This paper was presented at the Association for Education in Journalism and
Mass Communication in Toronto, Canada, August 2004.
If you have questions about this paper, please contact the author
directly. If you have questions about the archives, email
[log in to unmask] For an explanation of the subject line, send email to
[log in to unmask] with just the four words, "get help info aejmc," in the
body (drop the "").
Exploring a Reinforcement Model of Perceived Media Influence on Self and Others
Mary Beth Oliver
Pennsylvania State University
Paper submitted to the Comm Theory and Methodology Division of the
Association for Education in Journalism and Mass Communication.
Correspondence concerning this manuscript should be sent to Mary Beth
Oliver, College of Communications, 210 Carnegie Building, University Park,
PA, 16801, Phone: (814) 863-5552; email: [log in to unmask]; FAX: (814) 863-8161.
AV Needs: Overhead Projector
Exploring a Reinforcement Model of Perceived Media Influence on Self and
An experiment was conducted to explore the idea that when making estimates
of media influence on self and others, individuals employ a heuristic of
assuming reinforcement of existing attitudes. Support for the reinforcement
model was obtained, and was further applied to perceptions of the effects
of media violence on the self versus others. Results are discussed in terms
of providing a framework for interpreting third-person perceptions.
Exploring a Reinforcement Model of Perceived Media Influence on Self and
When making judgments about media influence on individuals' attitudes,
beliefs, and behaviors, what types of variables do most people consider? If
the typical person considered all of the variables that researchers have
pointed to as important in moderating and mediating media influence,
hundreds of variables would be considered, including historical contexts,
viewing situations, and individual differences, among others. However,
judgments of media influence, as with many judgments, likely fall prey to
heuristic processing, with individuals employing simple "rules of thumb" to
guide their estimates (Tversky & Kahneman, 1974). In this regard, research
on third-person perceptions suggests that one "rule of thumb" that people
may employ focuses on the recipient of media influence: the self versus
others. In general, a substantial and growing body of literature provides
consistent evidence for the tendency of individuals to estimate that other
people are more susceptible to harmful media influence than the self (see,
for example, Davison, 1983; Perloff, 1993, 2002).
Given that numerous studies have demonstrated that third-person
perceptions are robust and consistent, what mechanisms might explain
people's tendencies to distinguish between the self and others? To date,
numerous variables have been identified, including variables relating to
self-esteem, variables pertaining to estimates of media exposure, and
variables pertaining to perceived social distance of others, among many
others (for reviews, see Perloff, 1993, 1999, 2002). The purpose of the
present research is to contribute to the literature on explanatory
mechanisms by suggesting a model (or heuristic) that individuals may use
when estimating media influence on the self and others. Namely, this paper
explores the idea that people use a "reinforcement model" when making
judgments, using their own attitudes and their perceptions of others'
attitudes as baselines, and using these baselines to make estimates
concerning the strength and magnitude of media influence that acts to
reinforce existing attitudes.
Self-Enhancement and Third-Person Perceptions
In considering the multitude of possible explanations for third-person
perceptions that have been explored by researchers, Perloff (2002)
suggested that explanations focused on self-enhancement motivations have
received the preponderance of support. Of course, judgments of differential
media influence on the self versus others in the enhancement of self image
can take many forms, including perceptions that others are less critical or
literate media consumers, perceptions that others are simply more
vulnerable or susceptible to media influence, and perceptions that others
are more frequent consumers of media content that may cause harm, among
others (Brosius & Engel, 1996; Eveland, Nathanson, Detenber, & McLeod, 1999).
Consistent with all of these motivations related to self-enhancement,
research that has explored perceived effects of a wide variety of "harmful"
content (e.g., violence, pornography, misogynistic rap music) has reported
a tendency for individuals to perceive that others are more influenced, and
particularly when the "others" are socially distant from the self and/or
when others are assumed to have greater exposure to the media content in
question (Eveland et al., 1999; Gunther, 1995; Gunther & Hwa, 1996; Hoffner
& Buchanan, 2002; Lo & Wei, 2002). In contrast, this same body of
literature suggests that under some circumstances, individuals may perceive
that they are more influenced than others (a first-person perception),
though these estimates appear to be most common in situations in which
media influence is thought to have beneficial outcomes (Brosius & Engel,
1996; Duck, Hogg, & Terry, 1999; Gunther & Mundy, 1993; Henriksen & Flora,
1999). In general, then, research concerning perceived media influence of
content that is "harmful" or "beneficial" provides support for the general
idea that individuals tend to perceive media influence on the self versus
on others in ways that likely serve ego-enhancing ends.
Viewers' Models of Media Influence
In addition to suggesting that perceived media influence may reflect
self-enhancement motivations, several researchers have suggested that
people tend to employ short-cuts or heuristics in guiding their estimates.
Namely, one underlying assumption that people appear to employ is the
notion that media may have powerful or direct effects – but particularly on
others' attitudes. For example, McLeod, Detenber, and Eveland (2001)
employed a similar interpretation in their research concerning individuals'
estimates of the effects of exposure to violent and/or misogynistic music.
Although individuals' estimates of the influence of such music on the self
depended on their assessment of their own level of "common sense," their
estimates for others were predicted by the extent to which they believe
that others were exposed to such content. These authors interpreted their
findings as consistent with the fundamental attribution error and as
suggestive of the idea that people perceive media influence on others in
"magic bullet" terms: "With regard to media effects on others, people seem
to use a very simple heuristic: Exposure equals influence" (p. 692).
Individuals' tendencies to believe that media exposure has a strong impact
on others' attitudes is a concept that has been explored extensively by
Gunther and his colleagues (see, for example Gunther, 1998; Gunther & Chia,
2001; Gunther & Christen, 2002; Gunther, Christen, Liebhart, & Chia, 2001).
Namely, these authors have argued that when estimating public opinion,
individuals assume that the media are "persuasive" in affecting public
opinion. Consequently, if individuals perceive the media as representing a
given slant on an issue, the direction of this perceived slant should
predict the position that the public is thought to hold, as the public is
assumed to be more affected by media content than is the self.
The Importance of Existing Attitudes and Perceptions of Media Reinforcement
Whereas the notion that the media play a powerful, direct role in
affecting others' attitudes is consistent with public concerns surrounding
"harmful" media content (e.g., violent video games, television violence,
sexuality explicit media), additional researchers have highlighted the
importance of viewers' existing attitudes as moderators of estimated
influence. For example, Cohen and Davis (1991) explored responses to
negative political campaign advertisements among supporters and
non-supporters of the candidate being attacked. In general, their results
support the idea that people tend to perceive that media have little
influence on the self relative to others when the message is opposed to
their own position, but perceive a stronger influence of media on the self
relative to others when the message is consistent.
Similar results were reported by Driscoll and Salwen (1997) in their
research concerning perceived effects of media messages regarding O.J.
Simpson. Namely, third-person perceptions were smallest when the media
message matched the respondents' pre-existing beliefs concerning Simpson's
guilt or innocence. As these authors noted, "If there is a generalized
tendency to perceive the news from a third-person effect perspective, …this
tendency may be diminished, altered, or superseded by individuals' opinions
about the messages" (p. 551).
In general, these studies appear to imply that an additional schema or
heuristic of media influence may be at work when individuals make estimates
of media influence. Namely, studies that have examined existing attitudes
suggest that individuals may assume that media serve to reinforce existing
attitudes. This tendency appears to be particularly evident for issues on
which media may be perceived to have possible bi-directional influence
(i.e., leading to favorable or unfavorable attitudes on a given issue).
However, in the case of "harmful" content (e.g., media violence), the
reinforcement model may still be at work, as individuals may assume that
others are more unfavorably situated (e.g., aggressive) in general to begin
with, thereby making it more likely that the media will serve to reinforce
their unfavorable dispositions.
If people do employ a reinforcement schema in estimating media influence,
what might this tendency imply for third-person perceptions (or differences
in estimates of the effects of media on others versus the self)? This
tendency would imply that when estimating media influence, people first
make judgments concerning their own and others' "baseline" or existing
attitudes. Then, if individuals perceive that their attitudes are shared by
others, differences in estimated effects of media on the self versus others
should be minimal, though the strength and magnitude of the estimated
influence should vary as a function of perceived baseline attitudes.
However, when individuals perceive that others hold different attitudes,
estimated media influence on the self versus others should begin to
diverge, with the largest differences occurring when individuals perceive
that others generally have attitudes that are "opposite" from their own. In
this latter situation, individuals should perceive that media exposure
reinforces their own attitude in one direction (e.g., pro or con), whereas
exposure reinforces others' attitudes in the opposite direction.
Authors (2004) recently explored the notion of a "reinforcement model" in a
survey that assessed individuals' perceptions of media influence on the
issue of affirmative action in higher education. In their study,
respondents first reported their own attitudes about affirmative action,
their estimates of others' attitudes, and their perceptions of news
coverage of the issue. Subsequently, respondents reported perceived
influence of media coverage on their own and others' attitudes on scales
ranging from 1 (The News Makes Me More AGAINST Affirmative Action) to 9
(The News Makes Me More IN FAVOR OF Affirmative Action), with a mid-point
of 5 (The News Doesn't Affect My Attitudes Either Way). Their results
revealed that although people generally perceived that media coverage was
"neutral" in its treatment of affirmative action, media influence was
thought to affect attitudes (for both the self and others) in the direction
of existing attitudes. That is, influence was perceived to be in the "pro"
direction for individuals who were initially favorable toward the issue,
and in the "anti" direction for individuals who were initially unfavorable.
Consequently, differences between estimates of media influence on the self
versus others were minimized when individuals believed that other people
shared their attitudes, and were maximized when they believed that others
held opposing positions.
Present Study and Hypotheses
Although the notion of a reinforcement model in estimating media influence
has received some empirical support, the exploration of this model is
exploratory, at best. In addition, the use of correlational data presents a
series of potential problems in both analysis and interpretation. For
example, Authors (2004) reported that the respondents in their sample had a
strong tendency to perceive that others shared their own attitudes toward
affirmative action (a false consensus effect, see Gunther & Chia, 2001;
Marks & Miller, 1987). Consequently, exploring situations in which
respondents perceived that others held attitudes opposite to that of their
own was tenuous. In addition, although these authors interpreted their data
as supporting the idea that perceptions of existing attitudes predicted
media influence, an equally plausible interpretation is that the strength
and direction of estimated media influence predicted perceptions of
existing attitudes. For example, individuals who estimated that media
exposure caused more favorable attitudes may have then been more likely to
believe that the public, at large, consequently possessed more favorable
Because the problems inherent in correlational designs make exploration of
causal mechanisms difficult (if not, at times, impossible), the present
research attempted to examine the reinforcement model of perceived media
influence in an experimental context. Namely, rather than asking
respondents to estimate others' attitudes on given issues, the present
study manipulated others' attitudes by describing hypothetical persons as
having favorable, unfavorable, or neutral attitudes. Based on the notion
that people estimate influence of media on the basis of existing attitudes,
the following hypothesis was examined:
H1: Differences in perceived media influence on the self versus others will
vary as a function of perceptions of others' existing attitudes.
In addition to manipulating perceptions of others' attitudes, as in the
Authors (2004) study, the present investigation measured respondents' own
self-reported attitudes on given issues. Given that individuals'
perceptions of media influence on the self are thought to vary as a
function of their existing attitudes, an additional hypothesis was examined:
H2: Differences in perceived media influence on the self versus others will
be largest when self and other attitudes are most divergent and smallest
when they are most similar.
Whereas the previous two hypotheses were explored in the context of news
issues for which both "pro" and "anti" attitudes were plausibly expected
(e.g., affirmative action, attitudes toward the war in Iraq, animal
rights), it is unclear if the reinforcement model could be applied to
perceptions of media content that is understood to have decidedly harmful
or unfavorable effects on viewers (i.e., media violence increasing
aggression). Consequently, this study examined the following research
question to explore whether the reinforcement model could be extended to a
type of media content that has arguably generated the greatest research
attention in the third-person literature:
RQ1: Will differences in perceived influence of media violence on self
versus others vary as a function of individuals' perceptions of their own
and others' aggressive dispositions?
A convenience sample of 212 students at a large northeastern university
participated in the study. Respondents were approached in public places on
the university campus (e.g., library, student union, coffee shops) and were
given the questionnaire to complete at that time. The median age of the
sample was 21 years, with 101 males and 111 females agreeing to participate.
Design and Materials
A three-way within-subjects experimental design that manipulated the
attitudes of a hypothetical viewer was employed to test Hypotheses 1 and 2.
Within each of the three conditions, respondents were first presented with
a description of media coverage of a news-related issue. Following the
description of media coverage, respondents were then presented with a
description of a hypothetical viewer who was characterized as being likely
to possess favorable attitudes toward the issue described, unfavorable
attitudes, or neutral attitudes. The order in which the news issues were
presented was varied, and the issue / viewer attitude pairings were
The three issues employed in the description of media content included
affirmative action policies in higher education, U.S. involvement in Iraq,
and the use of animals in scientific research. These issues were selected
on the basis of past research and on timeliness. Within each description,
media coverage was described as presenting both "pro" and "anti" positions
on the issue (see Gunther & Chia, 2001; Authors, 2004). For example, the
description of media coverage of affirmative action read:
Media Content: News Coverage of Affirmative Action in Higher Education
This type of news content covers a balance of viewpoints on affirmative
action in higher education. Some content presents arguments that
affirmative action in higher education is a bad thing because it's unfair
to White applicants and it's unnecessary. Other content presents arguments
that affirmative action in higher education is a good thing because it
increases diversity and it helps to remedy present and past discrimination.1
Following the description of media content was a description of a
hypothetical viewer who was described as favorable, unfavorable, or neutral
toward the issue. For example, the viewer associated with the description
of affirmative action was described in one of the following ways:
(Pro): Miller is a young adult who grew up in a household that was
generally politically independent. However, since coming to college, Miller
has become more active in left-wing politics and has strong opinions about
a variety of social and political issues related to race and gender. Miller
identifies with being called a "liberal."
(Anti): Miller is a young adult who grew up in a household that was
generally politically independent. However, since coming to college, Miller
has become more active in right-wing politics and has strong opinions about
a variety of social and political issues related to race and gender. Miller
identifies with being called a "conservative."
(Neutral): Miller is a young adult who grew up in a household that was
generally politically neutral. Since coming to college, Miller has stayed
abreast of political news, and makes it a point to listen to a variety of
opinions about social and political issues related to race and gender.
Miller doesn't identify with being called a "liberal" or a "conservative,"
but can understand arguments from both perspectives.
A three-way between-subjects design that manipulated the aggressive
dispositions of a hypothetical viewer was employed to test the research
question concerning perceived influence of media violence. Respondents were
first presented with a description of violent media:
Media Content: Crime/Action Drama.
This type of media entertainment typically features some type of criminal
behavior (e.g., murder, robbery) that places the cast of characters in a
dangerous situation, often with the characters fighting for their lives.
The entertainment often features a great deal of suspense and action, such
as gun fights, car chases, and explosions.
Following this description was a description of a hypothetical viewer who
was described as possessing a disposition that was consistent with the
message (i.e., aggressive), was inconsistent (i.e., non-aggressive), or was
"neutral" (i.e., neither aggressive nor non-aggressive).
(Aggressive Viewer): Smith is a very high-strung individual who is often
quick to anger. Smith often has a bad temper, and has had several occasions
to get into fights with other people. While most of these fights have
involved only exchanging threats, some of them have involved physical
(Non-aggressive Viewer): Smith is a very mild-mannered person. Smith is
very understanding and peaceful, and goes to great lengths to avoid
conflict with other people. The thought of fights and arguments is
something that Smith finds disturbing and upsetting, preferring to talk
things out rather than resort to violence.
(Neither Aggressive nor Non-aggressive Viewer): Smith is generally an
even-tempered sort of person. Although Smith can sometimes get angry when
other people are rude or assertive, Smith is generally able to keep this
irritation in check. Although Smith generally avoids situations involving
conflict, Smith occasionally gets into arguments, though these arguments
never last long.
The first page of the questionnaire booklet asked respondents to indicate
their own attitudes about affirmative action in higher education, U.S.
involvement in Iraq, and the use of animals in scientific research on
scales ranging from -4 (Very Against) to 4 (Very In Favor Of), with a
midpoint of 0 (Neither in Favor Of nor Against). This front page also asked
respondents to indicate their level of aggressiveness on scales ranging
from -4 (Not at All Aggressive) to 4 (Very Aggressive).
The following pages of the questionnaire presented respondents with the
descriptions of media content and the hypothetical viewers. Following each
description, respondents were first asked to rate the viewer's
attitude/disposition on the issue in question (e.g., affirmative action,
aggression) on scales ranging from -4 (Very Against/Not at All) to 4 (Very
In Favor Of/Very). These questions were used as a manipulation check for
the description of the hypothetical viewers' attitudes/disposition.
After rating perceptions of the hypothetical viewer, respondents were then
asked to estimate how much they believed that the described media content
would affect the hypothetical viewer, and then how much it would affect
themselves. Both of these questions employed scales ranging from -4 (Would
Make Them/Me More Against) to 4 (Would Make Them/Me More In Favor Of), with
a midpoint of 0 (Would Have No Effect). In the case of media violence,
these scales ranged from -4 (Would Make Them/Me Less Aggressive) to 4
(Would Make Them/Me More Aggressive).
Recoding and Manipulation Check
Before examining Hypotheses 1 and 2, a preliminary analysis was conducted
on perceived media influence on self versus others as a function of story
type (Affirmative Action, Iraq, Animal Rights). This analysis revealed a
main effect of self-other and a main effect of story type, Self-Other: F(1,
210) = 18.75, ? = .92, ?p2 = .08, p < .001; Story Type: F(2, 209) = 9.09, ?
= .92, ?p2 = .08, p < .001.2 However, the Self-Other X Story Type
interaction was not significant, illustrating that perceptions of media
influence on self versus others was not unique to any specific story, F(2,
209) = 1.41, ? = .99, ?p2 = .01, p = .25. Consequently, responses to each
of the three scenarios (Affirmative Action, Iraq, Animal Rights) were first
recoded to reflect responses according to the three hypothetical viewers'
described attitudes (pro, anti, neutral), resulting in each attitude
condition including all three news story topics.
Subsequently, to examine the effectiveness of the manipulation of viewers'
orientations, respondents' perceptions of viewers' attitudes were analyzed
in a three-way repeated-measures analysis. As expected, viewers' attitudes
were perceived as most favorable in the "pro" condition (M = 2.28, SE =
.12), most unfavorable in the "anti" condition (M = -1.55, SE = .17), and
as moderate in the neutral condition (M = 0.48, SE = .06), with each
condition significantly different from the other, F(2, 210) = 151.85, ? =
.41, ?p2 = .59, p < .001.3 In contrast, respondents' self-reported
attitudes did not differ according to viewer-attitude conditions, with
self-reported attitudes generally falling within the "neutral" range, F(2,
209) = 1.19, ? = .99, ?p2 = .01, p = .31 (other pro: M = -0.05, SE = .16;
other anti: M = 0.18, SE = .16; other neutral: M = 0.19, SE = .16).
Perceived Media Influence as a Function of Others' Attitudes
Hypothesis 1 predicted that differences in perceived media influence on
self and others would vary as a function of perceptions of others' initial
attitudes. To explore this hypothesis, a 2 (media influence: self, other) x
3 (perceived others' attitudes: pro, neutral, anti) repeated measures
analysis of variance was employed. This analysis revealed a main effect for
self-other, F(1, 210) = 18.75, ? = .92, ?p2 = .08, p < .001, with perceived
media influence on others reflecting an effect in the favorable direction
(M = 0.24, SE = .06), and perceived media influence on the self as
negligible (M = -0.08, SE = .07). A main effect was also obtained for
others' attitudes, F(2, 209) = 29.21, ? = .78, ?p2 = .22, p < .001. This
effect was obtained because when others' were described as "pro," media
influence was thought to be in a favorable direction (M = 0.57, SE = .09),
when others were described as "anti," media influence was thought to be in
an unfavorable direction (M = -0.40, SE = .11), and when others were
described as neutral, media influence was thought to be negligible (M =
0.06, SE = .01).
However, these main effects should be interpreted in light of the
hypothesized Self-Other X Others' Attitudes interaction that was also
obtained, F(2, 209) = 47.03, ? = .69, ?p2 = .31, p < .001. As Figure 1
illustrates, perceptions of media influence on others varied according to
the direction of others' attitudes toward the issue at hand, whereas
perceived media influence on the self did not vary substantially as a
function of others' attitudes. Because respondents generally did not
perceive that they were greatly influenced by media coverage of the issues,
this graph illustrates the clearest instances of third-person perceptions
in the other-pro and the other-anti conditions. In these two conditions,
media influence was perceived to be stronger on others than the self, but
it is important to note that the nature of the third-person perception
differed substantially. When others were perceived to be pro, the
third-person perception illustrates perceptions of stronger media influence
in a favorable direction on others than on the self, but when others were
perceived to be anti, the third-person perception illustrates perceptions
of stronger influence in an unfavorable direction on others than on the
self. These results are consistent with the idea that individuals appear to
assume a strength and direction of media influence on others that is
consistent with or reinforcing of others' perceived initial attitudes.
Perceived Media Influence as a Function of Self-Reported and Others' Attitudes
Hypothesis 2 predicted that differences in perceived media influence on
the self versus others would be largest when self and other attitudes were
most divergent and smallest when they were most similar. To explore this
hypothesis, mixed-model analyses of variance were conducted for each of the
three other-attitude conditions (pro, neutral, anti). Each analysis treated
perceived media influence on self versus others as a within-subjects
variable, and self-reported attitudes as a continuous between-subjects
Consistent with Hypothesis 2, analyses for each of the three conditions
revealed the expected Self-Other X Self-Reported Attitudes Interaction,
other pro: F(1, 209) = 71.52, ? = .75, ?p2 = .26, p < .001; other neutral:
F(1, 210) = 77.82, ? = .73, ?p2 = .27, p < .001: other anti, F(1, 209) =
29.42, ? = .88, ?p2 = .12, p < .001. To illustrate and explore the nature
of these interactions, respondents' self-reported attitudes on each of the
issues were categorized as anti (below a score of -1), neutral (between -1
and 1), or pro (above 1). Subsequently, a 2 (media influence on others,
self) X 3 (self-reported attitudes) mixed-model analysis of variance was
conducted for each condition.4 Figure 2 illustrates the three interactions
and shows that differences between perceived effects on others versus the
self tended to be smallest when the respondent and the described viewer
shared similar attitudes. When the respondent and described viewer differed
in their attitudes, differences in perceived media influence were not only
large, but they also tended to reflect differences in direction of media
Perceived Effects of Media Violence
The final analysis was conducted to explore whether or not the influence
of self-reported and perceptions of others' orientations on perceived media
influence could be extended to an issue for which there is clearly
undesirable direction of media influence. Namely, a mixed-model analysis of
variance was conducted to examine the perceived influence of media
violence, with Self-Other serving as a within-subjects factor, and others'
dispositions (aggressive, passive, neutral) and self-reported orientations
serving as between-subjects factors.5 This analysis revealed a main effect
of self-other, with respondents reporting that media violence results in
greater levels of aggression on others (M = 0.84, SE = .09) than on the
self (M = 0.26, SE = .06), F(1, 206) = 41.12, ? = .83, ?p2 = .17, p < .001.
A main effect was also revealed for others' dispositions, with respondents
perceiving media violence to result in greater levels of violence in the
aggressive condition (M = 0.92; SE = .11) and the neutral condition (M =
0.72; SE = .11), but not in the passive condition (M = 0.00; SE = .11),
F(2, 206) = 19.16, ?p2 = .16, p < .001.
However, this main effect should be interpreted in light of an Other-Self X
Others' Disposition interaction that was also obtained, F(2, 206) = 47.77,
? = .68, ?p2 = .32, p < .001. Figure 3 illustrates this interaction, and
shows that when others were perceived as aggressive or neutral, respondents
perceived that media violence would result in greater levels of violence on
others than on the self. However, when others were perceived as passive,
respondents perceived that media violence would result in lower levels of
aggression among others, and higher levels for the self. In other words,
this analysis revealed traditional third-person perceptions for media
violence only when others were perceived as neutral or aggressive.6
The present study offers support for the idea that when making judgments
about media influence on the self and on others, individuals tend to
perceive that media serve to reinforce existing attitudes or dispositions.
This reinforcement model suggests, then, that third-person perceptions may
reflect, at least in part, individuals' differential judgments of their own
versus others' initial attitudes, and an assumption that media influence
does not necessarily change attitudes, but rather strengthens dispositions
that are in place initially (i.e., a baseline). Consequently, when
individuals perceive that others share similar attitudes or dispositions,
third-person perceptions should be minimized.
It is important to recognize that although this reinforcement model
provides a framework for interpreting third-person perceptions, this model
is not inconsistent with existing interpretations of the third-person
phenomenon, but rather provides a schema for examining and predicting
perceptions of media influence. For example, explanations of third-person
perceptions focusing on self-enhancement motivations can be understood as
consistent with perceptions of media reinforcement. In terms of estimates
of initial (or baseline) attitudes for the self and others, research on
self-enhancement biases would generally suggest that individuals would be
likely to perceive their own attitudes/dispositions as favorable or
"correct" (Fiske & Taylor, 1984). Presumably, too, the more strongly an
individual endorses a given attitude, the more favorably it is perceived.
Consequently, in terms of media influence, changes in attitudes resulting
from media exposure should be considered "beneficial" insofar as these
changes result in movement toward the "correct" attitude (the one that is
endorsed), and "harmful" insofar as these changes result in movement toward
the "incorrect" attitude.
Although this motivational explanation is consistent with the data obtained
in this study, there are obviously numerous additional mechanisms that
likely play a role in perceptions of media reinforcement. For example,
research on biased processing of persuasive messages suggests that
individuals with strongly held attitudes tend to evaluate more favorably
arguments that are consistent with existing attitudes, to assimilate
arguments that are ambiguous, and to counter argue or discount arguments
that are inconsistent. Consequently, exposure to the same message featuring
arguments on both sides of an issue should result in greater polarization
of initial attitudes (Eagly & Chaiken, 1993; Lord, Ross, & Lepper, 1979;
Miller, McHoskey, Bane, & Dowd, 1993; Munro et al., 2002). In the case of
the present study, these biases in processing may not necessarily be
recognized by respondents as biases per se. Nevertheless, individuals may
have at least an intuitive understanding of individuals' tendencies in
processing confirming and disconfirming messages, thereby affecting their
perceptions of how the self and others are likely affected by exposure to
Whereas the previous discussion of possible mechanisms explaining media
reinforcement seems most relevant for controversial issues on which people
likely hold a diversity of opinion (both pro and con), similar reasoning
can be applied to messages for which only one position is advocated or one
direction of effect is expected or assumed. For example, in the case of
perceptions of the influence of media violence or of sexually explicit
materials, the preponderance of research has either assumed harmful
effects, has employed pretests to determine that individuals actually
perceive harmful effects, or has asked about harmful effects specifically
(e.g., Eveland & McLeod, 1999; Hoffner & Buchanan, 2002; Hoffner et al.,
2001; Lo & Wei, 2002; McLeod, Eveland, & Nathanson, 1997). In the case of
such content, then, perceived reinforcement would be expected to occur
primarily in one direction (e.g., violent dispositions associated with
increased estimates of the influence of media violence). However, attitudes
or dispositions contrary to the portrayed or advocated position (e.g.,
passive dispositions) may be associated with assumed counter arguing that
could result either in little or no influence of media or in some movement
in the opposite direction to that portrayed. Nevertheless, reinforcement of
dispositions that are contrary to the position of a unidirectional message
may not be as strong as reinforcement when the media content contains
messages that are both consistent and inconsistent with initial attitudes.
The application of the reinforcement model presented here to existing
third-person research suggests that it may provide a useful heuristic in
organizing a diversity of findings. For example, some studies have reported
larger third-person perceptions of media violence among female than male
respondents, and higher estimates of the influence of media violence on the
aggression of male than of female viewers (e.g., Hoffner et al., 2001;
Scharrer, 2002). In these instances, given generally higher levels of
aggression among males than among females, the reinforcement model would
suggest that these gender differences may reflect differential perceptions
of baseline attitudes/dispositions, resulting in differential levels of
expected media reinforcement. Similarly, research that has examined
variants of the social distance corollary (Cohen, Mutz, Price, & Gunther,
1988; Eveland et al., 1999; Gibbon & Durkin, 1995; Meirick, 2004) could be
explained by the idea that individuals tend to perceive similar others as
holding comparable attitudes, with media thereby reinforcing shared
attitudes to the same degree.
Although this proposed model is admittedly exploratory, the implications
that it has in terms of both perceptions of public opinion and of expected
media influence are interesting to consider. For example, if individuals
assume that media reinforce existing attitudes, then assessment of
perceived media influence may serve as an indirect measure of perceived
public opinion. For example, if individuals assume that news consumption is
making people "more liberal," the reinforcement model would suggest that
these same individuals may likely perceive that the public is more liberal
to begin with. Similarly, if individuals perceive that the public holds
attitudes that are contrary to their own beliefs or opinions, these same
individuals may assume or fear that exposure to media will serve to
reinforce others' attitudes, thereby exacerbating perceptions of a hostile
media (Vallone, Ross, & Lepper, 1985).
Certainly the application and implications of this model are far reaching
and complex and would benefit from further exploration and explication. In
addition, future research on this proposed model should attempt to account
for limitations that are present in this study. For example, this study
presented participants with a single hypothetical viewer to represent
"other viewers." While this method was employed to allow for the
manipulation of the others' attitudes/dispositions, self-reported estimates
for a hypothetical viewer likely differ considerably from estimates for the
public at large or for peers. Although, similar research employing
correlational data has reported consistent findings to those obtained in
this study (Authors, 2004), future research would clearly benefit from
exploring a variety of ways to both measure and manipulate perceptions of
An additional limitation with the use of hypothetical viewers employed in
this study concerns potential demand characteristics. That is, individuals
were asked to estimate perceived media influence immediately after reading
a description of the hypothetical viewers' attitudes related to the media
content. Consequently, correlations between perceived attitudes and
perceived media influence may have reflected only attempts to appear
consistent. Although the notion of reinforcement seems to imply perceived
consistency, future research would benefit from systematic steps to reduce
any effects of demand characteristics on respondents' reported judgments.
Finally, of course, this study employed college students as participants,
which always serves to limit the generalizability of findings. Although
there are no a prior reasons to suspect that the pattern of results
observed here are unique to college students, future research would
undoubtedly benefit from the use of a more diverse sample. Indeed, a
greater diversity of respondents would provide the added opportunity of
exploring how individuals from varying backgrounds perceive others'
attitudes as similar or dissimilar from their own, thereby optimizing the
possibility of finding variance in perceptions in both the strength and
direction of perceived media influence.
This paper began with a question concerning the variables that individuals
likely consider when estimating media influence. The reinforcement model
presented here suggests that one important variable considered may be
baseline attitudes or dispositions, with media exposure assumed to
strengthen or solidify existing attitudes. Insofar as people are thought to
employ this heuristic to estimate both effects of media on the self and on
others, the present study suggests that existing research on third-person
perceptions may say as much about individuals' perceptions of others and
their attitudes as it does about media influence.
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,
Cohen, J., & Davis, R. G. (1991). Third-person effects and the differential
impact in negative presidential advertising. Journalism Quarterly, 68, 680-688.
Cohen, J., Mutz, D., Price, V., & Gunther, A. (1988). Perceived impact on
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, 47, 1-15.
Driscoll, P. D., & Salwen, M. B. (1997). Self-perceived knowledge of the
O.J. Simpson trial: Third-person perception and perceptions of guilt.
Journalism & Mass Communication Quarterly, 74, 541-556.
Duck, J. M., Hogg, M. A., & Terry, D. J. (1999). Social identity and
perceptions of media persuasion: Are we always less influenced than others?
Journal of Applied Social Psychology, 29, 1879-1899.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort
Worth, TX: Harcourt Brace Javanovich.
Eveland, W. P., & McLeod, D. M. (1999). The effect of social desirability
on perceived media impact: Implications for third-person perceptions.
International Journal of Public Opinion Research, 11, 315-333.
Eveland, W. P., Nathanson, A. I., Detenber, B. H., & McLeod, D. M. (1999).
Rethinking the social distance corollary: Perceived likelihood of exposure
and the third-person perception. Communication Research, 26, 275-302.
Fiske, S. T., & Taylor, S. E. (1984). Social cognition. New York: Random House.
Gibbon, P., & Durkin, K. (1995). The third person effect: Social distance
and perceived media bias. European Journal of Social Psychology, 25, 597-602.
Gunther, A. C. (1995). Overrating the X-rating: The third-person perception
and support for censorship of pornography. Journal of Communication, 45(1),
Gunther, A. C. (1998). The persuasive press inference: Effects of mass
media on perceived public opinion. Communication Research, 25, 486-504.
Gunther, A. C., & Chia, S. C. Y. (2001). Predicting pluralistic ignorance:
The hostile media perception and its consequences. Journalism & Mass
Communication Quarterly, 78, 688-701.
Gunther, A. C., & Christen, C. T. (2002). Projection or persuasive press?
Contrary effects of personal opinion and perceived news coverage on
estimates of public opinion. Journal of Communication, 52, 177-195.
Gunther, A. C., Christen, C. T., Liebhart, J. L., & Chia, S. C. Y. (2001).
Congenial public, contrary press, and biased estimates of the climate of
opinion. Public Opinion Quarterly, 65, 295-320.
Gunther, A. C., & Hwa, A. P. (1996). Public perceptions of television
influence and opinions about censorship in Singapore. International Journal
of Public Opinion Research, 8, 248-265.
Gunther, A. C., & Mundy, P. (1993). Biased optimism and the 3rd-person
effect. Journalism Quarterly, 70, 58-67.
Henriksen, L., & Flora, J. A. (1999). Third-person perception and children:
Perceived impact of pro- and anti-smoking ads. Communication Research, 26,
Hoffner, C., & Buchanan, M. (2002). Parents' responses to television
violence: The third-person perception, parental mediation, and support for
censorship. Media Psychology, 4(3), 231-252.
Hoffner, C., Plotkin, R. S., Buchanan, M., Anderson, J. D., Kamigaki, S.
K., Hubbs, L. A., Kowalczyk, L., Silberg, K., & Pastorek, A. (2001). The
third-person effect in perceptions of the influence of television violence.
Journal of Communication, 51, 283-299.
Lo, V. H., & Wei, R. (2002). Third-person effect, gender, and pornography
on the Internet. Journal of Broadcasting & Electronic Media, 46, 13-33.
Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and
attitude polarization: The effects of prior theories on subsequently
considered evidence. Journal of Personality and Social Psychology, 37,
Marks, G., & Miller, N. (1987). Ten years of research on the
false-consensus effect: An empirical and theoretical review. Psychological
Bulletin, 102, 72-90.
McLeod, D. M., Detenber, B. H., & Eveland, W. P. (2001). Behind the
third-person effect: Differentiating perceptual processes for self and
other. Journal of Communication, 51, 678-695.
McLeod, D. M., Eveland, W. P., & 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.
Meirick, P. C. (2004). Topic-relevant reference groups and dimensions of
distance: Political advertising and first- and third-person effects.
Communication Research, 31, 234-255.
Miller, A. G., McHoskey, J. W., Bane, C. M., & Dowd, T. G. (1993). The
attitude polarization phenomenon: Role of response measure, attitude
extremity, and behavioral consequences of reported attitude change. Journal
of Personality and Social Psychology, 64, 561-574.
Munro, G. D., Ditto, P. H., Lockhart, L. K., Fagerlin, A., Gready, M., &
Peterson, E. (2002). Biased assimilation of sociopolitical arguments:
Evaluating the 1996 U.S. presidential debate. Basic and Applied Social
Psychology, 24, 15-26.
Perloff, R. M. (1993). Third person effect research 1983-1992: A review and
synthesis. International Journal of Public Opinion Research, 5, 167-184.
Perloff, R. M. (1999). The third-person effect: A critical review and
synthesis. Media Psychology, 1, 353-378.
Perloff, R. M. (2002). The third-person effect. In J. Bryant & D. Zillmann
(Eds.), Media effects: Advances in theory and research (2nd ed., pp.
489-506). Mahwah, NJ: Lawrence Erlbaum.
Scharrer, E. (2002). Third-person perception and television violence: The
role of out-group stereotyping in perceptions of susceptibility to effects.
Communication Research, 29, 681-704.
Tversky, A., & Kahneman, D. (1974, September 27). Judgment under
uncertainty: Heuristics and biases. Science, 185, 1124-1131.
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.
Authors (2004). The importance of allowing for diversity of opinion in the
examination of third-person perceptions. Paper to be presented at the
annual meeting of the International Communication Association, New Orleans.
Two versions of this explanation were created that varied the order in
which the pro-arguments and anti-arguments were presented within the
2 The main effect for story type reflected that participants reported
effects in a slightly favorable direction for the stories concerning
affirmative action (M = 0.29, SE = 0.09) and Iraq (M = 0.23, SE = 0.10) and
in a slightly unfavorable direction for the story concerning animal rights
(M = -0.29, SE = 0.10). The main effect for self-other occurred because
participants estimated that others would be affected in a slighted
favorable direction (M = 0.24, SE = 0.06), whereas the effects on the self
would be negligible (M = -0.08, SE = 0.07). This and all additional
repeated measures analyses reported here employed a multivariate approach
using Wilks' criterion
3 This and all pairwise comparisons reported here employed Holm's
sequential bonferroni post hoc procedures.
4 These analyses also revealed the expected Self-Other X Self-Reported
Attitudes interactions reported previously: other pro: F(2, 208) = 30.56, ?
= .77, ?p2 = .23, p < .001; other neutral: F(2, 209) = 37.45, ? = .74, ?p2
= .26, p < .001: other anti, F(2, 208) = 11.22, ? = .90, ?p2 = .10, p < .001.
5 The manipulation check for the descriptions of the hypothetical viewer
revealed that the aggressive viewer was rated as highest in aggression (M =
3.08, SD = 0.84), the non-aggressive viewer as lowest in aggression (M =
-2.94, SD = 1.26), and the neutral viewer as in between (M = 0.27, SD =
1.44), F(2, 209) = 438.98, ?p2 = .81, p < .001
6 It is curious why third-person perceptions would be observed in the
other-neutral condition. However, an examination of estimated levels of
aggression revealed that participants perceived that the "neutral"
hypothetical viewer was slightly more aggressive (M = 0.27, SD = 1.44) than
was the self (M = -0.18, SD = 1.44), t(70) = 1.59, p = .12.
Figure 1. Perceived Media Influence on Others and Self as a function of
Figure 2. Perceived Media Influence on Others and Self as a function of
Self-Reported and Others' Attitudes
Figure 3. Perceived Influence of Media Violence on Others versus Self as a
function of Others' Dispositions