Professional Documents
Culture Documents
R01 - Baldwin 1987 - Salient Private Audiences & Awareness of The Self
R01 - Baldwin 1987 - Salient Private Audiences & Awareness of The Self
0022-.*5l4/87/$00.'75
Mark W. Baldwin
Clarke Institute of Psychiatry, Toronto, Ontario, Canada
We used self-awareness and cognitive priming methodologies to test the hypothesis that important
aspects of the experience of self derive from the way one would be perceived and responded to by a
private audience of internally represented significant others. In the first study, 40 undergraduate
women visualized the faces of either two acquaintances from campus or two older members of their
own family. Later, when they rated the enjoyableness of a sexually permissive piece of fiction, they
tended to respond in ways that would be acceptable to their salient private audience. There was some
evidence that this effect was especially pronounced for subjects made self-aware by the presence of a
small mirror, whose responsivity to self-image concerns was presumably heightened. In the second
study, 60 undergraduate men were exposed to a failure experience, and their resulting self-evaluations were assessed. Self-aware subjects' responses reflected the evaluative style of a recently visualized private audience. Strong negative self-evaluative reactions on a number of measures were evident when the salient audience tended to make acceptance contingent on successful performances,
but not when the audience manifested relatively noncontingent acceptance. These results demonstrate the influence of internally represented significant relationships on the experience of self.
ker,
1980,
Eagle,
1984,
of self often takes the form of imagining how the self would be
This research was supported by Social Sciences and Humanities Research Council of Canada doctoral and postdoctoral fellowships to
Mark W. Baldwin and was the basis for his doctoral dissertation at the
University of Waterloo. Some of the data reported in this article were
presented at the June 1984 meeting of the Canadian Psychological Association, Ottawa, Ontario, and at the August 1984 meeting of the American Psychological Association, Toronto, Ontario.
We would like to thank Janet Hunt for serving as an experimenter in
Study 2. Michael Ross and Mark Zanna for valuable conversations during the earlier phases of the research, and Aaron Brower, Nancy Cantor,
John Ellard, Eric Lang, Mary Miller, Paula Niedenthal, Julie Norem,
Carolin Showers, and an anonymous reviewer for their helpful comments on drafts of the article.
Correspondence concerning this article should be addressed to Mark
W. Baldwin, Cognitive and Behavioural Therapies Unit, Clarke Institute of Psychiatry, 250 College Street, Toronto, Ontario, Canada M5T
1R8.
the
1087
1088
Study 1
Our first objective was to test the basic premise that the standards used by self-aware individuals often derive from salient
private audiences. Research has already supported the notion
that the behavior of individuals who focus on publicly observable self-aspects tends to conform to the evaluative standards of
an audience of onlookers (see Scheier & Carver, 1983b, for a
review). There seems to be an increased impact, when individuals are acutely aware of the public self, of the evaluative tendencies of others in the environment (Fenigstein, 1979). In contrast, individuals who instead focus on private aspects of self are
observed to respond to internal sources of evaluative standards
rather than to the demands of public audiences (e.g., Froming,
Walker, & Lopyan, 1982). The goal of the first study was to extend the analysis of audience processes to the more self-definitional, private aspects of self as well and to examine whether
personal standards might be influenced by internally represented audiences (cf. Diener & Srull, 1979; Wicklund, 1982).
The hypothesis was that subjects, particularly when privately
self-aware, would be responsive to evaluative standards linked
to some private audience. Guided visualizations were used to
prime selected private audiences to see whether this might highlight the evaluative standards associated with those representations.
Undergraduate women were asked to indicate their enjoyment of a sexually permissive piece of fiction (cf. Gibbons,
1978) after a subtle manipulation in a separate context that
primed either a presumably permissive private audience (campus friends) or a comparatively conservative one (older family
members). These two audiences were used because they were
expected to be emotionally significant but would represent di-
1089
asked if, duringa break in the study, they would mind helping out some-
one else who needed people to fill out some questionnaires for her study.
As the purpose of the study was to determine whether selfevaluative standards would derive from a salient private audience, the priming manipulation was designed to facilitate a
vivid experience of the person imagined while avoiding any di-
ror (16 in. X 12 in., or 40.5 cm X 30.5 cm) facing the subject, positioned
other half of the study. To diminish possible suspicion about the mirrors, the room was described as a communal laborator> and small explanatory signs were attached to the mirrors ("Please do not disturb
equipment. Perception experiment No. 1703. Dr. Ci. Zilstein"). Also,
was no evidence that subjects were aware of the actual purpose of the
procedures.
After they read and responded to a filler article about a nearby conser-
vation area, subjects were given the second passage, which was the criti-
cal one for the study. This was a sexually permissive piece of fiction,
borrowed from a popular women's magazine, depicting a woman having
a sexual dream about a man she was attracted to. The passage was se-
Method
Subjects
passage. They were asked to rate on 7-point scales how exciting, clear.
was. Half of these questions were measures of how much they liked the
passage, which was expected to be affected by the private-audience manipulation, and half asked for an evaluation of the quality of the writing
style, which was expected to be a dimension not affected by the treat-
Procedure
ment condition. These questions came roughly 10 minutes after the vi-
and that they would be asked to try to picture in their minds various
sures. They were asked to report their attitude toward premarital sex on
the following 7-point scales: bad-good, unacceptable-very acceptable,
think of two people they knew from campus, and the other half were
told to select two older members of their family. This served as the ma-
exposed to their mirror image throughout the time allotted for reading
the passage and responding to the questionnaire. All subjects were then
debriefed and thanked for their participation.
and try to get a visual image in their mind of the people chosen. Subjects' visualizations were guided by the experimenter, who read the following instructions to help them develop their images (instructions were
read verbatim, once for each target person selected; ellipsis points represent a pause of approximately 5 s):
Focus your attention on this person. . . . Picture the person's
face. Really try to get an experience of the person being with you.
. . . You may want to remember a time you were actually with the
person, or you may already have a clear experience of what this
person is like.. . . Just try to get a good image of this person. You
may find that you can see the color of their eyes or hair, or maybe
hear their voice. . . . Imagine that this person is right there with
you. . . Now once you "have an image of the person, try to zoom
in and get a close-up, focused impression.. . . Hold this image for
a little while.. . . Imagine talking with the person.. . . Try to feel
them there with you.
Results
The major dependent variable was how much the subjects
reported liking the sexually permissive passage. The three measures of how enjoyable, pleasurable, and exciting the passage
was were averaged to give a single score of liking. The average
pooled within-cell correlation of these measures was satisfactory, r(38) = .79. This score was analyzed in a 2 (private audience: family members, campus associates) x 2 (sclf-awarcncss:
mirror, no mirror) analysis of variance (ANOVA). The only significant effect was a main effect for the private-audience visualization, F(\, 36) - 4.46, p < .05, indicating that subjects who
had thought of friends from campus reported liking the story
After a short discussion about the vividness of the various images and
the helpfulness of instructions (to support the cover story), subjects were
1090
Table 1
Liking of Sexually Permissive Passage by Self-Awareness
story as significantly less enjoyable than did those who had visu-
Self- awareness
Campus
associates
Family
members
Mirror
No mirror
4.97
4.67
3.90
4.13
private-audience visualization did have an impact on how subjects responded to the sexually permissive passage.
The second hypothesis to be investigated involved possible
uative structures of their most salient private audience, they adjusted their behavior to be consistent with that audience's
standards.
Because our goal was to focus on private self-aspects and private audiences, we chose to use the mirror manipulation, which
is generally assumed to increase private self-awareness (Froming et al., 1982). It is perhaps possible that for some reason our
subjects, although seated alone in cubicles with mirrors, were
also sensitized to concerns of impression management in the
public domain, and this might explain the influence of socially
derived standards (following the studies reviewed earlier, e.g..
Carver & Humphries, 1981). This seems an unlikely explanation for our study, however, given the lack of precedent in the
literature for increased concern with public self-images result-
ing from similarly private mirror-exposure situations. We suggest instead that the private as well as the public self is ulti-
ence beyond this strong tendency to use their more general system of values as an anchor. Although the Self-Awareness X
sense of acceptance by a private audience. Moreover, this perspective can address the emotional distress often experienced
around negative self-evaluations, because falling short of stan-
ror while reading the passages and filling out the questionnaires
were affected by a recent visualization of either campus friends
our study, this would suggest that subjects responded to an image of self associated with the primed relationship structure, as
opposed to being motivated to maintain a position that would
style of the critical passage did not yield any significant effects.
to some extent.
An important question for additional research regards the
degree of awareness of private-audience dynamics. During the
Discussion
Subjects' responses to the sexually permissive passage were
affected by the manipulation of private audience. Those who
natives apart, and it seems likely that both factors arc involved
1091
more global evaluative style. Predictions were that self-awareness would lead subjects to evaluate their task performance in
a way reflecting the noncontingent or contingently accepting
style of their salient private audience.
dition. She spontaneously reported that when she went to respond to the statement "I believe that premarital sex is desir-
Method
Subjects
cle. It would seem that private-audience influences may function across a range of levels of consciousness, from information
processing directed by unconscious cognitive structures repre-
Subjects were 60 undergraduate men enrolled in an introductory psychology course at the University of Waterloo. Participation in the study
entitled them to course credits. They were run in sessions of I to 4
people by a male experimenter. Random assignment to the six experimental conditions took place before the subjects' arrival.
Study 2
Often when people fail to achieve certain standards of performance, they evaluate themselves negatively and experience
some degree of emotional distress. The second study was designed to investigate more directly the self-evaluations people
make, particularly following a failure experience. Other research has focused on specific parameters of assigning meaning
to outcomes, such as social comparison information (Festinger,
1954) and causal attribution (Weiner, 1974). One might predict
that the general style of making such construals (e.g., what outcomes and standards are considered, where responsibility is assigned, what conclusions are drawn) may be shaped by implicational rules associated with particular private audiences.
The hypothesis of this study was that the self-evaluative style
exhibited by self-aware individuals would reflect the style of
their most salient private audience. A key factor in this formulation is the degree to which acceptance by the private audience
is experienced as ultimately contingent on meeting certain evaluative standards. Specifically, one will take achievement or morality very seriously if he or she is felt to have implications for
acceptance by a significant private audience. When acceptance
is linked to the success of one's recent endeavors, an appraisal
of failure leaves one with a feeling of rejection by the significant
other, possibly recapitulating past emotional experiences with
similar features (Bower, 1981). Conversely, noncontingent acceptance has been posited as a hallmark of healthy functioning
(cf. Rogers, 1959, "unconditional positive regard"). It is generally assumed that this evaluative style essentially buffers a person from the emotional consequences of failure by diluting its
meaning in the broader context of life's events.
In Study 2 a visualization manipulation was used to prime
different audiences, which varied in the degree to which acceptance was made contingent on successful performance. Some
Procedure
After reading and signing a consent form, subjects were led to a laboratory containing four cubicles. Half the cubicles contained a small mirror, as the manipulation of self-awareness. As in Study 1. attaching explanatory notes to the mirrors and describing the room as "cluttered
with other experimenters' equipment" seemed to allay any concerns
about the mirrors, and no subject expressed suspicion.
Subjects then donned headphones and received all their instructions
from individual tape machines. They were asked to sit back, relax, and
keep their eyes closed most of the time while doing a number of visualization exercises. The first exercise then began, in which subjects were
instructed to try to get a quick visual image of a number of objects that
were named. There were 30 objects in all. presented at intervals of 4 s
each. After the final object, as a distractor task, subjects were immediately instructed to try to visualize the letters of the alphabet as they were
named at 2-s intervals.
The next exercise was made up of the manipulation visualizations of
private audience. Subjects' instructions were to imagine being in the
situation that was described and to try to experience it as if they were
there. After this visualization, there was a 30-s period during which they
were asked to think about the final visualization, any feelings that went
along with it, and any situations of a similar nature that they had
been in.
Next came the failure experience. It was explained that one aspect of
the visualization process was that it bore close links to memory. For the
next exercise they were to try to remember as many words as they could
from the list of objects they had heard earlier. They then took the questionnaire folder from the wall in front of them and began to write what
words they could remember. (It was only at this point that the mirror
became unobstructed for subjects in the self-aware condition.) As they
tried to write down the words, they were periodically instructed to draw
a line across the page under the words they had written. This served our
purpose of making salient the fact that they were remembering very few
words. Four min were allotted for this exercise. This task is actually
very difficult, and pilot subjects had confirmed that it made them feei
incompetent.
After completing the recall task, subjects were instructed to turn off
Note that although this woman was conscious of her imagined interaction with her mother, she had not considered that this might have
been brought on by the earlier priming visualization. She was therefore
not conscious of the determinants of her self-evaluative behavior (Bowers, 1985).
1092
the tape machine and complete the questionnaire containing the dependent measures (to be described). When they were finished the study was
explained to them, with emphasis on the fact that the recall task was
quite difficult and that it was constructed so that they would not be able
to recall many words.
Private-Audience Visualizations
The situations that subjects visualized before the recall task were chosen to vary the level of contingency of the evaluative structure of their
salient private audience. Each subject heard only one of three audience
visualizations: control, noncontingent, or contingent. The audience visualization lasted approximately 1 min.
Control. This visualization was intended to be neutral with respect
to self-evaluation processes and to have subjects absorbed in the experiences of the moment. Subjects were instructed to imagine themselves
walking down a sidewalk. Their instructions included such statements
as "picture the color of the grass" and "try to feel the concrete under
your feet."
Noncomingem. Subjects in the noncontingent acceptance condition
were asked to imagine having lunch with a good friend, someone who
would stick by them and support them through good times and bad.
They were encouraged to "feel the warmth and acceptance with this
person." The intention of this visualization was to have subjects feel
liked and accepted noncontingently.
Contingent. The contingent-audience visualization was also designed
to have people feel liked and appreciated, but this group was to feel that
the private audience's acceptance was contingent on performing well.
They imagined meeting and chatting with a new acquaintance about a
class assignment and then later overhearing this person say "he was really smart. . . . 1 like people like that." The goal of this visualization
was to have subjects feel liked and accepted by this person, with the
additional meta-level message that they were liked because of their
abilities.
Results
Private-Audience Assessments
Dependent Measures
Immediately after the recall task, subjects filled out a questionnaire
that included the following dependent measures.
Mood. The first page of the questionnaire was made up of 10 bipolar
mood scales. Each scale consisted of a 100-mm line with endpoints labeled according lo various positive moods and their negative counterparts (e.g., good-bad, contented-discontented, hopeless-hopeful). Subjects were asked to draw a slash across each line to indicate "how you
are feeling right now."
Self-esteem. Next came the Rosenberg (1965) Self-Esteem Scale,
which Robinson and Shaver (1973) have suggested measures the selfacceptance aspect of chronic self-esteem.
Performance attributions. Subjects were asked to indicate on a
7-point scale whether they felt that their recall performance reflected
"something to do with the situation (e.g., the task, the setting, etc.) or
something about yourself (e.g., personal characteristics)."
Social comparison. This question asked subjects to estimate how
many words on average "some people you know" would recall if they
did the memory task.
Overgeneralization. Many writers (e.g.. Beck, 1970; Buss, 1980;
Carver AGanellen, 1983; Ellis, 1977; Janoff-Bulman, 1979) have noted
that a major element in depressive states and in chronic self-esteem
problems is the tendency to draw global evaluative conclusions about
the self from specific outcomes or behaviors. Although such inferential
styles may be magnified in clinical samples, it seems likely that most
people have experienced to some degree contingent styles of evaluation
from significant others in their personal histories. A cognitive set repre-
Table 2
Private-Audience Assessments by Prnale-Audience Condition
Private audience
Measure
Noncontingent
Control
Contingent
Liked"
Evaluated11
Contingency'
2.30
2.80
2.85
3.50
3.35
4.2 l d
2.70
4.60
4.45
Mood
A single mood score was computed for each subject by determining how far (in millimeters) the subject's slashes were from
the affectively negative ends of the individual scales. These values were then summed over the 10 items to obtain a score between 0 and 1 ,000, with higher numbers indicating better overall mood. This mood scale was satisfactorily reliable, with an
average interitem r of .45 and alpha of .89. A 3 X 2 (Private
Audience x Self-Awareness) ANOVA yielded only an Audience X
Mirror interaction, F(2, 54) = 3.99, p < .05 (both main effects
p> .10). To interpret the means (see Table 3), we considered
the effects of each audience in turn when the failure was experienced under no-mirror compared with mirror conditions. Subjects in the control condition reported somewhat higher mood
when a mirror was present than when it was not, 1(54) = 1.87,
p = .08. The mood ratings by the noncontingent group did not
differ between mirror and no-mirror conditions, t < 1 . Individuals in the contingent condition showed a very different pattern.
They gave quite high ratings in the absence of a mirror, but their
overall mood was significantly lowerwhen a mirror was present,
1093
/(54) = 2.1 l,p < .05. Because the mood measure was intended
as one of a number of multiple indicators of a global self-evaluative style, the interpretation of these effects will be postponed
until the Discussion section, where the profiles of results on the
various measures can be considered together.
Although performance on the task was generally quite poor.3
there was some variability across subjects. An interesting question involves the degree to which mood is determined by number of words recalled. A central assumption of the notion of
contingency is that feelings should be tied closely to performance, that is, that mood should correlate with recall. Overall,
this correlation was marginally significant: pooled within-cell
r(58) = .25, p = .06. A strong association between mood and
recall should be expected only in the contingent audience condition, however, and this is what we found: control r( 18) - .05.
ns; noncontingent r(!8) - .28. nv, contingent r ( 1 8 ) = .54.
p<.05.
Overgeneralization
An Overgeneralization score was determined for each subject
by coding the most extreme statement endorsed for each hypothetical situation described (i.e., if only the first statement was
endorsed, the item was coded /, if the second statement was
endorsed the item was coded 2, etc.), and then summing across
the four situations. This yielded a score potentially ranging
from 4 to 20, where higher numbers represent a greater tendency toward drawing personal implications from negative information about the self. This Overgeneralization scale was of
satisfactory reliability given the small number of items: average
interitem r = .23, a = .51. Its validity was supported by a negative correlation with scores on the self-esteem scale, pooled
within-cell r(56) = -.25. p = .07, indicating that, as we would
expect, the process of Overgeneralization was related to lower
self-esteem.
The 3 X 2 ANOVA revealed no main effect for self-awareness
but a marginally significant effect for audience, h(2, 54) - 2.97,
p .06. As evident from Table 3, the pattern of means is more
appropriately interpreted in terms of the significant Audience X
Self-Awareness interaction, F(1. 54) = 4.45. p< .05. This interaction is mostly attributable to the contingent set, mirror group,
which showed significantly more Overgeneralization than did
the contingent set, no-mirror group. ((54) - 2.73. p< .01. There
were no differences within the control and noncontingent
groups.
1094
Table 3
Self-Evaluation Scores by Self-Awareness and Private-Audience Conditions
Private audience
Contingent
Control
Noncontingent
Measure
Mirror
No mirror
Mirror
No mirror
Mirror
No mirror
Mood'
Overgeneralizationb
Performance attributions e
Social comparison
496
9.60
3.20
5.50
517
10.00
4.40
8.70
599
8.90
5.10
7.90
491
10.00
1.60
6.00
528
12.10
5.30
8.11"
652
9.70
4.70
6.30
Performance Attributions
On a 7-point scale ranging from the situation to myself, subjects attributed responsibility for their performance on the recall task. The 3 X 2 ANOVA yielded a main effect for audience,
F(2, 54) = 3.92, p < .05. The noncontingent group attributed
the least responsibility to the self, M = 3.80, followed by the
control group, M = 4.35, with contingent subjects attributing
the most responsibility to the self, M 5.00. The individual
comparisons between experimental groups and control were
not significant, (s < 1.5.
The main effect for audience was moderated by an Audience X Self-Awareness interaction, F(2, 54) = 5.13,p < .01 (see
Table 3). Simple effects tests show that subjects in the control
condition attributed significantly more responsibility for the
performance to the self when there was a mirror present than
when there was not, ((54) = 2.47, p < .05. In the noncontingent
audience condition, the opposite pattern occurred: there was
less self-attribution when a mirror was present than when there
was not, ((54) = 1.98, p = .06. The contingent group did not
differ on this measure between self-awareness conditions, t< 1,
as their attributions of blame remained uniformly high.
Social Comparison
trol subjects may reflect Wicklund's (1978) contention that selfTo assess social comparison factors, subjects were asked to
estimate how many words "some people you know" would re-
words, or about 1.5 words more than subjects' mean recall. Neither main effect in the 3 X 2 ANOVA was significant, but the
ness aids the process by increasing access to other, positive selfaspects. In the absence of measures specifically designed to as-
interaction again was reliable, F(2, 53) = 5.08, p < .05 (see Ta-
Discussion
(p < .05) and more lenient than did their counterparts who had
not been made self-aware. One could argue that self-focused
subjects' social comparison estimate that others would be able
to recall 5.5 words is also more reasonable, as it comes close to
the actual overall mean recall of 5.58 words. An unexpected
finding was the pattern of relatively negative attributions and
social comparisons made by subjects in this audience condition
who were not made self-aware. It is not obvious to us why there
should be these apparently negative consequences of visualizing
an accepting friend. When self-awareness was induced, however,
the noncontingent visualization did seem to buffer the impact
of the failure by inducing a fairly forgiving if not excusing style
of self-evaluation. It was as if an internalized friend were saying
"don't worry about it, it was a difficult situation, and no one
else would have done any better." It is interesting to note that
when these subjects were asked in debriefing how they felt their
friends would react to their failure, they often suggested that
support of this type would be offered.
In the contingent-acceptance condition, subjects who were
not self-aware reported an elevated mood, probably resulting
from the guided fantasy of being admired and liked because
of a successful performance. In hindsight, it seems not at all
surprising that the positive aspects of this visualization might
buffer individuals who are not self-aware against a subsequent
failure (see Wright & Mischel, 1982). This result, however, underlines the strategic risk we took in priming contingency with
a success experience, because we are then in the position of having to show that the evaluative style of this private audience is
capable of undercutting this effect when self-focus is increased
by the self-awareness manipulation. The manipulation checks
clearly show that experiences other than feeling successful and
liked resulted: Individuals also reported feeling more evaluated
and sensed that acceptance depended more on doing well, compared with noncontingent subjects. It was these additional aspects of their cognitive set that we hypothesized would leave
them vulnerable to negative self-evaluative experiences.
Consistent with this notion, subjects in both contingent conditions showed the strongest tendency to blame themselves for
their poor performance and to link their reported mood directly
to their performance on the task (r = .54). These undercurrents
of self-doubt became manifest when subjects' self-focus was increased. Self-aware subjects reported significantly lower moods
than did their non-self-aware counterparts (p < .05), presumably as a result of applying the contingent style of self-evaluation
to the subsequent poor task performance (i.e., "because I am a
failure, people will not accept me"). The most striking evidence
of a contingent mode of evaluation comes from the overgeneralization measure: Self-aware subjects were much more likely
(p < .01) to overgeneralize from specific negative information
about the self. The fact that the overgeneralization measure was
made up of a number of different hypothetical situations (e.g.,
denying a transgression, being unable to learn a new sport) suggests that the insidious effects of a contingently accepting audience might generalize far beyond the particular domain in
which it is introduced.4
Although the pattern of the means on the various measures
was not always entirely consistent, it did appear to reflect
different styles of self-evaluation for subjects with different sa-
1095
lient private audiences under conditions fostering self-awareness. Individuals in the noncontingent group, whose private audiences did not base acceptance on successful performances,
did not display overly self-critical responses after the failure on
the memory task. Subjects with contingently accepting audiences, in contrast, seemed inclined to accept blame, draw negative evaluative implications about the self, and report moods
that closely reflected the adequacy of their performance. A multivariate analysis supported this view of the measures as multiple indicators of the effects of the conditions. An interesting aid
in interpreting the findings is to combine algebraically the four
measures by standardizing each and then summing for each
subject.5 A low score on this general index indicates a tendency,
following the failure experience, toward low mood, a high degree of overgeneralization, attributing the failure to something
about the self, and estimating that others would be able to perform better on the recall task. Psychologically, a low score would
seem to reflect a generally negative self-evaluative reaction to
the failure experience; high scores obviously reflect the opposite
tendencies. The results of this summary index are displayed in
Figure 1.6
The styles of evaluation engendered by private audiences displaying different levels of contingency have implications for a
wide range of self-evaluative phenomena. Striving to achieve in
order to attain the contingent acceptance of a private audience
could be the motivation underlying the psychological perspective that has been termed ego involvement (e.g.. deCharms,
own performance, this score was first expressed as a proportionaldifference score, that is, own recall minus estimate of others' performance, divided by own recall.
6
1096
Non-Contingent
<a
-a
c
o
<f--"^ Control
0
01
CO
Contingent
-2'
Mirror
Absent
Mirror
Present
figure 1, Index of self-evaluative reactions by self-awareness and private-audience conditions. (Lower numbers indicate a tendency toward
lower mood, a high degree of overgeneralization, self-attributions for
failure, and estimates that others would perform better. Means based on
n = 10, except contingent audience-mirror condition, for which n = 9.)
In Study 2, self-aware subjects' self-evaluative reactions to a failure experience reflected the evaluative style of a salient private
audience and were negative to the extent that the audience's
acceptance was contingent on success.
Some interesting issues remain for further study. First, the
experiments reported here obviously demonstrate only that salient audiences can affect the behavior of self-aware people but
not that they necessarily do have an impact in less contrived
situations. It is encouraging in this respect to note that subjects
were in no way forced to consider the reactions of the audiences
provided. Particularly in Study 1, self-aware individuals were
influenced by audiences that were only minimally primed in an
ostensibly unrelated context.
A second general issue involves the interface between private
and public audiences (Froming & Carver, 1981; Scheier &
Carver, 1983b; Schlenker, 1980). Behavior is often performed
for both types of audiences, as when an actress tries both to earn
her public audience's applause and to meet the demands of her
internalized drama coach. We have thus far been focusing on
the possibility of the evaluative structures of social relationships
becoming represented internally; another approach would be
to identify cases when private audiences might color the way
actual other people are experienced (cf. transference). A contingent private audience, for example, might be projected onto the
world, with the result that others are seen asevaluative and judgmental.
As these considerations allude, there may be very real benefits
in exploring the similarities and relevance of these notions to
analytic and neoanalytic models of the superego (e.g., Freud,
1923/1947), ego development (e.g., Mahler et al., 1975), and
object relations (e.g., Greenberg & Mitchell, 1983). Are audiences primed outside of awareness perhaps more influential, for
example, and is it possible to bring unconscious audiences into
focal awareness? If so, would this then allow the individual to
defend more actively against undesired influence and therefore
function more autonomously?
In the studies reported here, we tried to test some fairly basic
hypotheses about the experience of self. By abstracting and articulating principles from a number of theories, and integrating
these insights with the dynamics of self-awareness, we attempted to establish the basis for a useful perspective on selfprocesses. We see these studies as offering encouraging support
for a movement toward studying cognitive representations of
significant relationships, which may afford some insight into
the issue of how individuals' personal social histories define
their sense of self.
References
General Discussion
The results of the two studies were consistent with the assumption common to a number of theories that individuals
process self-relevant information according to patterns established in the context of significant relationships. In Study 1,
people acted in line with values held by audiences that were
rendered salient 10 minutes earlier by a short visualization, and
there was some evidence to suggest that this effect held particularlv for individuals whose self-awareness had been increased.
1097
Froming, W. J., Walker, G. R., & Lopyan, K. J. (1982) Public and private self-awareness: When personal attitudes conflict with societal expectations. Journalof Experimental Social Psychology, AS', 476-487.
Gibbons, F. X. (1978). Sexual standards and reactions to pornography:
Enhancing behavioral consistency through self-focused attention.
Journal of Personality and Social Psychology, 36. 976-487.
Goffman, E. (1959). The presentation of .self in everyday life. New York:
Buss, D. M., & Scheier, M. P. (1976). Self-consciousness and self-attribution. Journal of Research in Personality, W, 463-468.
Doubleday.
alytic theory. Cambridge: Harvard University Press.
on ego-involvement and self-awareness. In A. Hastorf & A. M, Isen
(Eds.), Cognitive social psychology (pp. 110-147). New York; Flscvier/North Holland.
deter-
Diener, E., & Srull, T. K. (1979). Self-awareness, psychological perspective, and self-reinforcement in relation to personal and social standards. Journal of Personality and Social Psychology, 37,413-423.
Duval, S.. & Wicklund, R. A. (1972). A theory of objective self-awareness. New York: Academic Press,
Duval, S., & Wicklund, R. A. (1973). Effects of objective self-awareness
on attribution of causality. Journal of Experimental Social Psychology. 9, 17-31.
Eagle, M. N. (1984). Recent developments in psychoanalysis: A critical
evaluation. New York: McGraw-Hill.
Ellis, A. (1977). Psychotherapy and the value of a human being. In A.
Ellis & R. Grieger (Eds.), Handbook of rational-emotive therapy (pp.
99-112). New York: Springer.
Federoff, N. A., & Harvey, J. H. (1976). Focus of attention, self-esteem,
and the attribution of causality. Journal of Research in Personality,
10, 336-345.
Fenigstein, A. (1979). Self-consciousness, self-attention, and social interaction. Journal of Personality and Social Psychology. 3 7, 75-86.
Fisher, W. A., Miller, C. T., Byrne, D., & White, L. A. (1980). Talking
dirty: Responses to communicating a sexual message as a function of
situational and personality factors. Basic and Applied Social Psychology. 1. 115-126.
Freud, S. (1947). The ego and the id. London: Hogarth Press. (Original
work published 1923)
Froming. W. J., & Carver, C. S. (1981), Divergent influences of private
and public self-consciousness in a compliance paradigm. Journal of
Research in Personality, 15, 159-171.
cago Press.
Milter, A. (\9%]). Prisoners of childhood. New York: Basic Books.
Newcomb, T. M. (1943). Personality and social change New York: Dryden Press.
Plant, R. W., & Ryan, R. M. (1985). Intrinsic motivation and the effects
of self-consciousness, self-awareness, and ego involvement: An inves-
1098