challenge on the third-person
effect
The Third-Person Effect:
Social Cognition or Academic Creation?
Yu-Wei Hu
Professor
Graduate Institute of Mass Communication
National Taiwan Normal University
Taipei, Taiwan, R.O.C.
0118862-23583404
[log in to unmask]
Paper submitted to the Communication Theory and Methodology Division of the
Association for Education in Journalism and Mass Communication, New Orleans
Aug. 4-7, 1999
Abstract
While most studies of the third-person effect assumed that people naturally
think about the influence of media while they are exposed to media
presentations, this study provides evidence to show that this assumption may not
always be true. Some individuals may have never considered the impact of media
on others until researchers ask them to make a judgment. In this case, their
third-person perception is actually an elicited social comparison rather than a
spontaneous social cognition. The mixture of the spontaneous and the elicited
third-person perception may have caused weak connections between the
third-person perception and its behavioral consequences as found in related
literature.
The Third-Person Effect:
Social Cognition or Academic Creation?
INTRODUCTION
It was in the evening on August 5, 1998, a theory and methodology session on
the topic of the third-person effect in AEJMC's annual convention was about to
end. The moderator invited each panelist to make closing remarks. A young
panelist from Asia started to complain about his research experience: "In order
to create an index score of the third-person effect, as many researchers usually
did, I used two questions with similar wordings to measure survey respondents'
assessments of the impact of media on both themselves and others. Strangely
though, while most respondents easily indicated how much they were affected by
media presentations, many of them felt it was difficult to predict the extent to
which media influenced others."
To the young researcher's surprise, this complaint was resonated by the other
panelists in the session. Some American panelists also acknowledged that it is
not easy to make survey respondents evaluate the influence of media on others.
One American panelist further pointed out that when respondents who are
repeatedly urged by interviewers to assess media's impact on others tend to
exaggerate its influence on society as a whole.
At first glance, the difficulty of collecting data for a third-person effect
study seems to be nothing but a trivial methodological problem. After all, it
is not unusual for researchers to find respondents who refuse to answer survey
questions in the process of interviewing. But, still, how can we explain the
phenomenon that survey respondents feel it is more difficult to predict media's
impact on others than to evaluate the impact on themselves? Is this phenomenon
only a question of the technique of writing survey questions or skills of
interviewing? Or, has it already brought up a challenge on the theoretical
assumption of the third-person effect that while attending media presentations,
most people will not only think about the influence of media on themselves but
also predict the impact of media on other members of society? In other words,
could it be the case that many survey respondents cannot predict the extent to
which others will be affected by the media simply because they really have not
thought about this matter?
If most people are not interested in predicting the influence of media on
others, the third-person effect could be more an academic creation than a
general social cognition. In other words, respondents may have not thought
about the impact of media on others, yet, after being urged or encouraged by
survey interviewers, they finally make a judgment on the extent to which others
could be affected by media. And, interestingly enough, according to at least
one researcher's experience, this elicited social cognition tends to exaggerate
others' susceptibility to media's effect.
Of course, this challenge on the assumption of the third-person effect needs
to be examined carefully, that is the main purpose of this study. In fact, if
predicting the influence of media on others is not a general by-product of
media exposure, the theoretical statements of the third-person effect will need
a revised version. In addition, researchers of the third-person effect may need
to go back to investigate why some people are interested in monitoring the
impact of media on others while some other people are simply careless about the
influence of media on their fellows.
To be more specific, the current study will raise the following research
questions:
Q 1: Do most people think about the impact of media on others as they attend
media presentations?
Q 2: No matter how often or how rare people have monitored the impact of media
on others, after being encouraged to predict the influence of media on other
people, will they, as the third-person effect hypothesis states, tend to assign
media's impact on others more than on themselves?
Q 3: If social detection of media's effect is not a general by-product of media
exposure, what social, psychological, and communicative factors could contribute
to this social detection ?
LITERATURE REVIEW
A number of empirical studies have evidence to support the basic hypothesis of
the third-person effect. In general, as Davison (1983) first hypothesized,
people tend to believe others are more susceptible than themselves to the
influence of media. Media content categories that could generate the
third-person effect include (1) news about political or social issues (Atwood,
1994; Cohen, Mutz, Price, & Gunther, 1988; Davison, 1983; Gunther, 1991; Gunther
& Mundy, 1993; Hu & Wu, 1996; 1997; 1998; Mutz, 1989; Perloff, 1989; Price,
Huang, & Tewksbury, 1997; Rucinski & Salmon, 1990; Salwen, 1997; Salwen,
Dupagne, & Paul, 1998; Tiedge, Silverblat, Havice, & Rosenfeld, 1991; Vallone,
Ross, Lepper, 1985; White, 1995 ), (2) Political debate (Rucinski & Salmon,
1990), (3) advertisements (Cohen & Davis, 1991; Davison, 1983; Gunther & Mundy,
1993; Gunther & Thorson, 1992; Rucinski & Salmon, 1990; Salwen, Dupagne, & Paul,
1998 ), (4) TV dramas (Lasorsa, 1989), and (5) pornography or violence (Gunther,
1995; Lee & Yang, 1996; Lo & Paddon, 1998; Rojas, Shah, & Faber, 1995).
During the past decade, researchers have also found that the third-person
effect could be enhanced by several contingent factors. These factors are: (1)
perceived bias (Cohen, et al., 1988), low credibility (Gunther, 1991), or
persuasive intention (Rucinski & Salmon, 1990) of information sources, (2)
perceived undesirability or harm of media content (Gunther & Mundy, 1993;
Gunther & Thorson, 1992), (3) the individual's social or psychological
characteristics including age, educational level, perceived self-expertise, and
issue involvement or ego-involvement (Cohen & Davis, 1991; Lasorsa, 1989; Mutz,
1989; Perloff, 1989; Vallone et al., 1985; Rucinski & Salmon, 1990; Tiedge et
al., 1991), (4) perceived remoteness of the third-person (Cohen et al., 1988;
Cohen & Davis, 1991; Gunther, 1991; Park, 1997; White 1995), (5) the
individual's communication behavior (Rucinski & Salmon, 1990; Salwen, 1997), and
(6) individualism as a cultural factor (Park, 1997). The behavioral
consequences of the third-person effect, as suggested by Davison (1983), has
also been partly confirmed in several studies (Gunther, 1995; Lee & Yang, 1995;
Rojas et al., 1995; Salwen, 1997)
In spite of these findings that show the third-person effect may be contingent
on these social, psychological, cultural, and communicative factors, the
majority of literature seems to suggest that-with the exception of Gunther and
Thorson who found that positive emotional reactions to TV ads could stimulate a
reversed third-person effect-- the third-person effect is a general response to
various media presentations. While many studies support the basic hypothesis of
the third-person effect, from the prototype presented by Davison in 1983 to
current studies completed by other scholars, none have questioned the assumption
of the basic hypothesis.
The hypothesis of the third-person effect assumes that people naturally assess
the impact of media on others while they are exposed to media presentations.
While the results of the assessment may vary from person to person, it is
assumed that people will evaluate the impact of media on others. Thus, for
Davison and the other scholars who tested the hypothesis of the third-person
effect, the typical procedure of this testing is simply to ask survey
respondents or experimental subjects to assess the impact of media on both
themselves and others, and then to see whether the perceived media effect on
others is significantly greater than the perceived effect on themselves. And,
as mentioned above, no matter what contingent factors that have been brought in
to test the hypothesis of the third-person effect, the assumption of social
detection of media's effect has never been questioned.
The third-person effect is not the only theoretical notion that bears the
assumption of social detection in mind. For example, Noelle-Neumann (1974;
1977) raised the theoretical concept of the spiral of silence and argued that
most people will use their quasi-statistical organs to detect the opinion
climate of public issues. In almost every study of the spiral of silence,
Noelle-Neumann not only measured survey respondents' opinions about specific
issues, but also asked them to predict the opinions of others towards the same
issues. This measurement procedure was actually based on the assumption of
social detection.
In addition to the spiral of silence, other theoretical constructs related to
the social-psychological approach of public opinion study, such as pluralistic
ignorance (Fields & Schuman, 1976), the false consensus (Ross, Greene, & House,
1977), and the looking glass perception (Fields & Schuman, 1976), also base
their thinking on social detection. Researchers simply assume that most people
will think about others' opinions on various issues. People may or may not
accurately predict the mainstream of public opinion, but they are always
expected to be in tune with the broader climate of opinion for various issues
from time to time. As Vincent Price and Hayg Oshagan (1995, p. 178) indicated,
applying the social- psychological perspective to the study of public opinion is
to understand " the various social forces that impinge upon individuals as they
think about, understand, and express their opinions." This definition implies
that the detection of others' thinking has a strong influence on the way
individuals establish their own opinions of public issues.
Still, it is quite unlikely that every person will be interested in every public
issue in any society. As James E. Grunig (1982) suggested, whether an
individual can be identified as a member of the active public is purely
"situational." According to Grunig's (1983; 1989) situational theory, when
people encounter a public issue, those who have a lower level of problem
recognition and issue involvement but a stronger perception of external
constraint are more likely to engage in information processing than information
seeking. Which means some people may pay little attention to information about
issues they consider to be trivial, uncontrollable or irrelevant.
Once a person holds a passive attitude toward a public issue, will it be likely
that he or she still cares about the effect information about the issue has on
others? Or, will members of the passive public spend less energy in detecting
the extent to which others could be affected by the media's coverage of an
issue? In other words, is the third-person effect a universal communication
effect or just a situational effect of media presentations? This question
involves the re-examination of the assumption that social detection of media's
effect is a general by-product of media exposure.
The present study raises three arguments based on the preceding discussion
which questions the theoretical assumption of the third-person effect:
1: People may not always think about the impact of media on others while they
attend media presentations. This argument challenges the assumption of the
third-person effect hypothesis.
2: Being asked to assess the influence of media on others and the self, most
people tend to indicate that others are more susceptible to media's effect than
themselves. While this argument is consistent with the third-person effect
hypothesis, it emphasizes that sometimes the third-person effect could be simply
an elicited rather than a spontaneous social cognition.
3: The social detection of the impact of media can be a function of an
individual's social background, psychological condition and communication
behaviors. Based on the literature of the third-person effect and the
situational theory mentioned above, the current study proposes that higher
social and economic status (SES) level, greater involvement and greater media
use or interpersonal communication may contribute to a stronger intention to the
social detection of media's effect. Just as Davison (1983) suggested, those who
are more knowledgeable or more concerned about public issues may be more
sensitive to the influence of media reports on other members of society.
These arguments will be tested empirically. Were they supported by empirical
data, the arguments will bring in critical revisions to the theoretical notion
of the third-person effect.
METHODS
Data Collection
A telephone survey was conducted in Taipei city, Taiwan three weeks before the
1998 mayoral election. The coverage of the mayoral race was the top story in
most local media from early October to late November. Since the Taipei mayoral
election was closely watched and the two leading candidates had issued formal
complaints about the unfairness of some campaign reports, it will be interesting
to see whether most residents were sensitive to the influence of campaign
stories on others' voting tendencies. Or, as the present study argues, those
who were not interested in the campaign may have never considered the potential
impact of campaign stories on other voters.
The city's residential phone directory was chosen as the sampling frame and a
total of 1,000 phone numbers were selected by using the method of systematic
sampling. The interviewers were graduate students from a research methodology
course at Taiwan Normal University. The survey generated 616 completed
questionnaires. All of the respondents were residents of Taipei city and
eligible voters.
Measurement of Variables
The main purpose of the study was to examine individuals' social detection of
media's effect. Four variables were used to measure the extent to which
respondents had thought about the impact of media on others' voting tendencies
in the mayoral electoral campaign of Taipei city. The operational definitions
of these variables are:
(1) Have you ever thought about the impact of TV election reports on others'
voting tendencies (1 = never; 2 = sometimes; 3 = quite often)?
(2) Have you ever thought about the impact of newspaper election reports on
others' voting tendencies (1 = never; 2 = sometimes; 3 = quite often)?
(3) Have you ever thought about the influence of media's coverage of political
debate on others' voting tendencies (1 = never; 2 = sometimes; 3 = quite often)?
(4) Have you ever thought about the influence of election poll reports on
others' voting tendencies (1 = never; 2 = sometimes; 3 = quite often)?
In addition to the four variables related to social detection of media's
effect, the present study also asked respondents, no matter how often or how
rare they had thought about the influence of media on others, to assess the
impact of TV stories, newspaper reports, political debate and election poll
reports on themselves and others. These variables were measured by asking
respondents the following questions:
(1) How much do you think the TV reports about the mayoral election have
affected your voting tendency (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(2) How much do you think the TV reports have affected others' voting
tendencies (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(3) How much do you think the newspaper reports have affected your voting
tendency (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(4) How much do you think the newspaper reports have affected others' voting
tendencies (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(5) How much do you think the political debate has affected your voting
tendency (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(6) How much do you think the political debate has affected others' voting
tendencies (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(7) How much do you think the election poll reports have affected your voting
tendency (1 = none; 2 = not much; 3 = some; 4 = a lot)?
(8) How much do you think the election poll reports have affected others'
voting tendencies (1 = none; 2 = not much; 3 = some; 4 = a lot)?
Respondents' social backgrounds include four measures: (1) gender ( 1 = male; 2
= female), (2) age, (3) educational level (1 = illiterate; 2 = primary school; 3
= junior high school; 4 = senior high school; 5 = college; 6 = graduate school),
(4) personal monthly income (1 = less than N.T.$ 5,000; 2 = N.T.$ 5,001~10,000;
3 = N.T.$ 10,001~20,000; 4 = N.T.$ 20,001~30,000; 5 = N.T.$ 30,001~ 40,000; 6 =
N.T.$ 40,001~50,000; 7 = N.T.$ 50,001~60,000; 8 = N.T.$ 60,001~70,000; 9 = N.T.$
70,001~80,000; 10 = N.T.$ 80,001~90,000; 11 = N.T.$ 90,001~100,000; 12 = more
than N.T.$ 100,000).
Respondents' psychological characteristics were measured by asking the
following questions:
(1) How important is the mayoral election to you personally? (1 = not important
at all; 2 = not very important; 3 = somewhat important; 4 = very important)
(2) How closely did you follow the mayoral election campaign? (1 = not
attentive at all; 2 = not very attentive; 3 = somewhat attentive; 4 = very
attentive)
(3) How interested are you in the mayoral election campaign? (1 = not
interested at all; 2 = not very interested; 3 = somewhat interested; 4 = very
interested)
The measures of these three variables were combined into an index of political
involvement. The internal consistency reliability of this index (Cronbach
Alpha) is .70.
(4) How much do you know about the mayoral election campaign? (1 = none; 2 =
not much; 3 = some; 4 = a lot)
(5) How much influence will your vote have on the mayoral election results? (1
= none; 2 = not much; 3 = some; 4 = a lot)
In addition to social backgrounds and psychological states, variables
concerning respondents' communication behaviors were another set of predictors
used to detect the impact of media on society. Respondents were asked how much
mayoral election information they obtained from TV and newspapers? (1 = none; 2
= not much; 3 = some; 4 = a lot). Respondents were also asked whether they
watched the political debate on TV (1 = yes; 2 = no) and how often they were
exposed to the media's election poll reports? (1 = never; 2 = not very often; 3
= sometimes; 4 = very often) The interpersonal communication variables were
measured by asking respondents how often they have talked to family, friends and
colleagues about the mayoral election? (1 = never; 2 = not very often; 3 =
sometimes; 4 = very often). The measures of these interpersonal communication
variables were then combined into an index with a .73 Cronbach Alpha.
Data Analysis
The first argument proposed in this study is that people may not always think
about the impact of media on others. The descriptive statistics of the four
social detection variables will provide evidence to support or reject this
argument.
The second argument is that no matter whether people had thought about the
impact of media on others or not, after being asked to assess its influence on
others and themselves, most people tend to feel others are more likely to be
swayed by media than themselves. This argument will be examined by running
paired t-tests to compare the perceived impact of campaign information on others
and the self.
The third argument of this study suggests that social detection of media's
effect is a function of individuals' social background, psychological
characteristics and communication behaviors. Four multiple hierarchical
regression analyses will be used to investigate the correlations between the
social detection variables and the variables concerning the individual's social
and psychological characters as well as communication behaviors.
RESULTS
The first argument raised in this study is that social detection of media's
effect may not be a general by-product of media exposure. In other words, when
people attend media presentations, most individuals may not think about the
impact of media on others. Table 1 presents some evidence to support this
argument.
As Table 1 shows, during Taipei city's mayoral election campaign, over 90
percent of respondents acquired "some" or "a lot" of campaign information from
television reports, yet, as many as 40.5 percent of respondents had never
considered the effect of television reports on others' voting tendencies.
Results reveal that 50.4 percent of respondents occasionally thought about the
effect of television reports on others' voting tendencies, yet, only 9.1 percent
of respondents said they frequently thought about the potential impact of
television on other voters. However, a more interesting point is that, no
matter how rare or how often respondents engaged in social detection of media's
effect, after being asked to assess the impact of television reports on others,
most respondents made a judgment that others could be affected by television
reports. Results show that only 10.6 percent of respondents insisted they could
not make this kind of prediction.
Table 1 also shows that while 78 percent of respondents have obtained "some" or
"a lot" of information about the election campaign from newspaper reports, 45.5
percent of respondents had never thought about the effect of these reports on
others' voting tendencies. Only 8.3 percent of respondents had frequently
detected the impact of newspaper reports on other voters. However, while being
asked by interviewers to evaluate the impact of newspaper reports on other
voters, most respondents finally made a judgment on the effect newspaper reports
had on others, while only 13.3 percent of respondents still could not make such
a prediction.
Respondents' social detection of media's effect of political debate and
election poll reports are quite similar to that of television and newspaper
reports. Table 1 indicates that while 88.4 percent of respondents watched the
political debate on television, 41 percent of respondents had never thought
about the potential impact of the debate on other voters. Those who had
frequently considered the effect of the debate on others only amount to 7.2
percent of respondents. Again, when asked to assess the impact of the debate on
others, 10.1 percent of respondents could not predict the impact of the debate
at a social level.
Finally, as Table 1 presents, while 70 percent of respondents had occasionally
or frequently been exposed to election poll reports during the mayoral election
campaign, 47.2 percent of respondents had never thought about the impact of the
reports on other voters. Only 2.8 percent of respondents had frequently thought
about the effect of the reports on others. Similar to the preceding findings,
after being asked to do so, most respondents were willing to predict the
potential impact of election poll reports on other voters except for 14.2
percent of respondents who did not make such a judgment.
Overall, Table 1 indicates that while most respondents had attended various
media presentations about the mayoral election campaign, almost half of them had
never thought about the impact of media on others' voting tendencies. However,
while being asked to assess media's effect of campaign information at a social
level, most respondents made such a judgment without difficulty. This finding
supports the argument of this study that social detection of media's effect
could be an elicited response to media presentations rather than a universal
reaction to mass communications.
The second argument of this study is that, no matter whether people have
engaged in the social detection of media's effect or not, while being asked to
assess media's impact on others and themselves, they tend to believe that others
are more susceptible to media's effect than themselves. While this argument is
consistent with the basic hypothesis of the third-person effect, it emphasizes
that sometimes the third-person effect could be simply an elicited response to
media presentations rather than a spontaneous comparison that individuals make
on their own.
The results of four t-tests presented in Table 2 support the second argument of
the present study. In general, most respondents believe that (compared to
themselves) other voters are more susceptible to the effect of campaign stories
covered by television (t = 22.77, df = 549, p < .001) and newspapers (t = 25.35,
df = 531, p < .001), political debate (t = 20.69, df = 551, p < .001) as well as
election poll reports (t = 30.94, df = 524, p < .001). Again, it should be
noted that since almost half of respondents had not thought about the influence
of media on other voters until the survey interviewers asked them to make a
judgment, the third-person effect found in this study may partly be an academic
creation rather than a spontaneous reaction from respondents to media campaign
information.
The third argument of this study is that social detection of media's effect
could be a function of an individual's social background, psychological
characteristic and communication behaviors. Based on the literature of the
third-person effect and the situational theory, higher SES level, stronger
involvement and greater media exposure as well as interpersonal communication
may contribute to a stronger tendency to monitor the influence of media on other
people. The results of four hierarchical multiple regression analyses presented
in Table 3 to Table 6 support this argument.
Table 3 shows that those who had spent more time to detect the impact of
television reports on other voters can be characterized as being younger, better
educated, having greater campaign involvement, stronger self- perception of
campaign expertise, a strong reliance on information from television reports and
more frequent conversation with others about campaign issues.
Table 4 also indicates that the characteristics of younger age, better
education, greater involvement and perceived self-expertise, more exposure to
newspaper reports and frequent interpersonal communication are significantly
correlated to social detection of the impact of newspaper reports.
The results presented in Table 5 and Table 6 are basically parallel to that of
Table 3 and Table 4. Those who had spent more time thinking about the impact of
political debate on others are more likely to be better educated, younger
voters. These people are also characterized as having greater campaign
involvement and spending more time discussing the campaign issues with others.
Higher educational level, greater involvement and frequent interpersonal
communication are also significant predictors of the energy people spend to
detect the effect of election poll reports on others' voting tendencies.
CONCLUSION
During the past decade, the third-person effect had become a significant
research topic that received much attention from many communication scholars.
While most studies of the third-person effect assumed that people naturally
think about the impact of media on others while they are exposed to media
presentations, the current study provides evidence to show that this assumption
may not always be true.
As the survey results indicate, during the mayoral election campaign of Taipei
city in 1998, most respondents obtained campaign information from television and
newspaper reports. In addition, most respondents watched the television program
of political debate and frequently took notice of election poll reports carried
by the local media. However, almost half of respondents had never thought about
the impact of various campaign information on other voters. Those who often
detected media's impact on other voters amount to less than 10 percent of the
respondents. It is especially interesting to note that while both of the two
leading candidates of the mayoral election publicly expressed their worries
about the potential impact of election polls on individuals' voting tendencies,
only 2.8 percent of respondents often thought about how others would be affected
by election poll reports.
Although almost half of respondents had never thought about media's impact on
other voters, the present study still asks respondents to assess the influence
of media's campaign information on others and themselves. In general, these
results suggest that-- parallel to the findings of many studies of the
third-person effect-- people believe others are more likely to be affected by
media's campaign information than themselves. However, since many respondents
had never thought about the impact of media on other voters until survey
interviewers asked them to do so, their perception of the third-person effect of
various campaign information is actually an elicited social cognition.
This elicited third-person perception may be quite different from the
spontaneous one of those who are engaged in the social detection of media's
effect. As the results of the regression analyses demonstrate, respondents who
are younger and better educated, and those who have greater involvement and more
exposure to campaign information from the media and interpersonal channels are
more likely to think about media's impact on other voters. Theoretically
speaking, this active public's spontaneous third-person perception is primarily
a reflection of their concern about media's influence on election results.
On the other hand, those who have spent less time monitoring the impact of
media on other voters tend to be older and less educated. They have less
involvement and less exposure to campaign information from media and
interpersonal channels. The passive public seems to be quite apathetic when it
came to details concerning the election campaign. Therefor, their elicited
third-person perception is less likely to indicate an innate concern about
media's influence on election results. These individuals may give a response in
order to maintain their self-esteem by overestimating others' susceptibility and
their own invulnerability to media's campaign information. Furthermore,
compared to the elicited third-person perception, the spontaneous third-person
perception may be a good reason to give premise to support the regulation of
media coverage during election campaign. In other words, the two kinds of
third-person perceptions may be driven by different psychological mechanisms and
result in different behavioral consequences.
Overall, the research findings of this study make some suggestions which
further discuss the theoretical notions of the third-person effect. First, it
should be noted that social detection of media's effect may not be a natural
by-product of media exposure. While previous studies assumed that people think
about the impact of media on others as they attend media presentations, this
study shows evidence to support the argument that social detection of media's
effect is a variable rather than an assumed phenomenon.
Second, although some people have never thought about the impact of media on
others, they may still be asked to make such a judgment. In this case, the
third-person perception is actually an elicited social cognition rather than a
spontaneous reaction to media presentations.
Third, since social detection of media's influence is more likely to be a
behavioral intention of the active public than the passive public, the active
public's spontaneous third-person perception should more likely reflect
individuals' genuine concern about the social consequences of media's
performances. This spontaneous third-person perception may be more likely to
generate some consequent behaviors as a means of solving social problems caused
by media. Simply testing the basic hypothesis of the third-person effect
without differentiating the spontaneous third-person perception from the
elicited one could impede the thorough understanding of the third-person effect.
In addition, the mixture of the spontaneous and the elicited third-person
perception may have caused weak connections between the third-person perception
and its behavioral consequences as found in related literature. Thus, it may be
worthwhile for future studies to investigate the interaction effect of social
detection and the strength of the third-person perception on the behavioral
consequences of the third-person effect.
TABLE 1. Media Exposure, Social Detection, and Perceived Impacts of Campaign
Information on Others
_____________________________________________________________________
television reports
____________________________________________________________________
information acquired social detection perceived impacts
% N % N % N
none 2.4 15 never 40.5 248 none 6.3 39
not much 5.0 31 sometimes 50.4 309 not much 10.6 65
some 32.8 202 quite often 9.1 56 some 50.8 313
quite a lot 59.7 368 quite a lot 21.8 134
cannot 10.6 65
predict
_____________________________________________________________________
newspaper reports
_____________________________________________________________________
information acquired social detection perceived impacts
% N % N % N
none 10.6 65 never 45.5 280 none 6.8 42
not much 11.4 70 sometimes 46.1 283 not much 11.5 71
some 29.5 182 quite often 8.3 51 some 52.8 325
quite a lot 48.5 299 quite a lot 15.6 96
cannot 13.3 82
predict
_____________________________________________________________________
political debate
_____________________________________________________________________
attendance to debate social detection perceived impacts
% N % N % N
yes 88.4 541 never 41.0 250 none 6.8 42
no 11.5 71 sometimes 51.8 316 not much 10.7 66
quite often 7.2 44 some 56.3 346
quite a lot 16.1 99
cannot 10.1 62
predict
_____________________________________________________________________ election
poll reports
_____________________________________________________________________
frequency of attendance social detection perceived impacts
% N % N % N
never 17.4 107 never 47.2 287 none 14.8 91
not much 12.5 77 sometimes 50.0 304 not much 13.9 85
some 50.2 309 quite often 2.8 17 some 51.4 315
quite a lot 19.8 122 quite a lot 5.7 35
cannot 14.2 87
predict
_____________________________________________________________________
TABLE 2. Perceived Impacts of Campaign Information on Others and the Self
_____________________________________________________________________
television reports (df = 549)
mean t-value
others 2.98 22.77***
self 1.93
_____________________________________________________________________
newspaper reports (df = 531)
mean t-value
others 2.89 25.35***
self 1.77
_____________________________________________________________________
political debate (df = 551)
mean t-value
others 2.91 20.69***
self 1.93
_____________________________________________________________________
election poll reports (df = 524)
mean t-value
others 2.56 30.94***
self 1.28
_____________________________________________________________________
*** p < .001
TABLE 3. Hierarchical Regression Analysis: Social Detection of Television
Effects
(Beta, N = 616)
_____________________________________________________________________
Predictor Regression 1 Regression 2 Regression 3
_____________________________________________________________________
gender -.05 -.05 -.06
age -.14** -.16*** -.16***
educate .18*** .13** .12**
income -.0003 -.008 -.03
involvement .13** .05
self-expertise .12** .09*
political efficacy .02 -.02
media exposure .10*
interpersonal .21***
communication
_____________________________________________________________________
R2 .07 .12 .17
Adjusted R2 .07 .11 .16
R2 Change .07 .05 .05
Sig. of Change .0001 .0001 .0001
_____________________________________________________________________
* p < .05; ** p <.01; *** p <.001
TABLE 4. Hierarchical Regression Analysis: Social Detection of Newspaper Effects
(Beta, N = 616)
_____________________________________________________________________
Predictor Regression 1 Regression 2 Regression 3
_____________________________________________________________________
gender -.04 -.04 -.06
age -.09* -.11* -.11**
educate .23*** .20*** .15**
income .04 .03 -.02
involvement .09* .02
self-expertise .09* .04*
political efficacy .02 -.02
media exposure .19***
interpersonal .16***
communication
_____________________________________________________________________
R2 .09 .11 .17
Adjusted R2 .09 .10 .16
R2 Change .09 .02 .06
Sig. of Change .0001 .001 .0001
_____________________________________________________________________
* p < .05; ** p <.01; *** p <.001
TABLE 5. Hierarchical Regression Analysis: Social Detection of Political Debate
(Beta, N = 616)
_____________________________________________________________________
Predictor Regression 1 Regression 2 Regression 3
_____________________________________________________________________
gender .01 .0008 -.01
age -.13** -.15*** -.15***
educate .15*** .11* .11*
income .007 .005 -.03
involvement .20** * .15**
self-expertise .02 .02
political efficacy .06 .04
media exposure -.02
interpersonal .18***
communication
_____________________________________________________________________
R2 .06 .11 .14
Adjusted R2 .05 .10 .13
R2 Change .06 .05 .03
Sig. of Change .0001 .0001 .0002
_____________________________________________________________________
* p < .05; ** p <.01; *** p <.001
TABLE 6. Hierarchical Regression Analysis: Social Detection of Election Poll
(Beta, N = 616)
_____________________________________________________________________
Predictor Regression 1 Regression 2 Regression 3
_____________________________________________________________________
gender -.02 -.03 -.04
age -.02 -.04 -.05
educate .23*** .19*** .18***
income .06 .06 .02
involvement .15** * .06
self-expertise .08 .06
political efficacy .01 -.01
media exposure .10*
interpersonal .20***
communication
_____________________________________________________________________
R2 .07 .12 .16
Adjusted R2 .07 .11 .15
R2 Change .07 .05 .04
Sig. of Change .0001 .0001 .0001
_____________________________________________________________________
* p < .05; ** p <.01; *** p <.001
REFERENCES
Atwood, L. E. (1994). Illusions of media power: The third-person effect.
Journalism Quarterly, 71, 269-281.
Cohen, J., & Davis, R. G. (1991). Third-person effects and the differential
impact in negative political advertising. Journalism Quarterly, 68, 680-688.
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.
Davison, W. P.(1983). The third-person effect in communication. Public Opinion
Quarterly, 40, 1-15.
Fields, J., & Schuman, H. (1976). Public beliefs about the beliefs of the
public.
Public Opinion Quarterly, 40, 427-448.
Grunig, J. E. (1982). The message- attitude-behavior relationship: Communication
behaviors of organizations. Communication Research, 9, 63-200.
Grunig, J. E. (1983). Communication behaviors and attitudes of environmental
publics: Two studies. Journalism Monographs, No. 81.
Grunig, J. E. (1989). Publics, audiences and market segments: Segmentation
principles for campaigns. In C. T. Salmon (Ed.), Information campaigns:
Balancing social values and social change (pp. 199-233). Newbury Park,
CA: Sage.
Gunther, A. (1991). What we think others think: Cause and consequence in the
third- person effect. Communication Research, 18, 355-372.
Gunther, A. (1995). Overrating the X-Rating: The third-person perception and
support
for censorship of pornography. Journal of Communication, 45, 27-38.
Gunther, A., & Mundy, P. (1993). Biased optimism and the third-person effect.
Journalism Quarterly, 70, 57-67.
Gunther, A., & Thorson, E. (1992). Perceived persuasive effects of product
commercials and public service announcements: Third-person effects in
new domains. Communication Research, 19, 574-596.
Hu, Y. W., & Wu, Y. C. (1996). 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.
Hu, Y. W., & Wu, Y. C. (1997). The third-person effect of election news: The
synthesis of contingent factors into a causal model. Paper presented at the
annual meeting of the Association for Education in Journalism and Mass
Communication, Chicago, IL.
Hu, Y. W., & Wu, Y. C. (1998). The "critics", "believers", and "outsiders" of
election polls: Comparing characteristics of the third-person effect,
first-person effect and consensus effect. Paper presented at the annual
meeting of the Association for Education in Journalism and Mass
Communication, Baltimore, Md.
Lasorsa, D. L. (1989). Real and perceived effect of 'Amerika'. Journalism
Quarterly,
66, 373-378.
Lee, C., & Yang, S. (1996). 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 Eduacation in Journalism and Mass
Communication, Anaheim, CA.
Lo, V., & Paddon, A. R. (1998). The third-person perception and support
for restrictions of pornography: Some methodological problems. Paper
Presented at the annual meeting of the Association for Education in
Journalism and Mass Communication, Baltimore, Md.
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, 3-23.
Noelle-Neumann, E. (1974). The spiral of silence: A theory of public opinion.
Journal of Communication, 24, 43-51.
Noelle-Neumann, E. (1977). Turbulences in the climate of opinion:
Methodological application of the spiral of silence theory. Public Opinion
Quarterly, 41, 113-151.
Park, E. (1997). Individualism/collectivism, cultural variations in self-concept
and their implications for the false-uniqueness hypotheses: The cases of
third-person effect and optimistic illusion. Paper presented at the annual
meeting of the Association for Education in Journalism and Mass Communication,
Chicago, IL.
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., Huang, L. N., & Tewksbury, D. (1997). Third-person effects of
news coverage: Orientations toward media. Journalism and Mass
Communication Quarterly, 74, 525-540.
Price, V., & Oshagan, H. (1995). Social-psychological perspectives
on public opinion. In T. L. Glasser, & C.T. Salmon (Eds.), Public opinion
and the communication of consent (pp. 249-277). New York:
The Guilford Press.
Rojas, H., Shah, D.V. & Faber, R. (1995). 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,
D.C.
Ross, L., Greene, D., & House, P. (1977). The 'false consensus effect':
An egocentric bias in social perception and attribution process. Journal
of Experimental Social Psychology, 13, 279-301.
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.
Salwen, M. B. (1997). Perception of media power and fairness in campaign '96:
The
third-person effect and support for press restrictions. Paper presented at the
annual meeting of the Association for Education in Journalism and Mass
Communication, Chicago, IL.
Salwen, M. B., Dupagne, M., & Paul, B. (1998). Perceptions of media power
and moral influence: Issue legitimacy and the third-person effect. Paper
presented at the annual meeting of the Association for Education in
Journalism and Mass Communication, Baltimore, Md.
Tiedge, J. T., Silverblatt, A., Havice, M. J., & Rosenfeld, R. (1991).
Discrepancy between perceived first-person and perceived third-person mass
media effects.
Journalism Quarterly, 68, 141-154.
Vallone, R. P., Ross, L., & Lepper, M. R. (1985). The hostile media phenomenon:
Biased perceptions and perceptions of media bias in coverage of the Beirut
massacre. Journal of Personality and Social Psychology, 49, 577-588.
White, H. A. (1995). Issue involvement and argument strength as mediating
factors
in the third-person effect. Paper presented at the annual meeting of
Association
for Education in Journalism and Mass Communication, Washington, D.C.
|
