Perceptions of Media Power and Fairness in Campaign '96:
The Third-Person Effect and Support for Press Restrictions
By
Michael B. Salwen
University of Miami
School of Communication
PO Box 248127
Coral Gables, FL. 33124-1230
[log in to unmask]
WORK: 305-284-2265
FAX: 305-284-3648
Paper submitted to the Association for Education in Journalism and Mass
Communication, Communication Theory and Methodology Division, Chicago, IL.,
August 1997.
The author is a professor of communication at the University of Miami, School
of Communication, Coral Gables, Florida. He gratefully acknowledges funding from
a School of Communication Task Force grant to conduct this study.
Abstract
Perceptions of Media Power and Fairness in Campaign '96:
The Third-Person Effect and Support for Press Restrictions
This study applied the third-person effect approach to investigate
public perceptions of media influence and support for press restrictions
during the 1996 presidential campaign. The findings provided compelling
evidence for individuals to perceive greater media influence on people
other than themselves. The findings also supported the behavioral
component hypothesis for third-person perception to lead to support for
restrictions on "unfair" election news coverage. While the order of the
self and others effects questions did not consistently indicate that
third-person perception was an artifact of question order, some discrepant
findings warranted further investigation. The implications for the press
are discussed.
Third-Person Perception and Beliefs About Media Influence:
Support for Press Restrictions in Campaign '96
I often wonder whether we do not rest our hopes too much upon
constitutions, upon laws and upon courts. These are false hopes; believe
me, these are false hopes. Liberty lies in the hearts of men and women;
when it dies there, no constitution, no law, no court can ever do much to
help it
Judge Learned Hand (1959, pp. 89-90).
While constitutional guarantees grant the American press considerable freedom,
Judge Learned Hand and others observed that press freedom depends on public
support. In theory, the courts should not heed public opinion in press freedom
cases; in practice, court decisions often reflect public opinion (Black, 1960;
Marshall, 1988; Mishler & Sheehan, 1993). As journalist Finley Peter Dunne's
street-smart character Mr. Dooley wryly observed, "th' supreme coort follows th'
iliction returns" (Bander, 1963, p. 52).
This study investigated public perceptions of media influence and public
support for press restrictions during the 1996 presidential election. It applied
the third-person effect approach, which contends that people perceive media
messages to exert greater persuasive influence on other people than themselves.
The press is especially reliant on public support during what legal scholar
Vincent Blasi (1985) labeled "pathological periods" when "repressive dynamics
may penetrate the judicial psyche and cause judges to interpret the First
Amendment restrictively" (p. 450).
Finding not public support but condemnation of the press during many of the
press-government tussles of the 1960s, some of the nation's highest officials
attacked the press with "frequency and zeal." Officials proposed remedies, such
as blanket subpoenas, that "would have been unthinkable a decade ago." (Balk,
1972, p. 13)[1] More recently, Rucinski and Salmon (1990), reporting a
third-person effect study during the 1988 presidential campaign, warned of
public anger toward campaign news "that obfuscated the underlying issues of the
campaign" and fueled "calls for campaign reform and stricter controls on
campaign content. . . ." (p. 345)
The 1996 campaign had its share of obfuscating issues. Repeated intimations
about Clinton's wrongdoings and a leaked report about a prostitute scandal that
forced the resignation of the Democratic Party's top campaign advisor were among
the issues that provoked public outrage against politicians and the journalists
who covered -- and some claimed participated in -- the political process
(Garneau, 1996; Goldstein, 1996). As one example, former New York Times national
correspondent Dudley Clendinen (1997) claimed the disclosure of Newsweek's Joe
Klein as the anonymous author of a gossipy, critical novel about the Clintons
fueled public anger toward the double standard by which the press does not hold
its members to the standards it holds public officials: "What is the public to
think of the character of public figures who lie? Or [in the case of Klein] who
manipulate the public in order to profit from their cleverness? Two answers, at
least, are that the voters re-elected Clinton, and the New Yorker hired Joe
Klein to write about him" (p. 95).
Until the early 1980s, many scholars and civil libertarians believed that
public support for censorship was attributable to intolerance toward unpopular
expression (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950; Stouffer, 1955;
Zalkind, 1975; Zalkind, Gaugler, & Schwartz, 1975). This perspective, popular
among post-war scholars mindful of the lessons of Nazi Germany, suffered from an
elitist streak. Formal education was considered essential for increased
tolerance because with education comes the ability and sophistication to
understand the principles underlying freedom of expression (Prothro & Grigg,
1970). Zellman (1975) claimed that true tolerance required citizens to link
abstract principles to concrete situations. "This link requires some degree of
deductive reasoning and hence some minimal level of cognitive sophistication.
For most people, such a linkage does not occur" (p. 45).
Immerwahr and Doble's (1982) important multi-method study challenged the
intolerance perspective. They reported that public beliefs about freedom of the
press are dominated by principles of fairness, rather than closed-mindedness or
abstract principles. They chastised scholars for dismissing the public's demand
for fairness and claimed most people support unpopular expression "when the
objectionable views will be presented as part of a spectrum of diverse points of
view" (p. 182):
If the public is calling for an increase in press freedom as defined
from the listener's perspective, if the public's support for fairness laws
stems from a belief in the importance of a "marketplace of ideas," then
the tendency among elites to dismiss calls for fairness as calls for
censorship and repression represents a fundamental misunderstanding of the
public's position, a misunderstanding that could have potentially serious
consequences for a democratic society. (pp. 186-187, original emphasis)
Third-Person Perception
Recently, Davison (1996), who conceived the third-person effect (see Davison,
1983), approved of Perloff's (1993) description of the effect as a generalized
form of "third-person perception." Third-person perception is sometimes
explained by attribution theory (Gunther, 1991a; Rucinski & Salmon, 1990), which
contends that people attribute their behaviors and responses to situations while
they attribute other people's behaviors to irrational dispositions (Jones &
Nisbett, 1972).
As Rucinski and Salmon (1990) observed, attribution theory applied to
third-person perception is problematic because media content is a situational
factor and thus would suggest that a rational person would evaluate media
content to influence him or herself more than other people. As a result, they
offered a revised "self-serving bias in attribution" (p. 348). According to
self-serving bias, people systematically perceive themselves as better able to
discern false messages than other people who are easily aroused by emotional
appeals. An implicit -- and sometimes explicit -- assumption underlining the
third-person effect is that messages must be perceived as undesirable to believe
(Gunther, 1991; Gunther & Mundy, 1993; Gunther and Thorson, 1992; Cohen, Mutz,
Price, & Gunther, 1988). Rucinski and Salmon (1990) noted that for a
self-serving attribution approach to work in a political campaign context the
communications must be "viewed as threatening or dangerous to a political
system" (p. 348).
Most studies support the third-person "perceptual component" hypothesis
(Lasorsa, 1992; Perloff, 1993). A "behavioral component" hypothesis, which
enjoys less empirical support, predicts that third-person perception causes
people to support censorship to "protect" other people from perceived harmful
messages (Baughman, 1989; Rojas, Shah, & Faber, 1995, Salmon, 1989). Blasi
(1985), for example, in his theory of "pathological periods," noted that the
judiciary's resistance to freedom of expression to protect the public from
dangerous messages was "an exercise in self-paternalism" (p. 512).
The best evidence for the behavioral component comes from studies on
pornography and violent and misogynic rap lyrics (Gunther, 1995; Lee & Yang,
1996; McLeod, Eveland, & Nathanson, 1997; Paxton, 1996a; Rojas, Shah, & Faber,
1995). Behavioral component studies are often tested with "deviant" expression,
on the assumption that it is undesirable to believe or be influenced by deviant
messages. Political campaign news coverage is a form of nondeviant expression --
in fact, a socially beneficial expression that contributes to a functioning
democracy (Berelson, 1952; Manheim, 1976). Nevertheless, there are times when
political news reports involve sex scandals and other events that may be
perceived as undesirable to believe and deleterious to the political system.
Rucinski and Salmon (1990) found evidence for third-person perception during
the 1988 presidential campaign. But third-person perception did not lead to
support for restrictions on campaign coverage (measured as support for an
independent commission to regulate five types of political communication). They
noted, however, that perceived message harm was a significant predictor of
restrictions. Salwen and Driscoll (1996) examined the third-person effect and
news of the O.J. Simpson criminal trial, another sort of news essential for a
functioning democracy. They also found a third-person effect but failed to find
clear evidence for a behavioral outcome. Overall, behavioral component research
suggests that respondents are more tolerant of expression necessary for
democratic maintenance than messages perceived as deviant.
Question-Order Effects
It has been suggested that the third-person effect may be a methodological
artifact explained by the self-serving comparisons people make when they
appraise media effects on themselves and on other people during surveys.[2]
Studies that reversed the order of the self and others questions concluded that
the perceptual component is not merely a methodological artifact (Mason, 1990;
Tiedge, Silverblatt, Havice, & Rosenfield, 1991; Gunther, 1991b, 1995). Price
and Tewksbury (1996), who conducted the definitive study to date on
question-order effects, also found that question order did not significantly
impact third-person perception. But their findings left open the possibility
that question order may still be germane. Specifically, in experimental
conditions when the others question preceded evaluation of self impact,
respondents' evaluations of media influence on themselves decreased among those
with high levels of general political knowledge. The fact that political
knowledge emerged as an interaction term calls attention to recipient factors in
third-person perception.
Recipient Factors
Even if third-person perception contributes to people's willingness to restrict
messages, other factors affect perceptual processes. Education and the related
variables of self-perceived knowledge and expertise have repeatedly emerged as
factors that influence self-other perceptions (Glynn & Ostman, 1988; Gunther,
1995; Hu & Wu, 1996; Lang, 1995; Lasorsa, 1989; Rucinski & Salmon, 1990; Salwen
& Driscoll, 1996; Tiedge, et al., 1991; Willnat, 1996). Education is thought to
provide people with the ability to distinguish themselves as smarter than other
people -- or others as less smart than themselves, accounting for the perceptual
component (Salwen & Driscoll, 1995). Brosius and Engel (1996), however, found
evidence for education to be negatively associated with third-person perception.
Similarly, McLeod, Eveland and Nathanson (1997) failed to find consistent
evidence for a "knowledge corollary" in a third-person effect study involving
the perceived effects of misogynistic rap lyrics.
Tiedge and his colleagues (1991) reported that age contributed to third-person
perception. Older respondents evaluated less persuasive effects on themselves
than they did on others.[3] Tiedge et al. explained their age findings by noting
that older respondents may have developed access to more media sources over the
years. Rucinski and Salmon (1990) found age unrelated to third-person
perception; however, age was a negative predictor of effects on both self and
perceptions of others. Although age emerges as a factor in third-person effect
studies, the explanations often seem inductive. With age can come wisdom with
years and life experience or an unmovable adherence to the old ways and
resistance to new ideas. The Tiedge et al. and Rucinski and Salmon studies
examining the independent contributions of self and others effects, rather than
simply the self-other differential, underscore the importance of examining the
independent contributions of self and others components.
Media use is an intriguing recipient factor for communication scholars. As with
education, media use should provide individuals with confidence in their
knowledge relative to other people and thereby increase third-person perception.
As Rucinski and Salmon (1990) argued:
[I]t seems plausible that the stronger one's orientation to the mass
media, where orientation can be treated as either gross exposure or the
related notion of attention to media content, the greater the likelihood
of a perception of differential effects. It is likely that a person who
avidly reads public affairs information, for example, is more likely to
feel resistant to persuasive attempts because s/he possesses a substantial
knowledge base that serves as a defense (p. 349).
Studies that examined media use report conflicting evidence. Rucinski and
Salmon (1990) found media use patterns to be important predictors of media
influence on self and perceived others' voting decisions. Exposure to public
affairs television and attention to newspaper public affairs content led to
greater evaluations of media influence on oneself. On the other hand, newspaper
exposure was a negative predictor on self effects and estimations of media
influence on others. They argued that newspaper reading and attention to public
affairs news in general may be perceived as desirable to believe. Paxton (1997),
however, failed to find a correlation between media use and perceptual
discrepancy. Innes and Zeitz (1988) reported no support for the proposition that
radio listening or newspaper and magazine reading would increase third-person
perception. Television viewing, however, was positively associated with
third-person perception. Brosius and Engel (1996) did not find a systematic
relationship between media use and third-person perception and concluded that
"media use must therefore be regarded as weak, and its direction unpredictable"
(p. 157).
Research Questions and Hypotheses
This study investigated two research questions and five hypotheses about mass
media influence and public support for press restrictions during the 1996
presidential election:
RQ1: Will the order of the self and others effects questions affect judgments of
media influence on self and on others?
RQ2: What are the independent contributions of self and others effects as
predictors of support for restrictions on unfair election news coverage?
H1: Respondents will judge mass media messages about the presidential candidates
to exert a greater influence on other people than on themselves.
H2: Education will be a positive predictor of third-person perception.
H3: News media use (and interest) will be a positive predictor of third-person
perception.
H4: Education will be a negative predictor of support for restrictions on unfair
election news coverage.
H5: Magnitude of third-person perception (i.e., perceptions of greater effects
of election news coverage on other people than on oneself) will be a positive
predictor of support for restrictions on unfair election news coverage.
Method
Overview
A representative sample of 549 adults (age 18 and older) in the continental
United States was contacted by telephone during weekday evenings from Monday,
October 21, 1996, to Monday, November 4, the day before the election.[4]
Screening questions eliminated respondents under 18 and those who had not made
up their minds to vote for one of the three major presidential candidates
(Clinton, Dole, or Perot) "if the election were held today." This decision was
made because issue involvement -- in this case campaign involvement -- is a
factor in third-person effect research (Mutz, 1989; Perloff, 1989; Vallone,
Ross, & Lepper, 1985). A decision on candidate choice would represent a minimal
level of campaign involvement for increased likelihood of developed beliefs
about media effects (Alvarez, 1990).
Trained graduate and undergraduate student callers conducted the interviews
from a centralized phone bank. After four attempts to reach each number, the
rate of completed interviews to valid contacts was 70%.[5] A Spanish-language
instrument was used by bilingual callers when needed. The gender, race, and age
characteristics were within sampling error of known population parameters.
Fifty-four percent of the respondents preferred Clinton; 39% preferred Dole, and
7% preferred Perot, compared to 50%, 41%, and 9%, respectively, in the election.
The sample was 56% female and 44% male. Ages ranged from 18 to 89, with a median
age of 42. Among the seven education categories, the mode response was completed
high school (30%), followed by some college (25%), and completed college (21%).
Race and Hispanic ethnicity were measured by separate questions, so the totals
do not sum to 100%. Eighty-seven percent of the respondents were white, 10% were
African American, 2% were Asian and 1% were other race. Eight percent were
Hispanic.
Effects and Restrictions
To appraise respondents' evaluations of news media influence of election news,
respondents were asked two sets of questions. Media influence on others
consisted of four questions pertaining to respondents' perceptions of newspaper,
television, radio, and "all news media together including newspapers,
television, and radio" in influencing "other people's opinions" about the
presidential candidates. Respondents answered using a 1-10 scale, where 1 was
extremely ineffective and 10 was extremely effective. Media influence on self
consisted of parallel-phrased questions asking about news media influence "on
your own opinions." Callers alternated survey instruments with the set of self
questions asked before and after the others questions to test for question-order
effects.
Support for restrictions on unfair election news consisted of five questions
about support for a government commission "with the legal power to ban unfair" .
. . newspaper stories/ television news stories/ news stories in all the news
media, of television, newspapers, radio, and newsmagazines/ television
advertisements about the presidential candidates/ and political poll stories
about the candidates. Respondents answered using 1-10 scales, where 1 was
strongly disagree and 10 was strongly agree. The word "unfair" was inflected in
each question because third-person perception is not due solely to media power
or influence, but perceived message harm (Rucinski & Salmon, 1990).
Demographics and Psychographics
Respondents were also asked general news media use and campaign news media
interest questions. These consisted of three sets of three questions each about
newspapers, local television news, and network news. In the first question in
each set, respondents were asked how many days a week, if any, they read a
newspaper/ watched a local news program/ watched a network news program. Those
who used the medium one or more days a week were then asked whether they read a
newspaper/ watched a local news program/ watched a network news program
yesterday or today. After these questions in each set, respondents were then
asked how interested they were when they came across news about the presidential
election in each medium. Interest was measured with a 1-10 scale, where 10 was
extremely interested.
Respondents were asked their highest level of education. Their answers were
coded into seven categories ranging from 8th grade or less to graduate degree.
Respondents were also asked their year of birth. Age was coded as 1996 minus the
year. They were also asked their race (White, African American, Asian, or other)
and a separate question whether they were of Hispanic ancestry. Interviewers
coded respondents' gender.
Findings
Third-Person Perception
As a general test of third-person perception, Table 1 reports the means of the
self and others items. As, hypothesized, respondents evaluated greater media
influence on other people than on themselves (p < .001).
Table 1 Goes About Here
The self and others questions were subjected to a principal components factor
analysis. The results, explaining almost 70% of the variance, indicated clear
evidence for self and others dimensions (See Table 2). The four "self" items
(alpha = .86) and four "others" items (were alpha = .84) were reduced to
summated scales. A pairwise two-tailed t-test between the self and others scales
indicated a significant difference (Self = 4.47, sd = 2.26; Others = 6.37, sd =
1.81, t-value= 15.65, p < .001).
Table 2 Goes About Here
Question-Order Effects
The next step was to test the first research question examining whether
respondents' evaluations of the self and others questions may have been
influenced by the question order. Table 3 reports the pairwise comparisons of
the question order effects.
Table 3 Goes About Here
Six of the eight conditions did not indicate a question-order effect. However,
the fact that two conditions attained significance at p <.05 in the direction
expected by a self-serving comparison explanation means we cannot reject the
possibility of some influence attributable to question order.
For a self-serving explanation to be plausible in the "effects on self"
conditions, the mean "others first" value must be lower than the "self first"
value. In the newspaper effects on self condition, respondents evaluated less
newspaper influence on themselves after they estimated effects on others.[6] A
self-serving comparison would suggest that after estimating newspapers'
influence on others, respondents made themselves appear comparatively smarter by
evaluating themselves as less influenced than others. Although none of the other
"effects on self" conditions is significant, all are in the direction explained
by a self-serving comparison.
For a self-serving explanation to be plausible in the "effects on others"
conditions, the mean "self first" value must be higher than the "others first"
value. In the radio effects on others condition, respondents estimated effects
on others significantly higher after they first evaluated radio's effects on
themselves. According to a self-serving comparison explanation, respondents made
them appear comparatively smarter than others. Unlike the values in the "effects
on self" conditions, the nonsignificant findings do not even suggest a
self-serving comparison trend. This would indicate an advantage to first asking
the others' effects questions.
Perceptual Bias and Restrictions
To measure third-person perception, a "perceptual bias" transformation variable
was created by computing the difference between each individual's others-scale
and self-scale. Perceptual bias ranged from a 9 (10 on others minus 1 on self)
to -9 (1 on self minus 10 on others), where positive values indicated greater
magnitude of perceived effects on others than evaluations of effects on self, or
"third-person perception"; negative values indicated the reverse "first-person
perception." A zero value indicated no perceptual bias. As further evidence for
the third-person perception hypothesis (H1), more than three-quarters of the
respondents (76.8%) exhibited some magnitude of third-person perception. Only
16.3% exhibited first-person perception and 6.9% indicated no perceptual bias.
The overall perceptual bias mean was 1.90 (sd = 2.61).
A principal components factor analysis on the restrictions items (varimax
rotation) yielded a single factor explaining 84% of the variance. The
correlation matrix among the items accounted for substantial variance, ranging
from r =.73 (newspapers and polls) to r =.90 (television and media and
television and newspapers). The items were reduced to a summated scale (alpha =
.95).
Education and Media Use
The second and third hypotheses, respectively, predicted that education and
news media use (and interest) will be positive predictors of third-person
perception. As a reliability check of the daily news media use predictors, the
correlations between each pair of media use and yesterday/ today questions
accounted for considerable variance (newspaper, r = .70, p < .001; local news, r
= .72, p < .001; network news, r = .83, p < .001).
A factor analysis on the three campaign news interest items (varimax rotation)
indicated a single dimension explaining 81% of the variance. The items were
reduced into a summated scale (alpha = .88). The three media use questions,
however, did not achieve adequate reliability (alpha = .51).
Predicting Third-Person Perception
Table 4 reports the hierarchical regression predicting third-person perception
and separate regressions predicting the independent contributions of self and
others effects on third-person perception.
Table 4 Goes About Here
Educational level was not a significant predictor of third-person perception or
of self or others effects. In the first block of demographic variables, age was
a negative predictor of both effects on self and perceived effects on others. We
noted the lack of a theoretical rationale for age in third-person research. As
an inductive explanation, it may have been that the negative influence of age (p
< .05) on self effects was due to older respondents believing they attained
enough wisdom with the years to counter media influence. The negative predictor
of age on perceptions of others, so that as age increased perceptions of media
influence on others decreased (p < .001), cannot be easily explained. Hispanic
ethnicity also emerged as a negative predictor of effects on self, with
Hispanics evaluating less media influence on themselves (p < .05). There is no
rationale for Hispanic ethnicity to emerge, especially since the related
variable of race was not a significant predictor. With Hispanic ethnicity, the
finding may have been due to chance.
Candidate choice, the sole item in the second block, was not a significant
predictor of third-person perception. Candidate choice was assessed as a base
measure of campaign involvement for respondent inclusion in the study. There was
no variation in terms of having made up one's mind for a candidate. The
variation was only in which candidate was preferred; and there is no reason to
suspect that favoring either Clinton or Dole (only Clinton and Dole were
analyzed) should predict third-person perception. Thus, this finding does not
negate previous research that interest may be a factor in third-person
perception (Mutz, 1989; Perloff, 1989). In fact, given that campaign news
interest emerged as a significant predictor across the three regressions
suggests that involvement may have mattered. As interest in campaign news
increased, respondents perceived less third-person perception. In addition, as
campaign news interest increased, evaluations of media influence on self and on
others also increased.
The significant finding with campaign news interest was in the opposite
direction predicted in the third hypothesis. It may have been that respondents
did not perceive it undesirable to believe media messages about the candidates.
If so, there is nothing surprising in respondents interested in campaign news
perceiving media influence on themselves; likewise, it is not unreasonable that
increased campaign news interest predicts perceived effects on others, on the
assumption that people perceive others to be similarly interested in campaign
news.
While campaign news interest was contrary to the news media use (and interest)
hypothesis, there may have been important differences in how media use and
interest predicted third-person perception. As predicted in the media use
hypothesis, newspaper reading was a significant positive predictor of
third-person perception (and, it follows, a negative predictor of effects on
self). In line with the theoretical rationale for media use, it may be that
reading the newspaper -- a medium that is more closely associated with public
affairs news than television -- provides people with the confidence to evaluate
themselves smarter than other people. The expected positive predictor on
perceived others did not attain significance.
Since it was suggested that some of the counterintuitive findings predicting
third-person perception may have been due to the neutral -- or perhaps even
desirable -- nature of the messages (i.e., to believe news about the
candidates), it is important to examine predictions of support for restrictions
on "unfair" media messages. Here there can be no doubt that it is undesirable to
believe unfair messages. And here the findings were more clear and in line with
the hypotheses.
Predicting Restrictions
Table 5 reports hierarchical regressions predicting support for restrictions on
unfair messages. The table includes two regressions, differing in the last block
of perceptions variables. The first regression on the left includes perceptual
bias. The other regression includes the components of perceptual bias -- self
and others effects.
Table 5 Goes About Here
As hypothesized (H4), education was a negative predictor (p < .001) of support
for restrictions on unfair campaign coverage. Also as hypothesized (H5),
perceptual bias was a predictor of support for restrictions, although education
was a better predictor. The second research question examined the independent
contributions of self and others effects as predictors of support for
restrictions on unfair messages. The regression reporting the independent
contributions of self and others effects indicated that third-person perception
was better explained by perceived influence on others than estimated influence
on self. That is, as perceptions of media influence on others increased, support
for restrictions increased. While self effects was negative, as would be
expected, it was not significant. Overall, the restrictions models yielded a
clean picture, with support for restrictions on unfair election coverage
explained by lower education and higher third-person perception -- with the
caveat that third-person perception was largely a function of perceptions of
effects on others.
Conclusions
Recent years have witnessed a public backlash against a perceived irresponsible
press and a slew of unfavorable legal decisions that have caused anxiety in
journalistic circles.[7] Although most of these decisions were reversed or
reduced in appeal, they underscore the public's anger and willingness to
restrain the press. The press, which prides itself as a defender of the public's
"right to know," is in an awkward position when it depends upon the court for
protection from a vengeful public. This study applied the third-person effect
approach to investigate public perceptions of media influence and support for
press restrictions during the 1996 presidential campaign.
The findings provide compelling evidence for people to perceive greater media
influence on other people than themselves. However, some of the predictors of
third-person perception were counterintuitive. It was suggested that this was
because this study did not deal with a topic that was inherently undesirable to
believe. The fact that the perceptual component hypothesis was supported with a
neutral or positive topic raises the possibility that third-person perception
may be a general perceptual tendency, enhanced by negative messages but not
eliminated by positive or neutral messages (Brosius & Engel, 1996; Duck &
Mullin, 1995).
This study also supported the behavioral component hypothesis for third-person
perception to lead to support for restrictions on unfair election news, although
education was a better (negative) predictor. Previous studies with "nondeviant"
messages, such as Rucinski and Salmon's (1990) study of the 1988 presidential
campaign and Salwen and Driscoll's (1996) study of the O.J. Simpson criminal
trial, failed to report consistent evidence for a behavioral outcome.
Support for the behavioral component in this study may have been due to the
emphasis on "unfair" election news. While the public may view election news as a
legitimate topic, it may see unfair election news as less deserving of
protection than "fair" election news. The fairness perspective may be especially
germane to political communication. Rucinski and Salmon's election study, for
example, found that perceived message harm was a significant predictor of
support for restrictions. Thus, researchers may want to study alternative
messages with varying degrees of harm or fairness to investigate whether
unfairness is associated with increased support for restrictions.
The examination of the independent contributions of the self and others effects
that comprise third-person perception indicated that people's motives for
supporting restrictions on unfair messages were better explained by perceived
effects of unfair messages on others. This supports a paternalistic explanation
for censorship to "protect" others from harmful messages. Censorship is an ugly
word, so much so that we often substitute the word "restrictions"; no one wants
to defend such an indefensible act. We suspect that had we used "censorship" in
place of "restrictions," there would have been far more support for the
press.[8] Third-person perception, however, may make censorship appear less
undesirable -- and perhaps even necessary for the common good.
In addition to addressing theoretical concerns, this study also addressed
methodological issues in third-person perception. The self and others effects
questions were reversed to investigate whether third-person perception is a
methodological artifact resulting from the self-serving comparisons. Based on
previous studies, there was no reason to suspect that question order influenced
third-person perception. Although the test for question order generally
indicated that third-person perception was not an artifact of question
sequencing, we did not reject the possibility of some influence attributable to
question order. Researchers need to further investigate the influence of
question order; and if question-order effects exist, researchers need to devise
procedures to deal with this problem. If the sequence of the self and others
questions poses problems, the findings indicate that asking the others question
first mitigates -- but does not eliminate -- question order problems.
Methodological questions in third-person effect research may involve more than
just the order of the self and others effects questions. It may be that in
responding to the self and others effects questions, respondents pause and
reflect on the issues under study. This self-reflection may influence
respondents' opinions and behavioral intentions about the issues. If so, this
can have consequences for the behavioral component, and it underscores the need
for more methodological studies. Researchers should consider examining
behavioral intentions before and after administering the sets of self and others
effects questions.
A weakness of this study was our simple measure of political involvement as
having made up one's mind to vote for one of the major candidates. Involvement
may be a more important and more complex variable. It deserves more
sophisticated attention. The fact that interest in campaign news was a useful
predictor of third-person perception underscores the importance of involvement.
Another drawback was this study's measure of support for restrictions.
Researchers need to obtain behavioral measures of the behavioral component,
rather than stated behavioral intent. Finally, this study assumed that the
public perceives some issues as legitimate and illegitimate. Yet we did not
measure perceived issue legitimacy.
The findings have implications for the press during this time when the press is
under attack for its practices. The press, like other institutions, requires
public support. If there is any good news for the press in this study, it is
that more highly educated respondents were less willing to restrict the press.
Thus, it is to the advantage of the press to have an informed, educated
electorate. Educated respondents probably have a better understanding of the
rationale for a free press. Contrary to expectations, educated respondents did
not exhibit greater third-person perception. Third-person perception may be a
general social trait.
If it is to the advantage of the press to have an educated public, the question
is: What can the press do to educate the public? The press, it is often said,
can contribute to an informed electorate. For its own good, it must educate the
public about the value of expressive rights in general and, more specifically,
the role of the press in society. There is much to defend. But the press must
also concede and condemn press behaviors that are unfair and indefensible,
rather than leave the impression that the press supports its own out of blind
loyalty. Just as important, and probably more difficult, the press must convince
the public that people other than themselves are not all gullible dolts who
naively believe everything they read and hear in the press.
Footnotes
References
Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., & Sanford, R. N.
(1950). The authoritarian personality. New York: Harper.
Alvarez, R. M. (1990). The puzzle of party identification: Dimensionality
of an important
concept. American Politics Quarterly, 18, 476-491.
Balk, A. (1972). The media under attack. In The Twentieth Century Fund, A
free and responsive press. The Twentieth Century Fund task force report
for a national news council. (pp. 13-16). New York: The Twentieth Century
Fund.
Bander, E. J. (comp.) (1963). Mr. Dooley on the choice of law.
Charlottesville, VA: The Michie Company, Law Publishers.
Baughman, J. L. (1989). "The world is ruled by those who holler
the loudest": The third-person effect in American journalism history.
Journalism History, 16(1/2), 12-19.
Berelson, B. (1952). Democratic theory and public opinion. Public Opinion
Quarterly, 16, 313-330.
Blasi, V. (1985). The pathological perspective and the first
amendment. Columbia Law Review, 85, 449-514.
Brosius, H-B., & Engel, D. (1996). The causes of third-person
effects: Unrealistic optimism, impersonal impact, or generalized
negative attitudes towards media influence? International Journal of
Public Opinion Research, 8, 142-162.
Budiansky, S. (1995, January 9). The media's message. U.S. News & World
Report, 119(1), 45-47.
Clendinen, D. (1997). Politics, journalism and the scandal glut. Media
Studies Journal, 11(1), 89-96.
Cohen, J., Mutz, D. C., Price, V., & Gunther, A. (1988). Perceived impact
of defamation: An
experiment on third-person effects. Public Opinion
Quarterly, 52,
161-173.
Columbia Journalism Review (March/April 1997). Punishing the
Press [Special issue], 36(2).
Commission on Freedom of the Press (1947). A free and responsible press.
Chicago: University of Chicago Press.
Davison, W. P. (1983). The third-person effect in communication. Public Opinion
Quarterly, 47, 1-15.
Davison, W. P. (1996). The third-person effect revisited. International
Journal of Public
Opinion Research, 8, 113-119.
Duck, J. M., & Mullin, B. (1995). The perceived impact of mass
media: Reconsidering the third-person effect. European Journal of
Social Psychology, 25, 77-93.
Garneau, G. (1996, September 14). Scandalous dilemma. Editor &
Publisher, 129, 14-16, 24-25.
Geimann, S. (1997, March). It's fair to ask whether the ends
justify the means. Quill, 85(2), 43.
Glynn, C. J., & Ostman, R. E. (1988). Public opinion about public
opinion. Journalism Quarterly, 65, 299-306.
Goldstein, M. L. (1996). Guide to the 1996 presidential election.
Washington, D.C.: Congressional Quarterly, Inc.
Gunther, A. C. (1991a). What we think others think: Cause and
consequence in the third-person effect. Communication Research, 18,
355-372.
Gunther, A. C. (1991b). The third-person effect and support for
restrictions on pornography. Paper presented at the annual meeting of
the Association for Education in Journalism and Mass Communication, Boston.
Gunther, A. C. (1995). Overrating the x-rating: The third-person
perception and support for censorship of pornography. Journal of
Communication, 45(1), 27-38.
Gunther, A. C., & Mundy, P. (1993). Biased optimism and the
third-person effect. Journalism Quarterly, 70, 58-67.
Gunther, A. C., & Thorson, E. (1992). Perceived persuasive effects of
commercials and
public service announcements. Communication Research, 19,
574-596.
Hand, L. (1959). The spirit of liberty: Papers and addresses of
Learned Hand. New York: Vintage Books.
Hu, Y. W., & Wu, Y. C. (1996, August). Testing a theoretical
model on the third person effect: Perceived impacts of election polls.
Paper presented at the annual meeting of the Association for Education in
Journalism and Mass Communication, Anaheim, CA.
Immerwahr, J., & Doble, J. (1982). Public attitudes toward freedom of the
press. Public
Opinion Quarterly, 26, 172-180.
Innes, J. M., & Zeitz, H. (1988). The public's view of the impact of the
mass media: A test of the 'third person' effect. European Journal of
Social Psychology, 18, 457-463.
Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer:
Divergent
perceptions of causes of behavior. In E. E. Jones, D. F.
Kanouse, H. H.
Kelly, R. E. Nisbett, S. Valins, and B. Weiner (Eds.),
Attribution:
Perceiving the causes of behavior (pp. 44-63). Morristown,
WI: General Learning
press.
Lang, A. (1995, May). The third person effect and political
advertising: Truth, effectiveness, regulation and usefulness. Paper
presented at the annual meeting of the International Communication
Association, Albuquerque, NM.
Lasorsa, D. L. (1989). Real and perceived effects of "Amerika".
Journalism Quarterly, 66, 373-378, 529.
Lasorsa, D. L. (1992). How media affect policy-makers: The third- person
effect. In J. D. Kennamer (Ed.), Public opinion, the press and pubic policy
(pp. 163-175). New York: Praeger.
Lee, C., & Yang, S. (1996, August). Third-person perception and
support for censorship of sexually explicit visual content: A Korean
case. Paper presented at the annual meeting of the Association for
Education in Journalism and Mass Communication, Anaheim, CA.
Manheim, J. B. (1976). Setting the political agenda: Can democracy
survive television?
Journal of Communication 26(2), 84-90.
Marshall, T. R. (1988). Public opinion, representation, and the
modern Supreme Court. American Politics Quarterly, 16, 293-316.
Mason, L. (1990). Faculty factfinders: The self-other bias in
resolution of ambiguous messages. Unpublished doctoral dissertation,
Stanford University.
McLeod, D. M., Eveland, W. P. Jr., & Nathanson, A. I. (1997).
Support for censorship of violent and misogynic rap lyrics: An
analysis of the third-person effect. Communication Research, 24, 153-174.
McMasters, P. (1997, March). It didn't have to come to this.
Quill, 85(2), 18-21.
Mishler, W., & Sheehan, R. S. (1993). The Supreme Court as a
countermajoritarian institution? The impact of public opinion on
Supreme Court decisions. American Political Science Review, 87, 87-101.
Mutz, D. C. (1989). The influence of perceptions of media influence:
Third person effects
and the public expression of opinions. International
Journal of Public
Opinion Research, 1, 2-23.
Paxton, M. (1996a, April). Sex, violence and the third-person
effect: Correlating media effects and attitudes toward censorship.
Paper presented at the annual meeting of the Broadcast Education
Association, Las Vegas, NV.
Paxton, M. (1996b, August). The third-person effect and attitudes toward
expression. Paper
presented at the annual meeting of the Association for
Education in
Journalism and Mass Communication, Anaheim, CA.
Paxton, M. (1997, March). Media use, 'experts" and the third-
person effect. Paper presented to the Association for Education in
Journalism and Mass Communication, Southeast Colloquium, Knoxville, TN.
Perloff, R. M. (1989). Ego-involvement and the third-person
effect of televised news coverage. Communication Research, 16,
236-262.
Perloff, R. M. (1993). Third-person effect research, 1983-1992: A review
and synthesis. International Journal of Public Opinion Research, 5,
167-184.
Price, V., & Tewksbury, D. (1996). Measuring the third-person
effect of news: The impact of question order, contrast and knowledge.
International Journal of Public Opinion Research, 8, 120-141.
ProCD (1996). Select Phone. Danvers, MA.
Prothro, J. W., & Grigg, C. W. (1960). Fundamental principles of
democracy: Bases of agreement and disagreement. Journal of Politics
22, 276-294.
Rojas, H., Shah, D. V., & Faber, R. J. (1995, August). For the
good of others: Censorship and the third-person effect. Paper
presented at the annual meeting of the Association for Education in
Journalism and Mass Communication, Washington, DC.
Rosen, J., & Merritt, Jr. D. (1994). Public journalism: Theory
and practice. Dayton, OH: Kettering Foundation.
Rosen, J. (1993). Community connectedness passwords for public
journalism. St. Petersburg, FL.
Rucinski, D., & Salmon, C. T. (1990). The "other" as the vulnerable
voter: A study of
the third-person effect in the 1988 U.S. presidential
campaign.
International Journal of Public Opinion Research, 2, 345-368.
Salmon, C. T. (1989). Campaigns for social "improvement": An
overview of values, rationales, and impacts. In C. T. Salmon (Ed.),
Information campaigns: Balancing social values and social change (pp.
19-53). Newbury Park, CA: Sage.
Salwen, M. B., & Driscoll, P. D. (1995). Feeling informed? The
"assurance function" of the mass media. International Journal of
Public Opinion Research, 7, 270-275.
Salwen, M. B., & Driscoll, P. D. (1996, August). Self-perceived
knowledge and the third-person effect: Media influence during the
O.J. Simpson trial. Paper presented at the annual meeting of the
Association for Education in Journalism and Mass Communication, Anaheim,
CA.
Schramm, W. (1957). Responsibility in Mass Communication. New
York: Harper & Row.
Seigenthaler, J., & Hudson, D. L., Jr. (1997, March). Going
undercover. Quill, 85(2), 17.
Stouffer, S. (1955). Communism, conformity and civil liberties.
New York: Peter Smith.
Tiedge, J. T., Silverblatt, A., Havice, M. J., & Rosenfield, R.
(1991). Discrepancy between perceived first-person and perceived
third-person mass media effects. Journalism Quarterly, 68, 141-154.
The Twentieth Century Fund (1972). A free and responsive press.
The Twentieth Century Fund task force report for a national news
council. New York: The Twentieth Century Fund.
Vallone, R. P., Ross, L., & Lepper, M. R. (1985). The hostile
media phenomenon: Biased perception and perceptions of media bias in
coverage of the Beirut massacre. Journal of Personality and Social
Psychology, 49, 577-585.
Willnat, L. (1996). Mass media and political outspokenness in
Hong Kong: Linking the third-person effect and the spiral of silence.
International Journal of Public Opinion Research, 8, 187-211.
Zalkind, S. S. (Ed.). (1975). Civil liberties [Special issue].
Journal of Social Issues, 31(2).
Zalkind, S. S., Gaugler, E. A., & Schwartz, R. M. (1975). Civil
liberties attitudes and personality measures: Some exploratory
research. Journal of Social Issues, 31(2), 77-91.
Zellman, G.L. (1975). Antidemocratic beliefs: A survey and some
explanations. Journal of Social Issues, 31, 31-52.
Table 1: Self and Other Media Effects Comparisons*
Others Self
N M sd M sd T-Value
Newspapers 485 5.97 2.22 4.13 2.65 12.63
Television 498 7.18 2.29 4.93 2.78 15.59
Radio 475 5.23 2.20 3.70 2.55 11.73
Media 492 7.09 2.19 5.04 2.69 14.65
* All pairwise comparisons significant at p < .001.
Table 2: Factor Analysis of Self and Others Questions (Varimax Rotation)
Factor Factor
1 2
TV-Self .87 .10
Media-Self .87 .13
News-Self .84 .05
Radio-Self .78 .05
Media-Others .05 .87
TV-Others .07 .86
News-Others .06 .83
Radio-Others .13 .71
Eigenvalue 3.33 2.25
Variance 41.7% 28.2%
Table 3: Effects Question Order (Two-tailed significance).
Effects on Self
Newspaper on Self N Mean (sd) T-Value Sig.
Self First 262 4.32 (2.63)
Others First 276 3.83 (2.64) 2.18 p <.03
Television on Self N Mean (sd) T-Value Sig.
Self First 264 4.99 (2.63)
Others First 274 4.81 (2.89) 0.78 p <.44
Radio on Self N Mean (sd) T-Value Sig.
Self First 254 3.74 (2.69)
Others First 264 3.53 (2.47) 0.94 p <.35
Media on Self N Mean (sd) T-Value Sig.
Self First 266 5.14 (2.71)
Others First 275 4.91 (2.71) 1.00 p <.32
Effects on Others
Newspaper on Others N Mean (sd) T-Value Sig.
Others First 260 5.95 (2.14)
Self First 229 5.97 (2.33) 0.09 p <.93
Television on Others N Mean (sd) T-Value Sig.
Others First 269 7.26 (2.29)
Self First 235 7.02 (2.34) 1.18 p <.24
Radio on Others N Mean (sd) T-Value Sig.
Others First 258 5.02 (2.15)
Self First 224 5.44 (2.26) 2.12 p <.04
Media on Others N Mean (sd) T-Value Sig.
Others First 261 7.21 (2.01)
Self First 235 6.91 (2.33) 1.51 p <.20
Table 4: Hierarchical Regressions Predicting Third-Person Perception, Estimated
Effects on Self and Perceived Effects on Others
Third Self Others
Beta r+ Beta r Beta r
Block 1
Age++ -.07 .00 -.11* -.12** -.19*** -.14**
Education .05 .07 -.02 -.04 .02 .06
Gender -.04 .01 .07 .02 .04 .04
Hispanic .00 .02 -.09* -.08 -.10 -.08
Race -.06 .01 .08 .02 .00 .00
R2 = 0.1+++ R2 = 2.5 R2 = 4.6**
Block 2
Candidate -.01 .01 -.01 -.02 .02 .01
R2 = 0.1 R2 = 2.5 R2 = 4.6*
Block 3
Interest -.23*** -.19*** .36*** .33*** .15** .16**
Newspaper .17** .08 -.10* -.07 .07 .07
Network News -.04 -.11* .02 .09 -.02 -.02
Local News -.06 -.09 .08 .12** .00 .04
Total R2 R2 = 7.4** R2 = 15.6*** R2 = 7.4**
+ Pearson-product moment correlation coefficient with third- person perception
++ Among the dummy variables, gender was male = 1, female = 0; Hispanic
was Yes = 1, No = 0; race was African Americans = 1, Whites = 0, and
"others" excluded from this and subsequent the regression analyses;
Candidate preference was 1 = Dole, 0 = Clinton. Perot excluded from the
analyses.
+++ R2 total at each step in the regression
* p < .05
** p < .01
*** p < .001
Table 5: Hierarchical Regressions Predicting Support for Restrictions With
Perceptual Bias and Self-Others
Support for Restrictions
Beta r+ Beta r
Block 1
Age++ -.06 .02 -.05 .02
Education -.37*** -.36*** -.36*** -.36***
Gender .02 .03 .02 .03
Hispanic -.09 -.08 -.08 -.08
Race .06 .04 .06 .04
R2 = 14.3***+++ R2 = 14.3***
Block 2
Candidate .04 .05 .04 .05
R2 = 14.5*** R2 = 14.5***
Block 3
Interest .05 -.04 .02 -.04
Newspaper .05 -.06 .05 -.06
Network News .05 -.04 .05 -.04
Local News -.01 .03 -.01 .03
R2 = 15.1*** R2 = 15.1***
Block 4
Perceptual Bias .12* .07 ----
Self Effects ---- -.06 .02
Other Effects ---- .14** .15***
Total R2 R2 = 16.4*** R2 = 16.9***
+ Pearson-product moment correlation coefficient with third- person perception
++ Among the dummy variables, gender was male = 1, female = 0; Hispanic
was Yes = 1, No = 0; race was African Americans = 1, Whites = 0, and
"others" excluded from this and subsequent the regression analyses;
Candidate preference was 1 = Dole, 0 = Clinton. Perot excluded from the
analyses.
+++ R2 total at each step in the regression
* p < .05
** p < .01
*** p < .001
[1]
Several nonpartisan groups have called for press oversight. The Hutchin's
Commission on Freedom of the Press (Commission on Freedom of the Press, 1947), a
watershed document that was generally supportive of press freedom, called for an
agency independent of the government and press to oversee press performance.
This recommendation sparked a harsh response from the press. Schramm (1957), who
used the Commission's report to enunciate a social responsibility theory of the
press, wrote of the Commission's recommendations: "The aspect of the
Commission's work which so irritated the press was not the positive doctrine
enunciated in regard to the responsibilities of mass communication, but rather
the fact that outsiders were criticizing the always sensitive press, and, above
all, the fact that the Commission dared to wave a big stick at the press" (pp.
90-91). Twenty-five years later, The Twentieth Century Fund (1972) task force
called for a privately operated national news council to review press coverage.
[2] As Price and Tewksbury (1996) wrote: "Is it indeed the case that people
generally perceive themselves little affected -- and others greatly affected --
by the mass media? Or is it merely that, when asked to make explicit comparisons
between themselves and other people, respondents tend to flatter themselves by
reporting that they are less vulnerable to media impact? If the tendency to
report less media impact on oneself than on others depends critically upon
inducing with the survey instrument an explicit self-versus-others contrast,
then the generality of the effect could well be called into question" (pp.
121-122, original emphasis).
[3] Throughout this paper we use the term "evaluations" to describe media
effects on "self" because respondents do not perceive themselves. We use
"estimations" and "perceptions" to refer to "others" because this involves
perceptions of what others think or believe.
[4] The sample was selected from the Select Phone CD Rom national telephone
directories database (ProCD, 1996). A sample of names and corresponding numbers
was randomly selected from local directories and typed in, with the cursor
moving to a different name with each keystroke, until no more names appeared
with additional keystrokes. When the cursor fell on a name and number that no
longer moved with additional keystrokes, the number was selected. If the cursor
fell on a nonresidential listing, the next residential listing down was
selected. The last digit was randomly changed to include unlisted numbers.
[5] This was the ratio of completed calls to known valid contacts and outright
refusals. The ratio to original selected numbers, the response rate, was about
25%. This excludes respondents rejected because they had not met criteria for
selection in the sample (age and candidate choice). Some attempts were made on
later days to contact respondents who initially refused to complete the surveys.
[6] The way to read the "newspaper on self" condition is respondents evaluated
newspaper effects on themselves Mean = 3.83 when the set of others effects
questions preceded the self effects questions. Respondents evaluated newspaper
effects on themselves Mean = 4.32 when the set of self effects questions
preceded the others effects questions. It is important not to interpret these
findings in the reverse as stating that respondents estimated greater newspaper
influence on others when they first estimated effects on themselves. "Newspaper
(influence) on others" is a separate condition in the table.
[7] See six stories on "punishing the press" that appeared in a special issue
of Columbia Journalism Review (1997). See also Budiansky (1995), Geimann (1997),
McMasters (1997), and Seigenthaler and Hudson (1997). There are several
explanations for the public anger. Rosen and Merritt (1994), in the "public
journalism" tradition, claim that the press has become disconnected with the
public it serves (see also Rosen, 1993).
[8] We can make this suggestion based on our pre-tests of the instrument.
|
