You are on page 1of 26

The Harvard International Journal of

Press/Politics
http://hij.sagepub.com/

Compelling Arguments and Attitude Strength: Exploring the Impact of Second-Level Agenda Setting on
Public Opinion of Presidential Candidate Images
Spiro Kiousis
The Harvard International Journal of Press/Politics 2005 10: 3
DOI: 10.1177/1081180X05276095

The online version of this article can be found at:


http://hij.sagepub.com/content/10/2/3

Published by:

http://www.sagepublications.com

Additional services and information for The Harvard International Journal of Press/Politics can be found at:

Email Alerts: http://hij.sagepub.com/cgi/alerts

Subscriptions: http://hij.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://hij.sagepub.com/content/10/2/3.refs.html

>> Version of Record - May 23, 2005

What is This?

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
10.1177/1081180X05276095 Kiousis
Press/Politics
/ Compelling
10(2) Spring
Arguments
2005 and Attitude Strength

Compelling Arguments
and Attitude Strength
Exploring the Impact of Second-Level
Agenda Setting on Public Opinion of
Presidential Candidate Images
Spiro Kiousis

This study explores the relationship between attribute agenda setting and public
opinion of political candidates. Specifically, media salience of presidential candidate
attributes across five national elections is compared to public opinion data measuring
perceived candidate salience and the strength of public attitudes regarding candidates.
Findings suggest that media salience of attributes is strongly linked with strengthened
attitudes and is moderately linked with perceived candidate salience.The implications
of the findings are also discussed.

Keywords: agenda setting; attributes; compelling arguments; attitude strength

For more than thirty years, mass communication scholars have used agenda-
setting theory as a core conceptual framework for understanding the effects of
news on public opinion (e.g., Althaus and Tewksbury 2002; Hester and Gibson
2003; Iyengar and Kinder 1987; Ku et al.2003; McCombs, 2004; McCombs and
Shaw 1972; Roberts et al. 2002; Soroka 2003; Wanta 1997; Weaver et al. 1981;
Werder 2002). Though primarily concerned with the transfer of issue salience
from the mass media to the public, a small body of empirical work from this
scholarly perspective has shown that changes in media salience not only impact
perceived public salience but public attitudes as well (e.g., Entman 1989; Smith
1987; Soroka 2003). Such analyses have probed how media attention toward
“objects” (e.g., issues, events, candidate images, etc.) in the news is related to

Press/Politics 10(2):3-27
DOI: 10.1177/1081180X05276095
© 2005 by the President and the Fellows of Harvard College
3

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
4 Press/Politics 10(2) Spring 2005

public attitudes toward those same (or related) objects. Different aspects of pub-
lic attitudes have been highlighted, but a particularly pronounced influence has
been observed between media salience and attitude strength (Kiousis and
McCombs 2004; Weaver 1984, 1991).
In addition to exploring the impact of object salience, recent agenda-setting
work has scrutinized the influence of what has been called “attribute” salience.
For example, researchers have found that the attributes of political candidates
emphasized in news coverage become the attributes emphasized by voters when
describing those candidates. Missing from this research stream, though, are
studies that have concomitantly looked at the influence of attribute salience on
both perceived object salience and public attitudes toward objects,although such
relationships have been suggested (e.g., Ghanem 1997; McCombs and Estrada
1997; McCombs and Ghanem 2001; Yioutas and Segvic 2003). As a conse-
quence, the purpose of this exploratory investigation is to begin probing how
news media coverage of candidate attributes is linked to perceived candidate
salience and the strength of public attitudes regarding candidates. More specifi-
cally, an examination of news coverage and public opinion trends across five
national elections (1980-1996) is completed using presidential candidate
images.

Theoretical Background

Second-Level Agenda Setting and Object Salience


The emergence of second-level agenda-setting theory has shifted the focus of
research away from investigating what topics news media cover to how they cover
them. Conceptualized as attribute salience, this “second” level of agenda setting
deals with the properties, qualities, and characteristics in news content that
describe objects (McCombs and Evatt 1995; McCombs and Ghanem 2001). An
object, defined in much the same way that an attitude object is thought of in psy-
chology, is the element to which attributes refer, such as issues, political candi-
dates, public relations messages, and so on (Lopez-Escobar, Llamas, and
McCombs, 1998). While the types of attributes examined in research vary,
Ghanem (1997) has identified four major classes that are germane to many
object types: subtopics, framing mechanisms, affective elements, and cognitive
elements.
Considerable debate exists in the scholarly literature regarding the relation-
ship between attribute agenda setting and framing (de Vreese 2003; Hester and
Gibson 2003; Kosicki 1993; Maher 2001; McCombs and Ghanem 2001;
Scheufele 1999). While some scholars have argued for their separation (e.g.,
Scheufele 2000), others have called for their convergence (e.g., Takeshita 1997;
Yioutas and Segvic 2003). In addressing some of these concerns, McCombs et al.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 5

(2000) make the distinction between frames as “aspects” of coverage and frames
as “central themes”of coverage.Using the metaphor of a picture,they argue that

in the case of the central theme, our concern is with the central focus of the pic-
ture. In cases of aspects, the frame distinguishes between what the picture
includes and what is outside,a use of the term very similar to the idea of framing in
photography. (P. 79)

By using facets of candidate images, the current application of attributes falls


more in line with the aspects definition.
Several empirical inquiries have offered robust support for the theorizing
behind second-level agenda setting (e.g., Benton and Frazier 1976; Blood and
Phillips 1997; Kim et al. 2002; King 1997; Lopez-Escobar, Llamas, McCombs,
and Lennon 1998; McCombs et al.1997; Shaw et al.1999; Takeshita and Mikami
1995). In a time-series analysis of the media, government, and voter agendas in
Canada, Soroka (2002), for instance, found that including issue-attribute indica-
tors for the topics of inflation, environment, and debt/deficit played a major
role in predicting agenda-setting effects. As a whole, such findings are central to
an understanding of how voters learn and think about candidates and issues in
the electoral process (McCombs and Reynolds 2002; Weaver 1996).
With respect to candidate images, prior research has submitted that some of
the more prominent attributes highlighted by media and voters include candi-
date qualifications and personality traits (Graber 1972; McCombs et al. 1997;
Nimmo and Savage 1976; Sigel 1964). Building on such scholarship, the present
study examines three general candidate attributes: intellectual ability (personal-
ity trait), moral quality (personality trait), and leadership ability (qualification).
According to Ghanem’s (1997) typology and prior empirical work (Golan and
Wanta 2001; McCombs et al. 2000), these attributes are best grouped under the
category of cognitive elements.
Although the evidence for the transfer of attribute salience from the news
media to the public is increasing, research has also posited that media attention
to attributes of objects can significantly shape public salience of those same objects
(e.g., Ghanem 1996, 1997; McCombs 1996). For example, when it comes to
political issues, news about unemployment (as a subtopic) should not just shape
the perceived importance of unemployment but also the perceived importance
of the more general issue of the economy. This has been called the “compelling
arguments” hypothesis, whereby media emphasis on attributes provides people
with cues to modify their perceived salience of objects (McCombs 2004;
McCombs and Ghanem 2001; Severin and Tankard 2001; Yioutas and Segvic
2003). Though not originally conceptualized as examining the compelling-
arguments hypothesis, past research has detected some empirical support for it
(e.g., Schoenbach and Semetko 1992; Williams et al. 1983).

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
6 Press/Politics 10(2) Spring 2005

In addition to issues, the compelling-arguments hypothesis also has applica-


tions for political candidates in that the agenda of attributes for candidates should
also affect the overall perceived salience of candidates (i.e., object salience), yet
little research has probed it using political candidate images. One exception is
Kiousis et al. (1999), who observed some marginal experimental support for it.
Based on the logic of compelling arguments,the following hypothesis is offered:

Hypothesis 1: Media salience of presidential candidate attributes will be positively


associated with the perceived public salience of those candidates, the proportion
of the public who recognize those candidates.

Second-Level Agenda Setting and Attitude Strength


Beyond the impact of attribute salience on object salience via the compelling-
arguments hypothesis, some second-level investigations have also reported link-
ages between media salience of attributes and respondents’ perceptions of affect
(Lopez-Escobar et al. 1999; McCombs et al. 2000). Accordingly, relationships
between attribute salience and public attitudes appear likely. McCombs et al.
(2000) observed a close connection between affective media portrayals (posi-
tive, neutral, and negative) of various candidate attributes (e.g., ideology/issue
positions, biographical details, perceived qualifications, integrity, and personal-
ity and image) and affective descriptions of those attributes by voters in the 1996
Spanish General Election. In contrast, Golan and Wanta (2001) found little asso-
ciation between affective attributes in news coverage and voter perceptions of
candidates.
Direct linkages between second-level agenda setting and public attitudes are
captured by Kim et al.’s (2002: 11–12) notion of attribute priming, which “deals
with the influence of mass media on the public’s evaluation of issues. Specifically,
attribute priming hypothesizes that certain issue attributes emphasized in the
media will become significant dimensions of issue evaluation among the public.”
Applying this concept to presidential images, Kiousis (2003) found evidence of
attribute priming when media salience of the Monica Lewinsky scandal was asso-
ciated positively with perceived job approval and negatively associated with per-
ceived favorability of President Bill Clinton.
Historically speaking, first-level agenda-setting investigations do show link-
ages between media salience and shifts in public attitudes (Leff et al. 1986; Smith
1987). A key distinction made in research is between attitude change and attitude
strength. In particular, prior examinations have demonstrated that shifts in atti-
tude formation and extremity may result from agenda setting (Weaver 1984).
According to Weaver’s (1991) study of the federal budget deficit issue, as per-
ceived salience of the issue rose, the strength of opinion concerning a possible

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 7

solution to the problem (cutting spending) also increased. Indeed, the relation-
ship was more robust for attitude strength than for attitude change.
Related research on priming also reveals linkages between media salience and
public attitudes (via evaluations of political leaders), though the emphasis there
has been between media attention toward issues and attitudes toward politicians
(e.g.,Iyengar and Kinder 1987; Soroka 2003). Some scholars have also identified
more of a “basic priming” effect where media salience is connected to the holding
of opinions about objects (see McCombs and Reynolds 2002 for discussion).
Kiousis and McCombs (2004) reported that increased media attention to politi-
cal leaders was not just positively related to perceived salience but also to
increased levels of opinion holding regarding those figures. The current research
expands on such scholarship by looking at how attribute salience is linked with the
holding of opinions regarding objects. In general, the relationship between
media salience and attitude strength seems logical using the following rationale:
since mass media (from an agenda-setting perspective) tend to stimulate more
thinking and learning about objects and attributes in people’s minds (Golan and
Wanta 2001; Wanta 1997), one might consequently expect that this increased
thinking would lead to strengthened attitudes.
Several studies suggest a linkage between media exposure and attitude
strength (Erber et al. 1995; Sapiro and Soss 1999; Zaller 1992). From a psycho-
logical viewpoint, increased thinking could function as a plausible mechanism
for explaining the proposed relationship between media salience and attitude
strength. Specifically, a plethora of empirical work supports the assertion that
any impetus (media, events, etc.) enhancing thinking should also prompt more
extreme and persistent attitudes (e.g., Elms 1966; Petty et al. 1995; Zaller
1992). Summarizing this view, Tesser et al. (1995: 75) conclude that “thought,
then, tends to make evaluations more extreme, more accessible and more
enduring.” Because media exposure is often viewed as an impetus for increasing
thought, a reasonable outcome to expect is strengthened attitudes.
Missing from earlier research on agenda setting and attitudes, though, is the
influence of attribute salience on attitude strength. Because media salience of
objects is linked with attitude strength regarding objects, I predict a similar pat-
tern for attributes based upon the conceptual framework outlined above. While
attitude strength has been monitored in numerous ways, it has frequently been
defined in terms of attitude extremity (Abelson 1995; Judd and Johnson 1981;
Tesser 1978). This is the approach adopted for this study. Based on prior empiri-
cal work probing the relationship between mass media and attitude strength
(e.g., Erber et al. 1995; Kiousis and McCombs 2004; Weaver 1991), I scrutinize
the holding of nonneutral attitudes or what has been called “attitude dispersion.”
The following hypothesis is offered to test the relationship between attribute
salience and attitude dispersion as a dimension of attitude strength:

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
8 Press/Politics 10(2) Spring 2005

Hypothesis 2: Media salience of presidential candidate attributes will be positively


associated with strengthened attitudes toward those candidates,the proportion of
the public who hold nonneutral attitudes about them.

Method

To explore the relationships between media salience of attributes, perceived


object salience, and attitude strength regarding objects, I completed an exten-
sive time-series investigation of media coverage and public opinion of presiden-
tial candidate images from 1980 to 1996 (specifically, 1980, 1984, 1988, 1992,
and 1996). To keep the study manageable, the analysis focused on the eventual
Democratic and Republican party nominees for each election year. The specific
candidates were Jimmy Carter, Ronald Reagan, Walter Mondale, George H. W.
Bush, Michael Dukakis, Bill Clinton, and Bob Dole, respectively. The primary
rationale for this selection was the availability of public opinion data from the
National Election Studies (NES) polls that could be used to measure public
salience and attitudes toward presidential candidates. In particular, the key mea-
sure of name recognition (for perceived candidate salience) was only available in
NES poll data sets beginning in 1980. A secondary reason for using these polls
was that the national nature of the data would help with the generalizibility of the
findings.

Media Content Data


Since public opinion during national elections was examined, national media
sources were also needed to make comparisons. The Lexis/Nexis Database,
which frequently has been employed in prior scholarship (e.g., Domke et al.
1999), was chosen because of its abundant volume of news content. A combina-
tion of national print media outlets was used for the study. The New York Times
and Washington Post are, by most descriptions, thought to be among the more
prominent major media outlets in the United States and were therefore
employed in this analysis to provide newspaper content (Dearing and Rogers
1996). Complementing newspapers, news magazine content (U.S.News & World
Report and Newsweek) was also integrated into the inquiry, resulting in a total of
two newspapers and two magazines to estimate news coverage trends. Conse-
quently, I should have obtained a good picture of print media content patterns
during the time-period analyzed and also enhanced reliability (Chaffee 1991).
An added benefit of using multiple news outlets was that it provided several rep-
lication opportunities for testing my hypotheses, thereby raising the confidence I
could have in the findings. That is, every hypothesis was probed four times on the
two major candidates for each of the three attributes, resulting in twenty-four
separate tests of the hypotheses.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 9

Whenever scrutinizing media and public opinion trends, pinpointing the


proper time span from which to locate relationships is paramount (Chaffee
1972). The time lags employed in agenda-setting research have varied widely,
ranging from a few days and weeks (e.g., Wanta and Roy 1995; Zucker 1978) up
to several months (e.g., Sohn 1978; Stone and McCombs 1981). To capture a
broad range of content, news articles were tracked from January through Octo-
ber of every year analyzed (the months prior to elections). All months were
summed to form the “media agenda.”
Before measuring attribute salience, media salience of “objects” was deter-
mined using the conventional agenda-setting indicator of news story frequency.
Individual stories were the unit of analysis for this project. Basically, keyword
searches identified the volume of media attention dedicated to presidential can-
didates (Democratic and Republican). The keywords were “president” and the
surname of the candidates running for president during that year.A subsample of
stories was double-checked by a trained human coder to determine that a reli-
able and valid sample was garnered from Lexis/Nexis. Agreement between the
human coder and computer was .83 (Holsti’s calculation).1
The next question to address was how to gauge attribute salience. As alluded
to in the literature review, candidate qualifications and personality traits are
among the most visible candidate attributes in empirical analyses. Hence, the
three broad candidate attributes used for this project were intellectual ability
(personality trait), moral quality (personality trait), and leadership ability (qual-
ification). Tracing attribute coverage was not as straightforward as the first-level
measures, though, because more precision was needed to zero in on these specific
aspects of candidate news stories. To begin the collection of these data, the sam-
ple base originated from stories collected in the “object” part of the content anal-
ysis. To find stories emphasizing particular attributes, lists of synonyms and ant-
onyms were developed for each attribute that would represent the keywords in
the main Lexis/Nexis search. The one exception was for leadership ability,
where antonyms did not easily emerge during the operationalization phase of the
study. Instead of antonyms, the supplementary list of keywords developed for
this attribute mainly focused on action verbs that would be associated with the
concept (e.g., command, propose, etc.).
It is important to note that both synonyms and antonyms were listed and tabu-
lated because I was looking at the total amount of coverage (i.e., salience), not
just positive or negative information about candidates. Thus, I wanted to com-
prehensively cover the entire range of meaning for the terms I was examining, an
approach that is common in second-level agenda-setting research (e.g.,
McCombs et al. 1997). A rule was established that the combination of attribute
keywords had to appear within fifteen words of the candidates’surnames (within
the actual body of the article text) so that the discussion of attributes in the story

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
10 Press/Politics 10(2) Spring 2005

related to the candidates, not other actors in stories. It was possible for multiple
attributes to be present in the same news story.
In summary, then, an attribute coverage search for Democratic candidate
intelligence in 1992 might appear something like this. First, the number of sto-
ries containing the keywords “president” and “Clinton” would be garnered.
Among these stories, the number of articles containing the keywords “Clinton”
and one of the listed keywords for intelligence (see below) within fifteen words
proximity of “Clinton” in the article would designate the presence of this
attribute in 1992.
The initial list of keywords was quite extensive but was narrowed down
through a series of pretests designed to ascertain which keywords generated rel-
evant stories and which did not. The pretests began by first compiling lists of
synonym and antonym keywords that were related to the attribute being scruti-
nized. The antonym and synonym lists were mainly drawn from conventional
thesauruses. Once the lists were generated, the different keywords were then
tried in Lexis/Nexis to assess their efficiency in pulling up relevant stories about
the election and the attribute they were expected to cover.The words that pulled
up relevant material were kept, and those that did not were dropped.
In the end, the final keyword lists were as follows. For moral quality, they
were “moral,” “ethical,” “compassion,” “righteous,” “integrity,” “immoral,”
“unethical,” “corrupt,” “dishonest,” and “liar.” For leadership quality, they were
“leadership,” “negotiate,” “confident,” “command,” “initiate,” “propose,” and
“commander.” Finally, for intellectual ability, they were “intelligent,” “smart,”
“clever,” “shrewd,” “mistake,” “dumb,” “unintelligent,” “bad judgment,” and “poor
judgment.” Similar to the first-level content analysis, a coder reliability check
was performed between a trained human coder and the computer to establish
the extent of agreement. The coder reliability figure for attribute content was
.92 (Holsti’s calculation). In total, 11,343 stories were content analyzed for this
investigation.

Public Opinion Measures


The 1980, 1984, 1988, 1992, and 1996 NES polls were the principal sources
of public opinion data for this study, the perceived salience and attitude dispersion
regarding presidential candidates.2 Ideally, I would have measured the salience of
these candidates using a direct indicator of importance or prominence, much
like the widely used “What’s the most important problem facing this country
today?” (MIP) question. Unfortunately, no such measure existed in the NES sur-
vey (or comparable surveys for candidate salience).3 As is the case in most sec-
ondary analyses, I could only use the data that were available. As an alternative, I
used an indirect indicator of salience that was available in the data. This indicator,
name recognition of candidates, frequently has been employed as a measure of
perceived salience in both academic and political campaign research (see Cover

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 11

and Brumberg 1982 for discussion). Logically, people must be able to recognize
candidates to consider them salient. In other words, name recognition is a neces-
sary condition for and can serve as an indirect measure of salience when other
options are unavailable.
The NES items measuring attitudes toward presidential candidates were 100-
point feeling thermometers. According to Mann and Wolfinger (1980: 622),
these NES thermometer questions also are valid indicators of name recognition
because “the respondents could indicate they did not recognize the name [of
political figures], that they recognized the name but could not rate the person, or
that they recognized and rated the person.” Thus, the proportion of survey
respondents who “did not recognize” the person about whom they were asked to
give an opinion was subtracted from 100 percent to create the salience measure
for each of the two major presidential candidates. Kim et al. (2002) applied a
similar measure to gauge attribute salience of issues. Attitude dispersion was
measured by the proportion of respondents who reported nonneutral attitudes
(i.e., anywhere on the scale except at the 50-point mark).4

Data Analysis Strategy


The research design allowed me to compare the media agenda of attribute
salience with public opinion concerning presidential candidates. Spearman’s rho
correlations operated as the chief statistical test that compared the media and
public agendas (e.g., McCombs and Bell 1996; McCombs and Shaw 1972). The
following example comparing media coverage of morality and attitude disper-
sion illustrates the process of analysis. Assume that the number of stories in a
newspaper about the morality of Republican candidates on a yearly basis was
twenty for 1980, thirty for 1984, forty for 1988, fifty for 1992, and sixty for
1996. That is, the rank-order for 1980 is fifth, 1984, fourth, and so on, down to
1996, which ranks first in the news coverage among these five election years. For
purposes of this illustration, also assume that the percentages of people who
expressed nonneutral attitudes about Republican candidates were 72 percent
for 1980, 74 percent for 1984, 76 percent for 1988, 78 percent for 1992, and 80
percent for 1996. The rank-order of these five public opinion percentages of
nonneutral attitudes is identical to the rank-ordering based on the news cover-
age. A Spearman’s rho rank-order correlation calculated from these two lists
would produce a coefficient of +1.00, evidence supporting Hypothesis 2 that
attitudes became more dispersed as media coverage of morality increased for
Republican candidates.
Due to its exploratory nature,the correlational approach implemented in this
study is appropriate for establishing the linkages among media salience of attrib-
utes, perceived object salience, and attitude strength. However, more accumu-
lated evidence would be needed across several studies to determine the extent of
causal relationships. The emergence of significant correlations in this project can

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
12 Press/Politics 10(2) Spring 2005

lay the foundation for future research looking at potential causal relationships,
but the absence of significant correlations would falsify the proposed theoretical
framework (McCombs et al. 1997).

Results

General Media Coverage Trends


Prior to investigating the hypotheses, I begin by first reporting the general
news coverage trends of candidate attributes to provide some backdrop for the
findings. To limit the redundancy among the news outlets, the New York Times and
Washington Post are employed to present visually the media content data; how-
ever, all outlets were utilized in the actual data analysis. Figures 1 and 2 display
some examples of the news patterns for moral quality.
News stories about candidates’ moral quality varied greatly across the two
political parties. For Democrats, media attention stayed fairly flat until shooting
up in 1996, possibly due to intense questions raised about Bill Clinton’s moral
and ethical character (though closer inspection would have to corroborate this
conjecture). Media coverage of Republican morality was low in 1980 and then
shot up in 1984, staying relatively constant until 1996, accompanied by a slight
lull in 1988. This generally high level of news concentration is not surprising
since moral issues and concerns have been a hallmark agenda item of the Repub-
lican Party for many years, especially with the punctuated “family values” theme
popularized during the Reagan administration. In any event, my concern here is
not the origin of these patterns but with their links to public opinion.
Media salience of intellectual ability for Democratic candidates takes the
form of a bimodal distribution with peaks early and late in the time period exam-
ined, accompanied by low points in the middle elections. Figures 3 and 4 depict
these patterns.
Underlying reasons for these trends do not readily arise. The high peaks in
1992 and 1996 are perhaps indicative of Bill Clinton’s extensive academic train-
ing. Further investigations could better resolve these observations. For Republi-
can candidates, the distribution shows an up-and-down pattern from year to
year, with the highest points occurring in 1984 and 1992 and the lowest point
occurring in 1980. Similar to Democratic news content, the primary forces
behind those shifts in the flow of news coverage are not easily conspicuous, yet
one can make some educated guesses. The high volume in 1984 possibly stems
from Walter Mondale’s ubiquitous questioning of Ronald Reagan’s knowledge
and intelligence during that campaign. The high positions in 1992 and 1996
could exemplify some kind of equilibrium with the comparably high volume of
Democratic candidate stories concerning that attribute within the same time
frame.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 13

200
180
160
140
120 New York Times
100
80 Washington
60 Post
40
20
0
1980 1984 1988 1992 1996
Figure 1
Newspaper Coverage of Moral Quality for Democratic Presidential Candidates

180
160
140
120
New York Times
100
80
Washington
60 Post
40
20
0
1980 1984 1988 1992 1996
Figure 2
Newspaper Coverage of Moral Quality for Republican Presidential Candidates

The distribution of media attention to leadership quality for Democratic can-


didates takes on a bimodal image. Figures 5 and 6 display the news coverage
trends.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
14 Press/Politics 10(2) Spring 2005

90
80
70
60
New York Times
50
40
Washington
30 Post
20
10
0
1980 1984 1988 1992 1996
Figure 3
Newspaper Coverage of Intellectual Ability for Democratic Presidential Candidates

140

120

100
80 New York Times

60 Washington
40 Post

20
0
1980 1984 1988 1992 1996
Figure 4
Newspaper Coverage of Intellectual Ability for Republican Presidential Candidates

Jimmy Carter’s handling of the Iran hostage crisis probably fueled the apex in
1980. The Republican distribution pattern is replete with peaks and valleys over
time. The zenith in 1984 was possibly shaped by the Reagan administration’s

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 15

700

600

500
400 New York Times

300 Washington
200 Post

100
0
1980 1984 1988 1992 1996
Figure 5
Newspaper Coverage of Leadership Quality for Democratic Presidential Candidates

emphasis on foreign policy during the intense cold war days. In 1992, the Bush
campaign’s emphasis on the president’s role in the Gulf War, coupled with his
prior foreign policy experience, may additionally explain this high point. The
low point in 1980 is puzzling, especially since the Iran hostage crisis and cold war
were among the hottest issues of the day.
To examine the coverage of attributes as a whole, Tables 1 and 2 display the
frequency of stories concerning the various attributes in the New York Times over
the time period analyzed for Democratic and Republican presidential
candidates.
The chi-square values indicate that there were significant differences in the
proportions of media content dedicated to the different attributes. In the next
section, I address whether the differences were important with respect to public
opinion of candidates.

Hypotheses
Simply stated, Hypothesis 1 predicted that increased media salience of candi-
date attributes would be linked with increased candidate salience (compelling-
arguments hypothesis). Hypothesis 2 predicted that increased media salience of
candidate attributes would be linked with increased attitude dispersion regard-
ing candidates (attitude strength hypothesis). Tables 3, 4, and 5 exhibit the data
for the attributes of leadership, intellectual ability, and moral quality,
respectively.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
16 Press/Politics 10(2) Spring 2005

900
800
700
600
New York Times
500
400
Washington
300 Post
200
100
0
1980 1984 1988 1992 1996
Figure 6
Newspaper Coverage of Leadership Quality for Republican Presidential Candidates

The higher significance level of p < .10 is included here because of the explor-
atory nature of this research and because of the effects of the relatively small n (5)
on the significance test for rank-order correlations.
The first hypothesis predicted a positive association between media salience
of candidate attributes and their perceived public salience. For leadership, there
was mixed support for the hypothesis. Four of eight comparisons did approach
or attain statistical significance. The median value for the correlations coeffi-
cients was +.70. For the attribute of intellectual ability, the hypothesis was not
supported. Only two of eight coefficients approached or attained statistical sig-
nificance, and the median value was +.61. The strongest evidence for Hypothe-
sis 1 was observed for the attribute of moral quality. Six of eight comparisons at
least approached significance, and the median correlation was a robust +.80.
Thus, media salience of morality appears to be tied to increased candidate
salience. In total, twelve of twenty-four comparisons supported the compelling-
arguments hypothesis, modest evidence for its theorizing.
The second hypothesis expected that media salience of candidate attributes
would be positively associated with strengthened attitudes toward candidates.
Compared to the compelling-arguments hypothesis, there was much stronger
support for this relationship. For the attribute of leadership, seven of eight com-
parisons approached or attained statistical significance. The median values for
these correlations was +.85 (the highest of all comparisons). For the attribute of
intellectual ability, the evidence was weak for Hypothesis 2. Only three of eight
correlations were significant, and the median value was +.57. Strong evidence

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 17
Table 1
Frequency of news coverage of Democratic candidate attributes in the New York Times from 1980
to 1996
Leadership Knowledge Moral Quality

Year n % n % n %

1980 438 77 69 12 64 11
1984 289 72 64 16 50 12
1988 138 64 26 12 53 24
1992 170 64 40 15 55 21
1996 455 66 58 8 177 26
Note: Total number of news stories = 2,146. c2 = 70.80, df = 8, p < .001.

Table 2
Frequency of news coverage of Republican candidate attributes in the New York Times from 1980
to 1996
Leadership Knowledge Moral Quality

Year n % n % n %

1980 195 74 35 13 32 12
1984 882 78 88 8 166 15
1988 213 59 71 20 75 21
1992 595 74 81 10 131 16
1996 176 57 47 15 88 28
Note: Total number of news stories = 2,875. c2 = 99.80, df = 8, p < .001.

was again exposed for the attribute of moral quality. Seven of eight comparisons
at least approached significance. The median correlation value was +.74. Over-
all, seventeen out of twenty-four comparisons supported the assertion that
increased media salience of attributes was associated with increased attitude dis-
persion. Thus, substantial evidence was obtained for the submitted relationship
between second-level agenda setting and attitude strength.

Discussion

Theoretical Implications
In general, media salience of attributes was modestly associated with per-
ceived candidate salience and was strongly associated with the development of
nonneutral attitudes. As such, it appears that simple media attention to a candidate
is not the only important factor in public opinion, but so is how that candidate is

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
18 Press/Politics 10(2) Spring 2005
Table 3
Spearman’s rho correlations between media coverage of the leadership attribute and public opin-
ion of Democratic and Republican presidential candidates
Media Outlet Democratic Democratic Republican Republican
Saliencea Dispersionb Saliencea Dispersionb

New York Times (n = 3,551) .87** .90** .50 .70*


Washington Post (n = 3,718) .87** .80* .80*** .70*
U.S. News & World Report (n = 363) .87** .89*** .31 .94***
Newsweek (n = 272) .60 .54 .31 .94***
Note: Total number of news stories = 7,904. Public opinion data are based on 1980 to 1996
National Election Studies questions asking respondents to rate presidential candidates on a 100-
point thermometer scale. The presidential candidates were Gerald Ford, Jimmy Carter, Ronald
Reagan, Walter Mondale, George H. W. Bush, Michael Dukakis, Bill Clinton, and Bob Dole.
a. Salience refers to the extent to which people recognize the political candidates they are being
questioned about. These data are converted from the amount of people who “don’t recognize”
the public figures they are being questioned about. Higher correlations indicate that as media
coverage rises, more people recognize the public figures.
b. Dispersion refers to the extent to which people develop nonneutral positions about political
figures. Specifically, higher correlations mean as media coverage rises, more people are moving
away from the 50-percent category of the thermometer scales.
*p < .10. **p < .05. ***p < .01.

portrayed in news coverage. That said, it is clear that the magnitude of those rela-
tionships is contingent upon the particular attributes being considered. Specifi-
cally, increased attention to moral quality was the attribute most strongly associ-
ated with both dimensions of public opinion regarding candidates, followed by
leadership ability and intellectual ability. One explanation for this pattern may
be that moral quality and leadership ability are attributes that intuitively have a
stronger valence component than intellectual ability. Notably, the distinction
between perceived qualifications and personality traits does not seem relevant
here. By examining several attributes in the same analysis, future research may
be able to arrive at more definitive conclusions than the current investigation.
Evidence for the compelling-arguments hypothesis was only strong for one of
the three attributes analyzed. This seems to suggest that media attention to cer-
tain attributes may resonate more with the public than others in terms of increas-
ing the salience of candidates. In attempting to explain these findings, previous
studies suggest that political context may serve as a critical factor. Compared to
the current findings, for instance, Golan and Wanta (2001) observed that moral
quality was not an important attribute, but their study was based on data col-
lected in the early primaries and focused on the transfer of attribute salience.
Perhaps the difference in the current study was that it probed the entire election
year and looked at linkages with object salience.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 19
Table 4
Spearman’s rho correlations between media coverage of the intellectual ability attribute and
public opinion of Democratic and Republican presidential candidates
Media Outlet Democratic Democratic Republican Republican
Saliencea Dispersionb Saliencea Dispersionb

New York Times (n = 579) .62 .50 .60 .60


Washington Post (n = 674) .98*** 1.00*** .70* .50
U.S. News & World Report (n = 80) .63 .75** .37 .54
Newsweek (n = 80) –.13 –.03 .41 .84**
Note: Total number of news stories = 1,413. Public opinion data are based on 1980 to 1996
National Election Studies questions asking respondents to rate presidential candidates on a 100-
point thermometer scale. The presidential candidates were Gerald Ford, Jimmy Carter, Ronald
Reagan, Walter Mondale, George H. W. Bush, Michael Dukakis, Bill Clinton, and Bob Dole.
a. Salience refers to the extent to which people recognize the political candidates they are being
questioned about. These data are converted from the amount of people who “don’t recognize”
the public figures they are being questioned about. Higher correlations indicate that as media
coverage rises, more people recognize the public figures.
b. Dispersion refers to the extent to which people develop nonneutral positions about political
figures. Specifically, higher correlations mean as media coverage rises, more people are moving
away from the 50-percent category of the thermometer scales.
*p < .10. **p < .05. ***p < .01.

Another explanation for the limited findings of compelling arguments lies in


the type of object analyzed. Prior investigations of this proposed agenda-setting
relationship have tracked the impact of issue attributes on issue salience (e.g.,
Ghanem 1996). As a result, the theorizing behind the compelling-arguments
hypothesis may be more appropriate for issues than political candidates. As
stated earlier, Kiousis et al. (1999) only found marginal support for it within the
context of political candidates. On the other hand, the hypothesis was supported
in half of the comparisons in this study. Inevitably, replication of the findings in
the present inquiry will be needed to shed more light on the application of
employing the compelling-arguments hypothesis with political candidate
images. Examinations of third-party candidates may be ideal for future research
because of the lower inherent public knowledge about them when compared to
traditional party candidates.
The consistent pattern observed for the connection between attribute
salience and attitude dispersion is paramount because it confers further evidence
of the linkages between agenda setting and attitude strength. In addition, it
extends the explication of second-level agenda setting by proposing that attitudi-
nal consequences may result from the process. This builds on literature examin-
ing attribute priming (Kim et al. 2002; Kiousis 2003). Indeed, recent research
has suggested that attitudinal consequences should be a priority in future

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
20 Press/Politics 10(2) Spring 2005
Table 5
Spearman’s rho correlations between media coverage of the moral quality attribute and public
opinion of Democratic and Republican presidential candidates
Democratic Democratic Republic Republican
Media Outlet Saliencea Dispersionb Saliencea Dispersionb

New York Times (n = 891) .87** .90** .80* .70*


Washington Post (n = 959) .98*** 1.00*** .80* .70*
U.S. News & World Report (n = 86) .95*** .93** –.09 .55
Newsweek (n = 90) .62 .71* .66* .77**
Note: Total number of news stories = 2,026. Public opinion data are based on 1980 to 1996
National Election Studies questions asking respondents to rate presidential candidates on a 100-
point thermometer scale. The presidential candidates were Gerald Ford, Jimmy Carter, Ronald
Reagan, Walter Mondale, George H. W. Bush, Michael Dukakis, Bill Clinton, and Bob Dole.
a. Salience refers to the extent to which people recognize the political candidates they are being
questioned about. These data are converted from the amount of people who “don’t recognize”
the public figures they are being questioned about. Higher correlations indicate that as media
coverage rises, more people recognize the public figures.
b. Dispersion refers to the extent to which people develop nonneutral positions about political
figures. Specifically, higher correlations mean as media coverage rises, more people are moving
away from the 50-percent category of the thermometer scales.
*p < .10. **p < .05. ***p < .01.

agenda-setting scholarship (e.g., McCombs and Reynolds 2002). As McCombs


and Estrada (1997: 247) assert, Bernard Cohen’s famous summary of the
media’s power in public opinion might be reformulated to state that “the media
may not only tell us what to think about, they may also tell us how and what to think
about it, and even what to do about it [italics added].” Based on the current data, it
appears this line of inquiry is fruitful, but the focus should probably be more on
attitude strength than attitude change. Several dimensions of public attitudes
that might be relevant include attitude extremity (Abelson 1995), attitude
importance (Boninger et al.1995), and attitude ambivalence (Bargh et al.1992).
Given the observed bivariate patterns, a natural question that emerges from
the data is what sequence of influence among variables best explains the relation-
ships? While I did not explore this question in this study, two plausible models
surface based on prior scholarship that are pertinent to future research (see
Kiousis and McCombs 2004 for discussion): (1) media salience of attributes
impacts public salience of objects, which impacts attitude strength regarding
objects; and (2) media salience of attributes leads to attitude strength regarding
objects, which leads to public salience of objects. The first model represents the
more conventional approach of media influence whereby a hierarchy of effects is
thought to exist among cognitions, attitudes, and behaviors (e.g., McGuire
1986, 1989; Perse 2001). In comparison, the second model is rooted in

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 21
Table 6
Heuristic framework for examining relationship between agenda setting and attitude strength
Level of Analysis

Theoretical Model Objects Attributes

Media Salience ® Public Salience ® Attitude Strength


Media Salience ® Attitude Strength ® Public Salience

literature showing that the traditional sequence of influence can shift substan-
tially (see Pfau and Parrott 1993 for discussion). For instance, scholars have
described situations in political communication settings where this transpires:

If a voter pulls open the curtain and votes in a race where candidates are unfamiliar
and no issues are paramount, the behavioral act precedes [italics added] cognition
and affect. This may be the case when voters make choices for minor candidates,
perhaps picking a candidate because the name sounds familiar or the ethnicity is
appealing. (Jeffres and Perloff 1997: 237)

As a consequence, future research should address the sequence of influence


question in more detail. Table 6 provides a heuristic framework for such
inquiries.
Beyond such conceptual concerns, the methodological ramifications of this
study are also critical. My strategy for looking at attitude strength as opposed to
attitude change allowed me to uncover relationships between the media and
public agendas that would have gone unnoticed had I taken the conventional
approach. The notion of attitude dispersion is quite encouraging as a more gen-
eral outcome of increased media salience. Ironically, it showed up with greater
frequency than public salience.
Replication and extension of the relationships examined here will surmount
the inevitable limitations of secondary analysis for this study. Although it is sel-
dom feasible to juxtapose data from national surveys with national media sam-
ples, analyses based on other news media and other settings for public opinion
will strengthen confidence in the results found here. The twenty-four replica-
tions of each hypothesis in this study are a strong opening gambit.

Practical Implications
The practical applications of this analysis are also noteworthy. From a public
relations standpoint, for instance, research in political communication measur-
ing the impact of news coverage on public opinion represents an indirect indica-
tor of the influence of public relations efforts on public opinion—since the
sources of many news stories originate from public relations practitioners in the

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
22 Press/Politics 10(2) Spring 2005

form of news releases, interviews, press conferences, and so forth (Aronoff


1976; Cameron et al. 1997; Cutlip et al. 1994). For strategic communication
purposes, such research could assist political public relations and advertising
professionals in designing messages for candidates. For example, this study
somewhat suggests that describing candidates in terms of their moral quality is
probably the best approach for making their images more visible in public opin-
ion; however, because the valence of coverage was not accounted for, this does
not mean that an emphasis on morality would lead to positive coverage. In addi-
tion, the recommended strategy based on such research would vary depending
upon the goal of the communicators. Thus, if the goal is to strengthen attitudes,
then focusing messages on leadership ability may also be an option.
This line of inquiry also is valuable to professional journalists. The national
media are often accused of uniformly swaying public opinion, but the associa-
tions with attitude dispersion show that media coverage of candidate attributes
probably did not push voters in one direction. Instead, it helped them develop
and structure their opinions toward candidates. This is positive from a tradi-
tional journalism standpoint because—while media may have played a role in
helping people arrive at their decisions (which is acceptable and even professed
by most conceptions of democracy)—they did not push them to the same decision.
The findings also suggest to journalists that the aspects of stories they empha-
size in the news paint different images in people’s minds about political candi-
dates. As such, journalists may want to carefully consider their decisions in how
political candidates are portrayed in news reports. If there are particular attrib-
utes that voters tend to gravitate toward, then highlighting those attributes may
help individuals become more engaged in the political process. This heightened
attention may then lead to increased civic participation. Thus, this kind of
research is germane for developing approaches and strategies that can better
connect media, voters, and policymakers, a view embraced by most theories of
healthy democracies.

Acknowledgments

The author would like to thank Dr. Robyn Goodman and Dr. Greg Selber for
their assistance on coding content for this project. He would also like to thank
Dr. Lynda Lee Kaid for her feedback on an earlier version of the article. Finally,
he would like to thank Xu Wu for his assistance.

Notes

1. Data from 1976 were included in the coder reliability check because this study was a part of
a larger project examining media coverage and public opinion of the president.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 23

2. Warren E. Miller and the National Election Studies. NATIONAL ELECTION STUDIES,
1980: PRE-/POST-ELECTION STUDY [dataset]. Ann Arbor, MI: University of Michigan,
Center for Political Studies [producer and distributor], 1999. Warren E. Miller and the
National Election Studies. NATIONAL ELECTION STUDIES, 1984: PRE-/POST-
ELECTION STUDY [dataset]. Ann Arbor, MI: University of Michigan, Center for Political
Studies [producer and distributor], 1999. Warren E. Miller and the National Election Stud-
ies. NATIONAL ELECTION STUDIES, 1988: PRE-/POST-ELECTION STUDY
[dataset].Ann Arbor,MI:University of Michigan,Center for Political Studies [producer and
distributor], 1999. Warren E. Miller, Donald R. Kinder, Steven J. Rosenstone, and the
National Election Studies. NATIONAL ELECTION STUDIES, 1992: PRE-/POST-
ELECTION STUDY [dataset]. Ann Arbor, MI: University of Michigan, Center for Political
Studies [producer and distributor], 1999. Steven J. Rosenstone, Donald R. Kinder, Warren
E. Miller, and the National Election Studies. NATIONAL ELECTION STUDIES, 1996:
PRE- AND POST-ELECTION STUDY [dataset]. 3rd release. Ann Arbor, MI: University of
Michigan,Center for Political Studies [producer and distributor],1998.These materials are
based on work supported by the National Science Foundation under Grant nos. SBR-
9707741, SBR-9317631, SES-9209410, SES-9009379, SES-8808361, SES-8341310, SES-
8207580, and SOC77-08885. Any opinions, findings and conclusions or recommendations
expressed in these materials are those of the author and do not necessarily reflect those of
the National Science Foundation.
3. Of course, public opinion data concerning the candidate attributes studied in this inquiry
are available in the National Election Studies survey, but the focus here was on public per-
ceptions of objects. Comparisons with the attribute indicators are reported elsewhere and
are available upon request.
4. For salience measures, the percentage of all responses to the questions was used. For atti-
tude dispersion, the percentage of those who expressed attitudes (i.e., on the 100-point
scale) was used.

References

Abelson, Robert P. 1995. “Attitude Extremity.” In Attitude Strength: Antecedents and Consequences,
ed. Richard E. Petty and Jon A. Krosnick. Mahwah, NJ: Lawrence Erlbaum.
Althaus, Scott I., and David Tewksbury. 2002. “Agenda Setting and the ‘New’ News: Patterns of
Issue Importance among Readers of the Paper and Online Versions of the New York Times.”
Communication Research 29(2):180–208.
Aronoff, Craig. 1976. “Predictors of Success in Placing News Releases in Newspapers.” Public
Relations Review 1(2):43–57.
Bargh, J. A., Shelly Chaiken, Rajen Govender, and Felicia Pratto. 1992. “The Generality of Auto-
matic Attitude Activation Effect.” Journal of Personality and Social Psychology 62:893–912.
Benton, Marc, and Jean P. Frazier. 1976. “The Agenda-Setting Function of the Mass Media at
Three Levels of Information-Holding.” Communication Research 3:261–74.
Blood, Deborah J., and Peter C. B. Phillips. 1997. “Economic Headline News on the Agenda:
New Approaches to Understanding Cause and Effects.” In Communication and Democracy, ed.
Maxwell McCombs, Donald Shaw, and David Weaver. Mahwah, NJ: Lawrence Erlbaum.
Boninger, David S., Jon A. Krosnick, Matthew K. Berent, and Leandre R. Fabrigar. 1995. “The
Causes and Consequences of Attitude Importance.” In Attitude Strength:Antecedents and Conse-
quences, ed. Richard E. Petty and Jon A. Krosnick. Hillsdale, NJ: Lawrence Erlbaum.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
24 Press/Politics 10(2) Spring 2005

Cameron, Glen T., Lynne M. Sallot, and Patricia A. Curtin. 1997. “Public Relations and the Pro-
duction of News: A Critical Review and Theoretical Framework.” Communication Yearbook
20:115–55.
Chaffee, Steven H. 1972. “Longitudinal Designs for Communication Research: Cross-Lagged
Correlations.” Presented at the Association for Education in Journalism annual conference,
Carbondale, IL, August.
Chaffee, Steven H. 1991. Communication Concepts 1: Explication. London: Sage.
Cover, Albert D., and Bruce S. Brumberg. 1982. “Baby Books and Ballots: The Impact of Con-
gressional Mail on Constituent Opinion.” American Political Science Review 76(2):347–59.
Cutlip, Scott M., Allen H. Center, and Glen M. Broom. 1994. Effective Public Relations. 7th ed.
Englewood Cliffs, NJ: Prentice Hall.
Dearing, James W., and Everett Rogers. 1996. Communication Concepts 6:Agenda-Setting. London:
Sage.
de Vreese, Claes H. 2003. Framing Europe: Television news and European Integration. Amsterdam:
Askant Academic Publishers.
Domke, David, Mark Watts, Dhavan V. Shah, and David P. Fan. 1999. “The Politics of Conserva-
tive Elites and the ‘Liberal Media.’ ” Journal of Communication 49(4):35–58.
Elms, Alan C. 1966. “Influence of Fantasy Ability on Attitude Change through Role-Playing.”
Journal of Personality and Social Psychology 4:36–43.
Entman, Robert. 1989. “How the Media Affect What People Think: An Information Processing
Approach.” Journal of Politics 51(2):347–70.
Erber, Maureen W., Sara D. Hodges, and Timothy D. Wilson. 1995. “Attitude Strength, Attitude
Stability, and the Effects of Analyzing Reasons.” In Attitude Strength: Antecedents and Conse-
quences, ed. Richard E. Petty and Jon A. Krosnick. Mahwah, NJ: Lawrence Erlbaum.
Ghanem, Salma. 1996. “Media Coverage of Crime and Public Opinion: An Exploration of the
Second Level of Agenda Setting.” Ph.D. diss., University of Texas, Austin.
Ghanem, Salma. 1997. “Filling in the Tapestry: The Second Level of Agenda-Setting.” In Commu-
nication and Democracy, ed. Maxwell McCombs, Donald L. Shaw, and David Weaver. Mahwah,
NJ: Lawrence Erlbaum.
Golan, Guy, and Wayne Wanta. 2001. “Second-Level Agenda Setting in the New Hampshire Pri-
mary: A Comparison of Coverage in Three Newspaper and Public Perceptions of Candi-
dates.” Journalism and Mass Communication Quarterly 78(2):247–59.
Graber, Doris. 1972. “Personal Qualities in Presidential Images: The Contribution of the Press.”
Midwest Journal of Political Science 16:46–76.
Hester, Joe B., and Rhonda Gibson. 2003. “The Economy and Second-Level Agenda Setting: A
Time-Series Analysis of Economic News and Public Opinion about the Economy.” Journalism
& Mass Communication Quarterly 80(1):73–90.
Iyengar, Shanto, and Donald Kinder. 1987. News That Matters. Chicago: University of Chicago
Press.
Jeffres, Leo W., and Richard M. Perloff. 1997. Mass Media Effects. 2nd ed. Prospect Heights, IL:
Waveland.
Judd, Charles M., and J. T. Johnson.1981.“Attitudes, Polarization, and Diagnosticity: Exploring
the Effects of Affect.” Journal of Personality and Social Psychology 41:25–36.
Kim, Sei-Hill, Dietram A. Scheufele, and James Shanahan. 2002. “Agenda-Setting, Priming,
Framing and Second-Levels in Local Politics.” Journalism & Mass Communication Quarterly
79(1):7–25.
King, Pu-tsung. 1997. “The Press, Candidate Images, and Voter Perceptions.” In Communication
and Democracy, ed. Maxwell McCombs, Donald L. Shaw, and David Weaver. Mahwah, NJ:
Lawrence Erlbaum.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 25

Kiousis, Spiro. 2003. “Job Approval and Favorability: The Impact of Media Attention to the
Monica Lewinsky Scandal on Public Opinion of President Bill Clinton.” Mass Communication &
Society 6(4):435–51.
Kiousis, Spiro, Philemon Bantimaroudis, and Hyun Ban. 1999. “Candidate Image Attributes:
Experiments on the Substantive Dimension of Second Level Agenda Setting.” Communication
Research 26(4):414–28.
Kiousis, Spiro, and Maxwell McCombs. 2004. “Agenda-Setting Effects and Attitude Strength:
Political Figures during the 1996 Presidential Election.”Communication Research 31(1):36–57.
Kosicki, Gerald. 1993. “Problems and Opportunities in Agenda-Setting Research: A 20-Year
Assessment.” Journal of Communication 43(2):100–128.
Ku, Gyotae, Lynda L. Kaid, and Michael Pfau. 2003. “The Impact of Web Site Campaigning on
Traditional News Media and Public Information Processing.” Journalism & Mass Communication
Quarterly 80(3):528–47.
Leff, Donna, David Protess, and Stephen C. Brooks. 1986. “Changing Public Attitudes and
Policymaking Agendas.” Public Opinion Quarterly 50:300–314.
Lopez-Escobar, E., J. P. Llamas, and M. E. McCombs. 1998. “Agenda Setting and Community
Consensus: First and Second Level Effects.” International Journal of Public Opinion Research
10(4):335–48.
Lopez-Escobar, Esteban, Juan Pablo Llamas, Maxwell McCombs, and Rey Lennon. 1998. “Two
Levels of Agenda Setting among Advertising and News in the 1995 Spanish Elections.” Politi-
cal Communication 15:225–38.
Lopez-Escobar, Esteban, Maxwell McCombs, A. Tolsa, M. Martin, and Juan Pablo Llamas. 1999.
“Measuring the Public Images of Political Leaders: A Methodological Contribution of
Agenda-setting Theory.” Presented at the World Association of Public Opinion Research
Regional Conference in Sidney, Australia, June.
Maher, T. Michael. 2001. “Framing: An Emerging Paradigm or Phase of Agenda Setting?” In
Framing Public Life: Perspectives on Media and Our Understanding of the World, ed. Stephen D.
Reese, Oscar H. Gandy Jr., and August E. Grant. Mahwah, NJ: Lawrence Erlbaum.
Mann, Thomas, and Raymond Wolfinger. 1980. “Candidates and Parties in Congressional Elec-
tions.” American Political Science Review 74:617–32.
McCombs, Maxwell. 1996. “The Pictures of Politics in Our Heads.” Keynote address presented
at the National Chengchi University, Taipei, Taiwan.
McCombs, Maxwell. 2004. Setting the Agenda: The Mass Media and Public Opinion. Malden, MA:
Polity.
McCombs, Maxwell, and Tamara Bell. 1996. “The Agenda-Setting Role of Mass Communica-
tion.” In An Integrated Approach to Communication Theory and Research, ed. Michael Salwen and
Donald Stacks. Mahwah, NJ: Lawrence Erlbaum.
McCombs, Maxwell, and George Estrada. 1997. “The News Media and the Pictures in Our
Heads.” In Do the Media Govern? ed. Shanto Iyengar and Richard Reeves. London: Sage.
McCombs, Maxwell, and Dixie Evatt. 1995. “Los temas y los aspectos: Explorando una nueva
dimension de la agenda setting” [Objects and Attributes: Exploring a New Dimension of
Agenda Setting]. Comunicacion y Sociedad 8(1):7–32.
McCombs, Maxwell, and Salma Ghanem. 2001.“The Convergence of Agenda Setting and Fram-
ing.” In Framing Public Life:Perspectives on Media and Our Understanding of the World, ed. Stephen
D. Reese, Oscar H. Gandy Jr., and August E. Grant. Mahwah, NJ: Lawrence Erlbaum.
McCombs, Maxwell, Juan Pablo Llamas, Esteban Lopez-Escobar, and Federico Rey. 1997. “Can-
didate Images in Spanish Elections: Second-Level Agenda-Setting Effects.” Journalism & Mass
Communication Quarterly 74(4):703–17.
McCombs, Maxwell, Esteban Lopez-Escobar, and Juan Pablo Llamas. 2000. “Setting the Agenda
of Attributes in the 1996 Spanish General Election.” Journal of Communication 50(2):77–92.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
26 Press/Politics 10(2) Spring 2005

McCombs, Maxwell, and Amy Reynolds. 2002. “News Influence on Our Pictures in the World.”
In Media Effects: Advances in Theory and Research, 2nd ed., ed. Jennings Bryant and Dolf
Zillmann. Hillsdale, NJ: Lawrence Erlbaum.
McCombs, Maxwell, and Donald Shaw. 1972. “The Agenda-Setting Function of Mass Media.”
Public Opinion Quarterly 36:176–87.
McGuire, William J. 1986. “The Myth of Massive Media Impact: Savagings and Salvagings.” In
Public Communication and Behavior, vol. 1, ed. George Comstock. Orlando, FL: Academic
Press.
McGuire, William J. 1989. “Theoretical Foundations of Campaigns.” In Public Communication
Campaigns, 2nd ed., ed. Ronald E. Rice and Charles K. Atkins. Newbury Park, CA: Sage.
Nimmo, Robert, and Robert L. Savage. 1976. Candidates and Their Images: Concepts, Methods, and
Findings. Santa Monica, CA: Goodyear.
Perse, Elizabeth M. 2001. Media Effects and Society. Mahwah, NJ: Lawrence Erlbaum.
Petty, Richard E., Curtis P. Haugtvedt, and Stephen M. Smith. 1995. “Elaboration as a Determi-
nant of Attitude Strength: Creating Attitudes That Are Persistent, Resistant, and Predictive
of Behavior.” In Attitude Strength:Antecedents and Consequences, ed. Richard E. Petty and Jon A.
Krosnick. Mahwah, NJ: Lawrence Erlbaum.
Pfau, Michael, and Roxanne Parrott. 1993. Persuasive Communication Campaigns. Boston: Allyn &
Bacon.
Roberts, Marilyn, Wayne Wanta, and Tzong-Horng Dzwo. 2002. “Agenda Setting and Issue
Salience Online.” Communication Research 29(4):452–66.
Sapiro, Virginia, and Joe Soss. 1999. “Spectacular Politics, Dramatic Interpretations: Multiple
Meanings in the Thomas/Hill Hearings.” Political Communication 16(3):285–314.
Scheufele, Dietram A. 1999. “Framing as a Theory of Media Effects.” Journal of Communication
49(1):103–22.
Scheufele, Dietram A. 2000.“Agenda-Setting, Priming, and Framing Revisited:Another Look at
Cognitive Effects of Political Communication.”Mass Communication & Society 3(2):297–316.
Schoenbach, Klaus, and Holli A. Semetko. 1992. “Agenda-Setting, Agenda Reinforcing or
Agenda-Deflating? A Study of the 1990 German National Election.” Journalism Quarterly
69(4):837–46.
Severin, Werner J., and James W. Tankard. 2001. Communication Theories: Origins, Methods, and
Uses in the Mass Media. 5th ed. San Francisco: Longman.
Shaw, Donald, Maxwell McCombs, David Weaver, and Bradley J. Hamm. 1999. “Individuals,
Groups, and Agenda Melding: A Theory of Social Dissonance.” International Journal of Public
Opinion Research 11(1):2–24.
Sigel, Robert S. 1964. “Effects of Partisanship on the Perception of Political Candidates.” Public
Opinion Quarterly 28:483–96.
Smith, Kim A. 1987. “Effects of Newspaper Coverage on Community Issue Concerns and Local
Government Evaluations.” Communication Research 14:379–95.
Sohn, Ardyth B. 1978. “A Longitudinal Analysis of Local Non-Political Agenda-Setting Effects.”
Journalism Quarterly 55:325–33.
Soroka, Stuart N. 2002. “Issue Attributes and Agenda-Setting by the Media, the Public, and
Policymakers in Canada.” International Journal of Public Opinion Research 14:264–85.
Soroka, Stuart N. 2003. Media, Public Opinion, and Foreign Policy. Harvard International Journal
of Press/Politics 8(1):27–48.
Stone, Gerald C., and Maxwell McCombs. 1981.“Tracing the Time Lag in Agenda-Setting.”Jour-
nalism Quarterly 58:151–55.
Takeshita, Toshio. 1997. “Exploring the Media’s Role in Defining Reality: From Issue Agenda-
Setting to Attribute Agenda-Setting.” In Communication and Democracy, ed. Maxwell
McCombs, Donald L. Shaw, and David Weaver. Mahwah, NJ: Lawrence Erlbaum.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014
Kiousis / Compelling Arguments and Attitude Strength 27

Takeshita, Toshio, and Shunji Mikami. 1995. “How Did Mass Media Influence the Voters’ Choice
in the 1993 General Election Campaign in Japan? A Study in Agenda-Setting.” Keio Communi-
cation Review 17:27–41.
Tesser, Abraham. 1978. “Self-Generated Attitude Change.” In Advances in Experimental Social Psy-
chology, vol. 11, ed. Leonard Berkowitz. San Diego, CA: Academic Press.
Tesser, Abraham, Leonard Martin, and Marilyn Mendolia. 1995. “The Impact of Thought on
Attitude Extremity and Attitude-behavior Consistency.” In Attitude Strength: Antecedents and
Consequences, ed. Richard E. Petty and Jon A. Krosnick. Mahwah, NJ: Lawrence Erlbaum.
Wanta, Wayne. 1997. The Public and the National Agenda. Mahwah, NJ: Lawrence Erlbaum.
Wanta, Wayne, and M. J. Roy. 1995. “Memory Decay and the Agenda-Setting Effect: An Exami-
nation of Three News Media.” Presented at the Association for Education in Journalism and
Mass Communication, Washington, D.C.
Weaver, David. 1984. “Media Agenda-Setting and Public Opinion: Is There a Link?” In Communi-
cation Yearbook 8, ed. Robert N. Bostrom. Beverly Hills, CA: Sage.
Weaver, David. 1991. “Issue Salience and Public Opinion: Are There Consequences of Agenda-
Setting?” International Journal of Public Opinion Research 3(1):53–68.
Weaver, David.1996.“What Voters Learn from the Media.”Annals of the American Academy of Polit-
ical and Social Science 546:34–47.
Weaver, David, Doris A. Graber, Maxwell McCombs, and Chaim Eyal. 1981. Media Agenda-Setting
in a Presidential Election. New York: Praeger.
Werder, Olaf. 2002. “Debating the Euro: Media Agenda-Setting in a Cross-National Environ-
ment.” Gazette 64(3):219–34.
Williams, Wenmouth, Jr., Mitchell Shapiro, and Craig Cutbirth. 1983. “The Impact of Cam-
paign Agendas in Perceptions of Issues in the 1980 Campaign.” Journalism Quarterly 60:226–
31.
Yioutas, Julie, and Ivana Segvic. 2003. “Revisiting the Clinton/Lewinsky Scandal: The Conver-
gence of Agenda Setting and Framing.” Journalism & Mass Communication Quarterly 80(3):567–
82.
Zaller, John. 1992. The Nature and Origins of Mass Opinion. Cambridge: Cambridge University
Press.
Zucker, Harold G. 1978. “The Variable Nature of News Media Influence.” In Communication Year-
book 2, ed. Brent D. Ruben. New Brunswick, NJ: Transaction Books.

Biographical Note

Spiro Kiousis (Ph.D., University of Texas at Austin) is an assistant professor in the Department of
Public Relations at the University of Florida.
Address: Department of Public Relations, 2028 Weimer Hall, College of Journalism and Com-
munications, University of Florida, P.O. Box 118400, Gainesville, FL 32611-8400; e-mail:
skiousis@jou.ufl.edu.

Downloaded from hij.sagepub.com at National Dong Hwa University on April 12, 2014

You might also like