See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245518065

Priming or Framing: Media Influence on Attitudes Toward

Foreign Countries

Article in Gazette · December 2003

DOI: 10.1177/0016549203065006005

CITATIONS READS

196 6,108

3 authors, including:

Joseph Graf Lars Willnat

American University Washington D.C. Syracuse University
13 PUBLICATIONS 527 CITATIONS 76 PUBLICATIONS 2,268 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Joseph Graf on 11 July 2017.

The user has requested enhancement of the downloaded file.

GAZETTE: THE INTERNATIONAL JOURNAL FOR COMMUNICATION STUDIES
COPYRIGHT © 2003 SAGE PUBLICATIONS
LONDON, THOUSAND OAKS & NEW DELHI, VOL 65(6): 493–508
[0016-5492(200312)65:6;493–508;038750]
www.sagepublications.com

PRIMING OR FRAMING
Media Influence on Attitudes toward Foreign Countries

Paul R. Brewer, Joseph Graf and Lars Willnat

Abstract / This study examines two routes for media effects on the standards by which people
evaluate foreign countries. The first is indirect: a news story about an issue in a domestic context
may heighten the cognitive accessibility of thoughts about the issue, thereby priming audience
members to base their evaluations of foreign nations on those thoughts. The second is direct: a news
story that presents a frame linking an issue to a foreign nation in a way that suggests a particular
evaluative implication may shape how audience members judge that nation. An experiment revolv-
ing around media coverage of two issues and attitudes toward four nations found evidence for media
influence along the second route but not the first.

Keywords / attitudes toward foreign countries / cognitive accessibility / media effects / news frames
/ priming

Most people are heavily dependent on the mass media for information about
international affairs. As a result, the media can play an important role in
shaping mass perceptions of other nations. Studies have found that exposure to
news coverage increases knowledge about and can significantly influence public
opinion toward foreign nations (Albritton and Manheim, 1983, 1985; Manheim
and Albritton, 1984; Perry, 1985, 1987). Such perceptions, in turn, have
important implications in a number of areas, ranging from the nature of
personal interactions among people of differing nations to mass attitudes about
foreign policy to the practice of public diplomacy (Bartels, 1995; Manheim,
1991, 1994; Peffley and Hurwitz, 1992). Thus, it is not surprising that actors
in the international arena often undertake considerable efforts in order to mold
the content of media coverage.
But how, exactly, do the media influence attitudes toward other nations, and
under what specific circumstances will they do so? To date, there has been rela-
tively little research into the mechanisms that govern media influence on these
attitudes. The present study takes a step to fill this gap by testing whether news
stories help determine which standards of judgment citizens use to form atti-
tudes about foreign nations. Specifically, it tests for evidence of two different
sorts of media effects. First, can the media prime standards for judging foreign
nations by highlighting issues on the domestic agenda? Second, can the media
alter the judgment process by providing news frames that directly link issues to
specific foreign nations? The evidence for answering these questions comes from
494 GAZETTE VOL. 65 NO. 6

an experiment conducted in the US that revolved around two prominent issues

– terrorism and drugs – and four countries that American media coverage has
linked to one issue or the other: Libya, Iran, Colombia and Mexico.

Two Routes for Media Influence

Our theoretical account begins with the associative network model, the
dominant framework within cognitive social psychology for understanding
human information processing and judgment. According to this model, memory
consists of an organized network of concepts (or nodes) that are linked through
associative pathways (Anderson and Bower, 1973; Collins and Loftus, 1975).
Individual nodes within memory can be more or less accessible (i.e. easy to
recall); similarly, the associations between nodes can be strong or weak. Studies
grounded in this model suggest two routes by which the media might influence
how citizens form judgments about foreign nations.
First, research on media priming raises the possibility that news stories may
influence such judgments indirectly. The notion of priming is built on the
assumption that a stimulus can activate previously learned cognitive structures,
thereby influencing the judgment process (Fiske and Taylor, 1984). Most
research on priming points to accessibility as the key mediator of priming effects
(Higgins and King, 1981; Wyer and Srull, 1986, 1989). According to this view,
a concept’s accessibility within memory is determined in part by the frequency
(Higgins et al., 1985) and recency (Herr et al., 1983; Higgins et al., 1977; Srull
and Wyer, 1980) with which it has been used in the past. When a node is acti-
vated in memory – or primed – it becomes more accessible and thus more likely
to play a role in the formation of subsequent evaluations.
Through a series of experiments, Iyengar and Kinder (1987) demonstrated
that news coverage of an issue can prime viewers to give that issue more weight
in their overall evaluations of public officials and political candidates. The
experiments showed, for example, that exposure to media coverage of national
problems such as energy, defense and inflation boosted the weight that Ameri-
cans assigned to US President Jimmy Carter’s performance on these particular
issues in forming their general evaluations of his performance. The authors
explained these effects by arguing that ordinary people, when facing complex
political issues or events, do not base their judgments on all of the relevant
knowledge stored in their memories. Instead, they adopt a shortcut strategy,
making evaluations based on the pieces of information most easily retrieved
from memory (see also Krosnick and Brannon, 1993). Since people typically
rely on the mass media for information about political events (Iyengar and
Ottati, 1994), the accessibility of such information is determined partly by
which stories the media choose to cover (Iyengar and Kinder, 1987; Krosnick
and Kinder, 1990). Exposure to media coverage of an issue tends to make that
issue more accessible in people’s minds; this heightened accessibility, in turn,
increases the likelihood that people will base subsequent evaluations on their
thoughts about the issue.
One implication of the accessibility-driven view of media influence is that
news coverage need not draw a direct association between an issue and a specific
BREWER ET AL.: PRIMING OR FRAMING 495

target of judgment in order to influence how people form evaluations of that

target. In the associative network model, the activation of concepts follows the
principle of spreading activation (Collins and Loftus, 1975): once one node is
activated, activation spreads along the associative pathways to other nodes in
the mental network. Thus activation may expand from one political concept to
other, indirectly associated concepts (Domke et al., 1998). This raises the possi-
bility that news stories about issues on the domestic agenda may affect how
citizens evaluate foreign nations. For example, a story about domestic terror-
ism may prime the issue of terrorism in citizens’ memories, thereby heighten-
ing the impact of attitudes about terrorism on judgments of foreign nations
previously associated with terrorism.
Yet there are plausible rationales for questioning the likelihood of such
indirect influence. One revolves around the role of applicability or relevance in
mediating the effects of primed thoughts on judgments. Price and Tewksbury
(1997) argue that accessibility of a concept is determined in part by its applica-
bility to the current stimulus. Along similar lines, Miller and Krosnick suggest
that ‘the impact of accessible attitudes may be great or negligible depending on
their perceived relevance to the judgment at hand’ (Miller and Krosnick, 1996:
82; see also Iyengar and Kinder, 1987). If these authors are correct, then news
stories may have to pass an additional hurdle in order to produce priming effects
on attitudes toward foreign countries: they may need to prime an issue that
people see as relevant or applicable toward the country in question. When
people do not or cannot draw associations between the issue and the nation,
there will be no route along which activation can spread.
It may also be that media priming effects are not, in truth, mediated by
accessibility. Miller and Krosnick (2000) conclude that such effects are instead
the result of learning from the media about what news reporters think are the
most important problems facing the country and thus merit focus in forming
political evaluations. The implication of their argument is that spreading acti-
vation does not produce media priming effects. If that is so, audience members
will not automatically apply their thoughts about an issue in one realm of
judgment to other realms.
Finally, there is the potential for ambiguity in the link between the issue
and the target of judgment. Oftentimes citizens can draw multiple connections
between one concept and another. For example, Americans could draw a
negative association between the ‘war on drugs’ and the nations of Mexico and
Colombia (with the thought in mind that these nations are prominent sources
of illegal drugs that reach the US) but also a positive association between the
‘war on drugs’ and the same nations (with the thought in mind that these
nations are key US allies in that war). In such instances, news stories may
activate contradictory associations and thus fail to guide citizens’ judgments in
any consistent direction even when they do make thoughts about an issue more
accessible or induce learning about an issue’s importance.
A second route by which the media could influence attitudes toward foreign
countries is through framing. A frame, according to Gamson (1992: 3), is a story
line or ‘organizing idea’. Framing, in turn, consists of ‘select[ing] some aspects
of a perceived reality and mak[ing] them more salient in a communicating text,
496 GAZETTE VOL. 65 NO. 6

in such a way as to promote a particular problem definition, causal interpre-

tation, moral evaluation, and/or treatment recommendation’ (Entman, 1993:
52). In short, frames may guide how people understand the world and thus form
judgments.
Studies of framing and public opinion offer two psychological mechanisms
for explaining framing effects. As with priming, some argue that framing works
though an accessibility-driven process (Cappella and Jamieson, 1997; Iyengar,
1991). In this explanation, frames influence the accessibility of associations in
memory, thereby increasing the likelihood that audience members will follow
those associations when thinking about the issue. The other explanation posits
that framing works through a more thoughtful process than the accessibility
model suggests. According to Nelson et al. (1997), framing works by telling
people which associations should receive greater weight and which should
matter less. The authors’ experiments showed that issue frames affect import-
ance judgments and that these importance judgments mediate the effects of
frames on opinion, whereas accessibility does not. Both accounts, however,
conclude that exposure to frames can shape which standards of judgments
people use to evaluate the subject of the frame.
Given that citizens are heavily dependent on the mass media for infor-
mation about the world, one might expect the media to play an important role
in framing foreign nations for the public. Indeed, previous research has shown
that news frames come in many forms and influence public opinion in a number
of domains. The media can frame political campaigns in terms of policy or
strategy, with the latter frame fostering public cynicism about the political
process (Cappella and Jamieson, 1997). Similarly, the media can frame social
problems in episodic or thematic terms, thereby shaping audience members’
attributions of responsibility for those problems (Iyengar, 1991). Particularly
relevant in the case at hand, however, are news frames that explicitly associate
a specific issue with a specific nation and suggest an evaluative implication for
that association. For example, a news story could frame a particular country as
a supporter of terrorism, thereby suggesting that if one supports a ‘war on
terrorism’ then one should evaluate that nation negatively.
How might audience members respond to frames of this sort? In the
associative network model, exposure to such a frame could activate and thus
heighten the accessibility of a particular association (be it positive or negative)
between a stance on an issue and a foreign nation (Iyengar, 1991; Price and
Tewksbury, 1997). This, in turn, could produce a framing effect on judgments
of the nation. Alternatively, such exposure could shape the importance attached
to the association in the evaluation process. Exposure to the frame could also
create an association where none previously existed, thereby producing the same
effect.
Note, then, that this form of influence is not dependent on the existence of
a previously learned and unambiguous association between an issue and a
nation. Nor is it necessarily contingent upon the mediation of accessibility and
spreading activation (Nelson et al., 1997). Thus, it may be that media framing
effects on attitudes toward other countries can take place in cases when priming
along an indirect route does not.
BREWER ET AL.: PRIMING OR FRAMING 497

With all of this in mind, we have two goals. One is to test whether spread-
ing activation can produce priming effects along indirect routes, as the simple
accessibility-based model of priming suggests. The other is to test whether the
media will influence how citizens form judgments about other nations when –
and perhaps only when – they provide news frames that explain why one should
evaluate a specific country in terms of a particular issue and how one should
connect them.

Cases and Hypotheses

To investigate the effects of news stories on attitudes toward foreign countries,
this study tests whether newspaper stories about illegal drugs and terrorism can
shape how Americans evaluate four nations that have played important roles in
US foreign affairs: Libya, Iran, Mexico and Colombia. Although we focus here
on a specific population, a specific medium, specific issues and specific target
nations, our central concern is the process by which people, regardless of their
own nationality, form evaluations of nations that are foreign to them, rather
than the nature of public opinion toward any particular nation. Although the
medium, the audience, the issues and the target nations may differ in other
circumstances, the psychological process underlying judgments of foreign
nations should follow general principles of human cognition.
Moreover, the cases at hand seemed to be particularly promising ones for
examining media influence. Throughout the past two decades, the American
public has been exposed to a steady stream of stories linking Libya and Iran to
allegations of state terrorism. An examination of New York Times coverage, for
example, reveals that Iran was mentioned in 831 stories in 1999 and that 94
of those stories (11 percent) contained references to terrorism. Libya was men-
tioned in 180 stories in 1999 and 49 of those stories (27 percent) mentioned
terrorism. In January and February 2000, shortly before our experiment took
place, 27 of 173 stories that mentioned Iran (16 percent) and four of 20 stories
that mentioned Libya (20 percent) included a reference to terrorism.1 It is also
worth noting that a particularly common story line within the media coverage
for each nation was one that framed it as a sponsor of terrorism.
Media coverage of Mexico and Colombia, meanwhile, has often portrayed
these nations as major battlefields in the war on drugs. The New York Times men-
tioned Colombia in 530 stories in 1999 and 182 of those stories (34 percent)
mentioned illegal drugs. For Mexico, 312 of 2516 stories in 1999 (12 percent)
included a mention of drugs. As before, this tendency persisted in the first two
months of 2000: 31 of 92 stories that mentioned Colombia (34 percent) and 56
of 391 stories about Mexico (14 percent) also included a reference to drugs. Many
of the stories that included such references framed Mexico and Colombia as US
allies in the war on drugs. Some stories, however, framed these nations as sources
of illegal drugs without mentioning their efforts to fight the drug trade.
Given the nature of the media coverage, it may be that audience members
among the American public have learned to associate terrorism with Iran and
Libya and illegal drugs with Colombia and Mexico. If so, then the logic of
spreading activation suggests that exposure to news about terrorism and illegal
498 GAZETTE VOL. 65 NO. 6

drugs – even news within a domestic context – could influence how accessible
thoughts about these issues are when these audience members evaluate foreign
nations. Thus, exposure to media coverage of domestic terrorism and domestic
illegal drug use may prime audience members to use their attitudes about terror-
ism and drugs to form judgments of the four target countries. Accordingly, we
suggest the following priming hypotheses:

Hypothesis 1: Compared to people not exposed to any news about terrorism, people primed
with news stories about domestic terrorism should be more likely to judge Iran and Libya
based on associations with anti-terrorism efforts.

Hypothesis 2: Compared to people not exposed to any news about illegal drugs, people primed
with news stories about domestic illegal drug use should be more likely to judge Colombia and
Mexico based on associations with the war on drugs.

In the case of Hypothesis 1, participants who read news stories about terror-
ism in the US should be more likely to evaluate Libya and Iran based on their
attitudes toward terrorism than participants who did not read any stories about
terrorism. Given the nature of prior media coverage, this would be reflected in
the activation of associations between terrorism and the target nations (with
citizens seeing these countries either as supporters of state terrorism or targets
of anti-terrorism efforts). In the case of Hypothesis 2, stories about illegal drugs
in the US might activate associations between the war on drugs and the target
nations (with citizens seeing these countries either as drug traffickers or allies
in the war on drugs).
A second set of hypotheses revolves around whether exposure to stories that
directly link the issues to the target nations will influence the judgment process.
More specifically, they center on the effects of media coverage framing Iran and
Libya as sponsors of terrorism or Colombia and Mexico as allies in the war on
drugs. Note that these frames not only suggest which issue audience members
should consider in forming their judgments of the target nations; they also
suggest particular evaluative implications. Consequently, we state the following
framing hypotheses:

Hypothesis 3: Compared to people not exposed to any news about terrorism, people exposed
to news stories framing Iran and Libya as sponsors of terrorism should be more likely to judge
these nations based on negative associations with anti-terrorist efforts.

Hypothesis 4: Compared to people not exposed to any news about illegal drugs, people
exposed to news stories framing Colombia and Mexico as allies in the war on drugs should
be more likely to judge these nations based on positive associations with the war on drugs.

It should be noted that unlike Hypotheses 1 and 2, Hypotheses 3 and 4

specify the direction of the expected effects, allowing the use of one-tailed tests
for these hypotheses.

Method and Measures

The study is based on a pretest/post-test experiment with four test groups. The
experimental treatment and post-test took place about two weeks after the
BREWER ET AL.: PRIMING OR FRAMING 499

pretest was administered. A total of 199 students enrolled in undergraduate

classes at a large, private university on the East Coast of the US completed the
experiment in early 2000. Our use of such a sample raises questions, of course,
about the external validity of our findings – a point we address further in the
conclusion.
Both the pretest and the post-test contained questionnaire items that asked
for the participants’ general evaluation of each nation and their assessments of
US foreign policy toward these nations. Participants were first asked to evaluate
on a seven-point scale ‘how favorably or unfavorably’ they felt toward each of
the four target nations (1 = favorably, 7 = unfavorably). Similarly, participants
were asked whether they ‘favor or oppose diplomatic relations’ with Iran and
then Libya ‘in the near future’ (1 = strongly favor, 7 = strongly oppose) and
whether they ‘favor or oppose US aid’ to Mexico and then Colombia (1 =
strongly favor, 7 = strongly oppose). The questions regarding the four nations
were interspersed throughout various unrelated questions about politicians and
political parties.
Measures of attitudes toward Libya and attitudes toward Iran were created
by adding the scores from the questions about how favorably the participants
saw each nation and how strongly they favored diplomatic ties with each nation.
For Libya, the correlation between the items was .46 in the pretest and .53 in
the post-test; for Iran, it was .39 in the pretest and .51 in the post-test (all
significant at p < .001). Measures of attitudes toward Mexico and attitudes
toward Colombia were created by adding the scores from the questions about
how favorably the participants saw each nation and how strongly they favored
US aid to each country. The correlations between the items in each index were
more modest here (for Mexico, .31 in the pretest and .24 in the post-test; for
Colombia, .40 in the pretest and .29 in the post-test) but nonetheless signifi-
cant at p < .001. All four resulting indices were transformed to range from 1
(the most favorable attitude possible) to 0 (the most unfavorable).
The pretest also contained a series of items designed to assess students’
general attitudes toward terrorism and illegal drugs. Attitudes toward terrorism
were assessed by asking students to judge on a seven-point scale (ranging from
‘strongly agree’ to ‘strongly disagree’) whether (1) ‘The spread of terrorism is
the greatest threat to our national security’, (2) ‘The US should do all it can to
eliminate terrorism, even if that means war with terrorist nations’ and (3)
‘People who resort to terrorism sometimes have legitimate grievances’. A
measure of support for a war on terrorism was created by summing the
responses to these three questions (after reverse-coding the third) and then
transforming this index so that it ranged from 0 to 1, where 1 indicated the
strongest possible support for fighting terrorism. The reliability of the scale was
.50 (Cronbach’s alpha); its mean was .51.
Attitudes toward illegal drugs were assessed by asking students to evaluate
on an identical seven-point scale whether (1) ‘It is not really a big problem if
people occasionally use marijuana’, (2) ‘The war on drugs should be a high
priority in the United States’ and (3) ‘The federal government should do more
to end illegal drug use’. A measure of support for the war on drugs was created
by summing the scores for these three questions (after reverse-coding the first)
500 GAZETTE VOL. 65 NO. 6

and transforming the resulting index so that 1 indicated the maximum possible
support for the war on drugs and 0 indicated the minimum. The reliability of
the scale was .81 (Cronbach’s alpha); its mean was .51.
In addition, the pretest measured students’ gender and foreign affairs
knowledge. The inclusion of controls for these variables did not influence the
results of the analyses and they are therefore not reported.
Participants were randomly assigned to a control condition or one of four
experimental conditions. The treatments were based on actual newspaper
articles, though these were rewritten to be between 400 and 500 words long.
Care was taken in making the articles appear as real as possible, and they were
presented with New York Times bylines. The participants were told that the
project dealt with student opinions toward the news; to further disguise the
purpose of the study, the post-test began with a set of unrelated questions about
how interesting, credible and objective the articles were.
Some participants were randomly assigned to read two articles about
terrorism; others read two articles about the war on drugs. More specifically,
some participants read stories that explicitly framed Libya and Iran as sponsors
of terrorism (framing group 1) or Mexico and Colombia as US allies in the war
on drugs (framing group 2), whereas other participants read stories about
domestic terrorism (priming group 1) or the domestic war on drugs (priming
group 2). Participants in the control condition read stories about computer use
and the Euro, topics presumably unrelated in citizens’ minds to either the issues
at hand or the target nations.

Findings
To reiterate, this study tested for two types of media effects on how participants
evaluated the target nations: effects produced by stories that focused on issues
within a domestic context without mentioning a foreign nation (media priming)
and effects produced by stories that linked the foreign nations to issues through
news frames (media framing). In the first analysis, which examined the impact
of the terrorism stories on post-test evaluations of Libya and Iran, the partici-
pants who read articles about drug use or the control articles were grouped into
the control condition. Table 1 reports the means and standard deviations for
the pretest and post-test measures of attitudes toward Iran and Libya by con-
dition. As the table shows, these attitudes tended toward negativity across the
board, although the post-test attitudes of participants who read the stories that
directly linked Libya and Iran to terrorism were particularly negative.
According to Hypothesis 1, participants who read news stories about
domestic terrorism should have been more likely to evaluate Libya and Iran
based on their attitudes toward terrorism than participants who did not read
any stories about terrorism. This should have led to interactions between
exposure to the domestic terrorism treatment and support for a war on terror-
ism (presumably negative interactions, given the likely nature of any previously
learned associations). Hypothesis 3 points to another expectation: compared to
participants in the control condition, participants who read the stories framing
Libya and Iran as sponsors of terrorism should have been more likely to draw
BREWER ET AL.: PRIMING OR FRAMING 501

TABLE 1
Pretest and Post-test Attitudes toward Libya and Iran, by Condition

Libya Iran
———————— ——————————
Condition Pre Post Pre Post

Stories about domestic terrorism .45 .45 .46 .43

N = 43 (.18) (.19) (.20) (.22)
Stories linking Libya and Iran to terrorism .40 .32 .40 .35
N = 42 (.18) (.19) (.20) (.21)
Control .43 .41 .41 .41
N = 114 (.19) (.20) (.20) (.21)

Total .43 .40 .42 .40

N = 199 (.19) (.20) (.20) (.21)

Note: Table entries are means within conditions. Standard deviations are in parentheses.

negative associations between anti-terrorist efforts and these nations. Thus, one
would expect negative interactions between exposure to this treatment and
support for a war on terrorism.
The following ordinary least square regression model provided a test of both
hypotheses:

Post-test Attitude toward Libya/Iran = Support for war on terrorism +

Exposure to stories about domestic terrorism + Support for war on terror-
ism
Exposure to stories about domestic terrorism + Exposure to stories
linking Libya and Iran to terrorism + Support for war on terrorism

Exposure to stories linking Libya and Iran to terrorism + Pretest Attitude
toward Libya/Iran + Constant + e

Exposure to stories about domestic terrorism was coded as 1 if the participant

read this treatment and 0 otherwise. A similar procedure was used to create
exposure to stories linking Libya and Iran to terrorism. The model also included
terms for pretest attitudes toward the country and support for a war on terror-
ism. The crucial terms in the model, however, were the interaction terms,
created by multiplying participants’ support for a war on terrorism score and
each of the condition variables. The first of these interactions captured the
degree to which the impact of anti-terrorism depended on whether participants
read the stories about domestic terrorism. The second captured the degree to
which the impact of anti-terrorism depended on whether participants read the
stories that explicitly linked Libya and Iran to terrorism.
As Table 2 shows, the results produced no support for Hypothesis 1.
Exposure to the stories about domestic terrorism did not alter the impact of atti-
tudes about terrorism on either attitudes toward Libya or attitudes toward Iran.
In other words, priming along an indirect route did not seem to take place here:
502 GAZETTE VOL. 65 NO. 6

TABLE 2
The Impact of Terrorism Stories and Attitudes about Terrorism on Attitudes toward
Libya and Iran

Independent variable Libya Iran

Support for war on terrorism –.14 (.07) –.10 (.08)

Stories about domestic terrorism –.01 (.07) –.003 (.07)
Support for war on terrorism
stories
about domestic terrorism .07 (.12) –.04 (.14)
Stories linking Libya and Iran to terrorism –.11 (.08) .07 (.09)
Support for war on terrorism
stories
about Libya and Iran .09 (.15) –.28* (.16)
Pretest attitude .69** (.06) .62** (.06)
Constant .19 (.05) .21 (.06)

R2 .52 .46
N 193 194

Notes: Table entries are unstandardized regression coefficients. Standard errors are in paren-
theses. Significance tests are based on two-tailed tests except where noted in the text.
** p < .01, * p < .05.

participants who read these stories did not differ discernibly from control
participants in the way that they used their attitudes about terrorism to form
evaluations of the nations.
On the other hand, the results produced partial support for Hypothesis 3.
Exposure to news stories that explicitly portrayed Libya and Iran as sponsors
of terrorism significantly altered the impact of anti-terrorist attitudes on partici-
pants’ evaluations of Iran (p < . 05, one-tailed test). Moreover, the interaction
had the anticipated negative sign. In this instance, participants did seem to
adopt the frame of reference provided by the media in evaluating Iran. The
same did not hold true in the case of Libya: here, the effect of support for a war
on terrorism did not significantly differ as a result of exposure to the stories
linking Libya and Iran to terrorism. In sum, for Libya and Iran there was little
evidence of priming along the indirect route and mixed evidence of influence
produced by news frames for the nations at hand.
The picture became clearer in the second analysis, which examined the
effects of the stories about illegal drugs on post-test attitudes toward Mexico
and Colombia. In this analysis, the participants who read articles about terror-
ism or the control articles were grouped into the control condition. Table 3
reports the means and standard deviations for the pretest and post-test
measures of attitudes toward Mexico and Colombia by condition. Participants
in all three conditions tended to rate Mexico rather favorably and Colombia
slightly unfavorably both before and after the experimental procedure.
By the logic of Hypothesis 2, participants who read news stories about the
domestic drug war should have been more likely to base their post-test evalu-
ations of Mexico and Colombia on their attitudes about the war on drugs than
BREWER ET AL.: PRIMING OR FRAMING 503

TABLE 3
Pretest and Post-test Attitudes toward Mexico and Colombia, by Condition

Mexico Colombia
———————— ——————————
Condition Pre Post Pre Post

Stories about domestic war on drugs .62 .59 .47 .44

N = 46 (.18) (.17) (.18) (.17)
Stories linking Mexico and Colombia to
war on drugs .61 .56 .48 .47
N = 42 (.19) (.19) (.20) (.20)
Control .61 .59 .46 .43
N = 111 (.14) (.15) (.17) (.17)

Total .61 .58 .47 .45

N = 199 (.16) (.16) (.18) (.18)

Note: Table entries are means within conditions. Standard deviations are in parentheses.

participants who did not read any stories about drugs. This could have
produced interactions between exposure to the domestic treatment and support
for a war on drugs. Hypothesis 4 leads to another expectation: compared to
participants in the control condition, participants who read the stories framing
Mexico and Colombia as allies in the war on drugs should have been more likely
to draw positive associations between this war and the two nations. Here, then,
one would specifically expect positive interactions between exposure and
support for a war on drugs.
A second regression model tested Hypotheses 2 and 4. Again, dichotomous
variables captured the effects of the treatments. The index for support for the
war on drugs was also included in the model, as were its interactions with the
two treatments. The pretest measure of the dependent variable completed the
model:

Post-test Attitude toward Mexico/Colombia = Support for war on drugs

+ Exposure to stories about domestic war on drugs + Support for war on
drugs
Exposure to stories about domestic war on drugs + Exposure to
stories linking Mexico and Columbia to war on drugs + Support for war on
drugs
Exposure to stories linking Mexico and Columbia to war on drugs
+ Pretest Attitude toward Mexico/Colombia + Constant + e

The results of the model appear in Table 4.

The first thing that stands out is that the stories about the domestic war on
drugs failed to produce discernible priming effects in either direction for either
target country. In each case, the interaction between exposure to these stories
and support for the war on drugs fell well short of statistical significance. Thus,
priming effects that followed indirect paths failed to emerge in any of the four
tests in this study.
504 GAZETTE VOL. 65 NO. 6

TABLE 4
The Impact of Drug Stories and Attitudes about Drugs on Attitudes toward Mexico
and Colombia

Independent variable Mexico Colombia

Support for war on drugs –.10 (.05) .001 (.06)

Stories about domestic war on drugs –.03 (.07) .04 (.06)
Support for war on drugs
stories about
domestic war on drugs .02 (.12) –.08 (.11)
Stories linking Mexico and Colombia to the war
on drugs –.14* (.05) –.09 (.06)
Support for war on drugs
stories
about Mexico and Colombia .21** (.15) .24** (.10)
Pretest attitude .61** (.06) .52** (.06)
Constant .28 (.04) .19 (.05)

R2 .39 .30
N 194 195

Notes: Table entries are unstandardized regression coefficients. Standard errors are in paren-
theses. Significance tests are based on two-tailed tests except where noted in the text.
** p < .01, * p < .05.

In contrast, the data support Hypothesis 4. Exposure to stories that framed

Mexico as an American ally in the drug war encouraged participants to draw a
positive association between the war on drugs and Mexico (p < .01, one-tailed
test). Exposure to stories portraying Colombia in the same way had a similar
effect on how participants formed evaluations of that nation (p < .01, one-tailed
test). Put simply, participants followed the connections provided by these news
frames.

Conclusion
The findings presented here show that the media can influence the standards
by which people evaluate foreign nations. At the same time, the data also
suggest that there are limits to this influence. When members of our audience
read stories that offered a direct link between an issue and a nation that carried
a specific evaluative implication, they tended to adopt this frame of reference
in their own thinking. In three out of four cases, frames that provided this sort
of link shaped how people formed judgments. On the other hand, none of our
four tests produced evidence that priming effects work along an indirect route.
When participants read about issues on the domestic front, they did not carry
their thoughts over to the international realm. Influence along one route suc-
ceeded, whereas influence along the other route failed.
Some caution is warranted in drawing these conclusions. The absence of
priming effects along the indirect path might be due in part to the statistical
power afforded by the sample size or the imperfections of the measures used to
BREWER ET AL.: PRIMING OR FRAMING 505

capture attitudes toward the target nations and attitudes about the issues at
hand. The participants in the experiment also read the treatments in the context
of an experiment; hence they may have behaved differently than they would
have in a more natural setting. Furthermore, participants in this experiment
were undergraduate students and not necessarily typical of American citizens,
let alone citizens of other nations. Still, the methods that failed to capture
priming effects along the indirect route succeeded in capturing influence
produced by more explicit story lines. Care was taken to craft realistic treat-
ments and to disguise the purpose of this study. And while students may differ
in how they think about foreign nations compared to the general public, this
study focuses on the psychological processes by which people form these evalu-
ations – which both students and the general public should share, regardless of
nationality.
What accounts, then, for the failure of priming along the indirect path? One
partial explanation could lie in the power of stories to activate conflicting associ-
ations between issues and nations. It may be that audience members drew both
positive and negative associations between the war on drugs and the two target
nations for this issue. Such ambiguity could have interfered with the emergence
of a clear, consistent priming effect for the domestic drug use stories. On the
other hand, it seems less likely that the domestic terrorism stories would prime
both positive and negative associations between terrorism and the target nations
for that issue. Thus, this explanation, by itself, is probably insufficient.
A second explanation fits both sets of cases. It may be that audience
members simply did not associate terrorism with Iran and Libya or illegal drugs
with Mexico and Colombia prior to reading the stories about domestic terror-
ism and drug use. Thus, they may not have seen their thoughts about these
issues as applicable to the target nations. This explanation would fit well with
recent accounts (Miller and Krosnick, 1996; Price and Tewksbury, 1997) that
emphasize the importance of applicability or relevance in mediating priming
effects. It would also suggest that priming along the indirect route is generally
unlikely to emerge in the realm of judging foreign nations. The examples in this
study were deliberately chosen to reflect linkages between issues and nations
drawn within a real-world mass media discourse. If even these associations were
too weak within the minds of citizens to allow for spreading activation, then it
seems doubtful that priming effects produced by stories about domestic affairs
would exert a pervasive influence on mass judgments of foreign nations.
An alternative (or additional) explanation is that the study produced no
evidence of spreading activation because this is not the mechanism that
produces priming effects in the first place. The failure to find priming along the
indirect route is also consistent with the argument made by Miller and Krosnick
(2000) that priming effects result from learning about what trusted media
sources see as important, rather than from the automatic machinery of cogni-
tive accessibility. Under this explanation, one would not necessarily expect the
media to influence how people think about foreign countries unless they provide
stories that explain why one should associate particular issues with particular
nations.
When the media provide such stories, audience members do respond to
506 GAZETTE VOL. 65 NO. 6

them. Reading stories that linked Iran to terrorism led participants to judge Iran
in terms of their attitudes about fighting terrorism; similarly, reading stories that
linked Mexico and Colombia to the war on drugs led participants to judge both
countries in terms of their attitudes about the war on drugs. Moreover, the
stories led participants to use their thoughts about the issues in a particular way.
In the case of Iran, the stories led them to see support for a war on terrorism as
incompatible with favorable attitudes toward the nation. Meanwhile, in the
cases of Mexico and Colombia, the stories led participants to see support for a
war on drugs as a reason for favoring the nations. Thus, a moral of this study
is that one needs to provide a clear story – a frame that explains what issue is
at stake and on which side of the issue the foreign nation stands – in order to
shape how a mass audience judges another country.
That lesson has important implications. One is that actors in international
affairs may not need to worry too much about the impact of domestic news on
how mass publics view other nations. The evidence presented in this study
suggests that media audiences do not necessarily connect domestic issues such
as illegal drugs to foreign nations unless the media explicitly draw that con-
nection. Thus, nations that fear the indirect consequences of ‘guilt by associ-
ation’ should rest more easily.
However, it may be in the interests of some nations to attend to – and
perhaps attempt to shape – the specific frames that the news media provide for
them. As our study has shown, news frames are readily adopted by media audi-
ences and can influence their views of foreign nations. While this can have
undesirable consequences for foreign nations, it also suggests that certain story
lines that tie popular (or unpopular) stands to a nation could be a useful tool
in public diplomacy (see Manheim, 1991, 1994).

Note
1. The content analysis of the New York Times was conducted using keyword searches of the
Lexis/Nexis database and included both full-length stories and news briefs. The search terms
used were ‘drugs’ with Colombia and Mexico, and ‘terrorism’ with Libya and Iran. A handful
of stories involving the taking of performance-enhancing drugs by athletes were ignored for
purposes of this analysis.

References
Albritton, R.B. and J.B. Manheim (1983) ‘News of Rhodesia: The Impact of a Public Relations
Campaign’, Journalism Quarterly 60: 622–8.
Albritton, R.B. and J.B. Manheim (1985) ‘Public Relations Efforts for the Third World: Images in
the News’, Journal of Communication 35: 43–59.
Anderson, J.R. and G.H. Bower (1973) Human Associative Memory. Washington, DC: Winston.
Bartels, L. (1995) ‘The American Public’s Defense Spending Preferences in the Post-Cold War Era’,
Public Opinion Quarterly 58: 479–508.
Cappella, J.A. and K.H. Jamieson (1997) Spiral of Cynicism. New York: Oxford University Press.
Collins, A.M. and E.F. Loftus (1975) ‘A Spreading Activation Theory of Semantic Processing’,
Psychological Review 82: 407–28.
Domke, D., D.V. Shah and D.B. Wackman (1998) ‘Media Priming Effects: Accessibility, Associ-
ation, and Activation’, International Journal of Public Opinion Research 10: 51–74.
Entman, R. (1993) ‘Framing: Toward Clarification of a Fractured Paradigm’, Journal of Com-
munication 43: 51–8.
BREWER ET AL.: PRIMING OR FRAMING 507

Fiske, S.T. and S.E. Taylor (1984) Social Cognition. Reading, MA: Addison-Wesley.
Gamson, W.A. (1992) Talking Politics. New York: Cambridge University Press.
Herr, P.M., S.J. Sherman and R.H. Fazio (1983) ‘On the Consequences of Priming: Assimilation
and Contrast Effects’, Journal of Experimental Social Psychology 19: 323–40.
Higgins, E.T. and G.A. King (1981) ‘Accessibility of Social Constructs: Information-Processing
Consequences of Individual and Contextual Variability’, in N. Cantor and J. F. Kihlstrom (eds)
Personality, Cognition, and Social Interaction. Hillsdale, NJ: Lawrence Erlbaum.
Higgins, E.T., W.S. Rhodes and C.R. Jones (1977) ‘Category Accessibility and Impression
Formation’, Journal of Experimental Social Psychology 13: 141–54.
Higgins, E.T., J.A. Bargh and W. Lombardini (1985) ‘The Nature of Priming Effects on Categoriz-
ation’, Journal of Experimental Psychology: Learning, Memory and Cognition 11: 59–69.
Iyengar, S. (1991) Is Anyone Responsible? Chicago, IL: University of Chicago Press.
Iyengar, S. and D.R. Kinder (1987) News that Matters: Television and American Opinion. Chicago,
IL: University of Chicago Press.
Iyengar, S. and V. Ottati (1994) ‘Cognitive Perspective in Political Psychology’, pp. 143–87 in R.S.
Wyer and T.K. Srull (eds) Handbook of Social Cognition, Vol. 2. Hillsdale, NJ: Lawrence
Erlbaum.
Krosnick, J.A. and L.A. Brannon (1993) ‘The Impact of War on the Ingredients of Presidential
Evaluations: George Bush and the Gulf Conflict’, American Political Science Review 87: 963–75.
Krosnick, J.A. and D.R. Kinder (1990) ‘Altering the Foundations of Support for the President
through Priming’, American Political Science Review 84: 497–511.
Manheim, J. B. (1991) All of the People, All the Time: Strategic Communication and American
Politics. New York: M.E. Sharp.
Manheim, J. B. (1994) ‘Strategic Public Diplomacy: Managing Kuwait’s Image during the Gulf
Conflict’, pp. 131–48 in W.L. Bennett and D.L. Paletz (eds) Taken by Storm: The Media, Public
Opinion, and Foreign Policy in the Gulf War. Chicago, IL: University of Chicago Press.
Manheim, J.B. and R.B. Albritton (1984) ‘Changing National Images: International Public
Relations and Media Agenda Setting’, American Political Science Review 78: 641–57.
Miller, J.M. and J.A. Krosnick (1996) ‘News Media Impact on the Ingredients of Presidential Evalu-
ations: A Program of Research on the Priming Hypothesis’, pp. 79–100 in D.C. Mutz, P.M.
Sniderman and R.A. Brody (eds) Political Persuasion and Attitude Change. Ann Arbor: The Uni-
versity of Michigan Press.
Miller, J.M. and J.A. Krosnick (2000) ‘News Media Impact on the Ingredients of Presidential Evalu-
ations: Politically Knowledgeable Citizens Are Guided by a Trusted Source’, American Journal of
Political Science 44: 295–309.
Nelson, T.E., R.A. Clawson and Z.M. Oxley (1997) ‘Media Framing of a Civil Liberties Conflict
and its Effect on Tolerance’, American Political Science Review 91: 567–83.
Peffley, M. and J. Hurwitz (1992) ‘International Events and Foreign Policy Beliefs: Public Responses
to Changing Soviet–US Relations’, American Journal of Political Science 36: 431–61.
Perry, D.K. (1985) ‘The Mass Media and Inference About Other Nations’, Communication Research
12: 595–614.
Perry, D.K. (1987) ‘The Image Gap: How International News Affects Perceptions of Nations’, Jour-
nalism Quarterly 64: 416–21.
Price, V. and D. Tewksbury (1995) ‘News Values and Public Opinion: A Theoretical Account of
Media Priming and Framing’, paper presented at the annual conference of the International Com-
munication Association, Albuquerque, NM.
Price, V. and D. Tewksbury (1997) ‘News Values and Public Opinion: A Theoretical Account of
Media Priming and Framing’, pp. 173–212 in G. Barnett and F.J. Foster (eds) Progress in Com-
munication Sciences. Greenwich, CT: Ablex.
Srull, T.K. and R.S. Wyer (1980) ‘Category Accessibility and Social Perception: Some Implications
for the Study of Person Memory and Interpersonal Judgments’, Journal of Personality and Social
Psychology 38: 841–56.
Wyer, R.S. and T.K. Srull (1986) ‘Human Cognition in its Social Context’, Psychological Review
93: 322–59.
Wyer, R.S. and T.K. Srull (1989) Memory and Cognition in its Social Context. Hillsdale, NJ:
Lawrence Erlbaum.
508 GAZETTE VOL. 65 NO. 6

Paul R. Brewer is Assistant Professor of Journalism and Mass Communication,

University of Wisconsin-Milwaukee. His research on the media and public
opinion has been funded by the National Science Foundation and the Russell
Sage Foundation, and he has published articles in such journals as the Journal
of Politics, Public Opinion Quarterly, Political Communication, Political Psy-
chology and the Harvard International Journal of Press/Politics. He received
his PhD in political science from the University of North Carolina-Chapel Hill
in 1999.
Address Department of Journalism and Mass Communication, University of
Wisconsin-Milwaukee, 117 Johnston Hall, P.O. Box 413. Milwaukee, WI 53201,
USA. [email: prbrewer@uwm.edu]
Joseph Graf is currently a PhD candidate (ABD) at Stanford University and
an assistant research professor in the Institute for Politics, Democracy and the
Internet at George Washington University. He received his master’s degree in
journalism from Ohio University. [email: jgraf@gwu.edu]
Address School of Media and Public Affairs, George Washington University,
805 21st Street, NW, Washington, DC 20052, USA.
Lars Willnat is an associate professor in the School of Media and Public Affairs
at the George Washington University. His research on public opinion and inter-
national media effects has been published in various book chapters and in
journals such as Journalism and Mass Communication Quarterly, Political
Communication, International Journal of Public Opinion Research, Journalism,
Gazette and the Asian Journal of Communication. He received his PhD in mass
communication from Indiana University in 1992.
Address School of Media and Public Affairs, George Washington University,
805 21st Street, NW, Washington, DC 20052, USA. [email: lwillnat@gwu.edu]

View publication stats