Content uploaded by Piers Robinson
Author content
All content in this area was uploaded by Piers Robinson on Dec 29, 2014
Content may be subject to copyright.
Theorizing the I n f l u e n c e of Media on
Wo r l d Politics
Mode ls o f Me di a In fl ue nc e on F o r e i g n P ol ic y
!
Piers Robinson
ABSTRACT
!
Debate over the extent to which the mass media serves elite interests or,
alternatively, plays a powerful role in shaping political outcomes has been
dogged by dichotomous and one-sided claims. Some attribute enormous
power to the news media (the so-called CNN effect) while others claim the
media ‘manufactures consent’ for elite policy preferences. This article
reviews existing theories of media–state relations, in particular the work of
Daniel Hallin and Lance Bennett, and highlights theoretical and empirical
shortcomings in the manufacturing consent thesis. The article then
outlines two models, a model of media influence and Gadi Wolfsfeld’s
‘political contest model’, that serve to reconcile contrasting claims over the
power of the news media. The model of media influence is then applied to
the Vietnam War in order to reconcile contrasting claims (Hallin vs David
Culbert) regarding the role of the media during this conflict. It is argued
that the two models, taken together, provide a starting point for a two-way
understanding of the direction of influence between media and the state
that builds upon, rather than rejects, existing theoretical accounts.
!
Key Words CNN effect, indexing, manufacturing consent, media
influence, media–state relations, world politics
Piers Robinson is Lecturer in Political Communication at the School of Politics
and Communication Studies, University of Liverpool, Roxby Building, Chatham
Street, Liverpool L69 7ZT, UK. [email piersgr@liv.ac.uk]
European Journal of Communication Copyright © 2001 SAGE Publications
(London, Thousand Oaks, CA and New Delhi), Vol 16(4): 523–544.
[0267–3231(200112)16:4;523–544;020063]
523
Overview
In recent years the need to understand the relationship between the media
and world politics has become more pressing. Many commentators
attribute enormous power to news media, claiming they have the ability
to ‘move and shake governments’ (Cohen, 1994: 9). Such views have
found clearest expression in the debate over the CNN effect concerning
the apparent ability of news media coverage to drive western intervention
during humanitarian crises. Although doubt has been cast on the validity
of the CNN effect thesis (Gowing, 1994; Strobel, 1997), politicians
continue to assert the importance of news media coverage in shaping
policy responses to humanitarian crises (Blair, 1999; Holbrooke, 1999).
At the same time, other commentators, and indeed much media theory,
continues to downplay or deny the possibility of the media shaping or
influencing government policy-making. In its most totalizing formula-
tion, the ‘manufacturing consent’ school of thought maintains that,
almost without exception, the media functions primarily to mobilize
support for the policy preferences of dominant elites (Chomsky and
Herman, 1988; Hammond and Herman, 2000; Herman, 1993). More
mainstream accounts, represented by the work of Daniel Hallin (1986)
and Lance Bennett (1990), still maintain that media have little, if any,
independent influence on elite policy debates over foreign policy (see also
Mermin, 1999; Zaller and Chui, 1996).
The aim of this article is to both review our current conceptual
understanding of media–state relations and to discuss two theoretical
advances which introduce greater scope for recognizing the importance of
media in shaping, and at times determining, political outcomes. In doing
so, the goal is to offer the beginnings of a theory of media–state relations
that can provide a two-way understanding of the direction of influence
between the media and world politics. I start by outlining manufacturing
consent theory, noting in particular two distinct versions of it. Empirical
and theoretical problems with manufacturing consent theory are then
discussed in more detail. In particular I note the inconsistency between
manufacturing consent and the wealth of evidence that highlights the
pivotal role news media coverage can, at times, play in shaping
government policy. I then outline two theoretical advances. The first, my
own work, builds directly upon the work of Hallin (1986) and Bennett
(1990) but specifies the conditions, via a policy–media interaction model,
under which the media might play a key role in influencing political
outcomes. The utility of the theoretical framework is demonstrated by
using it to reconcile contrasting claims made by Daniel Hallin (1986)
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
524
and David Culbert (1998) with regard to media influence on US policy
during the Vietnam War. The second advance is represented in Gadi
Wolfsfeld’s (1997) political contest model. The distinctions between
Wolfsfeld’s model and my own model are discussed. It is argued that both
theoretical advances provide the potential for generating a nuanced, two-
way understanding of the direction of influence between media and the
state that takes us beyond current, totalizing theories of media–state
relations.
Manufacturing consent
A wealth of critical literature written over the last 25 years maintains
that the political and economic positioning of major news media insti-
tutions leads to a situation in which news accounts tend to support
dominant perspectives. More specifically, this literature emphasizes the
ability of government to influence the output of journalists and the
tendency of journalists to both self-censor and perceive events through
the cultural and political prisms of their respective political and social
elites. I refer to this literature as ‘manufacturing consent’ only loosely.
While some arguments about manufacturing consent (see in particular
Chomsky and Herman, 1988) are controversial, the thesis that news
media coverage is ‘indexed’ (Bennett, 1990) to the frames of reference of
policy elites receives substantial empirical support (Bennett, 1990;
Entman, 1991; Hallin, 1986; Mermin, 1999; Sigal, 1973; Zaller and
Chui, 1996). Broadly speaking, this critical literature understands the
news media as being influenced by, and not influencng, government
policy (e.g. Bennett, 1990; Chomsky and Herman, 1988; Entman, 1991;
Hallin, 1986; Herman, 1993; Glasgow University Media Group, 1985;
Paletz and Entman, 1981; Parenti, 1993; Philo and McLaughlin, 1993;
Williams, 1993; Zaller and Chui, 1996).
Two implicit versions of the manufacturing consent paradigm can be
discerned, an executive version and an elite version.1The executive version
(e.g. Chomsky and Herman, 1988; Entman, 1991; Glasgow University
Media Group, 1985; Herman, 1993; Philo and McLaughlin, 1993)
emphasizes the extent to which news media content conforms with the
agendas and frames of reference of government officials where government
officials are understood as members of the executive. For example, Robert
Entman (1991) analysed the divergent US media framing of the Korean
Airline and Iran Air shoot downs which occurred during the 1980s. Both
of these international incidents were similar, involving mistakes by the
military leading to the destruction of civilian airliners and large loss of
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
525
life. However, the US news media framed the Iran Air shoot down, for
which the US was responsible, in terms of a technical failure, while the
Korean Airline shoot down, for which the USSR was responsible, was
framed as a moral outrage. According to Entman (1991: 10) overall
media coverage was consistent with the policy interests of the respective
US administrations. For example, with respect to the relatively high level
of news media coverage of the Korean Airline shoot down compared to
that of the Iran Air shoot down, Entman argues:
A continuing judgment of importance likely made a political difference: A
continuing high degree of mass awareness of KAL pressured potential elite
opponents to the Reagan administration to remain silent or hop on the
‘Evil Empire’ bandwagon . . . lower mass awareness of the Iran Air incident
diminished a political resource White House foes might otherwise have
used to convince other elites to abandon the administration’s Persian Gulf
policy. (Entman, 1991: 10)
Importantly, according to the executive version, the news media do
not function to criticize or challenge executive policy lines. Accordingly
this vein of the manufacturing consent literature makes a strong implicit
claim that the conformity between news media coverage and executive
policy interests prevents news media influence on executive policy.
The second, elite version of the manufacturing consent paradigm
(e.g. Bennett, 1990; Hallin, 1986) holds that news media coverage
conforms to the interests of political elites in general whether they are in
the executive, legislative or any other politically powerful position in
society. The seminal study of elite manufacturing consent theory is
Daniel Hallin’s The Uncensored War. Examining the claim that during the
Vietnam War the news media played an oppositional role to official US
policy, Hallin finds that critical news media coverage occurred only after
sections of the Washington political elite turned against the war. Hence,
perhaps the event most cited as a case of news media influence on
government policy actually turns out to be a case of political elites
becoming divided over policy with critical news media coverage merely
being a reflection of this. Drawing upon these findings, Hallin develops
the concept of three spheres, one of consensus, one of legitimate
controversy and one of deviance. These exist with regard to any given
political issue. He argues that news media coverage, taking its cues from
political elites, rarely produces coverage within the deviant sphere but
rather either reflects elite consensus on an issue or elite ‘legitimate[d]
controversy’ (Hallin, 1986). Hallin’s work receives further conceptual
clarification through the work of Lance Bennett (1990). Bennett argues
that ‘mass media news is indexed implicitly to the dynamics of govern-
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
526
mental debate’ (Bennett, 1990: 108). When news media coverage high-
lights executive policy problems or failures, that is to say it is critical of
executive policy, this simply reflects a ‘professional responsibility [on the
part of the journalist] to highlight important conflicts and struggles
within the centers of power’ (Bennett, 1990: 110). Bennett’s theory
receives substantial empirical support via Mermin’s (1999) study of news
media coverage, elite debate and post-Vietnam US military interventions
and Zaller and Chui’s (1996) analysis of news media coverage of foreign
policy crises between 1945 and 1991. Both Zaller and Chui (1996) and
Mermin (1999) find that, consistent with predictions of elite manu-
facturing consent theory, the news media have rarely moved beyond the
confines of ‘official’ Washington policy debates.
An important claim of elite manufacturing consent theory is that
news coverage which criticizes or challenges executive policy occurs when
there exists elite conflict with regard to that policy. Hence, contrary to
the executive version of the manufacturing consent paradigm, the
possibility that news media coverage might be critical of executive policy
is allowed for. An implication of this possibility is that news media
coverage might have the ability to influence executive policy processes
when there is elite conflict over an issue. This is an important implication
of elite manufacturing consent that contrasts with the implications of the
executive version and which will be returned to. However Hallin (1986),
Bennett (1990) and, more recently, Mermin (1999) do not explore this
possibility and tend to equate a passive news media with reliance on
political elite sources whether executive, legislative or otherwise.2For
example, having analysed the lead up to US intervention in Somalia
during 1992, Mermin (1999)3finds that news coverage followed, not led,
elite calls for intervention in Somalia. This leads him to de-emphasize the
importance of the news media in causing US policy-makers to intervene
in the crisis. He writes:
Stories on Somalia were broadcast just after the articulation of demands for
US intervention in Washington in the summer and fall of 1992. Journalists
made the final decision to cover Somalia, of course, but the stage had been
set in Washington. The case of US intervention in Somalia, in sum, is
not at heart evidence of the power of television to move governments; it
is evidence of the power of governments to move television. (Mermin,
1999: 137)
With respect to Hallin’s (1986) finding that news media coverage
followed elite division over Vietnam, although he is careful not to dismiss
out of hand the possibility of media influence, he too uses this finding to
de-emphasize the possible importance of the news media. He writes:
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
527
The behavior of the media . . . is intimately related to the unity and clarity
of the government itself, as well as to the degree of consensus in the society
at large. This is not to say that the role of the press is purely reactive.
Surely it made a difference, for instance, that many journalists were
shocked both by the brutality of the war and by the gap between what they
were told by officials and what they saw and heard in the field. . . . But it
is also clear that the administration’s problems with the ‘fourth branch of
government’ resulted in a large part from political divisions at home. . . .
In a sense, what is really remarkable . . . is that the press and the public went
as far with American Policy in Vietnam as they did. (Hallin, 1986: 213)
In short, in that both deny, or do not explore the possibility that
news media coverage might play a key role in policy formulation, the
elite version of the manufacturing consent literature is correctly located
alongside the executive version that denies the existence of any inde-
pendent news media effect on policy.
News sources do not disprove news effects
There exist, however, both theoretical and empirical problems with the
manufacturing consent paradigm. I deal with each in turn.
Theoretically speaking, manufacturing consent theory is limited in
two ways. First the theory, in particular as articulated by Bennett (1990),
is rooted in an understanding of the relationship between journalists and
official sources. As noted earlier, finding that news media coverage is
indexed to elite opinion is equated, to all intent and purpose, with a
passive and non-influential news media. News sources, however, do not
disprove news effects, and by focusing upon the relationship between
news sources and journalists, the theory ‘black boxes’ the dynamics
between media coverage and any given policy process.4To explain this
shortcoming more fully it is useful to consider Mermin’s (1997) study of
US intervention in Somalia. As noted earlier, Mermin equates media
reliance on elite sources with non-influence. In doing so, however, he
leaves two important questions regarding the role of the media
unanswered. First, Mermin cannot provide evidence either for or against
the thesis that, by compelling senior policy-makers to respond to emotive
reporting of suffering people, news media coverage actually plays a key
part in causing policy-makers to intervene during a humanitarian crisis.
This is particularly important for those in humanitarian circles who wish
to know why western states intervene during humanitarian crisis. Also,
while news media coverage has been associated with recent interventions
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
528
it has also accompanied instances of non-intervention, for example non-
intervention during the 1994 genocide in Rwanda. Analysing the sources
of news reports might be able to explain why journalists covered, for
example, both the Rwandan genocide and the civil war/famine in
Somalia. But it cannot explain why news media coverage appeared to
cause intervention in Somalia but appeared unable to trigger intervention
in Rwanda. Answering this question requires us to move beyond news
sources to an analysis of news media content and its influence on the
policy process.
The second theoretical limitation concerns how underpinning
manufacturing consent theory is the assumption that because journalists
tend only to replicate elite views, they cannot play an independent role
during debates between elites. While journalists are undoubtedly
subjected to a variety of pressures and rarely have an entirely free hand in
deciding what to cover and how, it seems unreasonable to assume that
they play no independent role during political debate. As Timothy Cook
(1998: 12–13) argues, journalists should not be considered the passive
recipients of official information but as active participants functioning as
a political institution in their own right. Important to our concerns here,
this assumption means that elite manufacturing consent theory tends to
ignore the possibility that journalists might actually take sides (either
consciously or unintentionally) during elite debates over policy, or even
take the side of non-elites, and in doing so become active and powerful
participants in a political debate.
Empirically speaking, the implication of both elite and executive
manufacturing consent that the media does not influence government
policy, at least to any significant extent, is challenged by a wealth of
anecdotal and research evidence that points to news media coverage
playing a key role in the creation of policy. For example, taking again
Mermin’s analysis of news media coverage and US intervention in
Somalia, while his study leads us to ignore the possible importance of
news media, several key policy-makers claim to have been heavily
influenced by news media coverage of suffering people in Somalia. For
example, White House press secretary Marlin Fitzwater has stated, ‘I was
one of those two or three that was strongly recommending he do it, and
it was very much because of the television pictures of these starving kids’
(‘Reliable Sources, How Television Shapes Diplomacy’, CNN, 16 October
1994, cited in Minear et al., 1997: 55). More recently, George Bush
Senior has pointed to the pivotal importance of news media coverage in
persuading him to launch the intervention:
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
529
Former President Bush conceded Saturday that he ordered US troops into
Somalia in 1992 after seeing heart-rending pictures of starving waifs on
television. . . . Bush said that as he and his wife, Barbara, watched television
at the White House and saw ‘those starving kids . . . in quest of a little
pitiful cup of rice,’ he phoned Defense Secretary Dick Cheney and Gen.
Colin Powell, chairman of the Joint Chiefs of Staff. ‘Please come over to the
White House,’ Bush recalled telling the military leaders. ‘I – we – can’t
watch this anymore. You’ve got to do something.’ (Hines, 1999: A11)
With respect to Hallin’s analysis of the Vietnam War, historian
David Culbert (1998) has directly engaged with Hallin’s non-effect claim
arguing instead that news media coverage played a crucial role in helping
to change the course of US policy towards Vietnam. Culbert asks: ‘how
does one reconcile the persuasive conclusions of Daniel Hallin . . . that in
general, television followed elite opinion, or had little demonstrable
impact on policy-making in Vietnam, with the testimony of those who
insist that . . . footage did affect them’ (Culbert, 1998: 430). Focusing
upon the impact of the infamous footage of General Loan (chief of police
for Southern Vietnam) summarily executing an armed civilian, which was
broadcast across the world, and the claims of various actors that they were
influenced by the coverage, Culbert (1998: 437) concludes:
The Loan execution is the most visually significant footage to come out of
the war; it merits careful attention precisely because it defines the potential
of the medium for influencing elite and mass opinion. . . . Its impact
related to a changing climate of opinion which found policy-makers as well
as average citizens worried as to whether the USA’s Vietnam policy merited
continued support. In this moment of doubt and uncertainty, a visual
microcosm purporting to show the actual practice of justice by the
government of South Vietnam offered persuasive . . . evidence which gave
people looking for factual reasons to justify a change in policy an oppor-
tunity to do so.
To summarize, manufacturing consent theory fails to examine
closely the link between media coverage and policy formulation. This
leaves a gap in our understanding of the role media might come to play
in policy formulation and exacerbates the tendency for debates over media
impact to be characterized by unproductive effect/non-effect dichotomies
– for example, that between historian David Culbert and Daniel Hallin
on the question of media influence on US policy during the Vietnam War.
What is badly needed, and to which we turn next, is a theoretical
understanding of media–state relations that can account for instances
when news media coverage can influence policy formulation.
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
530
Ways forward
In the latter part of this article I want to discuss two theoretical advances.
The first is my own proposal for a policy–media interaction model that is
designed to build directly upon the work of Hallin (1986) and Bennett
(1990). The second is a review of Gadi Wolfsfeld’s political contest
model. Both offer different ways of theorizing instances of media
influence and both offer us the potential to understand instances of media
influence by building upon, rather than rejecting, manufacturing consent
theory.
Developing a theory of media influence
We can start by theorizing that, in accordance with manufacturing
consent theory, when there exists elite consensus over an issue news media
are unlikely to produce coverage that challenges that consensus. To put
this in Hallin’s terms, critical journalism is unlikely to surface with
respect to issues that fall within a sphere of consensus. However, when
there exists elite dissensus with respect to an issue, as predicted by both
Hallin (1986) and Bennett (1990), news media coverage reflects this
debate and we can expect to observe a variety of critical and supportive
news media coverage. It is in this scenario that news media has the
potential, at least, to start to play a more active and influential role in
policy debate and formulation because the possibility exists for news
media coverage to actually take sides in the elite debate. When and if this
occurs, by promoting a particular policy line advocated either by elites
outside the executive or particular members of the executive itself, news
media can play a key role in causing policy change. Before proceeding,
however, we need to identify more precisely the conditions under which
news media coverage comes to influence elite debate during periods of
elite dissensus or ‘legitimate controversy’ (Hallin, 1986).
Media framing and elite debate
First and foremost, we need to explain precisely what is meant by news
media coverage ‘taking sides’ in an elite debate. Here the concept of
framing is useful. The concept of framing refers to the ‘specific properties
of . . . [a] narrative that encourage those perceiving and thinking about
events to develop particular understandings of them’ (Entman, 1991: 7).
As ‘mentally stored principles for information processing’ (Entman,
1991: 7), frames offer ways of explaining, understanding and making
sense of events.
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
531
Relating framing to elite debate, this concept enables us to under-
stand how news media texts do not simply replicate reality, but can
actually be constructed so as to produce a particular understanding or
perception of a problem. In terms of covering elite debates, the extent to
which news media coverage adopts one particular framing of a problem
can be said to be indicative of the extent to which media coverage has
taken sides in that debate. For example, during the early part of the
1992–5 war in Bosnia, US policy-makers sought to avoid direct
engagement in the conflict. In order to justify non-intervention, policy-
makers regularly employed a particular framing of the conflict. For
example, William Shawcross (2000) describes how US Secretary of State
Warren Christopher employed a frame of ‘ancient ethnic’ hatreds when
justifying US non-involvement during the early part of the Bosnian
War:
Christopher told Congress that the conflict had ‘evolved into a war of
all against all . . . as struggle between three groups . . . each possessing
deep distrust and ancient hatreds for each other’. He believed that the
Bosnian combatants were not ready to make peace and that it would
therefore be dangerous for Clinton to insert US troops between them.
(Shawcross, 2000: 83)
This particular framing left unclear the causes of the conflict and, because
all sides were perceived as being equally to blame, justified non-
intervention. At this time during the war, news media coverage tended to
reflect and reinforce this frame (e.g. Campbell, 1998: 51–4). As the war
progressed, however, increasing numbers of congress persons, the executive
and the public were outraged at the violence and brutality in Bosnia and
actively sought direct US involvement. Towards the end of the war, the
prevalent frame became one of Serbian nationalist brutality against the
Bosnian government and news media coverage started to reflect this
interpretation of events. Importantly, this framing, unlike that of ancient
ethnic hatreds, demanded that something be done and focused blame on
Serbian nationalism. To the extent that journalists framed media reports
in this fashion, they also contributed to the political ambitions of those
who sought greater involvement in Bosnia. In doing so, news media
coverage became an active player in redefining the conflict in Bosnia as
one the US should do something about. With respect to the impact of
this coverage on policy formulation, one leading US official involved in
Bosnia, Richard Holbrooke, recently stated in an article titled ‘No Media
– No War’: ‘Let’s be clear: the reason the West finally, belatedly inter-
vened was heavily related to media coverage’ (Holbrooke, 1999: 20).
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
532
Policy uncertainty
Media coverage that adopts the reference frames of one side of an elite
debate, however, is not sufficient alone to start to account for instances
when media coverage influences policy processes. Here the concept of
policy certainty can be employed in order to theorize the circumstances
under which the policy process is most susceptible to media influence.
Recent research into the CNN effect (Gowing, 1994; Strobel, 1997;
Shaw, 1996; Minear et al., 1997) has consistently pointed to the
importance of the level of policy certainty in determining whether or not
news media coverage can influence the policy process. For example,
journalist Nik Gowing quotes Kofi Annan on the impact of media
coverage at points of policy uncertainty: ‘When governments have a clear
policy, . . . then television has little impact. . . . When there is a problem,
and the policy has not been thought [through] . . . they have to do
something or face a public relations disaster’ (Gowing, 1994: 85–6).
Similarly, Martin Shaw (1996: 181) argues that a ‘loss of policy certainty’
in the ‘aftermath of the Cold War’ may have ‘opened up a particular
window for the media’, while Warren Strobel (1997: 219) notes that ‘the
effect of real-time television is directly related to the . . . coherence . . . of
existing policy’. The idea of media influence at points of policy
uncertainty is also consistent with the broader policy studies literature
that points to a correlation between dissensus among policy-making elites
and the ability of ‘external’ actors to influence policy formulation
(e.g. Baumgartner and Jones, 1993; Hall, 1993; Schattsneider, 1961). For
example, discussing economic policy formulation in Great Britain during
the period of economic crisis and uncertainty during the 1970s, Peter
Hall argues:
Finally, the 1970s saw a vast expansion in the outside market place for
economic ideas. The media and City brokerage houses were important
participants in this marketplace, but they were not the only ones. Fuelled
by the release of the Treasury’s econometric model, new research institutes
sprang up, and a host of pamphlets about economic policy began to
circulate. In short, where once there was virtually no external commentary
on macroeconomic policy, something similar to a ‘policy network’ or ‘issue
network’ sprang up to provide outsiders with influence over a formerly
closed policy process. (Hall, 1993: 67)
Substantial theoretical work is also available which helps us to
define more precisely what we mean by policy certainty. For example,
Hilsman has developed a political process model which ‘sees a number of
different individuals and organizations involved in the policy-making
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
533
process. Each of these has power. Some have more power than others, and
the power of each varies with the subject matter’ (Hilsman, 1987: 82). In
terms of policy creation, the different power centres ‘attempt to build
coalitions among like-minded power centres. . . . Sometimes they succeed
in getting their ideal solution adopted; sometimes they succeed in
getting the half also they estimated was the best they could do’ (Hilsman,
1987: 82–3). Importantly, Hilsman (1987: 82–3) notes that (1) some-
times, the outcome of this bargaining process is ‘a policy that none of the
power centres really wanted but a compromise that achieves something
less than half a loaf for all’ and (2) not always is the ‘resultant policy
always completely logical or internally consistent’. Similarly, Alexander
George argues that:
Efforts at rational calculation of policy take place in three interrelated
contexts or subsystems within the policy making system: the individual
context (e.g. the chief executive, secretary of state); the small group context
of the face to face relationships into which the executive enters with a
relatively small number of advisors; and the organizational context of
hierarchically organized and coordinated processes involving the various
departments and agencies . . . in the executive branch. (George, 1980: 11;
emphasis in original)
George also notes, however, that the central danger for ‘rational’
decision-making is that policies emerging from the ‘play of intra-
governmental politics within the executive . . . may be more responsive
to the internal dynamics of such a policymaking process than to the
requirements of the foreign-policy problem itself’ (George, 1980: 114).
Drawing upon a typology of ‘distorted’ policy developed by Schilling
(1962), he specifies the types of ‘faulty’ policy that might emerge from
intra-governmental politics. These include (1) no policy at all; (2) com-
promised policy, when the direction that policy should take is left
unclear, or the means for achieving a well enough defined objective are
left unclarified or unfocused; and (3) unstable or blind policy, when the
internal struggle over policy is not really resolved (George, 1980: 114).
In short, through the work of George (1980) and Hilsman (1987),
we can understand policy-making as the outcome of a complex bargain-
ing process between a set of subsystems in government. Building upon
this theory we can define policy certainty as a function of the degree of
consensus and coordination of the subsystems of the executive with
respect to an issue. If an issue suddenly arises and no policy is in place,
or if there is disagreement, conflict of interest or uncertainty due to an
ambiguous policy between the subsystems of the executive there can be
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
534
said to be policy uncertainty. Conversely, policy certainty is the result of
agreement and coordination between the sub-systems of the executive.
The policy–media interaction model
We are now in a position to conceptualize more precisely the conditions
under which news media coverage comes to influence the policy process.
When there exists elite dissensus with respect to an issue, there is the
possibility that news media coverage might actually take sides in that
elite debate by adopting the reference frames of one side of an elite
debate. In effect, journalists become promoters, either consciously or
otherwise, of one particular elite group. Set in this context of negative
news media coverage, government is confronted with (1) the possibility
that public opinion might be influenced by the negative media coverage,
(2) associated damage to government image and credibility caused by the
‘bad press’ and (3) policy-makers might themselves start to question the
cogency of existing government policy. Crucially, the greater the level of
uncertainty over policy within the executive, the more vulnerable the
policy process is to the influence of negative media coverage. In this
scenario, a number of factors related to the existence of policy uncertainty
might come into play. First, if it is disagreement among the executive
policy subsystems over policy that is the cause of policy uncertainty,
critical media coverage might provide additional bargaining power to
those policy-makers seeking a change in policy direction. Second, if it is
the case that policy uncertainty is the result of there simply being no
policy in place, policy-makers are liable to feel pressured to respond to
critical coverage or else face a public relations disaster and criticism for
being ‘caught on the hop’. Here policy might be formulated, at least in
the first instance, primarily as a way of counteracting negative publicity.
Finally, not only does policy uncertainty make policy-makers susceptible
to media influence, it also means that government is ill-equipped to
respond to journalists by drawing upon its substantial public relations
apparatus. In other words, without a clearly articulated policy line with
which to respond to critical coverage, policy-makers become even more
vulnerable to a hostile press.
Alternatively, if government policy is decided on, policy-makers are
likely to resist the pressures of negative media coverage. Indeed, policy-
makers are more likely to work harder to sell existing policy by drawing
upon their substantial resources and credibility as an information source
in order to influence media debate. Here we might expect the level of
critical media coverage to subside, although parts of the elite might
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
535
continue to help generate some critical news reports. A summary of the
model that highlights its relationship to manufacturing consent theory
can be found in Table 1.
By way of example, let us apply this theory of influence to US policy
towards the Vietnam War. Hallin’s analysis highlights how news media
coverage of the war in Vietnam, up until 1968, was largely supportive of
the war and rarely published material that criticized or questioned official
US policy. This, according to Hallin, reflected the elite consensus regard-
ing US policy towards Vietnam. During this period we can characterize
media coverage as manufacturing consent for official policy (row 1 of
Table 1). During 1967 and 1968, however, concern was growing among
foreign policy elites and within the US administration as to the viability
of US intervention in Vietnam. As Culbert points out, ‘Large numbers of
Americans – policy-makers, soldiers in the field and average citizens –
had serious doubts about the wisdom of America’s Vietnam policy by
Autumn 1967’ (Culbert, 1998: 434). More specifically Hallin (1986:
159–60) argues:
A basic disagreement had thus emerged over US strategy. Westmoreland
and the Joint Chiefs believed increased military pressure would raise North
Vietnamese and NLF losses to the point that they could no longer go on
with the war. Their civilian opponents, concentrated primarily in the
Office of Systems Analysis, argued that the North Vietnamese could sustain
indefinitely the losses they would suffer even with substantial increases in
US military activity.
Table 1 The policy–media interaction model and theories of media–state
relations
Level of elite consensus Media–state relationship Role of the media
Elite consensus Media operates within
‘sphere of consensus’
(Hallin)
Media ‘manufactures
consent’ for official
policy
Elite dissensus Media operates within
‘sphere of legitimate
controversy’ (Hallin)
Media reflects elite
dissensus as predicted
by Hallin and Bennett
Elite dissensus plus policy
uncertainty within
government and critically
framed media coverage
Media takes sides in
political debate and
becomes an active
participant
Media functions to
influence direction of
government policy
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
536
On 31 January 1968, forces from North Vietnam launched the Tet
offensive, an attempt to encourage a general uprising throughout South
Vietnam against the US-sponsored government. Although the offensive
was a military failure, it was an embarrassment for the US government,
that had been maintaining that the war in Vietnam was being won. The
offensive also provided a wealth of dramatic and shocking news reporting
as the Vietnam conflict spilled over onto the streets of Saigon. The
execution of the armed civilian was the most graphic and brutal image of
this offensive, and also perhaps of the whole war. One of the most notable
journalistic judgements during this period was made by CBS commenta-
tor Walter Cronkite:
To say that we are closer to victory today is to believe, in the face of the
evidence, the optimists who have been wrong in the past. To suggest we are
on the edge of defeat is to yield to unreasonable pessimism. To say that we
are mired in stalemate seems the only reasonable, yet unsatisfactory,
conclusion. But it is increasingly clear to this reporter that the only
rational way out then would be to negotiate, not as victors, but as
honorable people who lived up to their pledge to defend Democracy, and
did the best they could. (CBS special broadcast 27 February 1968, quoted
in Culbert, 1998: 430)
In short, during this period of crisis, elite dissensus started to be
reflected in media coverage (row 2 of Table 1). However, the afore-
mentioned divisions within the subsystems of the US executive suggest
that dissensus was also present within the US executive itself. In terms of
our theoretical model, this suggests that the policy process would have
been susceptible to outside (i.e. media) influence. With respect to media
coverage, while much coverage clearly challenged official policy, in effect
adopting the perspective of those opposed to escalation in Vietnam,
whether or not overall news media coverage took sides in the debate over
Vietnam is beyond the scope of this speculative article. However, some of
the evidence provided by Daniel Hallin (1994) does suggest media
coverage adopting the perspective of those opposed to official policy. For
example, Hallin’s framing analysis indicates twice as many unfavourable
editorial comments vis-a-vis administration supporters as there were
against the critics of the war (Hallin, 1994: 44). Also, after the Tet
offensive, there were 10 times as many negative references to the
democratic credentials of the US-sponsored government in South Vietnam
as there were positive and 5.8 times as many negative references to the
morale of US troops as there were positive (Hallin, 1994: 45). Hallin’s
figures also show that during the Tet offensive critics of the administra-
tion were quoted twice as often as supporters of the administration.
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
537
Overall, while this evidence is not conclusive, it does suggest a prima
facie case that media coverage during and after the Tet offensive took
sides in the elite debate over whether to escalate, or seek negotiation and
withdrawal. In short, the conditions of policy uncertainty and critically
framed media coverage would appear to be present in this case there-
fore indicating the possibility of media influence on policy (see row 3 of
Table 1).
We can now start to make sense of both Culbert’s evidence
regarding influence and Hallin’s thesis that media coverage only followed
elite cues. The combination of policy uncertainty and critical news media
coverage meant that policy-makers were susceptible to news media
influence during this period. Culbert documents the influence of the
media in the following quote:
Harry McPherson, counsel to the President, . . . feels that the Cronkite
special ‘had a huge impact on Johnson and his sense of crumbling public
support for the war’. McPherson feels that Johnson ‘liked and trusted’
Cronkite, a fellow Texan, . . . McPherson thinks that Johnson watched
television not so much for information as to ‘gauge what its impact on the
public would be’. (Culbert, 1987: 227, cited in Culbert, 1998: 432)
In doing so, media coverage, having passively reflected elite consensus
prior to 1968, became an active participant in elite debate by adopting
the side of those opposed to the war and, in the presence of executive
policy uncertainty, influencing key policy-makers to move to withdrawal.
In short, Hallin is most likely correct in arguing that critical news media
coverage followed rather than caused elite dissensus over Vietnam.
But Culbert might also be correct because this coverage actually took
sides during the elite debate over policy, and in doing so helped shift US
policy towards withdrawal. By theorizing the conditions under which
media influences policy but building upon manufacturing consent theory,
the policy–media interaction model enables us to make sense of both
arguments.
Wolfsfeld’s political contest model
In his 1997 work The Media and Political Conflict Wolfsfeld develops
a political contest model of the media. Similar to the policy–media
interaction model outlined in the preceding section, Wolfsfeld’s goal is to
identify the conditions under which news media coverage comes to play
an active role in the formulation of policy. In doing so, his goal is to
achieve a more nuanced understanding of the relationship between the
news media and government. It is not my intention here to offer a
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
538
detailed appraisal of the specifics of Wolfsfeld’s excellent work. Rather,
my aim is to provide the reader with a summary of his argument so that
I can (1) distinguish his work from my own and (2) in doing so offer a
more comprehensive overview of the ways in which the media might play
a more influential role than is suggested by manufacturing consent
theory.
Wolfsfeld’s focus of concern is the relationship between news media,
groups in society that seek to challenge authorities and political change.
His central claim is that while news media normally function to reflect,
and even mobilize support for, dominant views in society, there are times
when they serve the interests of marginalized groups. The bulk of the
first section of his book is devoted to specifying more precisely
the conditions under which marginalized groups, what he refers to as
‘challengers’, can come to both set the media agenda and influence
political outcomes. Out of a number of variables which determine
whether or not challengers are able to seize control of the media agenda,
he argues that ‘The authorities’ degree of control over the political
environment is the key situational variable that determines whether the
news media will play an independent role in a political conflict’
(Wolfsfeld, 1997: 24). He also employs the concept of framing in order to
highlight how media coverage can effectively take the side of challengers
by promoting their particular perception of the political issue at stake.
As one example of an instance during which a challenger was able to
attain control of both the media agenda and the way in which media
coverage was framed, Wolfsfeld analyses the case of the Palestinian
Intifada during 1987. Summarizing, Wolfsfeld (1997: 167–8) argues that
during this period of unrest in the occupied territories, the Israeli
government lost control of the media agenda because (1) ‘they were
unable to take control over the political environment’, (2) the inter-
nationalization of the Palestinians’ struggle levelled the balance of power
between Palestinians (the challengers) and the Israeli government (the
authorities) and (3) the access of journalists to the sites of civil unrest
meant that the resulting footage of unarmed Palestinians engaging with
Israeli soldiers cast the Israelis, on balance, in a negative light. As a
result, a frame of ‘injustice and defiance’ (Wolfsfeld, 1997: 168) prevailed
in media reports that favoured the Palestinians’ cause.
I wish to draw attention to several aspects of Wolfsfeld’s work. The
first concerns the question of media influence on policy. While Wolfsfeld’s
study aims to develop an interactive theory of media–state relations, his
political contest model focuses primarily on explaining when and how
challengers can come to set the media agenda. As such, the model relates
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
539
largely to the question of the relationship between news sources and
the news. For example, the entire chapter detailing his study of the
Palestinian Intifada deals with explaining how and why the Palestinians
were able to secure favourable media coverage. The question of whether
this favourable (from the Palestinians’ point of view) coverage actually
influenced political outcomes is tackled in Wolfsfeld’s conclusion.
However, even here Wolfsfeld does not consider in any significant depth
the question of whether official Israeli policy was changed by the pro-
Palestinian media coverage. He does highlight how the actions of the
Israeli authorities, in particular the military, were shaped by concern over
media coverage (Wolfsfeld, 1997: 206):
Israeli authorities were especially concerned with the damage the intifada
was doing to their image . . . the Israelis spent a considerable amount of
time and effort attempting to control the damage from the news reports
coming out of the territories. . . . One of the clearest examples of Israeli
adaptation to the news media occurred in the field. The presence of the
news media had a direct influence on restraining soldiers’ behavior.
Wolfsfeld also highlights the influence of the media coverage on the
Palestinians (Wolfsfeld, 1997: 207–8). However, Wolfsfeld (1997: 208–9)
only starts to tackle the question of media impact on policy on the last page
of his discussion regarding the Intifada when he claims that (1) media
coverage changed the balance of power between the Palestinians and the
Israelis and (2) that it was media coverage which caused the US to
intervene diplomatically in the crisis although he references another
researcher’s study as evidence for the latter claim. However, at no point
does Wolfsfeld make explicit the effect the media coverage had on the
actual substance of Israeli policy. In short, Wolfsfeld’s political contest
model and case studies provide a strong theoretical account that explains
why challengers can come to set the media agenda, but it does not
theorize the link between the resulting media coverage and actual policy
outcomes. This is where the policy–media interaction model can serve as
a useful additional component to Wolfsfeld’s work by theorizing precisely
when media coverage might influence policy outcomes.
The second point I wish to make concerns the scope of potential
media influence. Unlike my proposed policy–media interaction model
that posits the possibility of an influential media as part of the elite
debate, Wolfsfeld’s model highlights the processes by which non-elites
(i.e. the Palestinians) were able to secure favourable press coverage.
Although this is likely to occur only infrequently, it remains important to
understand the ways in which the news media can come to play a part in
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
540
broader scale social and political change during which non-elite groups
succeed in achieving political change. As such Wolfsfeld’s model plays
an important part in completing our theoretical understanding of the
media–state relationship.
Concluding remarks
Taken together, the policy–media interaction model and Wolfsfeld’s
political contest model provide us with the theoretical tools that can help
us explain when and how the news media can come to play an important
part in the formulation of government policy. At the same time, both
models build upon existing and well-tested media–state theories and in
doing so avoid an unproductive continuation of effect/non-effect argu-
ments. The theoretical models can be used, as demonstrated when
examining the Vietnam example, to reconcile contrasting claims regard-
ing the role of the media. With respect to the frequency and significance
of media influence, the policy–media interaction model would suggest
that media influence is likely to be a frequent occurrence within the
context of elite debate over policy. Contrastingly, Wolfsfeld’s analysis of
non-elites securing the media agenda is likely to occur more rarely
although when it does the significance of media influence (in terms of
causing large-scale political change) might well be argued to be greater.
In either instance, the importance of the media regarding political
outcomes is far greater than allowed for by existing manufacturing
consent theory.
Debate and analysis of the media–state relationship have been
dogged by simplistic and dichotomous argument. The theoretical models
outlined here should provide the starting point for a research agenda that
can move us beyond this state of affairs through the generation of a
nuanced, two-way understanding of the direction of influence between
media and the state.
Notes
Thanks to Eric Herring for comments on earlier drafts of this article.
1. It should be noted that the distinction introduced here, developed jointly
with Eric Herring, serves as a conceptual device to delineate two possible
aspects of the manufacturing consent paradigm. It is not suggested that the
work of the authors cited as examples either (1) fits neatly into either category
or (2) explicitly describes itself as belonging to one or other version of the
manufacturing consent paradigm.
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
541
2. In the book Media, Power, Politics, Paletz and Entman (1981: 20) do note that
elites can utilize the news media in order to redistribute power among
themselves, although this remains a relatively undeveloped hypothesis in
their work which, as with Bennett, Hallin and Mermin, tends to de-
emphasize the possibility of independent news media influence.
3. This argument was originally published in Mermin (1997).
4. Unlike Hallin (1986) and Bennett (1990), Chomsky and Herman (1988)
analyse in some detail the policy–media interface and as such cannot be
accused of black boxing this dimension of media–state relations. They do not,
however, consider in detail instances when media coverage might influence
and change policy, but rather focus on media coverage as a reinforcement of
government policy.
References
Baumgartner, Frank and Bryan Jones (1993) Agendas and Instability in American
Politics. Chicago, IL: University of Chicago Press.
Bennett, Lance W. (1990) ‘Toward a Theory of Press–State Relations in the
United States’, Journal of Communication 40(2): 103–25.
Blair, Tony (1999) Speech by the Prime Minister, Tony Blair, to the Economic
Club of Chicago, Hilton Hotel, Chicago, IL, 22 April; available at:
www/feo.gov.uk/news (downloaded May 1999).
Campbell, David (1998) National Deconstruction: Violence, Identity and Justice in
Bosnia. London and Minneapolis: University of Minnesota Press.
Chomsky, Noam and Edward Herman (1988) Manufacturing Consent. New York:
Pantheon.
Cohen, Bernard (1994) ‘The View from the Academy’, pp. 8–11 in W.L. Bennett
and D.L. Paletz (eds) Taken by Storm: The Media, Public Opinion and US
Foreign Policy in the Gulf War. Chicago, IL: University of Chicago Press.
Cook, Timothy E. (1998) Governing With the News. Chicago, IL and London:
University of Chicago Press.
Culbert, David (1987) ‘Johnson and the Media’, pp. 224–7 in Robert A. Divine
(ed.) The Johnson Years, Volume One: Foreign Policy, the Great Society and the
White House. Lawrence, KS.
Culbert, David (1998) ‘Television’s Visual Impact on Decision-Making in the
USA, 1968: The Tet Offensive and Chicago’s Democratic National Conven-
tion’, Journal of Contemporary History 33(3): 419–49.
Entman, Robert (1991) ‘Framing US Coverage of International News: Contrasts
in Narratives of the KAL and Iran Air incidents’, Journal of Communication
41(4): 6–27.
George, Alexander (1980) Presidential Decision Making in Foreign Policy: The
Effective Use of Information. Boulder, CO: Westview Press.
Glasgow University Media Group (1985) War and Peace News. Milton Keynes:
Open University Press.
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
542
Gowing, Nik (1994) Real-Time Television Coverage of Armed Conflicts and Diplomatic
Crises: Does it Pressure or Distort Foreign Policy Decisions?, Harvard Working
Paper. Cambridge, MA: The Joan Shorenstein Barone Center on the Press,
Politics and Public Policy, Harvard University.
Hall, Peter (1993) ‘Policy Paradigms, Social Learning, and the State’, Comparative
Politics 25(3): 275–96.
Hallin, Daniel (1986) The Uncensored War. Berkeley: University of California
Press.
Hallin, Daniel (1994) We Keep America on Top of the World: Television Journalism
and the Public Sphere. London and New York: Routledge.
Hammond, Phil and Edward Herman (eds) (2000) Degraded Capability: The
Media and the Kosovo Crisis. London: Pluto Press.
Herman, Edward (1993) ‘The Media’s Role in US Foreign Policy’, Journal of
International Affairs 47(1): 23–45.
Hilsman, R. (1987) The Politics of Policy Making in Defense and Foreign Affairs.
Englewood Cliffs, NJ: Prentice-Hall.
Hines, Craig (1999) ‘Pity, not US Security, Motivated Use of GIs in Somalia,
Bush Says’, The Houston Chronicle 24(October): A11.
Holbrooke, Richard (1999) ‘No Media – No War’, Index on Censorship 28(3):
20–1.
Mermin, Jonathan (1997) ‘Television News and American Intervention in
Somalia’, Political Science Quarterly 112(3): 385–403.
Mermin, Jonathan (1999) Debating War and Peace. Princeton, NJ: Princeton
University Press.
Minear, Larry, Colin Scott and Robert Weiss (1997) The News Media, Civil Wars
and Humanitarian Action. Boulder, CO and London: Lynne Rienner.
Paletz, David L. and Robert Entman (1981) Media, Power, Politics. New York and
London: Free Press.
Parenti, Michael (1993) Inventing Reality: The Politics of News Media. New York:
St Martin’s Press.
Philo, Greg and G. McLaughlin (1993) The British Media and the Gulf War.
Glasgow: Glasgow University Media Group.
Schattsneider, E.E. (1961) The Semi-Sovereign People: A Realist’s View of Democracy
in America. Fort Worth, TX: Harcourt Brace Jovanovich.
Schilling, Warner (1962) ‘Politics of National Defense: Fiscal 1950’, in W.
Schilling, P. Hammond and G. Snyder (eds) Strategy, Politics and Defence
Budgets. New York: Columbia University Press.
Shaw, Martin (1996) Civil Society and Media in Global Crises. London: St Martin’s
Press.
Shawcross, William (2000) Deliver Us From Evil: Warlords and Peacekeeper in a
World of Endless Conflict. London: Bloomsbury.
Sigal, Leon (1973) Reporters and Officials. Lexington, MA: D.C. Heath.
Strobel, Warren (1997) Late Breaking Foreign Policy. Washington, DC: United
States Institute of Peace.
R O B I N S O N : T H E I N F L U E N C E O F M E D I A O N W O R L D P O L I T I C S
543
Williams, Kevin (1993) ‘The Light at the End of the Tunnel’, in John Eldridge
(ed.) Getting the Message: News Truth and Power. London: Routledge.
Wolfsfeld, Gadi (1997) The Media and Political Conflict. Cambridge: Cambridge
University Press.
Zaller, John and Dennis Chui (1996) ‘Government’s Little Helper: US Press
Coverage of Foreign Policy Crises, 1945–1991’, Political Communication 13:
385–405.
E U R O P E A N J O U R N A L O F C O M M U N I C A T I O N 1 6 ( 4 )
544