|
Looking Beyond Job Approval: How Media Coverage of the Monica Lewinsky Scandal
Influenced Public Opinion of the Presidency
By
Spiro Kiousis
Doctoral Student
Department of Journalism
University of Texas
Austin, TX 78712
[log in to unmask]
(512) 719-5572
Submitted to the AEJMC annual conference in New Orleans, LA August 4-7, 1999
* The author would like to thank Dr. Maxwell McCombs & Dr. Bruce Buchanan for
their insightful comments on earlier drafts of this paper.
Looking Beyond Job Approval: How Media Coverage of the Monica Lewinsky Scandal
Influenced Public Opinion of the Presidency
Abstract
Last year's Executive scandal involving Monica Lewinsky perplexed many media
experts because a story of such magnitude would normally be expected to heavily
sway public opinion of the presidency, yet most media accounts described minimal
fluctuations. Anchored in agenda setting, priming, and the elaboration
likelihood model of attitude change, the purpose of this paper was to, over
time, trace and compare media coverage of the Monica Lewinsky scandal to public
opinion of the presidency. Findings suggest that media coverage of the scandal
preceded changes in public opinion about the presidency (job approval and
perceived favorability). The data also show that newspapers and television news
wield similar influence on perceived favorability, but newspapers are more
related to swings in job approval, albeit weakly. Additionally, the data
provide some evidence for second-level agenda setting, priming, and attitude
change. Finally, media influence on perceived favorability is shown to be
robust when controlling for external factors (time & economic indicators) but is
not robust with respect to job approval.
Submitted to the AEJMC annual conference in New Orleans, LA August 4-7, 1999
* The author would like to thank Dr. Maxwell McCombs & Dr. Bruce Buchanan for
their insightful comments on earlier drafts of this paper.
Looking Beyond Job Approval: How Media Coverage of the Monica Lewinsky Scandal
Influenced Public Opinion of the Presidency
Looking Beyond Job Approval: How Media Coverage of the Monica Lewinsky Scandal
Influenced Public Opinion of the Presidency
Background
Introduction
Last year's Executive scandal involving Monica Lewinsky perplexed many media
experts because a story of such magnitude would normally be expected to heavily
sway public opinion of the presidency, yet most media accounts described minimal
fluctuations (e.g., Mitchell & Elder, 1998). Surely, mass communication and
political science scholars must be equally bewildered, given the influence mass
media is acknowledged to exert on perceptions of politicians (e.g., Roberts,
1997). While much conjecture and speculation has surrounded the possible impact
of this story on public opinion, a systematic, empirical analysis has yet to be
conducted to assess its influence. Accordingly, the purpose of this paper is to
fill this lacuna in research by, over time, tracing and comparing media coverage
of the Monica Lewinsky scandal to public opinion of the presidency. Unlike many
press pundits and White-House spin doctors who have only searched for immediate
negative effects on presidential job approval, we suspect some media influence
to exist but can probably only be detected by appropriating a longitudinal
research design such as the one employed for this project.
Presidential Assessment and Public Opinion
Public evaluations of the presidency typically have been operationally defined
as the level of presidential job approval (e.g., Mueller, 1970). According to
some research in this area, the principal factors that alter presidential
popularity are new administrations, international events, the economy, and war
(Simonton, 1987). Indeed, Dick Morris's famous explanation for Bill Clinton's
1992 presidential election victory ("It's the Economy Stupid") is especially
telling in the context of presidential job approval. Notwithstanding the impact
of these forces, other factors are also central in shaping presidential job
approval, including time (e.g., Stimson, 1976)-especially the early-term
honeymoon--, audience demographics (Zukin & Carter, 1982), and party
identification (Edwards, 1983).
Of particular importance to public appraisals of the president is the impact of
mass media (Graber, 1982). For example, Brody & Page's (1975) news discrepancy
theory submits that Executive job approval is inextricably linked to the amount
of positive media coverage a president receives. Elsewhere, Brody (1991), in a
rigorous analysis of news and the presidency, reported that media coverage of
the White House significantly predicts job approval, even when controlling for
the early-term honeymoon, policy proposals, economic factors, international
events, and time. Finally, Edwards (1983) conveys that "studies of press
coverage of presidential elections have concluded that the public's information
on and criteria for evaluating candidates parallels what is presented in the
media" (p.165). Shanto Iyengar and his colleagues have formally developed this
broad notion into the theory of priming (e.g., Iyengar & Kinder, 1987), which
argues that the qualities and characteristics emphasized in media texts dictate
how politicians are judged by the public (to be discussed further below).
Ultimately, mass media do help cultivate impressions about politicians but
should not be misconstrued to be the predominant influence.
Conceptual & Methodological Question of Job Approval
While the lion's share of both political science and mass communication
literature assumes job approval to be the chief indicator of people's
assessments of the Executive Branch, there are a slew of conceptual and
methodological reasons for considering additional measures to evaluate the
president. First, public appraisals of government officials certainly encompass
more than mere job approval. Though perceived performance is critical,
perceptions of personality traits also contribute to overall images of
politicians (e.g., Popkin et al., 1976; Nimmo & Savage, 1976). For example,
McCombs et al. (1997a), in study of Spanish elections, found that newspapers and
voters used personality traits when describing political candidates more
frequently than either qualifications or issue positions. Some of the more
prominent personality traits to be affected by media coverage of government
officials are integrity, ideals, sincerity, and honesty (Kinder, 1986; Edwards,
1983). Hence, public evaluations of politicians are multifaceted and
operational definitions should therefore extend beyond simple job approval. In
fact, Zukin & Carter (1982) write that "other ways of assessing presidential
popularity may tap slightly different attitudes or may be more (or less)
sensitive to changes in public attitudes toward the president" (p.215). As a
result, favorability ratings and specific questions about personality traits,
such as honesty or judgment, may help us hone in on other aspects of
presidential evaluations-a strategy utilized in the current project.
Beyond such theoretical problems, researchers have also raised methodological
concerns regarding the job approval measure (Jones, 1994). Some of these
include the sensitivity of the instrument, the extent to which it is an
objective measure of popularity, its tendency to aggregate voters, and its heavy
reliance on the Gallup poll (Zukin & Carter, 1982). In addition, a
methodological obstacle not widely elaborated in the literature (but certainly
pertinent) is determining the proper time lag to measure effects from the
various factors, whether they be short-term (war), long-term (economic
conditions), or both (media). Pinpointing the proper time lag to locate
relationships is paramount because "there is always the danger that a causal
effect will 'dissipate' over time if the researcher waits too long to measure
it" (Chaffee, 1972, p.8).
Specific to press influence, agenda setting theory sheds some light on this
matter by suggesting that it takes approximately four to seven weeks for
people's perceptions to change in response to fluctuations in media coverage
(Winter & Eyal, 1981; Salwen, 1988), though this issue remains largely
unresolved. For example, Wanta (1997a), in a meta-analysis that focused on
differences in agenda setting effects among different media channels (print and
broadcast), encountered scholarship documenting effects with time-lags ranging
from a few days to six months. Meanwhile, Kepplinger et al. (1989) found that
time lags of three months produce significant changes in public opinion.
Consequently, this examination will integrate numerous lags to deal with the
time problem but only to the point where they do not greatly hinder proper
sample size.
At this juncture, we have confirmed the necessity of widening the
conceptualization of presidential evaluations beyond the domain of simple job
approval. Nevertheless, a more detailed consideration of the political
evaluation process is required. Primarily, theories that navigate the
relationship between mass media and public opinion of politicians are especially
helpful. Thus, priming, agenda setting, and attitude change theories are now
explained to converge on how media coverage of political scandals shapes
people's impressions of the president.
Priming
On a broad level, priming theory claims that media establish the standards that
people employ to judge public officials (e.g., Iyengar & Simon, 1993; Krosnick &
Kinder, 1990; Iyengar et al., 1982). As a consequence, priming theory is an
invaluable asset for deepening our understanding of how media influence may
operate during political scandals, such as the Monica Lewinsky affair. Miller &
Krosnick (1997), for instance, assert that
News media coverage of national policy issues_presumably affects evaluations of
the president's general performance because all these issues touch on domains of
presidential responsibility. Coverage of such issues may also sometimes affect
evaluations of presidential character, but probably not as much as it affects
presidential performance evaluations_News media coverage of national policy
issues probably has even less impact on evaluations of a president's integrity,
except when an issue is scandalous in nature (p.262, emphasis added).
This implicit distinction between personal assessments and performance
evaluations hints that media attention to scandalous political stories may have
divergent consequences on performance and personal ratings-a question pursued in
this examination.
In addition to simple attention to a topic, there are other factors that mediate
priming effects. Miller & Krosnick (1997) maintain that priming effects are
more pronounced for people with high knowledge of politics. Further, the
specificity of a story will govern the degree to which people employ the
guidelines communicated in that story to appraise politicians (Iyengar & Simon,
1993). Therefore, because the current presidential impeachment scandal has been
displayed in more concrete terms (about specific people and events) in the news,
we might expect some shift in public impressions of the presidency, although not
necessarily on job approval.
While priming is instrumental for outlining particular aspects of press
influence on people's judgments about public officials, it is somewhat limited
because much of the research has been experimental (i.e., artificial) and has
usually probed television news instead of other media outlets. The larger
theory of agenda setting bridges this gap by scrutinizing television and
newspaper reporting patterns. The affiliation between agenda setting and
priming is natural because "priming is really an extension of agenda-setting,
and addresses the impact of news coverage on the weight assigned to specific
issues in making political judgments" (Iyengar & Simon, 1993, p.250). Thus,
agenda setting should further enhance our ability to sketch out how the public
perceives politicians via mass media, especially with its recent expansion into
the "second-level" (McCombs & Estrada, 1997).
Agenda-Setting & Salience
In its original conception, agenda setting theory posited that media controlled
the priority of salient issues in public opinion (McCombs & Shaw, 1972).
Specifically, the inventory of topics stressed by mainstream news organizations
in their media texts corresponded to the hierarchy of "most important problems"
nominated by people in public opinion surveys (Funkhouser, 1973). Of course,
many factors mediate the size of agenda setting effects, such as education level
(Wanta, 1997b), political interest (Erbring et al., 1980), and voter need for
orientation (Weaver, 1977).[1] The need for orientation concept is particularly
insightful for clarifying how public opinion may not be adversely affected by
news about political scandals because people do not always have a pressing need
or desire to attend to the media frenzy. Although the early research in agenda
setting mainly concentrated on issues, candidate images have also been examined
(e.g., Weaver et al., 1981), particularly the president (Wanta, 1992; Wanta et
al, 1989)--a research stream relevant to the current project.
Contemporary historical reviews of agenda setting delineate the process on two
levels (McCombs et al., 1997b). At the first level, the degree of importance
placed on an "object" (e.g., an issue, candidate, product, or whatever) by the
media more or less influences the degree of prominence that "object" will
receive in the public arena. At the second level, the "attributes" of objects
emphasized in the media become central "attributes" in public conceptions of
those same "objects" (McCombs & Evatt, 1995). An attribute can be thought of as
a property, quality, or characteristic that describes an object (McCombs &
Ghanem, 1998). In addition to modifying attribute salience, press emphasis on
certain attributes is also believed to provide "compelling arguments" for
audiences to alter their public concern about objects themselves (Ghanem, 1996,
1997).
The evidence for the transfer of attribute salience from press to the public is
growing (e.g., Benton & Frazier, 1976; Takeshita & Mikami, 1993; King, 1997),
although it is not an unqualified media effect. Ordinarily, attributes have
been designated as either substantive (cognitively grounded) or affective
(emotionally grounded)--(e.g., Ghanem, 1997). The extent to which either type
of attribute shifts public opinion is still a matter of controversy, but some
literature suggests that affective characteristics may exert more influence
(e.g., McCombs et al., 1997a).
Agenda Setting & Attitude Change
The transfer of salience from the press to the public is the core mechanism
operating in the agenda setting process. Though salience remains important, an
implicit assumption in agenda setting (that has received more attention
recently) is that influencing an object's salience through attribute emphasis
can partially shape public attitudes about that object as well (e.g., McCombs, &
Shaw, 1993; McCombs et al., 1997). For example, McCombs & Estrada (1997) state
that the Bernard Cohen's classic summary of the agenda-setting function of the
press should be reformulated to say "the media may not only tell us what to
think about, they also may tell us how and what to think about it, and even what
to do about it" (p. 247; emphasis added). Of course, we must recognize that
attitude change by no means occurs for all people in all cases as has been shown
ad nauseum in research (e.g., Katz & Lazarsfeld, 1955; Hovland et al., 1953;
Lazarsfeld et al., 1948); however, it is still necessary to explore how the
process works if we are to further deepen the conceptualization of second-level
agenda setting.
One weakness with the current explication of second-level agenda setting is that
it does not supply a theoretical map indicating how media emphasis on attributes
might influence attitudes. For example, a vital question encompasses whether
press emphasis on attributes can impact attitudes directly or must it first be
mediated by affecting salience? To address this question, we must first scan
some of the attitude change research that has permeated the field since its
inception.
Some of the more current persuasion research illustrates that salience (a.k.a.
comprehension or learning) is integral to attitude change (e.g., Zaller, 1997)
in some scenarios, but in others, it is not (e.g., McGuire, 1985). In fact,
Petty & Priester (1994) argue that "the existing evidence shows that message
comprehension and learning can occur in the absence of attitude change, and that
a person's attitudes can change without learning the specific information in the
communication" (p. 97). This leads us to ask how can we explicitly chart out
the possible impact of second-level agenda setting on attitudes given these
apparently conflicting observations?
The attitude change perspective that seems best equipped to handle this question
is the Elaboration Likelihood Model (ELM) because it proposes that media
messages can transform attitudes with or without changing salience, depending on
the circumstances (Petty & Cacioppo, 1981; 1986). Specifically, two routes of
persuasion exist that shape attitudes: central and peripheral. With the central
route, people's attitudes are impacted by a high degree of cognitive activity,
thereby suggesting that salience precedes attitude change. Conversely, the
peripheral route of persuasion occurs when "simple cues in the persuasion
context influence attitudes" (Petty & Priester, 1994, p. 101). Personal
relevance has been identified as a driving force behind which path is likely to
occur in people's minds (Burnkrant & Unnava, 1989), although several other
variables are meaningful, including perceptions of credibility, liking, and
consensus (O'Keefe, 1990). If attitude change does occur at all, a central
route transpires when people perceive media messages as high in personal
relevance, while a peripheral route functions when media messages are deemed low
in personal relevance.
Returning to second-level agenda setting, the ELM model would infer that media
emphasis on attributes can modify salience and subsequently attitudes in some
cases, or can influence attitudes directly in others, depending on how
personally relevant people perceive the media messages to be. In the case of
the Monica Lewinsky scandal then, logic dictates that a peripheral route seems
more likely because, according to several media outlets, public opinion polls
showed that people did not deem the investigation highly relevant (e.g., Berke &
Elder, 1998; Grier & Thurman, 1998; Balz & Marcus, 1998). If this is true, we
might also dub this as a "compelling arguments" hypothesis (Ghanem, 1997)
because attribute salience in the media would be conjectured to be directly
connected to shifts in public attitudes about an object, similar to its
postulated link to object salience (Ghanem, 1996). Nevertheless, because we are
at an early stage in the evolution of research tying agenda setting to attitude
change, we must consider both central and peripheral routes plausible in this
investigation (to be discussed in the next section).
It is indeed ironic that agenda setting, which was initially conceived as a
research stream in response to the limited effects model regarding attitude
change (Lasorsa, 1997), is now coming full circle to address the question of
attitude change that sparked its origin. As a result, one of the key goals in
this study was to empirically test how the media's attribute agenda is connected
to public attitudes by examining how press emphasis on the Monica Lewinsky
scandal is linked to public evaluations of the presidency. In short, this study
explored traditional agenda-setting influence on salience but also transcended
its conventional boundaries into the sphere of attitudes.
Putting it All Together
In relation to the current project, the conceptual framework offered by merging
second-level agenda setting, priming, and the Elaboration Likelihood Model then
enables us to better grasp the dynamics of how media coverage of the presidency
might affect public opinion. This model is more expansive than other
theoretical maps because it submits that press attention to a story can
transform both overall opinions of the presidency on the first level (e.g., job
approval or favorability) and on perceptions about specific dimensions of the
presidency on the second level (e.g., personality traits). In order to depict
exactly how the model operates in the current study, Figure 1 is displayed
below.
- - - Figure 1 Here - - -
For the present study then, we might think of journalistic attention to the
Monica Lewinsky story as an attribute of the broader coverage of Bill Clinton
(the object). When probing the influence of this story on perceptions of
personality traits, we are testing the bottom horizontal path of influence.
Meanwhile, when probing the effects of this story on job approval, we are
testing the diagonal or "compelling arguments" path of influence for salience.
When exploring the direct path from attributes to favorability, we are looking
at a peripheral route of persuasion. Finally, when tracing the path of media
influence to favorability via job approval, we are looking at a central route of
persuasion. Ideally, all paths should be investigated, but time and data
constraints compelled us to only examine these four paths. Unfortunately,
methodological procedures cannot always accommodate conceptual explications.
Nonetheless, our approach should be sufficient in an exploratory study such as
this, yet we still wanted to acknowledge the study's limitations up-front.
For the purposes of this project, job approval was probably more of a salience
measure because it is purported to monitor performance (substantive) and is
certainly more of a cognitive measure since the salience of issues has been
linked to fluctuations in job approval (e.g., Brody, 1991). Alternatively,
favorability was probably more closely linked to attitudes because it seems to
represent more of a personal assessment of the president (affective). Even
though job approval is not a precise measure of "salience" because it entails
both emotional and cognitive perceptions, we argue that this substitution was
necessary because no standard salience question exists for politicians as it
does for issues-i.e., the Gallup "most important problem" questions (Dearing &
Rogers, 1996). On the other hand, favorability appears to be more clearly
associated to attitudes. As a result, a noteworthy contribution that this
analysis can make is to assess the value of using these measures in media
analyses of the presidency.
Agenda Setting Across Media Channels
A final question that agenda setting theory raises is how did different media
channels covering the Monica Lewinsky affair affect public opinion. In
particular, which medium (newspapers of television) wielded more influence? As
referred to earlier, priming theory's concentration on broadcast news suggests
that television might be stronger (e.g., Miller & Krosnick, 1997). In agenda
setting, the evidence is somewhat more mixed. For example, Wanta (1997a)
reasons that "no clear trend emerges from the comparison of the studies on the
magnitude of agenda-setting effects across media" (p.148). Still, the bulk of
agenda-setting work appears to verify stronger effects for newspapers (Benton &
Frazier, 1976; Schoenbach, 1982; Weaver, 1977). Indeed, Protess & McCombs
(1991) articulate that "most studies that compare different media find that
newspapers are more likely to produce issue salience" (p.98). As a result,
since most agenda-setting work has found such a trend for issue salience, we
suspect a similar pattern to surface for public opinion of the presidency as
measured by job approval and favorability.
By synthesizing the tenets of agenda setting, priming, and the Elaboration
Likelihood Model of attitude change, we tested how media coverage of the Monica
Lewinsky scandal was connected to fluctuations in public opinion of the
presidency.
Hypotheses & Research Questions
A series of hypotheses and research questions were formulated based on the
previous literature review. Anchored in the priming and agenda setting
literature that has depicted a hierarchical relationship between mass media
coverage of politics and public opinion, the following hypothesis is posed:
H1: Changes in public opinion about the president will be more strongly linked
to media coverage preceding public opinion polls than to simultaneous or
subsequent media coverage.
Based on the agenda-setting scholarship reporting a stronger relationship
between newspapers and public opinion than for television and public opinion
(e.g., Schoenbach, 1982), the following hypothesis is offered:
H2: Newspapers will be more strongly linked to public opinion shifts than
television coverage.
Based on our discussion of presidential evaluations, priming, attitude change,
and second-level agenda setting, the following research questions are submitted:
RQ1: Will media attention to the Monica Lewinsky scandal will be correlated with
shifts in public opinion about presidential attributes?
RQ2: Will changes in overall public opinion about the president be uniform or
divergent with regard to favorability and job approval?
RQ3: Will any observed attitude change best be characterized by a peripheral or
central route of persuasion?
RQ4: How strongly associated will media coverage be with public evaluations of
the president when integrating the impact of other factors?
Method
Sample
A longitudinal time-series design was employed to track the impact of media
coverage on public opinion. Media content and public opinion surveys were
monitored from Jan. 19-Nov 1., 1998. This beginning date was selected simply
because it was the first week that the Monica Lewinsky story appeared in elite
media. The Nov. 1 end date was more of a subjective decision but was made
because the coverage ensuing the 1998 elections held on Nov. 3 largely focused
on the upcoming impeachment hearings, which we believed would confound the
effects of the original story and therefore, damage the study's validity.
Furthermore, election day seemed like a natural point to complete such an
investigation. The time increment utilized for this project was one week,
simultaneously permitting precise measurement of time-lags and generating an
adequate sample of weeks (N=41).
Media Content
A content analysis of New York Times newspapers and ABC World News Tonight
broadcasts was performed to gauge media attention to the Monica Lewinsky
scandal. These two news organizations were chosen because they are prime
examples of elite, national media outlets (e.g., Zhu & Boroson, 1997; Dearing &
Rogers, 1996). The following search strings were used in the Lexis / Nexis
database to collect the total sample of stories. For the New York Times (n=282
stories), an example search string looked like this: "date is aft jan. 18 1998
and bef Nov. 2 1998 and section a or section 1 and page 1 and president or
clinton and lewinsky." A similar search string was used for ABC news (n=187
stories). Simple frequencies were employed to determine the volume of media
coverage per week (unit of analysis), a technique found to be fairly reliable by
other scholars (e.g., Stone & McCombs, 1981).
Before continuing, it should be noted that several steps were taken to
strengthen the study's validity and reliability. First, articles had to appear
on the front page of the New York Times to ensure they were prominently
displayed. Such a procedure was not utilized for ABC World News Tonight content
because the intrinsic time-constraint of television news predetermines that the
stories inserted in the daily broadcasts were the "most important" of the day.
Further, the number of stories from each news outlet becomes more balanced when
comparing content in this manner. Second, the use of general rather than
specific key words was intentional, allowing us to obtain an aggregate measure
of the story's impact. Finally, coder reliability was assessed on a subsample
of the stories by having an independent human coder validate the sample
collected by Lexis / Nexis. Holsti's intercoder reliability figures for the
subsample was 88 percent.[2] While not perfect, we believe the sample of
stories collected from both news organizations functioned as a fairly
representative portrayal of how mass media reported the scandal.
Public Opinion
To measure perceptions about the presidency, composite indices of several
public opinion polls were created. This was an integral step for securing an
appropriate sample of surveys. As touched on in the theoretical section, second
level agenda setting served as the core conceptual framework for this study.
Thus, object salience had to be operationally defined (first-level). Object
salience was gauged by presidential job approval as scored by ABC
News/Washington Post, CNN/Time, FOX, and Gallup polls (alpha=.96). These
numbers represent the percentage of people who "approve" of the job the
president is doing in running the country.[3] Beyond object salience, public
attitudes towards to presidency also had to be operationally defined. In this
instance, public attitudes were calibrated by presidential favorability ratings
as scored by ABC News/Washington Post, CBS/New York Times, CNN/Time, and Gallup
polls (alpha=.91). These numbers stand for the percentage of people who have a
"favorable" opinion of the president in general.[4]
In addition to first-level measures, it was imperative that we operationalize
attribute salience (second-level) as well. Gallup poll questions about the
president's personal characteristics acted as the attribute salience variables
because other data that might be conceptualized in this way were scant. Due to
the scarcity of this data (n=3 to n=5), no distinction was made between
substantive and affective attributes, yet future research should probably
integrate this dichotomy into similar analyses. Moreover, the problems with
this data forced us to only examine one of the peripheral and central paths of
persuasion. Nevertheless, confirmation or rejection of those paths will shed
some light on the matter of attitude change in second-level agenda setting. The
particular qualities measured for the study were honesty, competence, and
judgment.[5]
Time Lags
As outlined earlier, it was essential that we test for relationships at
multiple time-lags to locate any potential linkages between press attention and
public conceptions of the president. Though the range of optimal time-lags has
varied in research (e.g., Winter & Eyal, 1981; Jablonski & Hester, 1998), a
six-week window was adopted for this study for two key reasons: (1) most
scholarship detects at least some relationship within this amount of time (e.g.,
Salwen, 1988; Zucker, 1978; Wanta & Hu, 1994); and (2) venturing beyond this
boundary impairs sample size to the point where statistical reliability could
become a problem. While we surmised that media influence would precede changes
in public opinion, there was always the possibility that they would move
synchronously or that public opinion swings would precede shifts in media
coverage. As a result, a six-week time-lag was utilized in two directions to
ensure that all three types of relationships could be observed.
Data Analysis
Pearson cross-lagged correlations were calculated to ascertain whether a
linkage existed between media coverage of the scandal and public opinion of the
presidency, except when sample size was small in which case Spearmann rho
correlations were used (i.e., for attribute salience). McCall (1994) recommends
such nonparametric tests when dealing with smaller sample sizes. When suitable,
multivariate techniques (partial correlations) were employed to analyze a number
of variables concomitantly.
The study's research design allowed us to efficiently test the hypotheses and
explore the research questions. To test Hypothesis 1, for example, the number
and size of correlations revealed whether media coverage moved public opinion,
they moved together, or public opinion led media coverage. Likewise, for
Hypothesis 2, the number and size of correlations also exposed which news
channel was more closely related to fluctuations in public opinion.
Correlations between media coverage and Gallup poll data on presidential
characteristics addressed Research Question 1. Research Question 2 was
investigated by scrutinizing the general direction of first-level correlations.
For Research Question 3, we examined the direct relationship between media
coverage and attitudes to test the peripheral path of persuasion. Subsequently,
we tested whether the influence of media coverage on public salience mediated
the relationship between press coverage and perceived favorability (via partial
correlations), thereby exploring a central route of persuasion. In the words,
we checked if the relationship between media coverage and attitude change could
remain strong when the impact of media coverage on public salience was removed.
Lastly, partial correlations demonstrated how well press attention to the Monica
Lewinsky story was connected to public assessments of the president in the face
of other factors (time and economic indicators) for Research Question 4.[6]
Results
H1: Changes in public opinion about the president will be more strongly linked
to media coverage preceding public opinion polls than to simultaneous or
subsequent media coverage.
The findings support Hypothesis 1 for favorability and moderately support it
for job approval. Media coverage appears to precede shifts in public opinion
with more frequency than public opinion swings precede changes in press
attention (or fluctuate synchronously). Tables 1 and 2 are presented below
summarizing the data.
- - - Insert Table 1 Here - - -
- - - Insert Table 2 Here - - -
For presidential favorability, the ratio of significant (p<.05) cross-lagged
correlations with media coverage preceding public opinion changes to
correlations with media coverage changes following public opinion shifts is 2 to
0. For job approval, the evidence is more mixed because no correlations reach
statistical significance, although the relationship between media coverage three
weeks prior to public opinion polls does approach significance (r=.29; p<.10).
On the whole then, it seems that media coverage precedes changes in public
assessments of the presidency.
H2: Newspapers will be more strongly linked to public opinion shifts than
television coverage.
The data are mixed with regard to this hypothesis for favorability but appears
to be modestly supported for job approval. Referring back to Table 1, newspaper
and television coverage are both substantially connected to public opinion in
one instance each, thereby making their influence appear shared. When taking
into account the size and significance level of the cross-lagged correlations,
television is slightly stronger, but the difference is so minute that we
conclude their influence to be equal for favorability. However, more definitive
generalizations about the effects of different media channels cannot be drawn
until we look at the correlations for job approval.
A different image emerges from the job approval data. Returning the Table 2,
the only noteworthy correlation (p<.10) occurs for newspapers, hinting that this
was the more influential medium. Nevertheless, the lack of significance shows
the relationship is not all that strong. Generally then, it seems that
newspapers and television both play strong complementary roles with respect to
favorability, but newspaper coverage is marginally more related to job approval
shifts.
RQ1: Will media attention to the Monica Lewinsky scandal will be correlated with
shifts in public opinion about presidential attributes?
Before presenting these results, it should be remembered that the sample sizes
for the findings on presidential attributes were extremely small (between n=3 to
n=5) due to scarce data availability, making all those correlations highly
tentative. Therefore, we recommend that these findings only be used to suggest
whether a relationship exists or not, and certainly nothing should be
extrapolated about the direction of such linkages.
Similar to overall judgments about the president, the data imply that a
relationship operates between press attention to political scandals and people's
perceptions about a government official's personal characteristics. Unlike
general assessments of the presidency, the relationship seems to be more
reciprocal and shared. Tables 3, 4, and 5 reflect these trends.
- - - Insert Table 3 Here - - -
- - - Insert Table 4 Here - - -
- - - Insert Table 5 Here - - -
Collectively, it appears there are connections between media coverage and public
perceptions of presidential competence, honesty, and judgment. The mixture of
positive and negative correlations is confusing but could probably be explained
had we accounted for tone in the study's content analysis. Regardless, these
data lend some credence to second-level agenda setting and priming effects,
although additional research should be executed to confirm these associations.
RQ2: Will changes in overall public opinion about the president be uniform or
divergent with regard to favorability and job approval?
According to Tables 1 & 2, it seems that the shifts in public opinion of the
president move in divergent directions when responding to press coverage of the
Monica Lewinsky scandal. Specifically, favorability appears to be negatively
linked (all significant correlations are negative), while job approval seems to
be positively associated (the one meaningful correlations is positive). In
addition, the general direction of favorability correlations is negative, while
job approval correlations are positive. Hence, higher press scrutiny of the
scandal was related to negative perceptions of favorability but was associated
to positive opinions about job performance. The divergent trends demonstrate
that presidential evaluations are multidimensional-a point to be further
considered in the subsequent section.
RQ3: Will any observed attitude change best be characterized by a peripheral or
central route of persuasion?
Findings suggest that the observed public attitude change in the Monica
Lewinsky scandal is best characterized by a peripheral route of persuasion. To
answer this research question, we converged on the time-point when media
attention to the scandal was connected to a shift in public attitudes (perceived
favorability at time-lag 6). As stated for Hypothesis 1, zero-order
correlations between media coverage six weeks prior to changes in perceived
favorability are both significant (ABC News r=-.54, p<.01; NYT r=-.47, p<.05).
Thus, a peripheral (direct) path of attitude change appears plausible here, but
we still needed to look for intervening variables.
Two partial correlation coefficients (for television and newspapers), which
removed the impact of press coverage on job approval (i.e., public salience),
were computed to deal with this problem. If the correlations decreased, this
would imply that public salience mediates the relationship between media
coverage of attributes and shifts in public attitudes. Alternatively, if the
correlations remained stable or increased, this would verify a more direct
relationship.[7] Specifically, we controlled for the influence of media
coverage on job approval at lag 3 and 6: lag 3 because it was the only point
when media impact on salience approached statistical significance; and lag 6
because it represented the period that coincided with public attitude change.
Ideally, we would have liked to control for the influence of all prior media
coverage on public salience but could not because sample size became a concern
when integrating all lags.
In both cases, the correlations remained stable (ABC News r=-.59, p<.01; NYT
r=-.55, p<.05). Accordingly, a peripheral route of persuasion appears more
likely than a central route for explaining observed attitude change during the
Monica Lewinsky scandal.
RQ4: How strongly associated will media coverage be with public evaluations of
the president when integrating the impact of other factors?
Here, we converged on the time points when media coverage were most influential
(at time lag -6 for favorability and time lag 3 for job approval), allowing us
to comprehensively scrutinize this phenomenon. To review, time and economic
indicators (Consumer Price Index and Consumer Confidence) were employed as
control variables because of their known influence on presidential assessments
(Simonton, 1987; Stimson, 1976). Partial correlations (controlling for time,
Consumer Confidence, and the Consumer Price Index) report that the relationship
between media coverage and public opinion is robust for favorability (ABC News
r=-.42, p<.05; NYT r=-.40, p<.10) but weak for job approval (NYT r=-.03, n.s.).
Therefore, it seems that media coverage remains strongly associated to public
assessments of the presidency in terms of valence but not salience.
Discussion
Methodological Concerns
Prior to discussing the implications of this research, it would be helpful to
enumerate some of its methodological shortcomings. One caveat was that the
public opinion data were extracted from many different sources and may not be
comparable. Nevertheless, this step was necessary to obtain an adequate sample
of surveys. Further, the high alpha reliability scores (alpha>.90) demonstrated
that the surveys were essentially chronicling parallel trends and could be
aggregated legitimately. Another concern entailed the small sample of data for
monitoring public opinion about presidential attributes. Again, this was
inevitable given the dearth of this type of data. While the sample is small,
our tentative approach in drawing generalizations enabled this study to function
as pretest for future empirical work in this area and is valuable in this
capacity. A related concern involves our failure to directly compare all
central and peripheral routes of persuasion, but this initial look at attitudes
in second-level agenda setting, at least, hints at which route is more feasible.
Finally, the use of only front-page newspaper stories and national television
newscasts affords a constrained view of the wide ranging media content available
for the Monica Lewinsky scandal. On the other hand, this strategy supplied a
manageable sample of data and isolated salient news. Consequently, it was
appropriate for an exploratory study such as this.
Implications & Conclusions
There are many directions future researchers might embark upon in light of the
current findings, especially in the fields of mass communication and political
science. The fact that media coverage led shifts in public opinion reveals that
press attention to governmental scandals shapes public opinion of politicians.
Though many pundits dismissed such influence, the current project illustrates
that researchers must be willing to look beyond the short-term trends in public
opinion that media and Washington elites expect. Thus, this study confirms the
utility of time-series designs. While causality cannot be demonstrated, such
approaches are crucial for detecting important relationships because of their
rigorous and systematic nature. Had we only adopted a cross-sectional scheme,
we would have deduced that media influence was minimal and erred in the same way
as countless Washington and media pundits. As a consequence, researchers would
be wise to further explore how media coverage of scandals is connected to public
opinion in other settings, implementing time-series designs, to either verify or
disprove the patterns documented in this study.
The mutual impact of newspaper and television coverage on favorability is
pertinent because it implies a more complementary relationship between media
channels on public attitudes during political scandals. On the other hand, the
minimal association evident between newspaper coverage and job approval conforms
more to a traditional agenda-setting model with newspapers wielding more
influence. The discrepancy in the results may lie in the fact that our
hypothesis was grounded in agenda setting (which focuses on salience), but
favorability is more of an attitude. Consequently, the agenda-setting model
worked better for the salience measure (job approval) than for the attitude
measure (favorability). By building on the findings of the current
investigation, future work may want to further concentrate on the differences in
medium impact for political scandals with respect to attitudes and cognitions.
Moreover, expanding such work into the realm of radio and the Internet is
paramount for understanding the dynamics of how different media inform people's
impressions about political scandals.
As asserted earlier, the findings grant some support to the notion of
second-level agenda setting. Further, the influence of media coverage on
presidential evaluations mounts evidence for priming and attitude change as
well. The theoretical model advanced herein affords a framework for subsequent
research to continue and explicate the process. For example, the hybrid use of
ELM, priming, and second-level agenda setting helps explain why a peripheral
route to attitude change was not surprising given the low levels of personal
relevance that the public exhibited toward the story. Therefore, future
inquiries should probe how different types of media coverage shape both overall
impressions about politicians and perceptions of their personal attributes.
Isolating what exact aspects of Executive images are likely to be affected by
press coverage is important for both scholars and political consultants alike.
Combining such research with traditional agenda-setting scholarship on issue
salience will supply a more holistic image of how news forms social reality.
The divergent linkages among press scrutiny, favorability, and job approval
expose that presidential evaluations are multifaceted. We account for the
divergent findings by arguing that job approval is more of a substantive
(cognitive) evaluation, while favorability is more of an affective (valence)
judgment (as described in the literature review). In terms of comparing
substantive and affective elements, the stronger impact on favorability shows
that the effects of attribute salience on attitudes deserve more scrutiny. In
addition, it demonstrates the necessity of further bridging the ELM and
second-level agenda setting models to better sketch out this process. Political
scientists and mass communication scholars should begin examining both
dimensions of public assessments (cognitions & attitudes). Limiting research to
only one half of the equation precludes theoretical advancement of candidate
image scholarship.
The ability of media coverage to foster some changes in public assessments of
the president, even in multivariate analyses, corroborates the statistical
findings of other researchers (e.g., Brody, 1991. However, the weakness of
media influence on perceived job approval shows that this pattern only occurs in
some cases. It is imperative that analysts continue to employ multivariate
research designs because the relationships among media, government, and the
public are so highly complicated and intertwined that many relationships will go
unnoticed without such methods.
In sum, it is hoped that this project established the value of monitoring media
coverage of political scandals for better understanding how public assessments
of government officials are cast.
Notes
Figures & Tables
Figure 1: 1st & 2nd Level Agenda Setting for Monica Lewinsky Scandal
[--- Pict Graphic Goes Here ---]
Note: Lines reflect possible paths of salience transfer and attitude influence.
Table 1: Cross-Lagged Correlations Between Media Covg. & Favorability of
President
Media Precedes P.O. P.O. Precedes Media
-6
-5
-4
-3
-2
-1
0 wk
1
2
3
4
5
6
ABC
-.54
**
-.12
.15
-.17
.10
.22
-.09
.01
.13
-.17
-.10
.07
.00
NYT
-.47
*
-.30
-.11
-.30
-.24
-.03
-.31
-.27
-.18
-.35
-.25
-.05
-.06
*p<.05
**p<.01
Note: Values represent the correlation between the number of stories about the
Monica Lewinsky scandal and the percentage of people rating saying their overall
opinion of Bill Clinton is favorable. Time lags are computed in one-week
increments.
Table 2: Cross-Lagged Correlations Between Media Covg. & Job Approval of
President
Media Precedes P.O. P.O. Precedes Media
-6
-5
-4
-3
-2
-1
0 wk
1
2
3
4
5
6
ABC
.11
.07
.20
.12
.15
.25
-.21
.-.14
-.09
-.11
.06
-.03
-.17
NYT
.24
.21
.21
.29
*
.21
.22
.05
-.18
-.08
-.10
-.01
-.02
-.02
*p<.10
Note: Values represent the correlation between the number of stories about the
Monica Lewinsky scandal and the percentage of people rating saying they approve
of the way Bill Clinton is handling his job as president. Time lags are
computed in one-week increments.
Table 3: Cross-Lagged Spearmann Rho Correlations Between Media Covg. &
Competence ("Can get things done")
Media Precedes P.O. P.O. Precedes Media
-6
-5
-4
-3
-2
-1
0 wk
1
2
3
4
5
6
ABC
1
*
1
*
.87
.87
.5
-.32
-1
*
-.60
-.79
-.36
-.15
-.67
.11
NYT
.87
1
*
.5
.5
1
*
-.20
-.60
-.87
*
-.82
-.21
-.36
-.15
-.10
*p<.05
Note: Values represent the correlation between the number of stories about the
Monica Lewinsky scandal and the percentage of people rating saying Bill Clinton
can "get things done." Time lags are computed in one-week increments.
Table 4: Cross-Lagged Spearmann Rho Correlations Between Media Covg. & Honesty
Media Precedes P.O. P.O. Precedes Media
-6
-5
-4
-3
-2
-1
0 wk
1
2
3
4
5
6
ABC
-.50
.00
-.5
-1
**
-1
**
.94
**
.05
.87
*
.50
-.10
-.22
-.15
.62
NYT
.5
.5
-.5
-1
**
-.87
.32
.21
.87
*
.60
-.10
-.21
-.36
-.10
*p<.05
Note: Values represent the correlation between the number of stories about the
Monica Lewinsky scandal and the percentage of people rating saying Bill Clinton
is "honest and trustworthy." Time lags are computed in one-week increments.
Table 5: Cross-Lagged Spearmann Rho Correlations Between Media Covg. & Judgment
Media Precedes P.O. P.O. Precedes Media
-6
-5
-4
-3
-2
-1
0 wk
1
2
3
4
5
6
ABC
.50
.86
.5
-.5
-.5
.33
-.95
*
.00
-.32
-.95
*
-1
*
-.63
.32
NYT
1
*
1
*
.50
-.50
-.86
-.50
-.83
.00
.32
-.95
*
-1
*
-.50
-.63
*p<.05
Note: Values represent the correlation between the number of stories about the
Monica Lewinsky scandal and the percentage of people rating saying Bill Clinton
"shows good judgment." Time lags are computed in one-week increments.
References
Balz, D., & Marcus, R. (1998, July 29). Deal's effect on president, politics
is debated. Washington Post (Section A, page A01).
Benton, M., & Frazier, P. (1976). The agenda setting function of the mass
media at three levels of "information holding." Communication Research 3,
261-274.
Berke, R.L., & Elder, J. (1998, Sep.16). Testing of a president: The public;
keeping Clinton in office, most say in poll, but his image is eroding. New York
Times (section a, page 1, column 1).
Brody, R.A. (1991). Assessing the President. Stanford, CA: Stanford
University
Press.
Brody, R.A., & Page, B.L. (1975). The Impact of Events of Presidential
Popularity: The Johnson and Nixon Administrations. In A. Wildavsky (ed.),
Perspectives on the American Presidency. Boston: Little Brown.
Burnkrant, R., & Unnava, R. (1989). Self-referencing: A strategy for
increasing processing of message content. Personality and Social Psychology
Bulletin 15, 628-638.
Chaffee, S.H. (1972). Longitudinal designs for communication research:
Cross-lagged correlations. Paper presented at the Association for Education in
Journalism and Mass Communication, Carbondale, Illinois.
Chaffee, S.H. (1991). Communication Concepts 1 Explication. London: Sage.
Dearing, J.W., & Rogers, E.M. (1996). Communication Concepts 6 Agenda-Setting.
London: Sage.
Edwards, G.C. (1993). The Public Presidency. New York, New York: St. Martin's
Press.
Erbring, L., Goldenberg, E.N., & Miller, A.H. (1980). Front-page news and real
world cues: A new look at agenda-setting by the media. American Journal of
Political Science 24, 16-49.
Funkhouser, G.R. (1973). The issues of the sixties: An explicatory study of
the dynamics of public opinion. Public Opinion Quarterly 37, 62-75.
Ghanem, S. (1997). Filling in the tapestry: The second level of
agenda-setting. In M. McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication and
Democracy(pp.3-14). Mahwah, New Jersey: Lawrence Erlbaum Associates
Graber, D.A. (1982). Conclusions: Linkage successes and failures. In D.
Graber (Ed.), The President & the Public (pp.260-286). Philadelphia,
Pennsylvania: Institute
for the Study of Human Issues, Inc.
Grier, P., & Thurman, J.N. (1998, Aug. 31). Why the political split between
public and washington elite. Christian Science Monitor (Section United States,
Political Memo, page 3).
Hovland, C.I., Janis, I.L., & Kelley, H.H. (1953). Communication and
Persuasion. New Haven, Connecticut: Yale University Press.
Iyengar, S., Peters, M.D., & Kinder, D.R. (1982). Experimental demonstrations
of the 'not so minimal' consequences of television news programs. American
Political Science Review 76,4, 848-58.
Iyengar, S., & Kinder, D.R. (1987). News that Matters. Chicago, Illinois:
University
of Chicago Press.
Iyengar, S., & Simon, A. (1993). News coverage of the Gulf Crisis and public
opinion:
A study of agenda setting, priming, and framing. In S. Iyengar & R. Reeves
(Eds.) Do the Media Govern? Politicians, Voters, and Reporters in America (pp.
248-257). Thousand Oaks, CA: Sage.
Jablonski, P.M., & Hester, J.B. (1998, November). A Longitudinal Time Series
Analysis of Agenda Setting for the Education Issue: The Agendas of the
President, the Media, and the Public. Paper presented to the 1998 Southwest
Symposium of the Southwest Education Council for Journalism and Mass
Communication.
Jones, C.O. (1994). The Presidency in a Separated System. Washington, D.C.:
The Brookings Institution.
Katz, E., & Lazarsfeld, P.F. (1955). Personal Influence: The Part Played by
People in the Flow of Mass Communication. Glencoe, Illinois: Free Press.
Kepplinger, H.M., Donsbach, W., Brosius, H.,-B., & Staab, J.F. (1989). Media
tone
and public opinion: A longitudinal study of media coverage and public opinion of
Chancellor Kohl. International Journal of Public Opinion Research 1, 326-342.
Kinder, D. (1986). Presidential character revisited. In R. Lau & D. Sears
(Eds.),
Political Cognition. Hillsdale, New Jersey: Lawrence Erlbaum Associates.
King, P. (1997). The press, candidate images, and voter perceptions. In M.
McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication and Democracy (pp.29-40).
Mahwah, New Jersey: Lawrence Erlbaum Associates
Krosnick, J.A., & Kinder, D.R. (1990). Altering the foundations of support for
the president through priming. American Political Science Review 84, 2,
497-512.
Lazarsfeld, P.E., Berelson, B.R., & Gaudet, H. (1948). The People's Choice.
New York: Columbia University Press.
McCall, R.B. (1994). Fundamental Statistics for Behavioral Sciences. San
Francisco, CA: Duxburry Press.
McCombs, M.E., & Shaw, D.L. (1972). The agenda-setting function of mass media.
Public Opinion Quarterly 36, 176-187.
McCombs, M.E., & Shaw, D.L. (1993). The evolution of agenda-setting research:
Twenty-five years in the marketplace of ideas. Journal of Communication 43,
2,
58-65.
McCombs. M.E., & Evatt, D. (1995). Issues and attributes: Exploring a new
dimension in agenda setting. Communicacion y Sociedad 8, 1, 7-32.
McCombs, M., & Estrada, G. (1997). The news media and the pictures in our
heads. In S. Iyengar & R. Reeves (Eds.), Do the Media Govern? (pp. 237-247).
London: Sage
McCombs, M., Llamas, J.P., Lopez-Escobar, E., & Rey, F. (1997a). Candidate
images in Spanish elections: second-level agenda-setting effects. Journalism &
Mass Communication Quarterly 74,4, 703-717.
McCombs, M.E., Shaw, D.L., & Weaver, D. (1997b). Prologue: "The game is
afoot." In M. McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication and
Democracy (pp.ix-xii). Mahwah, New Jersey: Lawrence Erlbaum Associates
McCombs, M.E., & Ghanem, S. (1998). The convergence of agenda-setting and
framing. Unpublished manuscript. University of Texas at Austin.
McGuire, W.J. (1985). Attitudes and attitude change. In G. Lindzet & E.
Aronson
(Eds.), Handbook of Social Psychology (3rd ed., Vol 2., pp. 233-346). New York:
Random House.
Miller, J.M., & Krosnick, J.A. (1997). Anatomy of news media priming. In S.
Iyengar & R. Reeves (Eds.), Do the Media Govern? (pp. 258-275). London: Sage.
Mitchell, A., & Elder, J. (1998, Oct. 15) Congress's rating continues to slip
in inquiry's
wake. New York Times (Section A, Page 1, Column 6).
Mueller, J.E. (1970). Presidential Popularity from Truman to Johnson.
American
Political Science Review 64, 18-34.
Nimmo, D., & Savage, R.L. (1976). Candidates and their Images: Concepts,
methods, and findings. Santa Monica, California: Goodyear.
O'Keefe, D.J. (1990). Persuasion Theory and Research. Newbury Park, C.A.:
Sage.
Petty, R.E., & Cacioppo, J.T. (1981). Attitudes and Persuasion: Classic and
Contemporary Approaches. Dubuque: Wm. C. Brown.
Petty, R.E., & Cacioppo, J.T. (1986). Communication and Persuasion: Central
and Peripheral Routes to Attitude Change. New York: Springer/Verlag.
Petty, R.E., & Priester, J.R. (1994). Mass media attitude change: Implications
of the elaboration likelihood model of persuasion. In J. Bryan & D. Zillmann
(Eds.), Media Effects: Advances in Theory and Research (pp. 91-122). New
Jersey: Lawrence Erlbaum Associates.
Popkin, S., Gorman, J., Phillips, C., & Smith, J. (1976). Comment: what have
you done for me lately: toward an investment theory of voting. American
Political Science Review 80, 779-805.
Protess, D.L., & McCombs, M.E. (1991). Part III: The agenda-setting process.
In D.L. Protess & M. McCombs (Eds.), Agenda Setting. Hillsdale, New Jersey:
Lawrence Erlbaum Associates.
Roberts, M. (1997). Political Advertising's Influence on News, the Public and
Their
Behavior. In M. McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication
and Democracy (pp.85-98). Mahwah, New Jersey: Lawrence Erlbaum Associates.
Salwen, M.B. (1988). Effect of accumulation coverage on issue salience in
agenda
setting. Journalism Quarterly 65, 100-106, 130.
Schoenbach, K. (1982). Agenda-Setting Effects of Print and Television in West
Germany. Paper presented at the International Communication Association annual
conference.
Simonton, D.K. (1987). Why Presidents Succeed. New Haven, Connecticut: Yale
University Press.
Stimson, J.A. (1976). Public Support for American Presidents: A Cyclical
Model. Public Opinion Quarterly 40, 1-21.
Stone, G.C., & McCombs, M.E. (1981). Tracing the time lag in agenda-setting.
Journalism Quarterly 58, 151-155.
Takeshita, T., & Mikami, S. (1995). How did the mass media influence the
voters' choice in the 1993 general election in Japan? A study of
agenda-setting. Keiko Communication Review 17, 27-41.
Wanta, W., Stephenson, M.A., Turk, J., & McCombs, M.E. (1989). How president's
state of the union talk influenced news media agenda. Journalism Quarterly 66,
537-541.
Wanta, W., & Hu, Y. (1994). Time-lag differences in the agenda-setting
process: An
examination of five news media. International Journal of Public Opinion
Research 6, 225-240.
Wanta, A. (1992). The influence of the president on the news media and public
agendas. Mass Comm Review 19, 3, 14-21, 48.
Wanta, W. (1997a). The messenger and the message: differences across news
media. In M. McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication and
Democracy (pp.137-151). Mahwah, New Jersey: Lawrence Erlbaum Associates.
Wanta, W. (1997b). The Public and The National Agenda. Mahwah, New Jersey:
Lawrence Erlbaum Associates.
Weaver, D.H. (1977). Political issues and voter need for orientation. In D.L.
Shaw & M.E. McCombs (Eds.), The Emergence of American Political issues: The
Agenda-Setting Function of the Press (pp.107-120. St. Paul, Minnesota: West
Publishing.
Weaver, D.H., Graber, D.A., McCombs, M.E., Eyal, C.H. (1981). Media
agenda- setting in a presidential election. New York: Praeger.
Winter, J., & Eyal, C. (1981). Agenda setting for the civil rights issue.
Public Opinion Quarterly 45, 376-383.
Zaller, J.R. (1997). A model of communication effects at the outbreak of the
gulf war. In S. Iyengar & R. Reeves (Eds.), Do the Media Govern? (pp.
296-311). London: Sage
Zhu, J.H., & Boroson, W. (1997). Susceptibility to agenda setting: A
cross-sectional and longitudinal analysis of individual differences. In M.
McCombs, D.L. Shaw, & D. Weaver (Eds.), Communication and Democracy
(pp.69-84). Mahwah, New Jersey: Lawrence Erlbaum Associates.
Zucker, H.G. (1978). The variable nature of news media influence. In B. D.
Ruben
(Ed.), Communication Yearbook 2 (pp. 225-245). New Brunswick, NJ:
Transaction Books.
Zukin, C., & Carter, J.R., Jr. (1982). The measurement of presidential
popularity: old wisdoms and new concerns. In D. Graber (Ed.),The President &
the Public (pp.207-241). Philadelphia, Pennsylvania: Institute for the Study
of Human Issues, Inc.
[1] Voter need for orientation is comprised of three factors: "the degree of
(1) interest in the message content; (2) uncertainty about the subject of the
message; and (3) effort required to attend to the message (including the
perceived likelihood that a reliable source of information is available"
(Weaver, 1977, p.131).
[2] The subsample was comprised of every 10th story from ABC News and every
10th article from the New York Times.
[3] Normally, the question was worded in the following manner: "Do you approve
or disapprove of the way Bill Clinton is handling his job as president?"
[4] The basic wording for the favorability questions was: "What is your overall
opinion of Bill Clinton - is it favorable or unfavorable (or have you never
heard of him)?" It should be noted though that the range of questions varied
more for favorability than job approval.
[5] Honesty, competence, and judgment questions were worded as follows. First
interviewers said "I'm going to read off some personal characteristics and
qualities. As I read each one, tell me if you think it strongly applies to Bill
Clinton, somewhat applies, or doesn't apply. First_Next." The specific phrases
were "can get things done," honest and trustworthy," and "shows good judgment."
Gallup aggregated answers into "applies" / "does not apply" categories, which
were utilized in the data analysis phase of this project.
[6] The specific economic indicators used were consumer confidence and the
consumer price index in order to gauge both fiscal performance and perceptions
of economic health.
[7] Hypothetically, a spurious correlation between variables a (media coverage)
& b (public attitudes) might exist because both are correlated to a third
variable c (public salience) that intervenes between the two (McCall, 1994). A
partial correlation coefficient allows us to test how direct the relationship
between variables a & b is. Thus, we can deduce whether a peripheral route of
attitude change is more likely, based on the value of the correlation
coefficient.
|
