|
Self-esteem, self-affirmation and threats to self-worth:
Testing a motivational explanation for the third-person effect
Patrick C. Meirick
Assistant Professor
Department of Communication
University of Oklahoma
Submitted to the Communication Theory and Methodology Division of the
Association for Education in Journalism and Mass Communication for
presentation at the 2003 AEJMC Convention, Kansas City, MO.
Direct inquiries to the author at:
Dept. of Communication
University of Oklahoma
610 Elm Ave.
Norman, OK 73019
Phone: 405-325-1574
Fax: 405-325-7625
E-mail: [log in to unmask]
RUNNING HEAD: Third-person effect
Self-esteem, self-affirmation and threats to self-worth:
Testing a motivational explanation for the third-person effect
ABSTRACT
The self-enhancement explanation for third-person effects argues that
perceiving oneself as resistant to media messages enhances one's
self-esteem. The need to self-enhance can be increased by threats to
self-worth or reduced by self-affirmation (Steele, 1988). In Study One,
third-person effects did not vary by threat condition or by self-esteem,
although those high in self-esteem perceived smaller effects on themselves
and others. In Study Two, third-person effects were smaller among those
whose self-worth was affirmed.
The third-person effect is the tendency of people to believe that media
messages have greater effects on other people than on themselves (Davison,
1983; for a review, see Perloff, 1999). The effect is quite robust when the
messages in question are perceived as undesirable or the effects harmful –
pornography, libelous news stories, negative political ads, controversial
product ads, etc. – as has usually been the case in third-person research
(e.g., Cohen, Mutz, Price and Gunther, 1988; Rucinski and Salmon, 1990;
Shah, Faber and Youn, 1999; Thompson, Chaffee and Oshagan, 1990). But for
desirable messages, the third-person effect is attenuated (Brosius and
Engel, 1996; Duck and Mullin, 1995; Innes and Zeitz, 1988; Gunther and
Mundy, 1993) or reversed into a "first-person effect" in which people see
themselves as more influenced than others (Chapin, 2000b; Cohen and Davis,
1991; Gunther and Hwa, 1996; Hoorens and Ruiter, 1996; Meirick, 2000;
Price, Tewksbury and Huang, 1998).
This message desirability contingency has led some communication
researchers to posit a self-enhancement explanation for first- and
third-person effects (Gunther and Mundy, 1993; Hoorens and Ruiter, 1996;
Perloff, 1999; Meirick, 2000). Self-enhancement is described as "the
tendency to see one's own traits, abilities and prospects in a somewhat
exaggeratedly rosy light" (Taylor and Brown, 1988), and it is associated
with building ourselves up and with tearing others down (if we can justify
it) so we look good by comparison (Kunda, 1999). It is the Western self's
chief mechanism for maintaining and bolstering self-esteem. It would be
consistent with self-enhancement to think ourselves relatively invulnerable
to harmful media influence yet relatively receptive to pro-social media
messages. Some communication researchers have noted the similarity between
the third-person effect and unrealistic optimism (Weinstein, 1980, 1989; L.
Perloff and Fetzer, 1986), which pertains to people's belief that compared
to others, they are more likely to experience positive events and less
likely to experience negative events in the future. Unrealistic optimism,
also known as optimistic bias, is generally regarded as motivated by the
need for self-enhancement (Taylor and Brown, 1994), and it is presumed that
third-person effect is similarly motivated. However, few studies have gone
beyond manipulating message desirability in testing the self-enhancement
explanation for the third-person effect.
One experimental approach in social psychology may provide the means for a
more direct test of the self-enhancement explanation. Steele's work on
self-affirmation (Steele and Liu, 1983; Steele, 1988) suggests that people
respond to specific threats to self-worth by seizing on opportunities to
affirm valued (but sometimes unrelated) aspects of the self. While Steele
and Liu (1983) originally addressed such threats in the context of
dissonance studies, they and others researchers have applied the
self-threat approach to other areas of self-enhancement, such as
self-serving theories of success (Dunning, Leuenberger and Sherman, 1995),
derogating others (Fein and Spencer, 1997; Branscombe and Wann, 1994), and
exaggerating self-ratings of social qualities (Brown and Smart, 1991). In
each case, subjects responded to threats to the self-concept by manifesting
the various forms of self-enhancement at greater levels than subjects who
weren't threatened. Lest anyone argue that these responses do not
implicate self-esteem, there is evidence that the response to self-threat
interacts with the individual's self-esteem such that people with high
self-esteem enhance themselves more than those low in self-esteem do (Brown
and Smart, 1991). Moreover, when people are given an opportunity to affirm
the self through another activity, self-enhancing biases diminish or disappear.
Steele's approach has become established enough that Fein and Spencer
(1997) and Dunning et al. (1995) used self-threat and self-affirmation
manipulations to test the motivational underpinnings of outgroup derogation
and self-serving theories of success. This approach may provide a similar
test for the third-person effect. That is the purpose of the present
research. Study One will manipulate threats to self-worth, which may
increase third-person effects; Study Two will manipulate self-affirmation,
which should attenuate third-person effects.
Beyond message desirability: Recent third-person effect studies
The self-enhancement explanation for the third-person effect seems to pass
the message desirability test. But so far, few other tests have been
conducted. In what appears to be the only other experimental examination
of the self-enhancement explanation, Brosius and Engel (1996) attempted to
make admissions of influence upon the self more palatable in one condition
by casting the media effects questions in ways that offered the respondent
control over the effect, asking "Do you let yourself be influenced?" versus
asking "Do the media influence you?" in the other condition. They expected
to find a diminished third-person effect when the respondent could be an
active participant in his influence and thereby maintain
self-determination, but the differences didn't reach significance. This
was an intuitively appealing idea, but the experimental manipulation had no
track record. Moreover, while maintaining agency could be enhancing, it's
possible that it would be even less enhancing to conceive of oneself as
willingly allowing undesirable influence.
A few recent studies have explored correlations between the third-person
effect and self-esteem. This is relevant to the self-enhancement
explanation because those high in self-esteem are thought to use
self-enhancement mechanisms more than those low in self-esteem (Taylor and
Brown, 1988). However, the findings have been inconsistent.
David and Johnson (1998) examined the third-person effect of media images
of perfect bodies and found a stronger third-person effect for people with
high self-esteem. But Chapin (2000a) found a non-significant negative
correlation (r=-.14) between self-esteem and the perception that others
were more affected than oneself. However, it should be noted that the
stimuli Chapin used were two safer-sex messages, which many would consider
smart to be influenced by, so a negative correlation with the third-person
effect (or a positive correlation with the first-person effect) would be
consistent with the expectation that those high in self-esteem would have
more self-enhancing responses. Banning (2001) hypothesized that
self-esteem would have a positive linear relationship with the third-person
effect, but found no such relationship. He also expected to find a
curvilinear relationship in which those highest in self-esteem would have a
smaller third-person effect than those who scored just moderately high;
that did not materialize, either.
Finally, there have been two recent studies that examined the third-person
effect's correlation with optimistic bias, one form of
self-enhancement. Chapin (2000a) examined third-person effects among
teenagers in response to safe-sex messages. It also asked teens optimistic
bias questions about their relative likelihood of contracting
HIV/AIDS. Chapin expected to find a positive correlation between
third-person effect and optimistic bias - that teens who thought themselves
relatively less likely to be persuaded also would be perceive themselves as
less likely to be infected. Instead, he found a small negative
relationship. In hindsight, Chapin (2000a) attributed this result to the
pro-social nature of the safe-sex messages. Chapin (2000a) argued that
optimistic bias would correlate more strongly with the third-person effect
if the messages in question were undesirable messages, such as those
typically used in third-person effect studies.
Salwen and Dupagne (2000) set out to put that assertion to the
test. Toward the end of 1999, they conducted a survey asking people about
the likelihood of Y2K problems for themselves and for other people. They
also asked them how strongly they agreed or disagreed that news coverage
had had a significant impact on their (other people's) beliefs about
Y2K. Although Salwen and Dupagne (2000) found evidence of third-person
effect and of optimistic bias, they observed almost no correlation (r=.04,
ns) between the two. However, news coverage was not necessarily perceived
as undesirable. Perceived quality of news coverage had a -.21 beta
(p<.001) in predicting third-person effect, a finding is consistent with
message quality findings that support a self-enhancement explanation
(White, 1997). The fact that message desirability was ambiguous may again
explain why there was no correlation between third-person effect and
optimistic bias. The negative events described in the optimistic bias
portion of the study were hardly ambiguous as to their desirability. Thus,
a first-person effect among those who found the news coverage desirable and
a third-person effect among those who found it undesirable could both be
routes to self-enhancement, but they would look like noise if one was
expecting a clear correlation between third-person effect and optimistic bias.
Of course, as Salwen and Dupagne (2000) conceded, "a single-issue study
cannot adequately explain the relationship between judgments about
experiencing events and judgments of media effects about the events" (p.
22). Moreover, the relationship between optimistic bias and third-person
effect is not necessarily relevant to the relationship between
self-enhancement and third-person effect because it bypasses the presumed
underlying cause: the need to maintain and enhance self-esteem. A more
appropriate test might manipulate the extent to which people feel the need
to enhance self-esteem. And self-threat and self-affirmation may provide
such tests.
Threats to self-worth and self-enhancement
Steele (1988) developed self-affirmation theory to explain how people
protect themselves from threats to self-worth. The central idea of
self-affirmation theory is that people may counter a threat to one aspect
of the self by affirming an unrelated aspect. The means of self-affirmation
vary, but these processes and strategies appear to serve largely the same
function and can, in fact, be substituted for one another to accomplish the
goal of enhancing self-esteem. Tesser (2000; 2001; Tesser and Cornell,
1991) describes this as the "confluence of self-esteem maintenance
processes."
For example, Fein and Spencer (1997) argued that stereotyping and prejudice
may be a common way for people to maintain their self-image. To test this
argument, they manipulated threat to self-worth. Subjects were more likely
to evaluate a gay man or a Jewish woman stereotypically or negatively if
they had received negative feedback (rather than neutral or positive) on an
intelligence test. Among those who had been threatened, derogating an
outgroup member mediated an increase in self-esteem. Similarly, other
researchers (Sinclair, 1998; Kunda and Sinclair, 1999) have found that
white men derogated black and female evaluators when the evaluators
evaluated them poorly, but not when they evaluated them highly.[1]
Dunning, Leuenberger and Sherman (1995) applied the self-threat paradigm to
Dunning's ongoing study of self-serving theories of success (e.g., Dunning,
Perie and Story, 1991; Dunning and Cohen, 1992). This tendency to describe
prototypes of excellence in ways that reflect one's own traits appears
similar to overly positive views of self, one of the self-enhancement
biases identified by Taylor and Brown (1988). Dunning, Leuenberger and
Sherman (1995) found that subjects who were told they failed at an
intellectual task 1) had self-serving theories about the attributes
essential to a successful marriage and 2) evaluated targets similar to
themselves more favorably than they did targets unlike
themselves. Subjects who were told they had scored highly on the
intellectual task showed no such self-enhancement.
The evidence is quite clear that threats to self-worth lead people to
perceive themselves in a self-enhancing light (Brown and Smart, 1991;
Dunning et al, 1995). It simply makes sense that self-enhancement can
achieved through positive perceptions of the self.
H1: Compared to those who aren't threatened, people whose self-worth is
threatened will perceive undesirable media messages to have less influence
on themselves.
There is some evidence that self-threat leads to less favorable evaluations
of others (Fein and Spencer, 1997; Beauregard and Dunning, 1998) or
outright sabotage of others' performance (Tesser and Cornell, 1991), but
the evidence has qualifiers. Tesser and Cornell's (1991) study involved a
special set of conditions: a close other who outperforms the self at a
self-relevant task. In other studies where people derogated others after a
threat to self-worth, the others who were derogated were members of
outgroups; ingroup members weren't derogated (Sinclair, 1998; Kunda and
Sinclair, 1999). So, to the extent that self-enhancement can be derived
from believing in the gullibility of one's fellow human, we would expect that
H2: Compared to those who aren't threatened, people whose self-worth is
threatened will perceive undesirable media messages to have more influence
on others.
Given these predictions for perceived effects on self and others, the
prediction for the third-person effect becomes clear:
H3: Compared to those who aren't threatened, people whose self-worth is
threatened will exhibit a stronger third-person effect; that is, they will
perceive undesirable media messages to have less influence on themselves
compared to others.
Self-esteem and responses to self-threat
Much of the work on self-enhancement seems predicated on the assumption
that these processes are universal. While they are certainly prevalent, at
least in Western culture, there are individual differences in the extent to
which people employ them. Taylor and Brown (1988) reviewed numerous
studies demonstrating that people with high self-esteem tend to actively
cope with failure, while those with low self-esteem tend to accept
it. Those high in self-esteem are more likely than those low in it to
claim more responsibility for success than failure (see Blaine and Crocker,
1993, for a review). However, Campbell (1986) found that while those high
and low in self-esteem both take credit for success, those high in
self-esteem are more likely to attribute failure to temporary situations,
while those low in self-esteem tend to blame themselves (Campbell, 1986).
Campbell's (1986) finding suggests that in dealing with self-esteem,
differences come to the fore when dealing with events that threaten
self-worth. A number of studies employing the self-threat approach have
borne out this differential response. Brown and Smart (1991) found that
people with high self-esteem who failed at a purported intelligence test
later exaggerated their social (but not intellectual) abilities on a
questionnaire and were more likely to help a graduate student. By way of
contrast, those with low self-esteem who failed the test denigrated their
social skills and were less helpful than low self-esteem subjects who
succeeded. This is consistent with other findings that for people high in
self-esteem, their personal strengths become highly cognitively accessible
in the wake of personal failure, while people with low self-esteem showed
no evidence of similar enhancement after failure (Dodgson and Wood, 1998).
These studies suggest that threats to self-worth have greater positive
consequences for self-perceptions upon those high in self-esteem than those
low in self-esteem. Resistance to undesirable messages is a valued trait
and perceiving oneself to have great resistance would be self-enhancing, so
it should show the same pattern with regard to threat and self-esteem.
H4: Among those who are threatened, people with high self-esteem will
perceive undesirable media messages to have less influence on themselves
than will those with low self-esteem.
As discussed earlier, there is some qualified evidence that people may
self-enhance by derogating others in response to a self-threat. Some of
that evidence also suggests that this response is more likely among those
who are relatively high in self-esteem. Beauregard and Dunning (1998) found
that subjects' evaluations of another person's intelligence were more
negatively related to their own after a threat to self-esteem than after
self-affirmation; this difference was greater among those with high
self-esteem than those with low self-esteem. Attributing gullibility to
people is analogous to attributing stupidity to people, and it appears that
those high in self-esteem may be more likely to do so to self-enhance when
threatened than those low in self-esteem. Therefore:
H5: Among those who are threatened, people with high self-esteem will
perceive undesirable media messages to have more influence on others than
will those with low self-esteem.
If we expect those high in self-esteem to perceive less media influence on
themselves and more influence on others than do those low in self-esteem,
the prediction for the third-person effect is determined:
H6: Among those who are threatened, people with high self-esteem will
exhibit a stronger third-person effect - that is, they will perceive
negative media messages to have less influence on themselves compared to
others – than will those with low self-esteem.
S
Key variables
Threat to self-worth: Threat to self-worth was operationalized with two
versions of a purported reasoning test and feedback thereon. To enhance
the perceived validity of feedback where success and failure are randomly
assigned, McFarlin and Blascovich (1984) created two tests out of the 10
easiest and 10 hardest items from the Remote Associates Test (Mednick,
1962), a word-association task originally designed to test
creativity. After taking one of the tests, as in Brown and Smart (1991),
subjects were told they'd scored in either the 85th percentile (not
threatening to self-worth) or 35th percentile (threatening to
self-worth). The manipulation appeared successful. Those given the easy
version of the test scored much higher (6.76 out of 10) than those given
the hard version (1.42 out of 10, t145= 17.850, p<.001). Prior to getting
feedback, subjects were asked to rate their performance and their own level
of "cognitive flexibility," the trait purportedly being measured. Those
given the easy version rated their performance higher (4.39 on a 1-to-7
scale) than those given the hard version (2.29, t145=8.859, p<.001), and
they rated their ability higher as well (4.85 vs. 3.77, t145=4.343, p<.001).
Self-esteem: A seven-point version of the 10-item Rosenberg (1965)
self-esteem scale was used to measure global self-esteem. Respondents
indicate their level of agreement or disagreement with questions such as
"On the whole, I feel satisfied with myself" and "At times I think I am no
good at all" (reverse-scored). The scale was highly reliable, with a
Cronbach's alpha of .90. Scales scores ranged from 70, the highest
possible, to 13. The mean scale score was 55, with a median of 57 and a
standard deviation of 9.45. Where required for analysis, the scale was
dichotomized and those scoring in the middle quintile were omitted. The
low self-esteem group consisted of those who scored 54 or lower, while the
high self-esteem group consisted of those scoring 63 or higher.
Third-person effect: An overall scale was created to assess the magnitude
of the third-person effect for each subject. The third-person effect was
assessed with sets of questions about the impact of different types of
undesirable media messages on themselves, others in their class and the
public in general. Specifically, subjects were asked about: cigarette
advertising effects on attitude toward brand and likelihood to smoke; diet
pill infomercial effects on belief in the effectiveness of the product and
likelihood to buy; psychic hotline ad effects on belief in the
effectiveness of the psychics and likelihood to call; and Sept. 11 news
effects on attitudes toward flying and attitudes toward Arabs and Muslims.
The format for the answers was typically a bipolar scale from –3 (much less
likely/favorable) to +3 (much more likely/favorable)[2]. The rationale for
a –3 to +3 scale was twofold. First, it avoids the potential pitfall of
assuming the direction of an effect. Second, it recognizes that people
desire autonomy in their attitudes and behaviors. They resist influence
through counterarguing and source derogation (Wright, 1973; 1980) such that
their attitudes may move in the opposite direction from the persuasive
intent of the message (Hoorens and Ruiter, 1996).
To explore the specific hypotheses concerning perceived effects on self and
others, scales were created to assess these effects. Alphas were .77 for
the 16-item scale of effects on others and .59 for the eight-item scale of
effects on self. To address the hypotheses concerning the third-person
effect itself, the 16 self-students and self-public difference scores were
summed to create a third-person effect scale. The questions about Sept. 11
news were reverse-coded so that positive scores reflected greater
message-consistent persuasion (fear of airlines and antipathy toward
Arabs). For ease of interpretation, the scale score sums were divided by 16
to show a mean self-other difference on the –3 to +3 scale.[3]
Method
Subjects: The 151 subjects were recruited from introductory, intermediate
and advanced classes in journalism. The average age was 20.9, 73.5 percent
were female, 83.2 percent were white, and 64.9 percent were sophomores or
juniors. Initially, a $250 cash drawing was offered as the lone incentive
for participation. Recruitment was eventually extended to classes that
offered extra credit, which accounted for 85.4 percent of the
subjects. Four subjects who scored between 0 and 2 on the easy version of
the test (and were therefore unlikely to believe they had scored in the
85th percentile) were removed from analyses for hypothesis testing.
Procedure: Sessions were scheduled over the course of two months. The
experimenter greeted the subjects, who arrived one or two at a time, and
then took them to individual rooms. He then explained the study, went over
the consent forms and gave them the first questionnaire, a 10-item
self-esteem measure (Rosenberg, 1965). A few minutes later, the
experimenter returned to introduce the "Cognitive Flexibility Test," which
was described to them as a measure of reasoning ability that was related to
college GPA and scores on standardized tests like the GRE and the
LSAT. Subjects were randomly assigned[4] to take an easy or difficult
10-item version of the Remote Associates Test (Mednick, 1962) and allowed
five minutes to complete it. After taking the test, subjects were given a
feedback form that showed the correct answers, the number of correct
answers they had, and their purported percentile rank among college
students nationwide. Those given the easy test were told they'd placed in
the 85th percentile; those given the hard version were told they'd placed
in the 35th percentile. Then, subjects evaluated the impact of different
types of media messages on themselves, other students in their class, and
the public in general. That questionnaire concluded with demographics
questions and a probe for what they study was about. Finally, subjects
were debriefed. In all, the procedure took 25 to 30 minutes.
Results
The third-person effect: An overall test of the third-person effect showed
that the mean perceived effect on others (.56 on a –3 to +3 scale) was
greater than the mean perceived effect on self (-.91, t150=26.866,
p<.001). Examined individually, effects on others were significantly
greater than effects on self for each of the 16 self-other
comparisons. All paired samples t's were greater than 6, and all p levels
were less than .001.
Hypotheses 1, 2 and 3: The first set of hypotheses predicted that,
compared to those who weren't threatened, people whose self-worth was
threatened would exhibit smaller perceived effects on self, greater
perceived effects on others, and a stronger third-person effect. However,
as Table 1 shows, none of these predictions were accurate. Threat
condition had no effect on perceived effects on self, contrary to H1 (means
of -.91 for nonthreatening vs. -.87 for threatening, t145=.32, ns). Nor
did threat affect perceived message effects on others, contrary to H2
(means of .573 for non-threatening vs. .574 for threatening, t145= .01,
p>.99).[5] Similarly, there were no differences between the threat
conditions in the size of the difference between effects on self and
effects on others, so H3 finds no support. In the threat condition, the
average effect on others was 1.45 points higher than effects on self; the
self-other difference in the non-threatening condition was 1.48, virtually
equal, t145=.288, ns.
Hypotheses 4, 5 and 6: This set of hypotheses predicted that, among those
who were threatened, people with high self-esteem would exhibit smaller
perceived effects on self, greater perceived effects on others, and a
stronger third-person effect than would those with low self-esteem. See
Table 2. Those high in self-esteem perceived smaller harmful media effects
on themselves and others than did those low in self-esteem. Zero-order
correlations found that among those who were threatened, self-esteem was
negatively correlated with both perceived effects on self (r=-.21, p<.10),
in marginal support of H4, and on others (r=-.28, p<.05), in contradiction
of H5. To better illustrate these relationships (and to eliminate those
neither high nor low in self-esteem), t-tests were run comparing high and
low self-esteem groups. Among those who were threatened, those high in
self-esteem expressed greater resistance to harmful media influence (-1.11)
than those low in self-esteem (-.67), t57=2.757, p<.01. Likewise, they
perceived smaller effects on others (.40 vs. .81), t57=2.681, p=.01. Among
those whose self-worth was not threatened, self-esteem had no significant
effect on perceived effects on self, (-.99 for high vs. -.76 for low,
t56=1.249, p>.20) or others (.55 for high vs. .59 for low, t56=.253, ns).
As for the third-person effect, there were no differences to be found. The
simplest test, a zero-order correlation using only subjects who were
threatened (N=73), made use of the full range of the self-esteem scale, but
found no relationship (r=-.05, ns). To more clearly illustrate the effect
(or lack thereof), t-tests were run using dichotomized self-esteem with the
threatened subjects. Those high in self-esteem showed a third-person
effect with a mean self-other difference of 1.50, virtually the same as the
1.48 difference for those with relatively low self-esteem (t57=.120,
ns). H3 is not supported.
Discussion for Study 1
The third-person effect manifests itself in the absence of self-threat, but
it would appear that the third-person effect is not affected by threats to
self-worth, nor are perceived effects on self and others. Yet a
self-process appears to play a role; self-esteem is negatively related to
perceived effects of undesirable messages on both self and others. This
was the case across experimental conditions, but particularly in the
presence of a threat to self-worth, where the differences between
self-esteem conditions became statistically significant. Those high in
self-esteem did not perceive significantly smaller effects on themselves
under threat, nor did those low in self-esteem perceive significantly
greater effects on others under threat, but threat amplified the
differences between the two groups.
The result here is somewhat consistent with David and Johnson's (1998)
finding that self-esteem was negatively correlated with perceived effects
on both self and others, although they also found a greater third-person
effect for those with high self-esteem, which was not at all the case
here. The subject of David and Johnson's (1998) study, pictures of models
with unattainable shapes, may have been particularly apt for self-esteem to
influence the third-person effect. It may also be that women with high
self-esteem are accurate in their perceptions that other women are more
likely to be affected by such pictures. This latter interpretation would
be consistent with Peiser and Peter's (2001) "limits/possibilities
perspective" stating that one's perceptual position, vis a vis education or
media use, for example, may affect the tendency one has to manifest the
third-person effect.
The perception that one has great resistance to harmful messages would be
consistent with self-enhancement, and it makes sense that those high in
self-esteem would be more likely to manifest this belief. But from a
self-enhancement perspective, high self-esteem people perceiving relatively
small effects on others wouldn't be expected. It might make sense to think
of it from the other perspective, that people low in self-esteem perceive
greater effects on others. People with low self-esteem in the threat
condition perceived mean effects on others of .81, which post-hoc LSD tests
showed was significantly higher than the means for people with high self
esteem in both the non-threatening (.55, p<.10) and threatening (.40,
p<.01) conditions. The threatened high-self-esteem group differed
significantly only from the threatened low-self-esteem group.
It seems that people with high self-esteem enhance themselves through low
estimates of negative effects on self, which can be easily reconciled with
their positive self-beliefs. People with low self-esteem, on the other
hand, may find it harder to believe in their own efficacy and resistance to
persuasion, so enhancement through perceived effects on self is more
difficult. Indeed, people tend to prefer consistent self-conceptions
(Swann, 1982; Secord and Backman, 1965).
There is some evidence that people with low self-esteem may instead seek
self-enhancement indirectly, for instance, themselves by enhancing groups
with which they are closely affiliated (Brown, Collins and Schmidt,
1988). Are the higher estimates of negative effects on others on the part
of low self-esteem people a similar attempt at indirect
self-enhancement? Unlikely. Self-enhancement through derogating others is
practiced more by those high in self-esteem (Beauregard and Dunning, 1998).
In any case, it appears likely that self-enhancement processes do play some
role in perceptions of media effects. But if so, why didn't a threat to
self-worth affect these perceptions? A possible explanation is that there
may be a ceiling on people's vigilance against influence from undesirable
messages, so that the threat manipulation simply couldn't move them much
further in that direction. After all, the third-person effect has been
found in dozens of studies that did not experimentally threaten people's
self-worth. People don't need any prompting to consider themselves more
resistant than others to harmful messages. Furthermore, the messages used
in this study were chosen for their high degree of undesirability. They
would have been found noxious whether or not one's self-worth was
threatened, and that may have contributed to a ceiling effect. When
perceived resistance to an undesirable message is already very high, there
may be no self-enhancement to be derived from perceiving more
resistance. Indeed, extreme "knee-jerk" resistance is a trait that is
frowned upon. There just may have been no room for perceived effects on
self to decrease, or the third-person effect to increase, in this study.
However, a self-affirmation manipulation would not be hampered by a ceiling
effect on vigilance. Quite the opposite. Prior self-affirmation would
presumably reduce the need to self-enhance through believing oneself
resistant to harmful media messages. That is one reason that a
self-affirmation study is necessary.
Self-enhancement and self-affirmation
Some of the first work addressing the reduction of need for
self-enhancement was done by Steele (Steele and Liu, 1983), who
reinterpreted dissonance studies from a self-affirmation perspective. In
this view of dissonance reduction, an individual's positive self-image is
threatened by the awareness of having done something foolish or wrong, and
an attitude change can make the errant behavior seem more reasonable, thus
diminishing the blow to self-worth. Steele and Liu (1983) made their case
for this view of dissonance reduction by providing another path to
self-affirmation – calling to mind valued aspects of the self – for some
subjects who had been induced to write an essay favoring a policy they did
not support. In one set of conditions, subjects were given a questionnaire
about politics and economics prior to the attitude questionnaire. For
those subjects who cared about politics and economics, attitude change was
minimal; they had already enhanced themselves by answering the questionnaire.
Further support for the notion of self-affirmation through alternate paths
came within the context of Tesser's self-evaluation maintenance model
(Tesser, 1980; Tesser and Campbell, 1983; Tesser and Cornell, 1991), which
Tesser was putting forward as Steele was formulating his self-affirmation
model. According to Tesser's model, people will act to minimize the threat
to self-worth posed by someone else's superior performance by becoming less
close to that person, reducing the self-relevance of the performance domain
or minimizing the other's performance. In Tesser and Cornell (1991), when
subjects believed that a word game was related to intelligence (i.e., that
it was self-relevant) – and that a friend and a stranger had outperformed
them in the first round (i.e., self-worth was threatened) – subjects aided
the stranger rather than the friend. They helped the friend when they
believed the game was just a game, presumably to bask in the friend's
reflected glory. But both patterns of differential helping were eliminated
when another opportunity for self-affirmation was provided.
Prior self-affirmation may also serve to pre-empt self-enhancing biases in
the absence of a threat to the self. Fein and Spencer's (1997) work on
prejudice also included a study in which half of the subjects had an
opportunity for self-affirmation via writing a few paragraphs about a value
important to them, and half did not. The self-affirmed subjects evaluated
a Jewish and a non-Jewish job candidate equally favorably, while those who
were not self-affirmed evaluated the Jewish candidate more
negatively.[6] Positive reinforcement through a high score on an
intellectual task kept people in the theories-of-success study by Dunning
et al. (1995) from concocting self-enhancing theories of success. This
came in lieu of a threat and seemed to have served the same purpose as the
prior self-affirmation described here. Moreover, Dunning has documented
self-serving theories of success in the absence of self-threats; people
employ them spontaneously, as they do with other self-enhancing mechanisms
like optimistic bias (Weinstein, 1980; 1989) and self-serving attribution
bias (Snyder, Stephan and Rosenfield, 1976). Yet getting prior affirmation
apparently obviated the need to self-enhance. This suggests that people do
not self-enhance exclusively to heal a wounded ego, nor do they
self-enhance at every possible opportunity. Apparently threat is not a
necessary prerequisite for self-enhancing behavior; in some cases, it may
merely whet an already healthy appetite for self-enhancement. But it also
seems that people's appetite for self-enhancement has limits, that they can
be temporarily sated with self-affirmation so that they pass on seconds, so
to speak.
The results of Dunning et al. (1995) suggest that when the need to
self-enhance is reduced, perceptions of the self won't be as exaggeratedly
positive. It follows that a similar intervention may leave people less in
need of viewing themselves as invulnerable to media influence.
H7: Compared to those who aren't given an opportunity to self-affirm,
people who self-affirm will perceive undesirable messages to have greater
effects on themselves.
There is evidence that self-enhancement via outgroup derogation (Fein and
Spencer, 1997) or sabotage of rivals (Tesser and Cornell, 1991) is lessened
when an opportunity to self-affirm is provided first. This suggests that
the need to perceive others as vulnerable to media influence might be
smaller for those who self-affirm.
H8: Compared to those who aren't given an opportunity to self-affirm,
people who self-affirm will perceive undesirable messages to have smaller
effects on others.
To the extent that the previous two hypotheses are upheld, we would also
expect that:
H9: Compared to those who aren't given an opportunity to self-affirm,
people who self-affirm will exhibit a weaker third-person effect.
S
Key variables
Effects on self and effects on others: The scales were calculated as
described in Study 1. Alphas were .82 for undesirable messages on others
and .62 for undesirable messages on self.
Third-person effect: The scale was calculated as in Study 1. Using Cohen
and Cohen's (1983) formula for difference-score scales, reliability was .52
for the third-person effect scale.
Self-affirmation: Randomly selected students were given the opportunity to
self-affirm using a procedure devised by Fein and Spencer (1997). In the
self-affirmation condition, subjects were asked to circle one of four
values from the Allport-Vernon values inventory that meant the most to
them, then write a few paragraphs about why it was important to them. In
another version, subjects were asked to write about the value that was
least important to them and write about why it might be important to
someone else.
This mode of self-affirmation is a variation on the procedure used by
Steele (e.g., Steele and Liu, 1983; Steele, 1988), which had all students
write on a given value, which would be important to some and not to
others. Fein and Spencer's (1997) procedure ensures that assignment to the
affirmation conditions is random and not potentially confounded with a
given value.
Method
Subjects: Subjects (N=75) were recruited from an introductory psychology
class. The subjects had a mean age of 19.2, 58.1 percent were female, 82.4
percent were white and 66.2 percent were first-year students. Subjects
were given credit toward a research experience class requirement. One
subject in the self-affirmation condition failed to write anything and was
excluded from the analysis.
Procedure: Data were gathered in two large group sessions after class,
once in mid-March and once in early May. The two versions of the
questionnaire were placed in random order in advance. Subjects were given
the questionnaire to fill out in their seats. The first section asked them
to circle the value (business/economics, art/music/theater, social
life/relationships or science/pursuit of knowledge) that was most important
(self-affirmation condition) or least important (no affirmation) to
them. In the remaining space on the page, subjects were asked to explain
why the value they circled was important to them (self-affirmation) or
might be important to someone else (no affirmation). After that, the
subjects all filled out the message evaluation questionnaire and
demographic questions that were used in Study 1. They then filled out the
paperwork required to receive participation credit and were
debriefed. Most people finished the questionnaire in 15 minutes.
Results
The third-person effect: An overall third-person effect was found. The
mean perceived message-consistent effect of undesirable messages on self
was -.84 on a –3 to +3 scale, much lower than the mean of .66 for
others. The resulting difference, 1.51, was significant (t73=16.236, p<.001).
Looking at each of the 16 self-students and self-public comparisons
individually, all showed effects on others significantly greater than
effects on self, with all t's greater than 4.9 and all p levels less than
.001.
Hypotheses 7, 8 and 9: It was expected that, compared to those who weren't
given an opportunity to self-affirm, people who self-affirm would exhibit
greater message-consistent effect on self, smaller effects on others, and a
weaker third-person effect. See Table 3. It appeared that those who
self-affirmed may have perceived themselves as less resistant to
undesirable messages (-.67) than those who did not (-1.00), a difference
that approached significance (t72=1.947, p=.055), a qualified confirmation
of H7. However, there was no significant difference in perceived effects
on others between those who self-affirmed (.61) and those who did not
(.72), (t72=.61, ns), so H8 finds no support.
While subjects in both conditions believed undesirable media messages would
affect others more than themselves, the self-other gap was indeed smaller
for those who did self-affirm (1.28) than those who did not (1.72), a
significant difference (t72=2.412, p<.05) in support of H9.
Discussion for Study 2
People who self-affirmed had a smaller third-person effect than those who
didn't, which supports a self-enhancement explanation. With their recent
self-affirmation, these subjects perhaps didn't feel as great a need to
believe themselves relatively invulnerable to harmful media influence. The
difference appeared to stem mainly from perceived effects on the self,
which is consistent with the findings of Dunning et al. (1995) and the
evidence that self-perceptions are usually the conduit for self-enhancement
(Brown and Smart, 1991; White, 1997). This result also supports the notion
that while self-enhancement need not be triggered by a threat to
self-worth, the need for it is finite. If people had an inexhaustible need
to enhance themselves, recent self-affirmation would not affect subsequent
opportunities to self-enhance. Here, it did.
The findings in Study Two also lend support to the "ceiling effect"
explanation for the non-findings in Study One. A threat to self-worth may
well have fueled the need to self-enhance. But people already are quite
resistant to undesirable messages, and perceiving oneself as rabidly
resistant may not have not have presented a viable avenue toward
self-enhancement. But Study Two would have avoided any such ceiling effect
because the self-affirmation manipulation tends to remove some of the usual
pressure to self-enhance. People who self-affirmed perceived themselves to
be only moderately resistant to undesirable messages, rather than strongly
resistant like the unaffirmed in Study Two (and both groups in Study One).
O
Based on the work of Steele (1988) and others who have employed his
methodology, it was expected that changing the need for self-enhancement
would change the size of the third-person effect in Studies One and
Two. To the extent that the third-person effect is driven by
self-enhancement, increasing or decreasing the extent to which people seek
enhancement was expected to increase or decrease the degree to which people
perceive themselves as resistant (and others as vulnerable) to undesirable
media messages. As it turned out, this expectation was half-right. Threat
to self-worth in Study One had no effect on the size of the third-person
effect, but the third-person effect in Study Two was indeed smaller among
people who had the opportunity to affirm themselves than among those who
didn't. The latter result supports the self-enhancement explanation for
the third-person effect. As discussed above, a ceiling on the third-person
effect (or more to the point, on people's willingness to perceive
themselves as resistant) might explain why changing the need for
self-enhancement changed the size of the third-person effect under
self-affirmation but not under threat.
Based on research showing that people high in self-esteem self-enhance more
than those low in self-esteem, particularly under threats to their
self-worth, it had been expected in Study One that people high in
self-esteem would demonstrate a larger third-person effect when threatened
than would people low in self-esteem. In fact, there was no difference in
the size of the self-other gap between the two groups. Those high in
self-esteem did perceive themselves to be more resistant to undesirable
messages than did those low in self-esteem, but the highs also perceived
smaller effects on others than did the lows. These differences between
self-esteem groups were significant for those who were threatened but not
those who weren't threatened. So in a condition in which we had reason to
expect greater self-enhancement among those high in self-esteem than among
those with low self-esteem, the only difference that was consistent with
the expectation was that highs perceived smaller media effects on
themselves than did lows. This suggests that perceptions of effects on
oneself may be a more important source of self-enhancement than perceived
effects on others or the self-other gap – the third-person effect --
itself. Supporting this suggestion is the fact that where significant
differences in the third-person effect were found in between the affirmed
and unaffirmed in Study Two, the differences were due almost entirely to
differences in perceived effects on self. These findings are consistent
with evidence that perceived effects on self are quite affected by
manipulations of message desirability, while evaluations of others appear
less affected (Gunther and Mundy, 1993; Salwen and Dupagne, 2000; White,
1997). In short, the case for the self-enhancement as an explanation for
perceived effects on self finds support in the current research.
However, the already shaky case for self-enhancement's role in perceived
effects on others finds no further evidence here. As noted before,
derogating others seems to be a relatively rare road to self-enhancement, a
road that is followed only if it has been paved with negative stereotypes
about the others being derogated (Kunda, 1999). In this research,
manipulating subjects' need for self-enhancement had no effect on how they
viewed the susceptibility of fellow students or people in general to
harmful media messages. It may well be that different results could be
found for socially stigmatized "others."
The current research suggests that self-enhancement, through its influence
on perceived effects on self, does have a role to play in the third-person
effect. This is not to rule out other explanations, especially ones that
bear on perceived effects on others. The relative roles of effects on
others and effects on self have long been debated in third-person effects
research. Perloff's (1993) first review of third-person effect research
offered overestimation of media effects on others and/or underestimation of
media effects on self as the two explanations for the third-person
effect. In studies that compare attitude change in the third-person study
with those in a control group, evidence has been found for both
explanations, but more consistently for overestimation. Some studies have
found support only for overestimation (Gunther, 1991; Lasorsa, 1989; Price
et al., 1998), while others have found both overestimation of effects on
others and underestimation of effects on self (Cohen et al., 1988; Gunther
and Thorson, 1992). The current findings suggest that self-enhancement is
a reason that people underestimate media effects on themselves, and that
lessening the need for self-enhancement can reduce the size of the
third-person effect.
This research has a number of limitations. It relies on a student
convenience sample, which could inflate the size of the third-person effect
(Perloff, 1999) and limit generalizability. Also, the reliability of the
third-person effect difference-score scale was rather low. But these
limitations are not crippling. The study was not attempting to measure
population parameters in the size of the third-person effect, in which case
a student sample would be suspect, but rather the effects of manipulating
basic psychological processes. There is no reason to believe that
different results would have been found in a sample of the general
public. As for the difference score scale, a more reliable scale would
have been preferable, but some significant findings were obtained despite
the subpar reliabilities.
This is perhaps the first study of the third-person effect in which
subjects' need to self-enhance was experimentally manipulated, a procedure
that confirms the findings of years of message desirability studies and
offers a level of causal inference not permitted by correlational
studies. We can say with more confidence than ever before that
motivational processes play a role in the third-person effect.
R
Mass Communication and Society, 4 (2), 127-147.
Beauregard, K.S. and Dunning, D. (1998). Turning up the
contrast: Self-enhancement motives prompt egocentric contrast effects in
social judgments. Journal of Personality and Social Psychology, 74 (3),
606-621.
Blaine, B. and Crocker, J. (1993). Self-esteem and self-serving biases in
reactions to positive and negative events: An integrative review. In R.
Baumeister (Ed.), Self-esteem: The puzzle of low self-regard (pp.
55-85). New York: Plenum.
Branscombe, N.R. and Wann, D.L. (1994). Collective self-esteem
consequences of outgroup derogation when a valued social identity is on
trial. European Journal of Social Psychology, 24, 641-657.
Brosius, H. and Engel, D. (1996). The causes of third person
effects: Unrealistic optimism, impersonal impact, or negative attitudes
toward media influence. International Journal of Public Opinion Research,
8, 142-162.
Brown, J.D., Collins, R.L. and Schmidt, G.W. (1988). Self-esteem and
direct versus indirect forms of self-enhancement. Journal of Personality
and Social Psychology, 55, 445-453.
Brown, J.D. and Smart, S.A. (1991). The self and social conduct: Linking
self-representations to prosocial behavior. Journal of Personality and
Social Psychology, 60, 368-375.
Campbell, J.D. (1986). Similarity and uniqueness: The effects of
attribute type, relevance and individual differences in self-esteem and
depression. Journal of Personality and Social Psychology, 50, 281-294.
Chapin, J.R. (2000a) Third-person perception and optimistic bias among
urban minority at-risk youth. Communication Research, 27, 51-81.
Chapin, J.R. (2000b). Not like me: Self vs. other distinctions in
first-person perception. Communication Research Reports, 17, 320-330.
Cohen, J. and Cohen, P. (1983). Applied multiple regression/correlation
analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence
Erlbaum.
Cohen, J. and Davis, R.G. (1991). Third-person effects and the
differential impact in negative political advertising. Journalism
Quarterly, 68, 680-688.
Cohen, J., Mutz, D., Price, V. and Gunther, A. (1988). Perceived impact of
defamation: An experiment on third person effects. Public Opinion
Quarterly, 52, 161-173.
David, P. and Johnson, M.A. (1998). The role of self in third-person
effects about body image. Journal of Communication, 48 (4), 37-58.
Davison, W.P. (1983). The third-person effect in communication. Public
Opinion Quarterly, 47, 1-15.
Dodgson, P.G. and Wood, J.V. (1998). Self-esteem and the cognitive
accessibility of strengths and weaknesses after failure. Journal of
Personality and Social Psychology, 75, 178-197.
Duck, J.M. and Mullin, B. (1995). The perceived impact of the mass
media: Reconsidering the third-person effect. European Journal of Social
Psychology, 25, 77-93.
Dunning, D. and Cohen, G.L. (1992). Egocentric definitions of traits and
abilities in social judgment. Journal of Personality and Social
Psychology, 63, 341-355.
Dunning, D., Leuenberger, A. and Sherman, D.A. (1995). A new look at
motivated inference: Are self-serving theories of success a product of
motivational forces? Journal of Personality and Social Psychology, 69, 58-68.
Dunning, D., Perie, M., and Story, A.L. (1991). Self-serving prototypes
of social categories. Journal of Personality and Social Psychology, 61,
957-968.
Fein, S. and Spencer, S.J. (1997). Prejudice as self-image
maintenance: Affirming the self through derogating others. Journal of
Personality and Social Psychology, 73, 31-44.
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. and 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. and Mundy, P. (1993). Biased optimism and the third-person
effect. Journalism Quarterly, 70, 1, 58-67.
Gunther, A.C. and Thorson, E. (1992). Perceived persuasive effects of
product commercials and public service announcements: Third-person effects
in new domains. Communication Research, 18 (3), 355-372.
Hoorens, V. and Ruiter, S. (1996). The optimal impact phenomenon: beyond
the third person effect. European Journal of Social Psychology, 26, 599-610.
Innes, J.M. and Zeitz, H. (1988). The public's view of the impact of the
mass media: A test of the 'third person' effect. European Journal of
Social Psychology, 18, 457-463.
Kunda, Z. (1990). The case for motivated reasoning. Psychological
Bulletin, 108, 480-498.
Kunda, Z. (1999). Social cognition: Making sense of people. Cambridge,
Mass.: MIT Press.
Kunda, Z. and Sinclair, L. (1999). Motivated reasoning with
stereotypes: Activation, application and inhibition. Psychological
Inquiry, 10, 12-22.
Lasorsa, D.L. (1989). Real and perceived effects of "Amerika." Journalism
Quarterly, 66, 373-378, 529.
McFarlin, D.B. and Blascovich, J. (1984). On the Remote Associates Test as
an alternative to illusory performance feedback: A methodological
note. Basic and Applied Social Psychology, 5, 223-229.
Mednick, S.A. (1962). The associative basis of the creative
process. Psychological Review, 69, 220-232.
Meirick, P. (2000). Political distance and message desirability: Three
studies of political advertising and the third-person effect. Presented at
the Association for Education in Journalism and Mass Communication annual
convention, Phoenix, AZ, August 2000.
Peiser, W. and Peter, J. (2001). Explaining individual differences in
third-person perception: A limits/possibilities
perspective. Communication Research, 28 (2), 156-180.
Perloff, L.S. and Fetzer, B.K. (1986). Self-other judgments and perceived
vulnerability to victimization. Journal of Personality and Social
Psychology, 50, 502-510.
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.
Price, V. Tewksbury, D. and Huang, L-N. (1998). Third-person effects on
publication of a Holocaust denial advertisement. Journal of Communication,
48 (2), 3-26.
Rucinski, D. and 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.
Salwen, M.B. and Dupagne, M. (2000). Optimistic bias and the third-person
effect: Public estimations of Y2K effects on self and others. Presented
at the Association for Education in Journalism and Mass Communication
annual convention, Phoenix, AZ, August 2000.
Secord, P.F. and Backman, C.W. (1965). An interpersonal approach to
personality. In B. Maher (Ed.), Progress in experimental personality
research (Vol. 2). New York: Academic Press.
Shah, D.V., Faber, R.J., and Youn, S. (1999). Susceptibility and
severity: Perceptual dimensions underlying the third-person
effect. Communication Research, 26, 240-267.
Sinclair, L. (1998). Justifying desired impressions of
evaluators: Motivated activation, application, and inhibition of
stereotypes. Unpublished doctoral dissertation, University of Waterloo.
Snyder, M.L., Stephan, W.G. and Rosenfield, D. (1976). Egotism and
attribution. Journal of Personality and Social Psychology, 33, 435-441.
Steele, C.M. (1988). The psychology of self-affirmation: Sustaining the
integrity of the self. In L. Berkowitz (Ed.) Advances in experimental
social psychology (Vol. 21, pp. 261-302). New York: Academic Press.
Steele, C.M. and Liu, T.J. (1983). Dissonance processes as
self-affirmation. Journal of Personality and Social Psychology, 45, 5-19.
Swann, W.B. (1982). Self-verification: Bringing social reality into
harmony with the self. In J. Sols and A.G. Greenwald (Eds.), Psychological
perspectives on the self, (Vol. 2), Hillsdale, NJ: Lawrence Erlbaum.
Taylor, S.E. and Brown, J.D. (1988). Illusion and well-being: A social
psychological perspective on mental health. Psychological Bulletin, 103,
193-210.
Taylor, S.E. and Brown, J.D. (1994). "Illusion" of mental health does not
explain positive illusions. American Psychologist, 49, 972-973.
Tesser, A. (1980). Self-esteem maintenance in family dynamics. Journal of
Personality and Social Psychology, 39, 77-91.
Tesser, A. (2000). On the confluence of self-esteem maintenance
mechanisms. Personality and Social Psychology Review, 4 (4), 290-299.
Tesser, A. (2001). On the plasticity of self-defense. Current Directions
in Psychological Science, 10 (2), 66-69.
Tesser, A. and Campbell, J. (1983). Self-definition and self-evaluation
maintenance. In J. Suls and A. Greenwald (Eds.) Social psychological
perspectives on the self (pp. 1-31). Hillsdale, NJ: Erlbaum.
Tesser, A. and Cornell, D.P. (1991). On the confluence of self
processes. Journal of Experimental Social Psychology, 27, 501-526.
Thompson, M.E., Chaffee, S.H., and Oshagan, H.H. (1990). Regulating
pornography: A public dilemma. Journal of Communication, 40, 73-83.
Weinstein, N.D. (1980). Unrealistic optimism about future life
events. Journal of Personality and Social Psychology, 39, 806-820.
Weinstein, N.D. (1989). Optimistic biases about personal risks. Science,
Dec. 8, 1989, 1232-1233.
White, H.A. (1997). Considering interacting factors in the third-person
effect: Argument strength and social distance. Journalism and Mass
Communication Quarterly, 74, 3, 557-564.
Wright, P. (1973). The cognitive processes mediating acceptance of
advertising. Journal of Marketing Research, 10, (Feb.), 53-62.
Wright, P. (1980). Message evoked thoughts: Persuasion research using
thought verbalizations. Journal of Consumer Research, 7 (Sept.) 151-175.
Wylie, R. (1979). The self concept. Lincoln: University of Nebraska Press.
T
Table 1: Perceived effects of undesirable messages by threat condition
Threat
No threat
t value
S
-.87
-.91
.320
O
.57
.57
.010
Third-person effect
1
Note: ***p<.001 **p<.01 *p<.05 ^p<.10
N=73 for threat. N=74 for no threat.
Table 2: Perceived effects of undesirable messages under threat by self-esteem
Low self-esteem
High self-esteem
t value
S
-.67
-1.11
2.757**
O
.81
.40
2.681**
Third-person effect
1
Note: ***p<.001 **p<.01 *p<.05 ^p<.10
N=31 for low self-esteem. N=28 for high self-esteem.
Table 3: Perceived effects of undesirable messages by affirmation condition
Not affirmed
Self-affirmed
t value
S
-1.00
-.67
1.947^
O
.72
.61
.610
T
1
Note: ***p<.001 **p<.01 *p<.05 ^p<.10
N=38 for not affirmed. N=36 for self-affirmed.
[1] Note that non-minorities were not derogated, even when self-worth was
threatened. Kunda (1990) suggests that negative stereotypes provide
justification for derogation. Third-person effect does on some level
suggest a derogation of others, but only relative to the self: others are
not as sharp as I am. Also, to the extent that the "third persons" can be
considered part of an outgroup, derogation would perhaps be justified.
[2] Two sets of purchase behavior questions gave response options on a 1
(never) to 7 (definitely) scale. They were worded "After seeing (diet pill
infomercials/ psychic hotline ads), how likely do you think it would be
that (you/ other students in this class/the public in general) would (buy
the pills/call a psychic)?" In calculating the mean scores for the
scales, these responses were transformed to a -3 to +3 metric.
[3] Reliability of difference score scales can be calculated as the
average reliability of the components minus their correlation, all of which
is divided by 1 minus the correlation (Cohen and Cohen, 1983). Using this
formula, the reliability of the third-person effect scale is
low: .42. Reliability of the self-other difference score is not usually a
preoccupation of third-person effect researchers. Typically, the
third-person effect is assessed either by examining self-other differences
on an item-by-item basis (e.g., Hoorens and Ruiter, 1996) or by examining
difference scores of self and other indexes without reporting the
reliability of the difference score (e.g., Chapin, 2000a; Shah et al.,
1999). However, McLeod et al. (1997) managed to achieve respectable
reliabilities (.60 to .74) using Cohen and Cohen's (1983) formula on their
three-item scales of effects of rap lyrics on target groups' knowledge,
attitudes and behavior. In the present research, the topics vary much more
widely, which makes high reliabilities difficult to obtain. But despite
the diversity of the scale's topics, it may provide additional power
through aggregation. Any statistically significant results found using the
overall scale will be in spite of its low reliability, not because of it.
[4] Random assignment in all studies was done using the Research
Randomizer at www.randomizer.org, a Web site run by social psychologists
Geoffrey C. Urbaniak and Scott Plous. The Randomizer uses a Java applet
that taps an algorithm to produce near-random numbers.
[5] The two groups of others did differ in perceived influence, but threat
had no effect on perceived effects for either classmates (t 145=.152,
p=.88) or the public (t 145=.165, p=.87).
[6] Fein and Spencer (1997), citing Spencer and Steele (1990), are careful
to note that the self-affirmation manipulation they used, "in the absence
of self-image threat, … does not affect participants' state self-esteem"
(p. 32).
|
