESSAY
795
There are some powerful forces that pull us together. For example, we feel a
duty to vote as citizens at one level. We recognize the necessity of a collectivity.
We also have many competing interests. These are alternative communities of
interest, and any time spent with an alternative community necessarily takes the
time away from the more general civic community. Earlier we called the civic
community vertical because it represents all the people, and alternative communities horizontal because they pull us toward people or interests that are
more personal. So the ingredients for melding agendas are 1) information about
the civic community, 2) information about personal communities, and 3) our
personal interests, experience and beliefs. Figure 2 shows these three elements.
How can we blend these three elements? We argue that the forces that
urge us toward agendamelding can be stated in the following way:
Agenda Community Attraction (ACA)
¼ Vertical Media Agenda Setting Correlation (Squared)
þ Horizontal Media Agenda Setting (Squared) + Personal Preferences.
Vertical media agenda setting is the square of the level 1 agenda-setting
correlation for the social system correlation being measured. Let’s say that
the figure for a given social system is .80. The square of .80 is .64. Of course
at least .20 (1.00 minus .80) is not accounted for by vertical media. In our
formula, we allocate this ‘‘missing’’ correlation to horizontal media. The
square of .20 is .04. Horizontal agenda setting measures the level of agreement between citizens and more personalized media sources, including other
people and specialized media, such as magazines, Websites, or other desired
media. So if we cannot measure horizontal media–which is a very broad
array of sources indeed–we may, like Eratosthenes, estimate it. If we know
FIGURE 2 The elements of agendamelding.
796
MCCOMBS, SHAW, WEAVER
two parts of the variance, .64 to vertical media and .04 to horizontal media,
we can compute how much is personal.
If the audience-vertical media correlation were 1.00 and the audiencehorizontal media correlation (if we could completely measure it) were 1.00,
then we could predict audience issue saliences perfectly. But it is difficult
to imagine that audiences are ever that passive. We do not make all our judgments on what media are saying, vertical or horizontal, and so we can take
the vertical media correlation we do know, and make reasonable guesses
about both horizontal media and audience preferences. A more exact
description of personal is this:
h
i
Personal Preferences ¼ 1:00  ðAS1Þ2 þð1  AS1Þ2
The residual reflects the individual values and preferences not accounted for
by our responses to vertical and horizontal media. Psychologists have long
known that personal predispositions are an important part of our makeup.
Our preferences and experiences influence the media we choose and the messages to which we give attention and what we take away from the message.
The concept of need for orientation, which was discussed in the previous section
of this essay, takes these individual differences into account. But our personal
preferences are not fully engaged or satisfied by our responses to the media.
Here we are able to look clearest at vertical media and estimate horizontal media
and personal influences. We have used a political election here as a test but we
theorize that a similar process occurs with all major choices we have to make.
In our hypothetical example, we use .64 (.80 squared) for the vertical issue
correlation, leaving .04 (.20 squared) for the alternative ‘‘missing’’ mediated
issue agenda. Adding .64 and .04 results in .68, leaving .32 for residual personal
preferences not reflected in the media agendas. For an agenda community with
a correlation of .80, therefore, the ACA formula would look like this:
ACA ¼ .64 þ .04 þ .32 ¼ 1.00. This accounts for all the collective variances in
the social=political system. The challenge will be to determine and measure
the contribution of alternative horizontal communities. Of course, these results
are hypothetical. Shortly, we will present some empirical results.
Dynamics of Agendamelding and Civic Balance
One would expect that audiences would meld vertical and horizontal
agendas with their own experiences differently across groups and over time.
Figure 3 demonstrates the theoretical dynamics as the correlations vary
between vertical and horizontal media in a given social system. These
dynamics describe agendamelding and civic balance at any point in time,
DEUTSCHMANN SCHOLARS ESSAY
797
FIGURE 3 The dynamics of agendamelding and civic community balance.
the balance among vertical and horizontal media and personal preferences.
As vertical media agendas go up and down, so do horizontal media and
personal inputs.
The figure suggests there is a dynamic at work and that political=social
systems are constantly in transition. On the left side of the model, we argue
that systems with correlations that are reasonably high with vertical media
suggest a relatively stable social system. But as the power of alternative
community agendas rises, there is a transition period of instability. To the
far right, stability reemerges but around an alternative community. Then
the process continues with a new stability or a new shift to another alternative community. A community with an agenda setting vertical media
correlation of .75 (squared, .56) and a horizontal media correlation of .25
(squared, .06) would represent a residual public involvement of .38. This
seems like a stable civic balance among communities of media and public
because there is a dominant civic thread with reasonable alternative agenda
communities providing challenging views to the dominant vertical agenda.
In terms of percentages, the contributions of the elements are 56% vertical,
6% horizontal, and 38% personal. Protecting these alternative community
agendas is what the First Amendment is about. Of course this model
remains to be tested with agenda-setting data across national states.
798
MCCOMBS, SHAW, WEAVER
A Test of Agendamelding and Civic Balance
We conducted a short test of this model in our 2008 replication of the 1968
Chapel Hill study. In 1968, we content analyzed area media — newspapers,
television, and other media — and conducted door-to-door interviews with a
random sample of 100 voters (McCombs and Shaw, 1972). In 2008, we interviewed 70 Chapel Hill voters in depth over the June-early August period in
the summer of 2008 when voters finally settled on the candidates, Barack
Obama for the Democrats and John McCain for the Republicans. We content
analyzed newspapers and five network evening news programs, ABC, NBC,
CBS, CNN, and Fox News (here we are using television data only). These
represented the vertical media. In the 2008 study, unlike 40 years earlier, we also
content analyzed samples of horizontal broadcast radio and television hosts.
Among these were Stephanie Miller, Rush Limbaugh, and Jon Stewart. This
was our attempt to compare voters with at least a small sample of alternative
news programs that represented alternative communities, primarily liberal or
conservative. Of course we also asked voters about other ways they sought
information, including other people, but we limit our analysis here to these
broadcast hosts. So in this study we compare the melding of vertical and horizontal agendas by different groups of voters in the 2008 presidential election.
The first six columns of Table 1 demonstrate the correlations among all
the voters, and then among Democrats, Republicans, and Independents
with vertical and horizontal media. The voters of different parties did relate
to the vertical and horizontal media differently. Table 1 shows that Republicans and Democrats related in a similar way to vertical media, and neither
were linked very strongly to horizontal media. Independents fell below both
TABLE 1
2008 Agendamelding, Actual and Theoretical
Actual
Vertical Horizontal
1
2
3
AS AS2 AS
All
.77 .60 .07
Democrat
.79 .63 26
Republican .75 .57
.19
Independent .69 .47
.00
Theoretical
Sums
Personal Horizontal
4
5 (2 þ 4)
6
7
8
AS2
.01
.07
.04
.00
.61
.70
.61
.47
.39
.30
.39
.53
AS
.23
.21
.25
.31
AS2 (2 þ 8)
.05
.65
.04
.67
.06
.63
.10
.57
9
Personal
Difference
Actual vs.
Personal Theoretical
10
11
.35
.33
.37
.43
.04
.03
.02
.10
Note: The agenda setting correlations in columns 1 and 3 are partial correlations in which the
influence of the other agenda (horizontal or vertical) has been partialled out.
DEUTSCHMANN SCHOLARS ESSAY
799
Democrats and Republicans in their relationship to vertical media.
Although not displayed in Table 1, we also compared the 41 men voters with
29 women voters in this study and found that they did not use vertical and
horizontal media significantly differently, and we compared the 47 voters
who were 40 or older with the 23 voters younger than 40 and found that
older voters related somewhat more to vertical media and younger voters
to horizontal media, perhaps reflecting changing media use habits. These
are the results from an actual measure of the agendas of some alternative
community news agendas.
We can compare these 2008 results, where we measured some horizontal
media, with our ACA formula prediction. Columns 7–11 in Table 1 show
the results expected with the estimated correlations of audience to the alternative communities. Column 11 shows the difference in personal inputs
between the hypothesized ACA model and the actual data, and we see that
the ACA formula underestimates the attraction of alternative communities.
With both measured and theoretical calculations, there is evidence of agendamelding, as all voters did not use the vertical and horizontal=alternative
media agendas the same way. Whatever made voters identify as Democrats,
Republicans or Independents also seemed to lead then to mix media
messages differently.
Extending Agendamelding Studies
One might also speculate that the dynamics of agendamelding shift over time.
For example, the young soldiers of the so-called greatest generation who
fought in World War II were socialized in a period in which daily newspapers
were powerful, along with local and network radio. These vertical media
from the 1920s and 1930s did not confront the same level of horizontal media
competition as they now do. That would suggest that the social system in the
late 1930s and early 1940s was anchored in stability on the left side of our figure of the dynamics of agendamelding and civic balance. Times change, of
course. Today we find a lack of consensus about the conflicts in which we
are involved, such as Afghanistan, Iraq, and potentially in Syria and other
troubled spots in the world. This is not just true in the United States but also
around the world where various religious or sectarian groups have challenged
the institutional vertical agenda. Often social media provide routes of information, coordination, recruitment, and social support. The vertical agenda
provides a flagpole around which citizens can gather, whether or not they
agree on what to do about the issues. Vertical media present the big acts in
a circus while the sideshows entice us to see the tallest man or fastest turtle.
If media agenda community is community, as we argue, then the lack
of consensus around a civic agenda means there is a decentralization of
800
MCCOMBS, SHAW, WEAVER
authority from the core toward the periphery. This seems to be a worldwide
phenomenon. The evolution of social=political community civic balance does
not seem to come directly from changing media technology but from how we
are melding the emergent media agenda communities. There is a caution in
this for us. More than at any time in history, we have the opportunity to find
satisfying personal community, but also the opportunity to divert attention
from the civic community that sustains us.
CODA
As agenda-setting research approaches the half-century mark, there are
exciting and challenging times ahead. In the words of that famous, albeit
fictional, investigator Sherlock Holmes, ‘‘Come, Watson, come. . . . The
game is afoot.’’2
ACKNOWLEDGMENT
Dr. Shaw would like to thank Dr. Chris Vargo of the University of
Alabama, Dr. Sherine El-Toukhy of the National Institutes of Health,
and Dr. Joan Cates of the University of North Carolina, as well as Dean
Brad Hamm of the Medill School of Journalism at Northwestern University
and Dr. Jean Folkerts, now at the University of Kansas.
REFERENCES
AAPOR (2014). Social media in public opinion research: Report of the AAPOR task force
on emerging technologies in public opinion research. Deerfield: IL, American Association for
Public Opinion Research.
Atkinson, M. L., Lovett, J., & Baumgaartner, F. R. (2014). Measuring the media agenda.
Political Communication, 31, 355–380.
Bantimaroudis, P., Zyglidopoulos, S., & Symeou, P. (2010). Greek museum media visibility and
museum visitation: an exploration of cultural agenda setting’, Journal of Communication, 60,
743–57.
Blumler, J. G. (1979). The role of theory in uses and gratifications studies. Communication
Research, 6, 9–36.
Camaj, L. (2014). Need for orientation, selective exposure and attribute agenda-setting effects.
Mass Communication and Society, 17, 689–712.
Camaj, L., & Weaver, D. H. (2013). Need for orientation and attribute agenda-setting during a
U.S. election campaign. International Journal of Communication, 7, 1442–1463.
2
Arthur Conan Doyle, ‘‘The Adventure of the Abbey Grange,’’ The Return of Sherlock
Holmes (1904).
DEUTSCHMANN SCHOLARS ESSAY
801
Chernov, G., Valenzuela, S., & McCombs, M. (2011). An experimental comparison of two
perspectives on the concept of need for orientation in agenda-setting theory. Journalism &
Mass Communication Quarterly, 88, 142–155.
Duggan, M., & Smith, A. (2013). Social media update 2013. Pew Research Center. Retrieved from
http://www.pewinternet.org/files/oldmedia/Files/Reports/2013/Social%20Networking%
202013_PDF.pdf
Gerbner, G., Gross, L., Morgan, M., Signorielli, N., & Shanahan, J. (1994). Growing up with
television: Cultivation processes. In J. Bryant & D. Zillman (Eds.). Media effects: Advances in
theory and research (pp. 43–68). Mahwah, NJ: Lawrence Erlbaum.
Guo, L. (2014). Toward the third level of agenda setting theory: A Network Agenda Setting
Model. In T. Johnson (Ed.), Agenda setting in a 2.0 world: New agendas in communication
(pp.12–133). New York, NY: Routledge.
Guo, L., & McCombs, M. (2011). Network agenda setting: a third level of media effects. Paper
presented to the International Communication Association, Boston, 2011.
Kelly, R. (2009). Twitter Study. San Antonio, TX: Pear Analytics.
Kim, Y., Kim, Y., & Zhou, S. (2014). Theoretical and methodological trends of agenda setting
theory: a thematic meta-analysis of the last four decades. Working paper, University of
Alabama.
Kim, K., & McCombs, M. (2007). News story descriptions and the public’s opinions of political
candidates. Journalism & Mass Communication Quarterly, 84, 299–314.
Kliger-Vilenchik, N. (2011). Memory setting: applying agenda-setting theory to the study
of collective memory. In M. Neiger, O. Meyers & E. Zandberg (Eds.), On media memory:
Collective memory in a new media age (pp. 226–237). London, UK: Palgrave Macmillan.
Lippmann, W. (1922). Public opinion. New York, NY: Macmillan.
McCombs, M. (2014). Setting the agenda, 2nd edition. Cambridge, UK: Polity Press.
McCombs, M., & Shaw, D. (1972). The agenda-setting function of mass media. Public Opinion
Quarterly, 36, 176–187.
McCombs, M., & Stroud, N. J. (2014). Psychology of agenda-setting effects: Mapping the paths
of information processing. Review of Communication Research, 2(1), 68–88.
McCombs, M., & Weaver, D. H. (1973). Voters’ Need for Orientation and Use of Mass Communication. Paper presented to the International Communication Association, Montreal,
Canada, April 25–28.
Matthes, J. (2006). The need for orientation towards news media: Revising and validating a
classic concept. International Journal of Public Opinion Research, 18, 422–444.
Matthes, J. (2008). Need for orientation as a predictor of agenda-setting effects: Causal
evidence from a two-wave panel study. International Journal of Public Opinion Research,
20, 440–453.
Meraz, S. (2011a). Using time series analysis to measure intermedia agenda setting influence
in traditional media and political blog networks. Journalism and Mass Communication
Quarterly, 88, 176–194.
Meraz, S. (2011b). The fight for ‘how to think’: Traditional media, social networks, and issue
interpretation. Journalism: Theory, Practice, and Criticism, 12, 107–127.
Meraz, S. (2013). Media Agenda Setting in a Competitive Environment: Examining the Role of
Sources in Setting the Topical Discussant Agenda in the Tea Party Patriots’ Facebook
Group. In Tom Johnson, (Ed.), Agenda setting in a 2.0 world: New agendas in communication
(pp. 1–27). New York, NY: Routledge.
Messner, M., & Garrison, B. (2011). Study shows some blogs affect traditional news media
agendas. Newspaper Research Journal, 32(3), 112–126.
Neuman, W. R. (1990). The threshold of public attention. Public Opinion Quarterly, 54,
159–76.
802
MCCOMBS, SHAW, WEAVER
Neuman, W. R., Guggenheim, L., Jang, S. M., & Be, S. Y. (2014). The dynamics of public
attention: Agenda setting theory meets big data. Journal of Communication, 64, 1936–214.
Noelle-Neumann, E. (1993). The spiral of silence: Our social skin (2nd ed.). Chicago, IL:
University of Chicago Press.
O’Connor, B., Balasubramanyan, R., Routledge, B. R., & Smith, N. A. (2010). From Tweets to
polls: Linking text sentiment to public opinion time series. Proceedings of the International
AAAI Conference on Weblogs and Social Media. Retrieved from http://www.aaai.org/ocs/
index.php/ICWSM/ICWSM10/paper/view/1536
Shaw, D., & Martin, S. (1992). The function of mass media agenda setting. Journalism
Quarterly, 69, 902–920.
Tolman, E. C. (1932). Purposive behavior in animals and men. New York, NY: AppletonCentury.
Vargo, C., Guo, L., McCombs, M., & Shaw, D. (2014). Network issue agendas on Twitter
during the 2012 U.S. presidential election. Journal of Communication, 64, 296–316.
Vu, H. T., Guo, L., & McCombs, M. (2014). Exploring ‘‘the world outside and the pictures in
our heads’’: a network agenda setting study. Journalism & Mass Communication Quarterly.
Wallsten, K. (2007). Agenda Setting and the blogosphere: an analysis of the relationship
between mainstream media and political blogs. Review of Policy Research, 24, 567–587.
Weaver, D. H. (1977). Political issues and voter need for orientation. In D.L. Shaw &
M.E. McCombs (eds.), The emergence of American political issues: The agenda-setting
function of the press (pp. 107–119). St. Paul, MN: West.
Weaver, D. H. (1980). Audience need for orientation and media effects. Communication
Research, 3, 361–376.
Weaver, D. H., Shaw, D., Wojdynski, B., & McKeever, R. (2010). Need for orientation, media
use and agendamelding, presented at the annual conference of the World Association of
Public Opinion Research, Chicago, IL, May 11–13.
Yagade, A., & Dozier, D. M. (1990). The media agenda-setting effect of concrete versus
abstract issues. Journalism & Mass Communication Quarterly, 67, 3–10.
Zucker, H. (1978). The variable nature of news media influence. In B. Ruben (ed.), Communication yearbook 2 (pp. 225–240). New Brunswick, NJ: Transaction Books.
Copyright of Mass Communication & Society is the property of Taylor & Francis Ltd and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder’s express written permission. However, users may print, download, or email
articles for individual use.
Journal of Health Communication, 15:76–94, 2010
Copyright # Taylor & Francis Group, LLC
ISSN: 1081-0730 print=1087-0415 online
DOI: 10.1080/10810730903460559
Antecedents to Agenda Setting and Framing in
Health News: An Examination of Priority, Angle,
Source, and Resource Usage from a National
Survey of U.S. Health Reporters and Editors
SHERRIE FLYNT WALLINGTON, KELLY BLAKE,
KALAHN TAYLOR-CLARK, AND K. VISWANATH
Harvard School of Public Health, Boston, Massachusetts, USA and
Dana-Farber Cancer Institute, Boston, Massachusetts, USA
The influence of news media on audience cognitions, attitudes, and behaviors in the
realm of politics, race relations, science, and health has been extensively documented
(Finnegan & Viswanath, 2002; Zhou & Moy, 2007). Agenda setting and framing studies show that news media influence how people develop schema and place priorities
on issues, with media stories serving as a major source of issue frames (Scheufele,
1999). News media are critical intermediaries for translating important health and
science information into forms easily disseminated to and understood by the public
(Viswanath et al., 2008). Dorothy Nelkin (2001) suggested that the media serve as
brokers between science and the public, framing the public’s social reality and shaping public consciousness about science-related events.
Shoemaker and Reese (1996) identified several factors contributing to the
making of news. Factors include social norms and values of journalists, organizational constraints such as deadlines and limits of time and space, pressures from
social movement organizations and interest groups, and reliance on government
and community leaders through journalistic beat systems that often demarcate
established hierarchies for source and resource usage in newsgathering. Early
research primarily studied general assignment and public affairs reporters; the
work of journalists who report on health and medicine is understudied, with a few
exceptions (Kaiser Family Foundation, 1999; Ranshoff & Ranshoff, 2001; Schwitzer
et al., 2005).
Recent health journalism work has begun to bridge research gaps using
systematic explorations of occupational practices of health and medical reporters
(Viswanath et al., 2008). We seek to build on this area of research to examine
how certain organizational and individual characteristics of health reporters
and editors differentially may influence factors that act as antecedents to media
Funding for this research has been given by NIH Grants (5 U01 CA114644;
5R25CA057711-14; OMB#0925-0541) and funding by the Dana-Farber=Harvard Cancer
Center to Dr. Viswanath.
Address correspondence to Sherrie Flynt Wallington, Lombard Comprehensive Cancer
Center, Georgetown University Medical Center, Research Bldg., W326A, 3970 Reservoir
Road, NW, Washington DC 20057. E-mail: sfw49@georgetown.edu
76
Antecedents to Agenda Setting and Framing
77
agenda setting and framing in health and medical science news (McCombs &
Ghanem, 2001). Based on the premise that the sources, resources, priorities,
and angles that journalists use in their reporting act as antecedents to media
agenda setting and framing, our study examines how source and resource reliance
and the selection of angles and priorities are differentially influenced both by
journalists’ individual characteristics and the structure of the news organizations
in which they work.
Theoretical Frameworks
Agenda setting and framing are two serviceable frameworks for examining both the
occupational practices of media professionals and the audience effects resultant from
media exposure. Therefore, it is instructive to examine factors that contribute to the
agenda and frames used in health and medical science reporting. One set of factors
that potentially influences the media agenda and media frames has been discussed in
research on community structure and its influence on the press. Developed by the
Minnesota community media studies team of Tichenor, Donohue, and Olien
(1973, 1980), and advanced by others (Demers, 1996; Demers & Viswanath, 1999;
Hindman, 1999; Pollock, 2007; Pollock & Yulis, 2004), the community structure
approach suggests that local mass media, for the most part, are supportive of and
dependent upon local institutions, and they are reflective of the balance of power
and status of different social groupings within the community (Olien, Donohue, &
Tichenor, 1995). By drawing attention to how news media interact with and constitute a subsystem within the larger community system, this approach identifies
important antecedents of news reporting, specifically the role of community characteristics as antecedents of news reporting (Riffe, Lacy, & Fico, 2005). Pollock (2007)
further suggests that the community structure approach goes beyond simply exploring the impact of media on society by also exploring the impact of society on media.
The idea that the social context in which reporting occurs influences health and
medical science reporters’ and editors’ occupational practices and values forms a
critical backdrop for this study. We add further nuance by simultaneously looking
at individual-level factors and contextual factors for independent influences on
media agenda setting and framing.
Agenda Setting
Agenda setting scholarship has focused mostly on the transfer of the salience of
topics covered in the media to the priorities people place on those topics (McCombs
& Shaw, 1972). Agenda setting may be categorized into three distinct but related
themes that examine the link between (a) public agenda setting—issues portrayed
in the media and the public’s priorities, (b) policy agenda setting—media coverage
and its influence on the legislative agenda of policymaking bodies, and (c) media
agenda setting—antecedents such as institutional roles and processes that are influential in the selection of issues and content covered in the media (Kosicki, 1993).
We focus on the latter, media agenda setting, to examine the ‘‘who’’ and ‘‘what’’ that
influence U.S. health journalists’ coverage of health and medical topics, and how
those influences may be differential across organizational and individual characteristics of reporters and editors.
78
S. F. Wallington et al.
To examine media agenda setting in health and science news, it is useful to
understand journalists’ reliance on news sources. Gans (1979) likened the journalist-source relationship to a dance, with sources seeking access to journalists and
journalists seeking access to sources. Gandy (1982) considered information provided by sources as a form of journalistic subsidy that influences what and how
journalists report. Journalists rely heavily on governmental or elite sources (Schudson, 2006). This primarily is due to reporters’ perceptions that official sources
know more and have more accurate information (Tuchman, 1972), accessibility
of these source types, and the source’s ability to articulate and provide credible
information (Fico, 1984).
Daily newsgathering realities and tight deadlines make journalists heavily
dependent on resources for reporting (Curtin, 1999). Having ready reporting
resources is part of ‘‘routinizing’’ newsgathering (Tuchman, 1997). A study of a television network affiliate found that the amount of available resources to cover a story
was as strong a determinant in deciding news coverage as standard news judgment
(Berkowitz, 1991). Thus, it is important to examine how news organizations’ structural characteristics may differentially affect the resources available to and used by
health and medical science reporters and editors.
One monumental change in the available toolbox of journalistic resources in the
past decade has been the use of the Internet (Hachigian & Hallahan, 2003). Internet
resources have made journalists’ jobs easier and improved their work quality
(Callison, 2003). Journalists use Internet resources to identify experts, gather background information, provide context, find facts and ready references, access
government and company information, stay abreast of current events, and identify
story ideas (Middleberg & Ross, 2002). Print journalists indicate that the
Internet allows them to compete with radio and television breaking news (Middleberg & Ross, 2002). To date, no other study has examined differential source and
resource usage as critical antecedents to media agenda setting in health and medical
science news.
Framing
Some suggest that early agenda setting literature, by focusing only on the transfer
of an issue’s salience, was too limiting, and instead have argued that news media
influence how people should think about specific topics, thus ‘‘framing’’ issues for
the public (Iyengar, Peters, & Kinder, 1982). Framing literature largely explains
the news media’s role not just in amplifying issues, but also in defining issues
for the public, thereby expanding agenda setting from merely drawing attention
to a topic to actually articulating points of view regarding that topic (Weaver,
2007). Reese (2001) characterizes frames as organizing principles that are socially
shared and persistent over time, working symbolically to meaningfully structure
the social world. Reese’s definition provides a useful framework for examining
several priorities and angles (antecedents to framing) that health reporters and
editors say they use when reporting on health. We argue that antecedents to news
framing may include priority setting by reporters and editors, and the angles or
lenses through which journalists choose to tell a story. That is, health and medical
science news may be constructed using a priori ideas about the goals of their
reporting (priorities) and how stories should be told in order to garner the most
impact (angles).
Antecedents to Agenda Setting and Framing
79
Although relatively unexamined, some research has been published on decisionmaking styles of editors and reporters. Early inquiries note that organizational and
personal background factors influence how media managers view decision making,
as well as the priorities they place on reporting (Hartman, Lundberg, & White,
1990). Research examining the media’s ‘‘gatekeeping’’ function reveals that reporters
and editors institute a set of shared news values when determining priorities for
reporting and newsworthiness. These news values, or priorities, typically have
included prominence, human interest, conflict, novelty, timeliness, and proximity
or local appeal (Gans, 1979). Sylvie and Huang’s (2008) study of 341 editors and
managerial staff from U.S. dailies revealed that several sets of values underlie
journalistic decision making and priority setting, including social values (e.g., tradition, group conflict); journalistic values (e.g., objectivity, responsibility); organizational values (e.g., motivation, company goals); and audience values (e.g., impact,
timeliness).
Our study builds on these broad categories to examine what factors predict
differential usage of selected priorities often used in health journalism as a way to
explore priority setting as an antecedent to news framing. Additionally, we explore
other angles that health reporters and editors may use to shape their stories as a way
of extending the exploration of differential usage of news frames. To date, no other
study has examined differential priority and angle selection as critical antecedents to
framing in health and medical science news.
The activities of interest groups, policymakers, and others interested in shaping
media agendas and issue frames may impact both the volume and character of news
messages about a particular issue (Scheufele & Tewksbury, 2007). Moreover, structural characteristics of newsrooms and the journalists that comprise them may differentially impact the usage of sources and resources in reporting, as well as the
priorities and angles that shape news coverage of a topic. Our study examines antecedents to agenda setting and framing in health and medical science news to explore
how journalists’ individual and organizational characteristics differentially impact
their choice of news sources and resources (factors that contribute to media agenda
setting), and priorities and angles (factors that contribute to framing).
Methods
Data
Study data were drawn from a 2005 national survey of U.S. health and medical
science reporters and editors supported by the National Cancer Institute, a component of the National Institutes of Health. The sampling frame was developed using
Bacon’s MediaSource, a comprehensive database of more than 80,000 national and
local print and broadcast media outlets, which includes contact information for
media newsroom personnel. The target population was editors and reporters
working in the following media: news services and syndicates, radio and television
programs, community newspapers, daily newspapers, magazines, radio stations=
networks, cable television stations=networks, and television stations=networks.
For all analyses, the news organization was the unit of analysis, and the units
of observation were reporters and editors. We identified 1,482 news organizations
that represented local and national print and broadcast media outlets of varying
size. Findings reflect responses from 468 reporters and editors from 463 news
80
S. F. Wallington et al.
organizations, yielding a response rate of 31.2%. Although the response rate may
seem low, post hoc analyses revealed that nonrespondent organizations were similar
to respondent organizations in terms of media type and size, with local television
showing only a small difference. Survey methodology is outlined further in an earlier
study using this dataset (Viswanath et al., 2008).
Measures
Dependent Variables. To explore antecedents to media agenda setting, we chose
outcome variables that reflect a variety of possible sources and resources journalists
may use when gathering news and developing health and medical science stories for
publication or broadcast. News media organizations rely on sources and resources
for diversity in perspectives to bring in views from those likely to lend insights to
events or issues (Reese, Grant, & Danielian, 1994). Respondents were asked how
often they rely on obtaining information from each of the following sources: (a)
government scientist or official, (b) industry scientist or spokesperson, (c) other
scientist or researcher, (d) health care provider, and (e) patient or advocacy organization representative. In addition, respondents were asked how often they rely on
each of the following tangible or electronic resources when working on a health or
medical science story: (a) government websites, (b) other websites, (c) news releases,
and (d) scientific journal articles.
To explore antecedents to framing in health and medical science news, we
chose outcome variables reflecting a variety of possible priorities and angles that
may be used in reporting. Respondents were asked how important each of the
following priorities are when developing a story: (a) disseminating new, accurate
information; (b) educating the public so people can make more informed decisions;
(c) providing entertainment; (d) developing public health and scientific literacy;
and (e) influencing the public’s health behavior. Also, respondents were asked
how often they choose the following angles: (a) public impact, (b) economic impact,
(c) controversial new information, (d) human interest, and (e) need to change
personal behavior.
Item responses for all outcome variables employed a scale of 1 to 5, with 5 being
‘‘very often’’ and 1 being ‘‘not at all.’’ For data analysis using logistic regression, scales
were dichotomized such that responses of 4 and 5 represented the response of ‘‘often.’’
Independent Variables. Media studies utilizing a structural approach to examine
journalistic practices and media effects have documented that organizational
characteristics may influence journalists’ newsgathering in several ways (Demers &
Viswanath, 1999; Tichenor et al., 1980; Weaver & Wilhoit, 1996). While there is
considerable controversy over the mechanisms through which the nature of news
organization ownership (e.g., private versus public) influences journalism, it is widely
agreed that media ownership matters (Bagdikian, 2000). The nature of ownership
structure likely influences media coverage through the provision of resources for
reporting and socialization in the newsroom.
We used three indicators as predictor variables characterizing the types of media
organizations represented in the survey: (a) whether the respondents’ media organization is owned by a public corporation whose shares are traded on an exchange, (b)
whether the organization is owned by a group or chain, and (c) the number of
full-time news and editorial staff employed by the organization. In addition to
Antecedents to Agenda Setting and Framing
81
organizational structure, individual characteristics of respondents were included as
predictor variables in our analyses of media agenda setting and framing. We utilized
four survey questions to reflect individual characteristics of respondents: two were
demographic variables (education level and years working as a journalist), and
two represented respondents’ perceptions of occupational autonomy. Below, we
offer background and explanations for the inclusion of selected demographic and
occupational autonomy variables.
We also dichotomized the education variable in order to examine how specialization may affect source and resource usage, as well as choice of news priorities
and angles. Seventy percent of respondents to our survey had a bachelor’s degree;
as such, we examined the role of education as a predictor of media agenda setting
and framing by assessing differential responses between respondents who had a
bachelor’s degree or less compared with those with a master’s degree or more, with
the assumption that advanced degrees offer greater specialization.
Researchers consistently have explored journalism’s organizational culture and
occupational autonomy (Johnstone, Slawski, & Bowman, 1976; Weaver, Beam,
Brownlee, Voakes, & Wilhoit, 2006; Weaver & Wilhoit, 1986, 1996). Scholars
characterize journalistic autonomy as journalists’ freedom to shape their work without being controlled by internal and external powers (Scholl & Weischenberg, 1999).
Pollard’s (1995) work illustrated that journalists reported higher job satisfaction if
they had autonomy, authority, and control over their work. Shoemaker and Mayfield
(1987) suggest, however, that the news media are far from autonomous, and content
decisions are swayed by social and institutional pressures within and outside the news
organization, by a desire to maintain the status quo, and by sources that have social
power within the community (Olien et al., 1995; Pollock, 2007). We sought to explore
the role of occupational autonomy in media agenda setting and framing by examining
the following predictor variables: (a) whether the respondent has the freedom to select
the stories that he or she thinks are important, and (b) whether the respondent has the
freedom to determine which aspects of the story should be emphasized. Responses
employed a 5-point Likert scale, with 1 being ‘‘do not agree’’ and 5 being ‘‘strongly
agree.’’ For analysis, we dichotomized responses of 4 and 5 to reflect ‘‘agree,’’ which
allowed us to create binary outcomes that could be used in logistic regression.
Data Analysis
We used multivariable logistic regression to model the fitted odds that survey
respondents’ organizational and individual characteristics independently and differentially predict the sources and resources journalists use to work on health news stories (factors that may be considered antecedents to media agenda setting) and the
priorities and angles used in reporting (factors that may be considered antecedents
to framing). Complete case analyses were utilized for each outcome variable
(n ¼ 419–421, depending on survey item), and one logistic regression model was
run per outcome variable. All independent variables were included in each model
to control simultaneously for the contribution of organizational and individual characteristics of respondents. Analyses also controlled for potential confounding by
respondent gender and age. Weights were added to reflect differential probabilities
of selection per stratum (e.g., national outlets were sampled with certainty, while
local television and radio were sampled using a simple random sampling method).
Statistical significance was tested at 95% confidence levels.
82
S. F. Wallington et al.
Results
Characteristics of the Sample
The largest numbers of respondents were from local, rather than national, media
outlets; local newspapers with a circulation of
28,300 composed 30.6% of the sample, while smaller local newspapers (with a circulation of 280,000 composed 11.8% of
the sample (see Table 1). Other sample characteristics and descriptive statistics have
been reported elsewhere (Viswanath et al., 2008).
Media Agenda Setting: News Sources
Our analyses revealed several differences among U.S. health and medical science
reporters and editors in the likelihood of using news sources from several sectors
(see Table 2). Individual characteristics such as education and years working as a
journalist were independent predictors of sourcing. Respondents with a bachelor’s
degree or less were 2.24 times more likely than respondents with a master’s degree
or higher to use health care providers as sources (p < 0.0001), and 2.33 times more likely to use patient or advocacy organization representatives (p < 0.0001). Additionally, health reporters and editors with a bachelor’s degree or less were significantly more likely than respondents with a master’s degree or more to use government scientists and officials (OR ¼ 1.52; p < 0.01), and significantly less likely to use ‘‘other’’ scientists or researchers (OR ¼ 0.56; p < 0.0001). Those respondents with 1–15 years’ experience working as a reporter or editor were significantly less likely than those working 16 or more years to use ‘‘other’’ scientists or researchers (OR ¼ 0.70; p < 0.01), but 31% more likely than the more seasoned respondents to Table 1. Distribution of respondents by medium and geographic level 2005 NCI survey of U.S. health and medical science reporters and editors Strata n % # of news organizations in each stratum N National TV broadcast National radio broadcast National news services National newspapers Newspaper publishers, circulation
100 K Local television Local radio Local newspapers, circulation 280 K Total 12 1 7 14 5 67 9 112 143 43 55 468 2.6 0.2 1.5 3.0 1.1 14.3 1.9 23.9 30.6 9.2 11.8 100.0 27 5 44 13 10 326 48 373 373 131 132 1,482 Respondents 83 (0.78, 1.31) (0.59, 0.95) 0.75 (0.89, 1.51) 1.16 Organization not owned by public corporation whose shares are traded on an exchange Organization not owned by a group or chain Small media organization ( .06,
p < .05,

p < .01,


p < .0001. Referent categories: Organization owned by a public corporation; organization owned by a group or chain; large media organization; master’s degree or more; 16þ years working as reporter or editor; does have freedom to select stories. 95% CI OR Government scientist or official (n ¼ 420) Independent variables Table 2. Continued Antecedents to Agenda Setting and Framing 85 use patient or advocacy organization representatives (OR ¼ 1.31; p < 0.05). Perceived occupational autonomy did not seem to contribute independently to sourcing in this analysis. Organization size, a structural characteristic, significantly and differentially contributed to news sourcing in several areas. Health reporters and editors working for small media organizations (defined as .06,  p < .05,  p < .01,  p < .0001. Referent categories: Organization owned by a public corporation; organization owned by a group or chain; large media organization; master’s degree or more; 16þ years working as reporter or editor; does have freedom to select stories. Organization not owned by public corporation whose shares are traded on an exchange Organization not owned by a group or chain Small media organization (defined as .06,  p < .05,  p < .01,  p < .0001. Referent categories: Organization owned by a public corporation; organization owned by a group or chain; large media organization; master’s degree or more; 16þ years working as reporter or editor; does have freedom to determine emphasis. Organization not owned by public corporation whose shares are traded on an exchange Organization not owned by a group or chain Small media organization (defined as