Professional Documents
Culture Documents
Censorship of mass media content is often justified by the argument that the
message will have undesirable effects on society. Oddly though, while believ-
ing that controversial forms of expression such as defamation, fictional vio-
lence, sedition or explicit sex should be censored for the good of the public,
many people have no qualms about viewing such content themselves.
This curious discrepancy, often called the third-person effect (Davison,
1983), has two components. The first suggests that people are prone to a per-
ceptual bias, leading them to estimate that a communication will have more
influence on others than on themselves. The second component proposes that
people may react in some way according to this estimate of larger effects on
others. In other words, people tend to consider themselves less vulnerable to
harmful influences, and their support for censorship may be rooted instead in
their concern about effects on others.
The first component, the perceptual bias, is a robust finding. There is plenti-
ful empirical evidence of the tendency to perceive greater media influence on
others than on the self. People have exhibited this phenomenon in judgments
Albert C. Gunther is a n associate professor in the Department of Agricultural Journalism at the Uni-
versity of Wisconsin-Madison, The author would like to express appreciation to the University of
Wisconsin L & S Survey Center for data collection, and to Lulu Rodriguez for assistance with data
analysis and preparation o f tables.
Copyright 0 1995Journal ofcommunication 45(1), Winter. 0021-9916/95/$5.00
27
Journal of Communication, Winter 1995
Among several explanations for the perceptual bias, perhaps the most funda-
mental is that of biased optimism-the tendency for people to think they are
not as likely to experience negative events and influences as are others (Wein-
stein, 1989). This bias, it is argued, stems from the need to sustain a positive
self-image. When individuals in a representative sample are asked to rate their
own driving, for example, between 75% and 90% say their driving is safer than
the average person’s (Svenson, 1981). People are motivated to reinforce self-
esteem by perceiving themselves as smarter or better off than the majority of
their peers, and thus they see others as more vulnerable to media content.
As for the consequences of these perceptions, theoretical arguments and
empirical evidence support the idea that a person’s opinions and behaviors are
sometimes “sociotropic”responses, that is, determined by his or her percep-
tions of others (Kinder & Kiewiet, 1981; Mutz, 1992). But the third-person
effect postulates a more specific phenomenon than the general sociotropic
case. That is, it is effect on others relative to effect on self that underlies the
third-person effect. And more specifically still, it is the magnitude of difference
between people’s estimations of negative influence on others and on them-
selves that leads to some reaction in attitude or behavior.
These two components of the third-person hypothesis-a perception, and an
effect on attitudes or behavior-can be linked in the following way. Prone to
biased optimism, people feel that whatever message effect they incur them-
selves is relatively benign, an acceptable personal response. The greater nega-
tive effect seen in others, however, is likely to be perceived as an unfortunate
and undesirable social effect. As the perceived distance between acceptable
I There is some irony in this outcome, in that the real communication effect “may thus be due not to
the reaction of the ostensible audience but rather to the behavior of those who anticipate, or think
they perceive, some reaction on the part of others” (Davison, 1983, p. 3).
28
Overrating the X-Ruting
effect on self and unacceptable effect on others grows larger, people will feel
that the social-level effect grows more harmful, and that something should be
done about it (Gunther, 1991).
Past studies have ventured explanations for responses based on personal vs.
social-level perceptions. Self-interest, it is argued, influences opinion and behav-
ior largely when the benefits are clear and “unobstructed by competing consid-
erations” (Green & Gerken, 1989, p. 12). By contrast, sociotropic responses
may result from altruistic concern for the welfare of others. But people may
also act on their social-level perceptions because of indirect benefits to the
self-the probability, for example, that improved social conditions will eventu-
ally improve one’s personal condition. In the latter case, self-interest is also
motivating social-level responses that would benefit the well-being of others
only incidentally (Kinder & Kiewiet, 1981).
Since the optimistic bias suggests that people will consider themselves rela-
tively immune to harmful media effects, the incentive to change attitudes or
take actions on behalf of personal well-being is probably not compelling. How-
ever, to the extent that people see negative social effects as resulting from
media content, one would expect them to favor steps to prevent such effects,
either out of concern for the well-being of others, or because failure to contain
such effects will make society a riskier environment for the self.
Backgrounding Pornography
29
.Journal of Communicataon, Winter 1995
Such a relationship will likely be moderated, however, by the fact that while people generally
agree strongly with the abstract principle of free speech, their support drops sharply when consider-
ing particular situations (McCloskey & Brill, 1983).
It should be noted that perceptual bias can only be measured at the aggregate level in this study.
Any individual estimating little effect on self and more effect on others may be quite correct in
doing so. But if a majority of people demonstrate the third-person perception, then some portion
of that aggregate judgment must be in error.
30
Overruting the X-Ruting
Respondents were 648 adults age 18 or over randomly sampled from the conti-
nental U S . population. The data were collected as part of a computer-assisted
telephone survey run by the University of Wisconsin L&S Survey Center, which
conducts about 50 interviews a week in a rolling cross-sectional sample. The
items included in this analysis ran from October 18, 1989, to January 23, 1990.
Trained interviewers administered the survey and made u p to seven callbacks
to numbers not answering. The response rate was a rather low 46%.* However,
comparisons based on key demographic variables such as age, gender, and
income indicated that these respondents did not differ significantly from those
in the General Social Survey and the National Survey of Families and House-
holds.
Midway through the survey the issue of X-rated media content was intro-
duced, and respondents were asked about their own exposure to such con-
tent-specifically how often, if ever, they had seen or come across X-rated
films or magazine^.^
Immediately following this item they answered a series of questions about
their perceptions of the effects of such content on themselves and on other
people in generaL6 Respondents were asked about the effect of X-rated films or
magazines on their own moral values concerning sex, and on other people's
moral values concerning sex. They also answered about the effect on their
own attitudes toward the opposite sex, about the attitudes of men in general
toward women, and about the attitudes of women in general toward men.
Each item was accompanied by the following response scale: (a) a large nega-
tive effect; (b) a small negative effect; (c) no effect at all; (d) a small positive
effect; (e) a large positive effect.' Although the two items measuring perceived
effects of X-rated content-moral values concerning sex and attitudes toward
the opposite sex-appear to tap different dimensions, they were nevertheless
' The denominator in this response calculation includes refusals, numbers that were busy or not
answered at each call, and numbers that were answered, but where the selected respondent was
not aVdihbk in any of the Seven callbacks
' The phrase "X-rated films o r magazines" wab incorporated to avoid connotations of the term por-
nography, and to accommodate time constraints in the survey. It is plausible that different respon-
dents made different interpretations of the phrase "X-rated" content. However, it is safe to say that
each was estimating effects of the Same content on self and others. Thus, while the general word-
ing might have affected the validity or reliability of individual responses, it is unlikely to have
affected the self-other difference scores, which are the important Factor in testing third-person phe-
nomena.
f> Questions about effects on one's self were posed slightly differently for people who said they had
never had any exposure to X-rated material. They were asked "If for Some reason you were
exposed to films or magazines with X-rated content . . . ."
' Questions focusing on more specific positive and negative effects of pornography would have
heen desirable, but again time constraints required framing them in this way. Interviewers were cau
tioned to be alert to difficulties because of the sensitive nature of the items concerning sex, but
they reported almost no unease or hesitation to answer. This experience is consistent with past
research (DeLamater, 1982).
31
Journal of Communication, Winter 1995
” A t-test showed no significant difference between those respondents omitted and those included in
the analysis in their attitudes toward restriction of pornography. Nor did they differ on most other
attributes, although they were more likely to be women, and to be somewhat less educated.
32
Overrating the X-Rating
Table 1. Mean Estimates of effect of X-Rated Content on Self and Others, and for
Subgroups Based on Exposure and Gender (Standard Deviations in Parentheses)
N Self Others Difference t-value
Effect on moral values 550 3.24 [.78) 3.55 [1.28) .31 5.19*
Effect on attitudes 525 3.17 1.66) 3.63 (1.13) .46 8.72*
Combined effect 492 3.23 1.64) 3.62 11.07) .39 8.36*
Exposure
None 115 3.37 1.60) 3.63 (1.19) .26 2.33**
Some 319 3.22 1.63) 3.63 11.02) .41 7.52*
Gender
Male 230 3.08 1.601 3.49 (1.071 .41 6.57*
Female 262 3.36 i.65j 3.73 ii .07j .37 5.42*
Note: Items are recoded so that 1 = large positive effect, 2 = small positive effect, 3 = no
effect at all, 4 = small negative effect, and 5 = large negative effect. Significance levels
were calculated using paired t-tests.
*p i,001 **p i,051.
retained in the regression analysis by substituting the mean score for their miss-
ing value.
Hypothesis 1 proposed that people will estimate X-rated media content to
have a greater negative effect on others than on themselves. The data sup-
ported this difference in judgments quite clearly. A majority of respondents,
61%, perceived others to be more negatively influenced by X-rated material
than themselves. Nineteen percent reported no difference in effect, and 20%
perceived more negative effect on themselves.
The significance of these differences is reported in Table 1. Paired t-tests
affirmed that people did perceive that others would be more influenced by por-
nography in both the combined effect index and in the individual measures.
Furthermore, the finding was consistent among subgroups. Both males and
females exhibited the perceptual bias, but males estimated less negative effect
overall, and they saw themselves as being only slightly negatively influenced
by pornography. Interestingly, respondents who reported no exposure to por-
nography demonstrated the smallest third-person pattern, primarily because
they estimated that they would be negatively affected by X-rated material quite
a bit if they were exposed to it themselves.
A major methodological challenge to third-person phenomenon findings lies
in the potential order effect when people knowingly compare themselves to
others. However, random assignment of question order in this design revealed
no order effect. The difference between estimated effects on self and others
was the same whether respondents were asked about themselves first (M = .41)
or others first (M = .37), t(490) = .49, ns. Respondents’ scale scores also did not
vary by question order. R e s t s o f the two groups showed no difference in esti-
mates of the effect of pornography on self whether respondents answered
about self first (M = 5.60) or about self after answering about others (M = 5.57)
t(605) = .24, ns. Similarly, there was no difference in estimates of the effect of
pornography on others whether respondents answered about others first ( M =
33
Journal of Communication, Winter 1995
7.02) or others after answering about self (M = 7.26) t(502) = .83, ns. Thus
there is good evidence that the side-by-side questions did not produce artifac-
tual contrast effects.
Hypothesis 2 proposed that the magnitude of the perceptual bias will be posi-
tively related to support for restrictions on pornography. This hypothesis was
tested with a multiple regression equation (see the first two columns of Table
2 ) assessing the role of various factors in predicting support for restrictions.
The independent variable of particular interest in this equation is the magni-
tude of the perceptual bias, calculated as the difference between perceived
effect on others and on self. But the independent influence of perceived effect
on self has also been partialled out, because one would expect self-interest,
and the mere extent of effect, to have some part in the outcome. An array of
control variables was included in the equation.
As expected, those who supported free speech opposed regulation. Females,
those with a religious affiliation, and those reporting no exposure to pornogra-
phy were more likely to favor its regulation-also no surprise. More perceived
negative effect on self also predicted more support for regulation. But when
controlling for these additional variables, the magnitude of perceptual bias was
significantly related to opinion favoring pornography restrictions-support for
Hypothesis 2 .
However, while it supports Hypothesis 2 , the role of the perceptual bias is
overshadowed somewhat. In particular, people were more responsive to per-
ceived effects on themselves when considering pornography restrictions. This
is an important point. In testing effect-on-opinion, the full sample may partially
obscure the third-person process, because such a test incorporates a substantial
number of cases that do not demonstrate the perceptual bias. That is, the full
sample includes many people (39%) who reported no difference between them-
selves and others, or even more negative effect on themselves, and thus failed
to demonstrate the first component of the third-person effect.
Therefore, the same regression analysis was applied to those people (three-
fifths of the sample) who did perceive that others would be more negatively
affected than themselves.’ The results (columns 3 and 4 in Table 2 ) are little dif-
ferent in general, but the second component of the third-person effect is far
more prominent. Among people who demonstrated the hypothesized third-per-
son perception, the magnitude of perceptual bias is a highly significant predic-
tor, with at least as much influence on support for restrictions as perceived
effect on self.
This subsample reported a higher level of education, tt646) = 4.15, p < ,001, but did not differ sig-
nificantly from the rest of the sample on other relevant variables, including the support for restric-
tions index.
34
Overrating the X-Rating
3rd-person
Full sample R2 change subsample R2 change
Sex - .56 (,19)* .05 - .38 (.22) .04
Education - .22 (.lo) .02 - .29 (.l1)* .02
Religion 1.33 (.26)** .03 .88 (.32)* .02
Attitude toward free expression - .27 (.08)** .03 - .27 (.08)* .03
Exposure to X-rated content - .72 (.14)** .04 - .80 (.16)** .05
Perc’d effect on self .97 (. 16)** .04 1.26 (.24)** .03
Perceptual bias .31 (.lo)* .01 1.33 (.20)** .08
Total R2 .22 .27
SE 2.28 2.13
N 632 41 5
Note: Higher score on the dependent variable (recoded)indicates greater support for regu-
lation of pornography. Other variables coded, or recoded, as follows: sex (1 = female. 2 =
male); religion (0 = no religious affiliation, 1 = any religious affiliation); support for free
expression (1 = low, 4 = high); exposure (1 = none, 2 = some); effect on self and others (1
= large positive effect, 3 = no effect, 5 = large negative effect).Variables that made a triv-
ial contribution (age, children in household, political affiliation) in initial analyses were omit-
ted from the final regression calculation.
* p < .01 * * p c ,001
These data offer compelling evidence o f the tendency for people to estimate
that some types of media content will have more negative effect on others than
on themselves. Confirming findings in smaller populations, the national sample
lends substantial external validity to the perceptual bias hypothesis. Beyond
that, controlling for question order effectively eliminates t h e possibility that the
finding is an artifact of measurement.
The data also indicate that better educated individuals are more prone to the
perceptual bias, a finding with theoretical significance. Attribution theory sug-
gests that people seek reasons for events?o r even understand them in terms of
their perceived causes. In some situations people can attribute their percep-
tions of greater influence on others to self-perceived expertise (Lasorsa, 1989),
in-group membership (Vallone et al., 19851, or apparent biases in the source
(Gunther, 1991). When situation-specific attributions are lacking, however, the
perceptual bias may result from attributions to education or other exogenous
factors.
The direction o f error in the perceptual bias cannot be determined in this
study because there is no way to gauge any actual effect of exposure to por-
nography. But it is certain that people in the aggregate are wrong-they d o sys-
tematically judge others to be more negatively influenced by pornography than
themselves. Either they are underestimating influence o n themselves, overesti-
mating influence on others, or doing some of both. Some past research sug-
35
Journal of Communication, Winter 1995
I"Two of the experimental studies (Gunther, 1991; Gunther & Thorson, 1992) have analyzed the
source of inaccuracy, and those findings favor the view that people err in overestimating effects on
others.
36
Overrating the X-Rating
References
Bryant, J. & Brown, D. (1989). Uses of pornography. In D. Zillmann & J. Bryant (Eds.), Pornogru-
phy: Research Advances andpolicy considerations (pp. 25-56). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Colien, J., Mutz, D., Price, V. & Gunther, A. (1988). Perceived impact of defamation: An experiment
in third person effects. Public Opinion Q u a ~ e d y52,
, 161-173.
Davison, W. P. (1983). ‘The third-person effect in communication. Public Opinion Quurterly, 47,
1-15.
DeLdmdter, J. (1982). Response effects of question content. In W. Dijkstra & J. van der Zouwen
(Eds.), Response Bebauior in the Survq Interview (pp. 13-48). London: Academic Press.
Donnerstein, E. 8r Berkowitz, L. (1981). Victim reactions in aggressive erotic films as a factor in vio-
lence against women. Journal of Personality and Social Psychology, 41, 710-724.
Downs, D. A. (1990). The newpolitics 0fpornograph.y Chicago: University of Chicago Press.
Green, P. & Gerkin, A. E. (1989). Self-interest and public opinion toward smoking restrictions and
cigarette taxes. Public Opinion Quarterly, 53, 1-16.
Gunther, A. C. (1991). What we think others think: Cause and consequence in the third-person
effect. Communication Research, 18, 355-372.
Gunther, A. C. & Mundy, P. (1993). Biased optimism and the third-person effect.Journalism Quar-
terly, 70, 5 8 4 7 .
37
Journal of Communication, Winter 1995
Gunther, A. C. & Thorson, E. (1992). Perceived persuasive effects of product commercials and pub-
lic service announcements: Third-person effects in new domains. Communication Research, 19,
574-596.
Herrman, M. S. & Bordner, D. C. (1983). Attitudes toward pornography in a Southern community.
Criminology, 21, 349-374.
Howard, J., Reifler C. & Liptzin, M. (1970). Effects of Exposure to Pornography. Technical Report of
the Commission on Obscenity and Pornography. Vol. 8. Washington, DC: U.S. Government Print-
ing Office.
Kernell, S. (1986). Going Public: New Strategies of Presidential Leadership. Washington, DC: Con-
gressional Quarterly Press.
Kinder, D. R. & Kiewiet, D. R. (1981). Sociotropic politics: The American case. BritishJournal of
Political Science, 11, 129-161.
Lasorsa, D. L. (1989). Real and perceived effects of “Amerika.”Journalism Quarterly, 66, 373-378
McCloskey & Brill (1983). Dimensions of tolerance: What Americans believe about civil liberties
New York: Russel Sage Foundation.
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, 1-21.
Mutz, D. C. (1992). Mass media and the depoliticization of personal experience. American Journal
of Political Science, 36, 483-508.
Newsweek. (1985, March 18). The war against pornography, pp. 5 8 4 6
Perloff, R. (1989). Ego-involvement and the third-person effect of televised news coverage. Commu-
nicution Research, lG, 236262.
Randall, R. S. (1990). Freedom and taboo: Pornography and thepolitics of a selfdivided. Berkeley,
CA: University of California Press.
Rucinski, D. & Salmon, C. T. (1990). The ‘other’as the vulnerable voter: A study of the third-person
effect in the 1988 campaign. International Journal of Public Opinion Research, 2, 345-368.
Svenson, 0 . (1981). Are we all less risky and more skillful than our fellow drivers? Acta Psycholog-
ica, 47, 143-148.
Thompson, M. E., Chaffee, S. H. & Oshagan, H. H. (1990). Regulating pornography: A public
dilemma.Journal of Communication, 40, 73-83.
U.S. Commission on Obscenity & Pornography. (1970). Report. Washington, DC: U.S. Government
Printing Office.
Vallone, R., Ross, L. & Lepper, M. (1985). The hostile media phenomenon: Biased perception of
media bias in coverage of the Beirut massacre. Journal of Personality and Social PKychology, 49,
577-585.
Weinstein, N. D. (1989, Dec. 8). Optimistic biases about personal risks. Science, pp. 1232-1233
Wilson, J. B., Linz, D. & Randall, B. (1990). Applying social science research to film ratings: A shift
from offensiveness to harmful effects. Journal of Broadcasting and ElectronicMedia, 34, 443-468.
Zillmann, D. & Bryant, J. (1989). Pornography;Research advances andpolicy considerations. Hills-
dale, NJ: Lawrence Erlbaum Associates.
38