ALTERNATIVE MODES OF SELF-ENEANCEMENT
NANCY LYNN ASHTON
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVESITY OF FLORIDA
Len Bickman first fostered my excitement about social psychology
and gave me invaluable research experience. He also gave me the im-
portant pointer, "Nancy, it doesn't take brains to get a Ph.D., just
perseverance." Mary Shaw ably served as chairman of my committee. In
this role he helped me think through the conceptualization and method-
ology, organize the manuscript and more times than he knew he lifted
my morale and made me see thirgs in the proper perspective.
Barry Schlenker, Jack Feldman, and Barbara Baldridge gave assis-
tance on methodological questions and writing style and James McClave
helped on the analyses. Franz Epting deserves special thanks for being
there at the end.
Thanks also go to Mary Brennan, Carol Cohen and Jennifer Thom for
serving as experimenters.
I'd like to thank the subjects, but given the results, it's hard
to do so!
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . .. . . . . . ii
LIST OF TABLES . . . . . . . . ... . . . . v
ABSTRACT . . . . . . . . ... . . . . . vi
ONE: INTRODUCTION. . . . . . . . ... .... 1
Responses to Performance Feedback . . . . . . 6
Self-esteem. . . . . . . .... .. .. 6
Attributions of Causality . . . . . . . 8
Evaluation of the Characteristic . . . . . 9
Evaluation of the Source of Feedback . . . .. 10
Choice of Social Comparison. . . . . . ... 10
Summary. . . . . . . . . . . ... 11
Influencing Variables and Empirical Predictions ... 11
Importance of the Characteristic . . . . .. 11
Nature of the Characteristic . . . . ... 16
Source of Feedback . . . . . . . ... 19
Feedback Outcome . . . . . . . ... 23
Summary and Hypotheses. . . . . . . . ... 29
TWO: METHODOLOGY . . . . . . . . ... . .33
Subjects. . . . . . . . . ... ...... .33
Design and Overview . . . . . . . . ... 33
Independent Variable Manipulations . . . ... 34
Procedure . . . . . . . .... .. . . .. 38
Dependent Variable Measures . . . .... . . 41
THREE: RESULTS . . . . . . . . . .... . 44
FOUR: DISCUSSION. . . . . . . . . ... .. .. 61
TABLE OF CONTENTS
APPENDICES ........................... 70
APPENDIX A: PRE-EXPERIMENTAL RATINGS OFTHE CHARACTERISTICS. 71
APPENDIX B: CONTENTS OF BOOKLET AT SESSION I. . . . .. 74
APPENDIX C: PAPER-AND-PENCIL DEPENDENT MEASURES . . ... .96
REFERENCES. . . . . ..... . . . . .... .104
BIOGRAPHICAL SKETCH . . . . . . . . .. . . .111
LIST OF TABLES
1. Mean Ratings for Independent Variable Characteristics . 35
2. Manipulation Check for Credibility Variable. . . . ... 37
3. Manipulation Check for Feedback Variable. . . . ... 38
4. Reliability and Validity Data for Self-esteem Scales. .. 42
5. 'Mean Ratings of Attribution Items for Hypothesis II . . 47
6. Mean Ratings for Hypothesis III . . . . . . . 48
7. Mean Ratings for Second Social Comparison Measure
(Hypothesis IV) . . . . . . . .... ..... 49
8. Mean Ratings for Second Social Comparison Measure
(Hypothesis V). . . . . . . . . . . . 50
9. Mean Ratings for Attribution Items (Hypothesis VII) . . 52
10. Mean Ratings for Hypothesis VIII. . . . . . 52
11. Mean Ratings for Evaluation of Source (Hypothesis X). . 54
12. Correlations > .36 by Condition and Related Contingency
Tables. . . .. . . . . . . . .. 58
13. Contingency Tables for Median Ranks . . .... . . 60
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
ALTERNATIVE MODES OF SELF-ENHANCEMENT
NANCY LYNN ASHTON
Chairman: Dr. Marvin E. Shaw
Major Department: Psychology
Numerous conceptual writings and experimental works dealing with
self-concept and self-esteem suggest that under certain conditions
individuals have a need to aggrandize and enhance their self-concept,
even to the point of denial or distortion of "reality." This is viewed
as a normal and valid part of self-evaluation and behavior, though it
can reach extreme and unproductive limits in certain personality dis-
The present investigation was designed to study this notion of
self-enhancement as an alternative form of behavior to accurate self-
assessment of abilities and opinions as presented in Festinger's theory
of social comparison processes. Seventeen hypotheses were tested
dealing with alternative responses to feedback about one's performance.
Data were collected on 96 female and 96 male college students from
introductory psychology classes at the University of Florida by two
teans of experimenters. Independent variable manipulations were: the
nature of the particular characteristic on which subjects were evaluated
(opinion or ability), importance of the particular characteristic (high
or low importance), credibility of the source of feedback (moderate or
low),and nature of the feedback (positive or negative). Response mea-
sures included self-esteem, attributions of causality for one's per-
formance, evaluation of the characteristic, evaluation of the source of
feedback, and a behavioral social comparison measure.
Results indicated little support for the experimental hypotheses,
largely due to the unexpected effect of the feedback variable. One
hypothesis on attributions of causality received some support in that
the two ability/feedback conditions differed in attribution to luck
while the two opinion/feedback conditions did not. The other hypotheses
were not supported. However, there were some indications of self-
enhancing responses, though in different ways than were predicted.
Results suggested that subjects did devalue an important characteristic
and credible source after receiving negative feedback. There was some
limited evidence for use of alternative modes of enhancement after
negative feedback and for opinions. Alternative explanations for the
results and implications for future research are discussed.
Much of human behavior and perceptions centers around making
judgments about the world; that is, evaluating persons and relevant
stimuli in order to effectively deal with the environment. But in-
dividuals not only form judgments of objects and other people, they also
go through a process of self-evaluation. As they do for objects and
other people, so individuals attach meaning and a sense of worth to
themselves. This brings forth an affective reaction and a cognitive
core of beliefs about one's worth. Such a process can be very important
in human development and interaction, and is related to such topics as
achievement motivation, affiliation, interpersonal relations and group
processes. In fact, every aspect of our experience--our behavior, per-
cepcions, thoughts and feelings--can be seen as influenced by and as
influencing our self-conceptions and evaluations of ourselves.
Festinger's theory of social comparison processes (1954a) includes
a series of hypotheses relating to self-evaluation and behaviors de-
signed to complete the self-assessment process. Festinger states that
there is a human drive to evaluate one's abilities and opinions, and
that we all seek information to assist us in an accurate assessment.
Humans have a positive motivation to know their environment and in fact
direct much of their behavior to finding out about the world around
them. More specifically, we try to find out what we can and cannot do,
and whether our opinions are valid. This has important implications
for our development and self-preservation. It becomes urgent to have
precise and valid knowledge about our abilities and perceptions. A
process of "reality testing" would enable us to assure ourselves that
our opinions are correct and our abilities properly appraised, but for
many judgments there exists no independent reality test. This is
especially true for opinions, and points up an important difference
between abilities and opinions that is addressed later. We must compare
our abilities and opinions to others' abilities and opinions in order to
complete our self-assessment. Festinger (1954b) says that in this way
a non-social motivation to know what we can and cannot do leads to
social behavior in the form of comparison with others. Nissen (1954)
cites comparative support for a universal drive to know, in that animals
seek precise information but depend more on their sense organs than do
humans for reality testing and accurate assessment. Due to less de-
veloped sense organs in humans, the social process becomes very impor-
Since the original paper on social comparison processes (1954a)
others have expanded Festinger's notions about opinion and ability
assessment to include evaluations of emotions (Schachter, 1959) and
personality characteristics (Hakmiller, 1966; Wheeler, 1966). The basic
assumption underlying these works is that when objective, non-social
means are not available as a standard, people compare themselves with
others on the salient opinion, ability, emotion or personality charac-
teristic, thus using social rather than physical reality as the basis
of their assessment. Building on the further assumption that accuracy
is the most salient goal for individuals, Festinger hypothesizes that
people compare themselves with a similar individual since the most
information can be obtained from that type of comparison.
There are two objections to Festinger's thesis. First, several
researchers have challenged the notion that a similar other will be
chosen for comparison. For one thing, similar others do not always
provide the most information (Gruder, 1971; Strauss, 1967; Wheeler,
Shaver, Jones, Goethals, Cooper, Robinson, Cruder and Butzine, 1969).
There are situations when more information can be obtained from com-
paring oneself to a dissimilar rather than similar comparison other.
One example of this is when the individual does not know the range of
a particular ability or opinion. The second criticism is more basic--
that individuals may not always desire accuracy. A parallel process to
accurate self-appraisal operates in the form of self-enhancement. Ob-
viously, there are times when accurate assessment of oneself is important.
For example, when about to undertake some task requiring certain skills,
it is advantageous to know if one has the necessary ability. Certain
short-term and long-term decisions in an individual's life necessitate
accurate evaluation of abilities and opinions. But humans also have a
desire for self-enhancement; they want to look good in front of others,
and furthermore often prefer to look good to themselves. This desire
may override the need for accuracy in presentation and evaluation.
Extrapolating from Festinger's framework, there are situations
under which individuals might be expected to choose a person who is
more negative than themselves, thus making themselves look good in com-
parison. This would occur in a threatening situation or at a time when
self-esteem is more important than accuracy. Hakmiller (1966) found
that subjects in a high-threat condition focused more upon dissimilar
others than did subjects in the low-threat condition, choosing comparison
others who were more likely to do worse than themselves. He calls this
phenomenon defensive social comparison, in that subjects reduced the
threat of negative information about themselves by choosing comparison
with someone who was worse. Darley and Aronson's (1966) affiliation
study showed a similar effect; subjects in a high-fear-uncertainty con-
dition chose to wait with someone who's fear reaction was lower than
their own rather than with someone who's fear reaction was similar to
their own. Rose (1966) states that a person controls the selection of
others and can therefore exercise the power to enhance the self.
Many conceptual writings expound on the process of enhancing the
self for the benefit of both others' opinions of an individual and in
terms of that individual's self-perceptions and evaluations. In a dis-
cussion of reference group phenomena, Hyman and Singer (1968) note that
in the process of self-appraisal the individual selects the social com-
parison standard. The choice of a comparison can maintain, enhance or
injure self-regard. Gergen (1971) refers to the emotional component of
self-evaluation and states that there is a strong tendency to select
cues providing positive confirmation of the self. We undertake per-
ceptual scanning that is biased toward seeing ourselves as we wish to
be. Allport (1968) listed ego-enhancement as one aspect of the proprium
or self. Wylie (1961) included references to self-enhancement and
Sullivan (1953) proposed that the self-system, which is concerned with
protecting the individual's self-esteem, is capable of selective in-
attention. In their theory, Combs and Snygg (1959) speak of the
important needs of maintenance and enhancement of the self and how
these affect the perceptions of individuals. For Heider (1958), when
the personal lots of p and o are compared, self-concept is directly
affected and in fact may be enhanced or undermined as a result of com-
parison with someone else. We try to preserve a good image and produce
a good opinion of ourselves in others by showing our positive side and
hiding what is negative. Heider (1944) says that in addition to the
tendency to see ourselves and others in certain ways, we have a tendency
to keep our ego-level high. Pepitone (1964) hypothesizes a unidirec-
tional pressure toward maintaining or enhancing self-esteem. Crowne
and Marlowe (1964) state that people attempt to validate their own
self-worth in their defensive efforts to cope with anticipated failure.
Rogers (1951) put forth a series of propositions about the self
and its growth. He stated that the organism has one basic tendency and
striving to actualize, to maintain and enhance the self. Most of the
ways of behaving which the organism adopts are those which are consis-
tent with the concept of the self. Any experience which is inconsistent
may be perceived as a threat. A larger number or greater intensity of
threat-perceptions leads to more rigid organization of the self-
structure to maintain itself. One defense under threat is selective
perception or the distortion and omission of data. Rogers also empha-
sizes the need for positive self-regard. In general, these and other
formulations about the self make the point that there is a desire to
protect one's ego by perceiving the world and presenting oneself so as
to perpetuate a favorable image in the eyes of others and in the mirror
Empirical work likewise relates to this notion of self-enhancement.
Schlenker (1975b) cites evidence that undergraduates do desire self-
enhancement and in fact ignore information known only to themselves
when selecting a public image. Jones (1973), in his review of the
self-evaluation literature, finds more support for the incentive posi-
tion that individuals prefer favorable evaluations than he does for the
consistency viewpoint that people most desire consistent information.
In Schneider's (1969) study, subjects were given bogus feedback on their
performance on a social sensitivity test. Subjects in the failure
condition were more self-enhancing in their presentations in that they
emphasized positive points about themselves rather than negative points.
There are numerous behaviors that relate to attempts at self-
enhancement and the process of ego-maintenance. We consider five such
behaviors as they relate to the independent variables under considera-
tion. Modes of self-enhancement include such responses as attributions
of causality for one's performance, evaluation of the characteristic on
which the individual is evaluated, evaluation of the source of feedback
and social comparison choice. Change in self-esteem, the fifth response,
is not a self-enhancement tactic, but rather happens when self-enhance-
ment does not occur. Some general discussion of these topics is in
order before we turn to the factors influencing self-enhancement.
Responses to Performance Feedback
Self-esteem can be viewed in terms of evaluative attitudes toward
the self; as an attitude of approval or disapproval it indicates the
extent to which individuals believe themselves to be capable, worthy
and successful (Coopersmith, ]967). Attitudes toward the self are
defined in the same way as attitudes toward other objects--as orienta-
tions toward or asay from some object or event, as predispositions to
respond favorably or unfavorably toward these and related objects and
events. And like other attitudes, self-evaluations carry affective
loadings and have motivational consequences. Argyle (1969) defines
self-esteem as the extent to which a person has favorable attitudes
toward the self. In other words, there is a favorable or unfavorable
attitude of "I" toward "me." Shaw and Wright's (1967) presentation of
attitude scales includes defining characteristics of attitudes that also
apply to self-evaluations. Attitudes are based upon evaluative concepts
regarding characteristics of the referent object and give rise to
motivated behavior. They vary in quality and strength on a positive-
neutral-negative continuum. Attitudes are learned, have specific
referents and possess varying degrees of interrelatedness to one another.
Though they are relatively stable and enduring, they are subject to
change. Shaw and Wright present attitude scales for self-acceptance
that measure attitude toward the self.
Coopersmith (1967) also points out some dissimilarities between
attitudes toward other objects and self-attitudes. Self-esteem differs
from other attitudes in that the self is the only person who formulates
the abstraction, appreciates its limits and content, and is in a posi-
tion to define and evaluate its characteristics. In addition, the
object under consideration (oneself) is different for each individual.
Because this work includes reference to opinions at a later point
in the discussion, it is necessary to give brief mention to the dis-
tinction between attitude and opinion in social psychological litera-
ture. Shaw and Wright (1967) note that though an opinion is similar
to an attitude in that both involv- cognitions about the wcrld, it
differs from an attitude in that chare is no commitment associated with
it. They cite other researchers who state that opinions are verbal-
izable while attitudes are sometimes "unconscious." Furthermore,
opinions are responses while attitudes are response predispositions.
Berkowitz (1972) states that opinions have no dynamic or driving pro-
perties and are assumptions or judgments that the individual believes
to be true. In cor.trast, attitudes have motivational impact. In gener-
al, attitudes have more influence on human behavior and hold more
emotional meaning for the individual.
The origins of one's self-image and self-esteem are probably the
reactions of others (Argyle, 1969; Symonds, 1951). Charles Cooley
(1968) discussed the notion of "social self," and the looking-glass
concept, which refers to the fact that the self reflects imagined
reactions and appraisals by others. The individual thus interprets
judgments of the self held by others. Continuing in this tradition,
Mead (1968) offered the proposal that the self is determined by actual
reactions of significant others to the self and by the individual's
imitation of these reactions. Much empirical work and observation of
behavior shows that our evaluations of ourselves are very much affected
by the implicit and explicit feedback we get from others about our
standing relative to others.
Attributions of Causality
Change in self-esteem is one possible consequence of feedback about
one's standing on a particular attribute, but there exist mechanisms to
deal with feedback such that self-esteem is not affected. By changing
attributions one can alter the degree to which a given performance is
seen as a reflection of the self. Heider (1958) went into detail about
the fact that observers are interested not only in what can do, but
whether p's performance is due to either personal characteristics or
the environment. In an earlier article (Heider, 1944) he quoted Hoppe's
point that not only is success or failure important, but the beloning-
ness of them to the person is crucial. In Heider's system, can ex-
presses ability, which is seen as a stable underlying disposition. But
ability may be attributed to person or the environment, in that it
refers to power and effective environmental force. By attributing a
performance to either person or environment, one is able to control the
presumed locus of causality, which may have strong effects on the judg-
ment of the actor as an individual. And we are prone to alter percep-
tions of causality so as to protect or enhance self-esteem, by attributing
good outcomes to our own characteristics and bad outcomes to external
sources. Thus egocentric perceptions reflect the tendency to structure
perceptions of causality such that we obtain credit for good actions
and avoid blame for bad outcomes. Feather (1968) states that it is
likely that there is an asymmetry in attribution of responsibility for
outcomes depending upon whether one succeeds or fails. This expectation
is based on Heider's balance theory (Feather, 1969).
Evaluation of the Characteristic
Another type of cognitive mechanism is available to a person
attempting self-enhancement under the guise of self-evaluation of per-
sonal characteristics--that is, one can re-evaluate the characteristic
to maximize self-aggrandizement. If one fails at a task or is inferior
to others on an attribute, one can devalue the importance of the attri-
bute and treat is as trivial. On the other hand, one can magnify and
emphasize the significance of attributes on which one excels. Heider
(1958) states that the value of something may be reassessed by the
person who lacks it, and Lewin, Dembo, Festinger and Sears (1944) speak
of switching standards to avoid feeling failure. Lecky (1969) cites as
one method of dealing with inconsistencies the notion that the individual
will interpret an incident in such a way that it can be assimilated.
A graphic picture of re-evaluation comes to mind in the Aesop fable
about the fox who devalues the grapes he cannot have.
Evaluation of the Source of Feedback
In addition to an evaluation of the attribute itself, an individual
makes a judgment of the source of feedback. Evaluations can mean dif-
ferent things depending upon the source from which they originate. An
individual often has available the mechanism of devaluing the credibility
of a source who says something negative, or of enhancing the credibility
of someone who says something positive. Argyle (1969) notes that this
option of devaluing the source of information or the carrier of infor-
mation is one way of preserving self-esteem.
Choice of Social Comparison
The final response measure to be considered in this paper is social
comparison choice. Previous discussion has noted that this is a tactic
one can use to make oneself feel better by comparing with someone who is
worse off than oneself (e.g. Hakmiller, 1966).
In discussing self-esteem, social comparison, attributions of
causality and evaluation of both the attribute and the source of feed-
back, we have considered five reactions to others' evaluations of us.
Four of these factors (attributions of causality, social comparison
choice, evaluation of the attribute, evaluation of the source of feed-
back) can be used to achieve self-enhancement, while self-esteem change
is another possible reaction. There are of course many variables that
pertain to and influence attempts at self-enhancement and changes in
self-esteem, but only a few of the important ones are considered here.
We deal with the importance of the characteristic on which the indi-
vidual is evaluated, whether it is an opinion or ability, the source
of feedback in terms of credibility, and whether the individual re-
ceives positive or negative feedback. In accord with the conceptual
and empirical evidence cited, specific hypotheses are then advanced.
Influencing Variables and Empirical Predictions
Importance of the Characteristic
This section includes material relevant to importance of the
particular characteristic from the areas of self psychology, attitude
change and social comparison processes. Allport (1968) distinguished
between matters that are vital and central to the self and those that
are peripheral. The assumption is that important characteristics are
vital to the self and it is more likely that self-esteem pressure will
arise with more important characteristics (Singer, 1966). Rose (1956)
stated that self-enhancement occurs more strongly when the contents are
very salient to the individual. He defines content salience as impor-
tance to 2 of a particular kind of evaluation. Rosenberg (1968)
emphasized the importance of knowing how much a person values a quality
and how good the person wants to be. In reading through the literature
one notes the many definitions of importance as well as the fuzzy dis-
tinctions between importance and related terms such as salience and
valence. In the conceptualization here a characteristic that is impor-
tant to someone is defined as being of greater value to that person.
Presumably the person cares more about being good on an important char-
acteristic than on an unimportant one. The drive to evaluate is thus
more pertinent for central, vital, important characteristics. Festinger
(1954a) says that in a group situation there is an increase in the
pressure to reduce any discrepancy of opinion as the importance of the
issue increases. Argyle (1969) points out that the self influences
behavior in situations which are ego-involving and which bring self-
esteem into operation. He also notes that importance to the performer
of the variable being assessed influences the extent to which self-
awareness is aroused.
Researchers in the area of attitude change have also considered
the impact of importance of the attitude object and related concepts to
behavioral response. In the social judgment theory of attitude change,
Sherif and Hovland (1961) consider the topic of involvement in an issue
and its relation to latitudes of acceptance and rejection in an attitude
change situation. They state that ego-involvement limits the range of
assimilation for attitude positions different from one's own. Rokeach
(1968) also notes that not all beliefs are equally important to the
individual. Beliefs vary along a central-peripheral continuum. The
more a belief is functionally connected with other beliefs, the more
central it is and the more implications it has for other beliefs. He
has experimental support for the idea that individuals show greater
resistance to change on central rather than peripheral beliefs.
Other theories dealing with attitude change recognize the impli-
cations of the importance of a belief or attitude. Festinger (1957)
states that the more important an attitude is to an individual, the
greater will be the resulting dissonance after exposure to contradictory
bits of information. And greater dissonance leads to a stronger moti-
vation to do something to reduce the dissonance. Abelson (1967) proposes
that when two cognitive elements stand in imbalanced relation to each
other, the tendency will be to apply bolstering toward the more in-
tensely affected element. Another possible response is denial of the
implications of the less intensely affected element and/or denial of
the relation between the two elements. Krech and Crutchfield (1948)
state that in understanding and predicting behavior it is as necessary
to know how important or central an attitude is as it is to know the
sign or extremity of the attitude. Eagly and Manis (1966) found that
issue involvement had a significant effect on junior high school stu-
dents' responses to messages challenging their views. For high involve-
ment the communicator and message were evaluated much less favorably
than when subjects were characterized by low involvement.
The above citations refer to importance of the characteristic
itself. While social comparison literature has not directly considered
this factor, the results of the Hakmiller (1966) and Wheeler (1966)
studies show that evaluation of the characteristic prior to receiving
feedback is influential. Wheeler found that when the characteristic was
positively valued subjects chose similar comparisons, but Hakmiller
dealt with negatively valued traits and found no support for the simi-
larity notion. Thornton and Arrowood (1966) conducted research on the
goal valence of an ability (strongly positive or strongly negative) and
found that a significantly greater number of subjects chose to compare
with someone better off than themselves in the positive than in the
negative conditions. Wheeler et al. (1969) found that the direction of
comparison (downward or upward) was related to positivity-negativity of
the characteristic. These experiments were not considering the impor-
tance of the characteristic per se but they do illustrate that prior
evaluations of the attribute can differentially affect behavior.
In accord with the conceptual and empirical work cited, several
hypotheses are proposed. It is hypothesized that negative feedback on
an important attribute is more likely to lower self-esteem than is
negative feedback on an unimportant characteristic. Likewise, reactions
to positive feedback differ based on the importance of the character-
istic, with positive feedback on an important attribute more likely to
raise self-esteem than positive feedback on an unimportant one (Hypoth-
In terms of attributions, the motivation is greater to attribute
negative feedback on an important characteristic to the environment and
positive feedback to the self than when the characteristic is less
important. This constitutes Hypothesis II. Schlenker (1975a) obtained
significant positive correlations between importance of doing well and
subjective ratings of personal responsibility in extreme success con-
ditions. There was not a corresponding negative correlation between
perceived responsibility and importance of doing well in failure
conditions. He suggests that perhaps subjects in failure conditions
were hesitant to attribute failure to the other members of their group
in the study. This could have been due to social constraints against
such behavior or because their identification with their group led
them to include the group in their egocentric attributions. The final
possible explanation is that subjects rationalized failure by denying
the importance of doing well.
Important and central beliefs are more functionally connected to
other beliefs (Rokeach, 1968). For this reason it is more difficult
to re-evaluate the importance of an important characteristic than it is
to change evaluation of an unimportant characteristic (Hypothesis III).
Prior research supports this hypothesis. Miller (1965) found that high
issue involvement consistently reduced the persuasive effect of dis-
crepant information on a variety of response measures. In an opinion
change experiment, subjects high in initial concern for the opinion
issue changed their opinions less than did subjects low in concern for
the issue (Fine, 1957). Similarly Snyder, Mischel and Lott (1960)
studied people who were either high or low on aesthetic value (Vernon-
Allport Study of Values) in a social influence situation. The high-
aesthetic-value subjects conformed much less than did the low-aesthetic-
value subjects. Thus resistance to re-evaluation should be especially
manifest for important characteristics. This parallels the asymmetry
observed in the level of aspiration studies with individuals (Lewin et
al., 1944) and with groups (Zander and Medow, 1963). In the latter
study members' evaluations of their own performance is more strongly
determined by their group's score when the group succeeds than when it
In line with the notion that one cares more about looking good on
an important characteristic, it follows that a self-enhancing comparison
other will more often be chosen on an important characteristic than on
an unimportant one (Hypothesis IV). This contradicts Festinger's
(1954a) hypothesized drive for accuracy, which no doubt should also be
greater for more important issues than for unimportant ones. In making
the enhancement prediction, we are not so much clearly discarding the
Festinger hypothesis as noting an important exception in that we all
seek self-enhancement at times, especially when accuracy is not important
for some decision or future activity.
Nature of the Characteristic
Another aspect of the characteristic is applicable and is in fact
a part of Festinger's (1954a) theory. This has to do with the focus of
evaluation---in Festinger's terms, whether one is being assessed on an
opinion or ability. Although both are considered together in most of
his discussion, Festinger does point out that there are two major dif-
ferences: there are non-social constraints that either make it impossible
to change an ability or make it much harder to change an ability than to
change an opinion, and there exists a unidirectional drive upward for
abilities but not for opinions. This latter point refers to the fact
that different performances on an ability usually have intrinsically
different values, and there is merit set on doing better and better
(at least in Western culture). The higher the score on a performance,
the more desirable it is. But with opinions there is no inherent basis
for preferring one over another. Value comes not from having greater
or less worth, but from the subjective feeling that the opinion is
valid and correct. This relates to Sullivan's (1953) process of the
consensual validation of beliefs and Blau's (1967) point that social
consensus defines beliefs as right or wrong. The distinction between
physical and social reality is relevant here. In determining where a
person stands on an ability, physical reality is a more likely recourse
for judgment, but it is not available in the case of opinions. Further-
more, in line with Gordon's (1966) conceptual and empirical work on
influence theory versus social comparison, an individual can attempt
to persuade someone with a different opinion, but this is not likely
to be an option for someone differing on an ability.
In light of the present conceptualization of self-esteem as an
evaluative attitude toward oneself, and pertinent to abilities and
opinions, it is assumed that there is a unidirectional drive upward.
In other words, it is "better" and more statistically common to have
a positive evaluation of oneself. Self-enhancement behaviors are ex-
amples of attempts to achieve and maintain a positive attitude about
oneself. In the present context one important difference between
opinions and attitude toward oneself is that the unidirectional drive
upward exists for attitudes about oneself. Festinger's (1954a) point
that opinions are valid based only on consensus with opinions of others
can be seen in terms of the self being enhanced if others have similar
opinions. Of course the self is also enhanced if others have favorable
opinions of that person.
Gerard (1968)'states that Festinger misled the field when he dis-
cussed ability and opinion as two distinct processes, since they are
"inextricably connected," with ability comparison being the more
fundamental. Gerard states that a person compares his or her own
opinion to that of someone else in light of what is assumed to be the
ability of the other person to form and hold an opinion on the issue
in question. There is some validity to this point though there seem
to be numerous counter-examples. Are movie stars in possession of
special abilities to hold opinions on breakfast cereals, detergents and
the like? Yet many Americans seem to be influenced by the prestige sug-
gestion that Madison Avenue uses in advertising campaigns. There are
also fundamental differences between opinions and abilities that have
ramifications in personal behavior. For instance, ability but not
opinion is demonstrated through performance. It can therefore be seen
as more of a reflection of the person, and thus is more related to self-
esteem and feelings of worth about oneself. And as Heider (1958) out-
lines so well, attributions of ability are a common part of our lives
in observing others' behavior as well as our own. There exists no such
notion of a stable, underlying disposition when evaluating opinions.
Therefore it is hypothesized that subjects are more likely to
choose a similar supporting social comparison person in the case of
evaluation of opinions than in the case of evaluation of abilities.
Hypothesis 'I follows from the absence of a physical reality to assess
the validity of opinions. Since the subjective feeling of accuracy of
opinions comes from social consensus, individuals are more likely to
look to others of the same opinion. In this manner, individuals can
attach validity to their own opinions.
Opinions have a social basis and therefore can be seen as being
less a reflection of the individual's inherent being than are abilities.
We have stated that there is no notion of an internal attribute of an
opinion that is comparable to the attribution of a personal ability
disposition. Though attributions are made about us on the basis of our
opinions, we have the option of opinion change, which is not as likely
with abilities. Therefore, ability issues are more closely associated
with the individual and have more influence on self-esteem. Thus we
would expect more change in self-esteem following assessment of an
ability than after opinion evaluation (Hypothesis VI). The non-social
constraints on abilities make change less likely as a response than is
Similar reasoning leads to the hypothesis that the attribution of
positive performance to the self and negative performance to the environ-
ment will be more probable for abilities than for opinions (Hypothesis
VII). This is because it is hypothesized that ability forms a more
permanent unit with the self than does opinion, and because self-
enhancement is more important with abilities. In addition, opinion
change is possible when one does not get confirmation of an opinion,
but ability change is not usually available to an individual. Because
of this asymmetry, re-evaluation of the characteristic's importance will
occur more often for abilities than for opinions (Hypothesis VIII).
Source of Feedback
Typically people have some notion of their own performance or
possession of an attribute and desire to know what that means in terms
of the standing of others on the same characteristic. The source of
any feedback people might get about their opinions, abilities, or other
characteristics will have some effect on their reactions and how they
deal with that information. Aristotle noted that the speaker is a force
just as important as the receiver or the message itself (Bettinghaus,
1968). Wylie (1961) states that the source of alleged failure or de-
valuation is important. There is a difference between the experimenter
judging subjects on the basis of a standard and subjects discovering for
themselves that they cannot do a task. Argyle (1969) notes that changes
of self-image are more likely to occur if the other person is regarded
as a valid source of information. Among Pepitone's (1958) hypotheses
about status relations serving to mediate attributions of causality,
one notion is that higher status people are seen as having more valid
and competent observations and criticisms. Heider (1944) cites the
phenomenon of prestige suggestion, where the acceptance of statements
and evaluations depends on the prestige of the person to whom they are
Several studies on attitude change have shown that communicator
variables have important influences on behavior. Sherif and Sherif
(1967) note the importance of source variables in that high status in-
creases the range of assimilation for a communication. Kelman (1961)
discusses the importance of such communicator variables as credibility,
power and attractiveness. Hovland, Janis and Kelley (1963) point out
important factors influencing communication effectiveness are the com-
municator cues as to the trustworthiness, intentions and affiliation of
the source. These presumably influence the evaluation of the communi-
cator's presentation and the receiver's acceptance. Hovland et al.
(1963) hypothesize no difference between different communicators in
receiver's attention and comprehension based on credibility. Rather,
the difference seems to come from the motivation to accept the conclusion
advocated. Arnet, Davidson and Lewis (1931). Haiman (1949) and Hovland
and Weiss (1952) demonstrated a positive relationship between the
credibility of a communicator and the extent of opinion change. The
same communication was judged less fair and less justified when given
by a low-credible source, although the effect had disappeared four
weeks later (Hovland and Weiss, 1952). It seems that the effect of the
source is maximal at the time of the communication. Aronson, Turner and
Carlsmith (1963) showed that a high-credible communicator was more suc-
cessful in inducing opinion change than a mild-credible communicator.
This occurred at every point of discrepancy between subject opinion and
communication. Johnson and Steiner (1968) examined the influence of
source of evaluation on fraternity members' responses to negative in-
formation about themselves. With a low-credible source there occurred
little or no conformity, assimilation or rejection of the source as
invalid. The more-credible source led to rejection and conformity.
The authors draw attention to the need to test readjustment of the sub-
jective scale of judgment such that "goodness" is represented by a lower
rating than it was previously. They suggest that this phenomenon may
occur instead of loss of self-esteem.
Some of the attribution literature relates to the effects of
communicator status. Kelley (1967) discussed information dependence of
A on B if B can raise A's level of information to a level higher than
A can attain from alternative sources. But A must make attributions
regarding B's message and arrive at a perceived validity using the
external criteria of distinctiveness, consistency over time, consistency
over modality and consensus. Communicator factors that are particularly
relevant to attributions A might make of B's message are B's expertness,
and trustworthiness. The former reflects contact with external causal
variables relevant to the message, while the latter refers to an absence
of irrelevant causal factors in B's statements.
The above discussion makes the point that communicator devaluation
is open to the individual who does not want to accept an unfavorable
communication and wants to continue to see the self in a prior favorable
light. If individuals get some negative feedback or learn of failure
on a characteristic from another individual, they may derogate the
communicator and try to minimize the importance and validity of the
It is hypothesized that the more credible source of feedback causes
more effect on the subject's self-esteem than does a less credible
evaluator (Hypothesis IX). This follows because the more credible
evaluator's statement carries more weight and appears more likely to be
accurate, and because it is harder to devalue a more credible person.
This latter point is an additional hypothesis stemming from a cost
framework in that it is harder (more cognitively costly) to lower evalu-
ation of a highly evaluated object than a low one (Hypothesis X).
The last hypothesis concerned with the credibility of the source
of feedback is that following feedback from the more credible person,
individuals are more likely to attribute the performance to the self
following positive feedback and to the environment following negative
feedback than they are for the low-credible source (Hypothesis XI).
This follows from the fact that a more credible source's communications
are viewed as more accurate than a low-credible source's communications.
Bettinghaus (1968) states that the perceived differences in status
between source and receiver may well lead to one individual being per-
ceived as more worth listening to, more believable or more influential.
A related influence is that a more credible person is probably further
removed from the receiver in status than the less credible person. The
comparison between self and communicator may be an important under-
lying mechanism in the process of acceptance of a communicator's
A crucial variable in considering reactions to feedback about
oneself is whether it is positive or negative. Wylie (1961) discusses
the effects of success and failure on task-relevant or global self-
esteem. Not only is a person's level of self-regard of great importance
in predicting behavior, but self-esteem can be temporarily raised or
lowered in response to experiences and feedback from others. Singer
(1966) states that in the social comparison process, one is implicitly
evaluating an opinion of oneself and one's self-esteem. Morse and
Gergen (1970) conducted research which lends support to this notion.
They based their study on the assumption that the social comparison
process is a source of instability in self-conception. Subjects were
exposed to either a very socially desirable or undesirable person ap-
plying for either the same or a different job. A dependent measure
showed that subjects exposed to the socially desirable person lowered
self-esteem while self-esteem was raised for subjects who met the un-
Argyle (1969) states that experiments on the effects of manipulated
success/failure on self-esteem generally show the expected effect of
lowered self-esteem following failure and raised self-esteem after
success. But there is more change of self-report upwards following
success than downwards after failure. Here is a manifestation of the
desire to protect and enhance the ego. Examples of studies showing the
influence of success and failure on self-ratings are Koocher (1971) and
Videbeck (1960). Koocher investigated swimming competence and per-
sonality change in adolescent boys. Development of swimming skill
resulted in the enhancement of self-concept, giving some support to
White's (1959) contention that a sense of competence is basic to the
development of self-esteem. Boys who did not increase in swimming com-
petence did not show a negative influence on their self-concept. Videbeck
found that feedback from a "speech expert" who evaluated the subject's
poetry reading resulted in the most change in self-rating on the same
scales the expert used. Less change was evident on related scales and
no change occurred for unrelated self-rating scales. Thus self-concept
change was more likely on relevant dimensions than on global self-
Several investigators failed to find evidence for such defensive
reactions to negative feedback. Regan, Gosselink, Hubsch and Ush (1975)
found no evidence for self-enhancement in their subjects. Students
were given either no feedback, praise or criticism about their own or
another student's performance and responded to the question, "How good
do you think your (her) answers were?". Actors rated themselves less
positively than observers rated them. The authors suggest the notion
of self-derogation in anticipation of evaluation to explain the results.
In addition, all subjects were females, which may be related to a similar
effect in a study (Nicholls, 1975) where females were more likely than
males to derogate themselves. In addition, Regan et al. only tapped
one possible mode of response to evaluation. Subjects may have employed
some other cognitive mechanism as a self-aggrandizing tactic. A further
possibility is that subjects did not take the situation as seriously as
is necessary to experience threat from negative feedback. Eagly and
Acksen (1971) found that subjects were more receptive to positive in-
formation about themselves than negative information when they expected
future evaluation. In the negative-information condition there was no
difference between expectancy levels. This offers some support for
Regan et al.'s notion of self-derogation in anticipation of evaluation.
Israel (1963) found a larger private acceptance of attitudes which
underrated the self than with those that over-evaluated the self. An-
other instance of failure to demonstrate self-enhancement is a study by
Howard and Berkowitz (1958) investigating reaction to one's evaluators.
The results suggest that the desire for accuracy sometimes outweighs
the desire for self-enhancement, as Festinger states. It is important
to point out that this was a role-playing study, which lessens one's
confidence in the findings.
There are two general areas of explanation for this series of
failures to demonstrate self-enhancement responses. The theoretically
more interesting possibility has to do with expectation of future evalu-
ation and the concomitant self-derogation in anticipation of evaluation.
This may constitute a limiting influence on the conditions under which
self-enhancement occurs. The other general explanation concerns method-
ological issues such as narrow sample, using only one response mode or
situation that subjects may not take seriously.
The level of aspiration studies are also relevant to reactions to
positive and negative evaluation. Lewin et al. (1944) elaborate on the
psychological feeling of success or failure and the cessation of
activities when possibilities of achieving further success are not good.
One can avoid feeling failure by switching the standard (changing the
goal line, in their terms) after the fact or by severing the relation
between performance and the self. The latter point of course refers to
attributional processes. A way to lift morale after failure is to
restructure the field so that defeat is not attributed to one's own
inferiority (Heider, 1944). Merton (1946) proposes that attribution
of causality may be a defense one is prone to use after failure. The
belief in chance enables one to preserve self-esteem following failure.
In short, one can make attributions so that self-esteem is not affected.
Weiner et al. (1972) suggest that in attempting to explain a suc-
cess or failure outcome, people assess own and other's ability level,
amount of effort, difficulty of the task and magnitude and direction of
experienced luck. The first two elements reflect personal locus of
causality while the latter two are external sources. The authors have
found that success is more likely to be attributed to personal factors,
and failure tends to be attributed to external sources. Such tactics
can be seen as ego-defensive and as manifestations of attempts at self-
enhancement. Streufert and Streufert (1969) examined the effects of
success and failure on attribution of actuality. Subjects in decision-
making teams attributed more causal effect to their own team under
success than under failure. Johnson, Feigenbaum and Weiby (1964) also
found support for attributing success (more than failure) to the self.
In Schlenker's (1975a) study there was a main effect for performance on
attribution of causality to luck, in that failure subjects stressed luck
more than did success subjects. Frieze and Weiner (1971) conducted a
simulation study in which successful subjects were more likely to
attribute performance to internal sources than were failure subjects.
Fitch (1970) found experimental support for the proposition that a
drive for self-enhancement disposes subjects to attribute success out-
comes to internal sources to a greater extent than is true of failure
outcomes. There was also partial support for the notion that self-esteem
consistency affects causal attributions, in that low self-esteem persons
who perceived failure feedback attributed more causality to internal
sources than did their high self-esteem counterparts. Fontaine's (1975)
research led him to conclude that ego-oriented motives were more im-
portant in self-attributions in a real task situation than was true
for other-attribution in a simulation study. He states that comparison
with a similar other might be one important factor increasing the
salience of ego motives.
Nicholls (1975) states that there are two bases for predicting an
individual's attributions. One is that people use information in a
logical fashion, while the other holds that attributions reflect self-
enhancement, approval-seeking and defensive motives. In research with
fourth-graders Nicholls found less evidence of bias after the test
trial than after the practice trial. In fact, failure was attributed
to poor ability more than success was attributed to good ability. This
suggests a derogatory bias rather than self-enhancement. Biased use of
information also varied with attainment value of the characteristic and
with sex of the subject. Luginbuhl, Crowe and Kahan (1975) found that
success was attributed to internal and unstable elements while failure
was attributed to stable elements. Successful subjects attributed their
performance to effort almost twice as much as to ability, while failing
subjects made attributions to ability four times as much as to lack of
effort. Another study not supporting the notion of self-enhancement
was conducted by Chaikin (1971). Subjects participated in a twelve
trial "game" with the dependent measure taken after trials one, five
and twelve. There were no differences between success and failure or
between ascending and descending feedback. The authors state that in cases
of failure subjects may see the possibility of bettering their per-
formance through persistence.
The last set of self-enhancing hypotheses to be presented here
concern the feedback variable and stem from the desire to make oneself
look good. In line with this, negative feedback tends to lower self-
esteem while positive feedback does not (Hypothesis XII). After positive
feedback subjects will give a source a higher evaluation than subjects
receiving negative feedback, who will tend to devalue the characteristic
as well (Hypotheses XIII and XIV). Continuing this attempt at self-
enhancement, a subject who has failed (gotten negative feedback) will
be more likely than a subject who has succeeded to choose a comparison
who is worse off, thus looking and feeling good in comparison (Hypoth-
esis XV). The probability of performance being attributed to the
environment is greater for negative feedback than for positive, which
will be attributed to the self (to ability and personal characteristics
and effort). This constitutes Hypothesis XVI and will be especially
true for an important ability. Finally, the lowering of self-esteem
following negative feedback is the least likely response since modes
of self-enhancement will be attempted first (Hypothesis XVII). But if
self-enhancement does not occur, self-esteem will drop.
An important part of this study, in contrast to many other experi-
ments on self-evaluation, is the measurement of several modes of
self-enhancement. A potential methodological problem for such a situa-
tion is that subjects may show an effect (work out self-enhancement)
on the first possible opportunity. It is thus important to vary the
order of measurement of dependent variables and to test for order
effects. This issue will be more fully addressed in methodology and
results sections. Another question, to be addressed in the analyses,
is the evidence for alternative use of self-enhancement responses; this
refers to the question of whether subjects will exhibit self-enhancement
on only one of the dependent measures.
Summary and Hypotheses
The basis of the previous material is that there will be some
effect on behavioral or cognitive reactions following information about
one's position on a characteristic relative to others' positions.
Specifically, in this context, the individual will exhibit some behavior
relating to enhancement or protection of the self, taking advantage of
positive feedback and compensating for negative feedback. This is
offered as a co-existent formulation to Festinger's (1954a) theory
about using comparison with others to achieve accurate appraisal of
oneself. In situations where self-esteem is more important than ac-
curacy, Festinger's notions will not apply. Rather, the process of
self-enhancement will occur. Secord (1968) lists two contexts evoking
self-evaluation: situations where one deviates from a self-standard
and at times when others give direct or implicit evaluation of the self.
In such situations the desire for self-enhancement is probably greater
than the desire for accuracy. The present methodology presents
individuals with evaluation of the self, presumably leading to self-
Cohen (1959) states that research is needed on the modes of
self-evaluation. There are a number of aspects of human behavior that
have the potential for reflecting self-enhancement efforts. We have
chosen five behaviors, with the implicit assumption that these are the
most common types of reaction. At a general level, the hypothesis is
presented that some reactions will be manifested in at least one of the
following areas: self-esteem, evaluation of the characteristic, attri-
butions of causality, choice of social comparison other and evaluation
of the source of feedback. We have stated that subjects will show some
response on at least one of these variables, after negative feedback
especially, but which response any one subject chooses cannot be pre-
dicted with much accuracy. In observing these behaviors only, we run
the risk that reaction may occur in some other area of the individual's
behavior or cognitions. But the thesis is that effects will be most
likely seen in the aforementioned behaviors. The following hypotheses
deal more specifically with the relationship between the independent
and dependent variables.
Hypothesis I Feedback on an important characteristic is more likely
to influence self-esteem than is feedback on an un-
important characteristic, in that negative feedback on
an important characteristic lowers self-esteem while
positive feedback raises self-esteem.
Hypothesis II More differentiation between attribution of causality
to the self and environment following positive and
negative feedback occurs when the characteristic is
important than when it is not important.
Subjects are more likely to have a low evaluation of
an unimportant characteristic than to have a low evalu-
ation of an important one following negative feedback.
Feedback on an important characteristic more often
leads to choice of a self-enhancing comparison other
than does feedback on an unimportant characteristic.
Negative feedback on an important characteristic more
often leads to choice of someone scoring more poorly
as a comparison.
When the characteristic is an opinion rather than an
ability, subjects more often choose a similar (and
thus supporting) social comparison following negative
Ability-feedback has more effect than opinion-feedback
The differentiation between attribution of causality
to the self and environment is greater in the case of
positive and negative feedback on abilities than for
Subjects are more likely to give a low evaluation of an
opinion characteristic than an ability characteristic
following negative feedback.
A credible evaluator's feedback has more effect on
self-esteem than does feedback from a less credible
Subjects are more likely to give a low evaluation of
the low credible evaluator than the more credible
evaluator following negative feedback.
Differentiation between attributing positive feedback
to self and negative feedback to environment is more
pronounced for a credible than for a less credible
In general, positive feedback tends to raise self-esteem
while negative feedback leads to lowered self-esteem.
Positive feedback leads to a higher evaluation of the
source of feedback than does negative feedback.
Positive feedback leads to a higher evaluation of the
characteristic than does negative feedback.
Following negative feedback motivation to choose a more
negative comparison is greater than after positive
Positive feedback leads to more attributions of causal-
ity for performance to the self and negative feedback
is attributed to the environment.
Lowered self-esteem is least likely as a response
following negative feedback, since modes of self-
enhancement will be attempted.
One hundred and ninety-two undergraduate students enrolled in
introductory psychology received experimental credit for participating
in the study. Three additional subjects had to be deleted from the
sample: two were suspicious and one had been told about the experiment
by a previous participant.
Design and Overview
The main design of the study was a 2 x 2 x 2 x 2 between-subjects
factorial design varying importance of the characteristic on which
subjects received feedback (high or low importance), the nature of the
particular characteristic on which subjects received feedback (ability
or opinion), source of feedback (moderate or low credible source), and
nature of the evaluation outcome (positive or negative). Twelve subjects
participated in each of the sixteen cells. Nested within the two
variables dealing with the characteristic were three different impor-
tant opinions, three different important abilities, three different
unimportant opinions and three different unimportant abilities. Two
pairs of female experimenters conducted the study, with one member of
each pair always playing the part of "project coordinator" (Experimenter
A) and the other being "interviewer" (Experimenter B). Each subject
was exposed to only one pair. Equal numbers of males and females served
in each condition.
Before presenting the procedure, an outline is given of checks
on the operationalization of the independent variables. Thus, when
reading the procedure, readers will be aware that independent variable
manipulations had been previously rated appropriately by similar college
Independent Variable Manipulations
For nature and importance of the characteristic a sample of 55
undergraduates from introductory psychology (males and females) rated
a list of 54 opinions or 56 abilities on their importance [see Appendix
A]. Based on these ratings the final 12 characteristics were chosen so
that 6 characteristics were rated significantly more important than 6
others, but such that differences in ratings did not exist within
ability or opinion categories. The first question asked subjects to
rate "how important you think it is for someone to have an opinion on
that object (be good on that ability)." Analysis of variance showed
that there were significant differences in importance between charac-
teristics, F (11,648) = 422.19, p < .01. Scheffd's multiple comparison
tests showed that important attributes in general differed significantly
from unimportant attributes. Unimportant attributes were similar,
except that the rating for cross-word puzzles was higher than the rating
for platform shoes (p < .05). The characteristics and their mean
ratings are given in Table 1.
Mean Ratings for Independent Variable Characteristics
Important Abilities Important Opinions
Clear thinking 5.46 Education 5.37
Reading comprehension 5.26 Law and justice 5.37
Decision-making 5.24 Self-reliance 5.26
Unimportant Abilities Unimportant Opinions
Building cupboards 2.32 Birthday parties 2.23
Flower arranging 2.37 Platform shoes 1.58
Cross-word puzzles 2.67 T-shirts in public 1.83
The second question asked raters how interested they would be in
finding out whether their expressed opinion or ability performance was
similar to that of experts in a relevant field. Again, analysis of
variance showed that significant differences existed, F (11,648) =
153.32, p < .01. Scheff6's multiple comparison procedures showed that
important characteristics differed from unimportant ones, but there
were no differences within categories.
For source of evaluation feedback, the two interviewers who gave
the feedback to the subjects were rated by nine introductory psychology
students prior to the main study as a check on the credibility manipula-
tion. Below is a description of the manipulation of credibility:
Moderate credibility. For moderate credibility the interviewer
dressed in a professional-looking pantsuit or dress, over which she
wore an unbuttoned white lab coat. On the wall above her desk was a
plaque signifying that she had received the Personality Interviewer
degree from the Institute for Assessment Research. When she first
addressed the subject in her office, she said, "Hello, I'm Miss Cohen.
I've had a lot of experience scoring and interpreting these scales and
I'm known for being quite effective at this. But I still enjoy talking
with our participants."
Low credibility. The low credible interviewer wore blue jeans and
a casual shirt, in the manner typical of undergraduate dress. On the
wall over her desk was a still life picture. When she first spoke to
the subject she said the following: "Hi, I'm Carol Cohen. I haven't
had much experience at scoring and interpreting these scales. I'm just
learning to do this interviewing, so I hope this comes out all right."
The three questions which the raters answered, and the results of
t-test comparisons are given in Table 2.
For performance feedback 18 pilot subjects (males and females)
were led through the procedure and gave ratings of the feedback they
received. They had randomly received positive or negative feedback
appropriate to their opinion or ability condition. The interviewer
gave one of the following four statements:
Positive ability: "The way you answered this scale is very similar
to the way experts in this area answered it. You did quite well com-
pared to other people who have completed this scale. Your score was
Positive opinion: "The views you expressed on this issue are very
similar to the opinions of experts in this area who completed this scale.
You did quite well compared to other people who have answered this scale.
Your score was 201."
Negative ability: "The way you answered this scale is not at all
Manipulation Check for Credibility Variable
1. "How much faith do you put in the interviewer's ability to score and
interpret the student's scale?" (Six point scale)
Mean Standard Deviation
Low credibility 2.0 1.4
t (7) = 1.93, p < .05
Moderate credibility 3.8 1.6
2. "How experienced is the interviewer at the task?" (Six point scale)
Mean Standard Deviation
Low credibility 1.75 1.5
t (7) = 1.96, p < .05
Moderate credibility 3.20 0.5
3. "How professional was the interviewer?" (Six point scale)
Mean Standard Deviation
Low credibility 2.0 0.57
t (7) = 2.29, p < .025
Moderate credibility 3.2 0.84
similar to the way experts in this area answered it. You did not do
very well compared to other people who have completed this scale. Your
score was 141."
Negative opinion: "The views you expressed on this issue are not
at all similar to the opinions of experts in this area who completed
this scale. You did not do very well compared to other people who have
answered this scale. Your score was 141."
Subjects responded to two questions concerning the feedback they
received. The results of the analyses can be seen in Table 3.
Manipulation Check for Feedback Variable
1. "Rate your standing anywhere from -3 to +3." (Seven point scale)
Mean Standard Deviation
Positive feedback 1.66 0.5
t (15) = 8.12, p < .001
Negative feedback -1.38 1.06
2. "What is your reaction to your feedback? Assign a number from +3
equals "good" to -3 equals "bad"."
Mean Standard Deviation
Positive feedback 2.00 0.86
t (16) = 3.53, p < .005
Negative feedback 0.22 1.17
In summarizing the manipulation checks, the analyses show that
the manipulations for the independent variables were successful in that
differences were perceived by pilot subjects. Important and unimpor-
tant characteristics differed, moderate and low credibility were seen
differently and the two feedback conditions elicited different reactions.
In six large group sessions of at least twenty people in each,
undergraduates convened for the first part of the study. The project
coordinator (Experimenter A) conducted the session, beginning with the
following introduction: "Hello, I'm the local representative for the
Institute for Assessment Research which is based in New York City.
The Institute is conducting a large nation-wide study on the assessment
of different abilities and opinions in many segments of the population.
Our local group is in charge of a study on university students. This
study involves two sessions--today I will administer the assessment
scales to you and a background information sheet. We are assessing
several abilities and opinions, but each of you will receive only one
scale, so do not be concerned if the person next to you has a different
scale. At the second session each of you will meet individually with
one of our staff who will tell you how you responded compared to other
people. This study is concerned with differences in opinions and
abilities based on background variabiles. When we evaluate your re-
sponse we make two comparisons: how you compared to people with similar
backgrounds and how you compared to a set of experts from the particular
ability or opinion field, who filled out our scales several weeks ago.
In the booklet I'm handing out to you, you'll find a cover letter ex-
plaining the project, a background information sheet and one of the
assessment scales. Please answer as completely as you can. On the
first page, please put your name and phone number so we can reach you
to schedule the second session. After you have been given feedback on
your scale response, your name will be separated from the booklet and
all later analyses will be done anonymously." Subjects then completed
their booklets. One of the twelve characteristics was thus randomly
given to each subject, making sure an equal number of males and females
received each characteristic. (See Appendix B for copies of the cover
letter, the background information sheet and the assessment scales.)
Within a week subjects were called and scheduled for a half-hour
session in the Psychology Building. When they arrived, the interviewer's
room was arranged for the appropriate credibility condition. After the
initial introduction and credibility manipulation, the interviewer
said, "I'm going to give you some information on how you performed
relative to experts who filled out this scale and compared to other
subjects. Then I have some forms for you to fill out to help us in
our research. But first your Psychology Department requires that we
get your informed consent." Subjects read the form, signed it and the
experiment continued. Subjects randomly received positive or negative
feedback appropriate to their ability or opinion condition. After this,
the interviewer said, "That's all I have on that right now. I'd like
you to complete these two forms for me." She then administered the
self-esteem measure and the attribution of causality measure (see
Appendix C). The order of administration was alternated so that some
subjects received the self-esteem measure first and others got the
attribution scale first. When the subject had completed these scales
the interviewer said, "The project coordinator wants to see you for a
few minutes, but she's late in getting here. We have copies of dif-
ferent levels of quality of responses to the scales in these folders.
Here are some people who scored above 210, here's 131-170, here's 171-
210 and here's below 130. You can look at these while I get the co-
ordinator." For each session the interviewer randomly varied the
placement of the folders and the order in which she described them.
As she stood to go to the door, she noted which folder the subject con-
sulted. As she went into the hall, the project coordinator (Experimenter
A) "just happened" to arrive at that moment. She introduced herself
and led the subject to another office. Experimenter A was blind to the
subject's condition. When seated in the other office, she said, "There
are a couple of research concerns I have about this project and I'd like
you to answer some questions that will help me. One concern has to do
with the notion of standardization. I don't know how much you know
about research, but what we strive for is to have each participant be
treated in the same way and go through the same process. This is a
little hard in this case where we have several interviewers, each of
whom sees only some of the participants. I'd like to get some idea of
how they are each conducting the sessions to see if they're pretty much
doing the same things the way they're supposed to be doing. So I have
a rating form I'd like you to fill out about the session you just had
with the interviewer. Please answer honestly and don't feel you're
putting someone's job in jeopardy because you're not. That's not how
I use this information. The other form is needed because we are thinking
of putting some new characteristics in the study. Before we do that
I'd like some idea of how people rate the characteristics. Included
on this form are some opinions and abilities we're using now as well as
the ones we hope to use in the future." Experimenter A varied the order
of administration of the Personnel Performance Sheet and the Attribute
Rating Scale. When subjects had finished, a debriefing session was
conducted and subjects were probed for suspicions. In addition, Ex-
perimenter A told them the nature of the study and asked them to maintain
silence about the study since more subjects from the psychology classes
would be participating.
Dependent Variable Measures
In order to tap several psychological processes that might be
influenced by the independent variables, measures were taken on five
general dependent variables. (See Appendix C for copies of these
1. The attribution of causality for individual performance measure
asked subjects to state percentages of their performance due to ability,
effort, task difficulty and luck. In addition, they rated each of
these four elements on a five point scale in terms of the degree to
which it was a cause of their behavior. A summary measure was calculated
for each of these two types of measures: the sum of task and luck
ratings was divided by the sum of ability and effort ratings giving an
indication of whether performance was attributed more to self or en-
2. Self-esteem was measured by the Janis-Field Feelings of Inade-
quacy Scale and the Rosenberg Self-esteem Scale. Reliability and
validity data for these scales are presented in Table 4. These scales
were untitled when presented to the subject. The scores were summed
giving a total self-esteem index score.
Reliability and Validity Data for Self-esteem Scales
Janis-Field Feelings of Inadequacy Scale Rosenberg Self-esteem Scale
Reliability Split-half .72 to .82 Test-retest .85
Validity Convergent .67 to .84 Convergent .56 to .83
Predictive .24 to .27
Note: See Robinson and Shaver (1973) and Rosenberg (1965).
3. Personnel Performance Sheet included ratings of the interviewer
on 11 adjectives. In addition, subjects gave an overall rating of the
interviewer, a rating of her clarity in giving instructions and an
assessment of how good she was for the job.
4. The Attribute Rating Scale consisted of ratings of 17 char-
acteristics, including the one on which the subject had been assessed.
Four questions tapped importance to the self, importance to the general
public, extent to which a friend would refer to the characteristic in
describing the subject and how interested the subject would be in getting
feedback on her/his standing relative to others. Included as ratings of
the characteristic were three items found on the Personnel Performance
Sheet. These were ratings of the initial assessment scale on its
validity, its accuracy as a measure and its usefulness.
5. Social comparison choice was behaviorally measured by noting
which of four folders each subject looked at in order to get an idea of
how others had performed. Folders were labelled with the scores in-
cluded in them: "Below 131", "131-170", "171-210", and "Above 210".
The interviewer unobtrusively noticed which folder each subject chose
Thirty-four dependent variables were used to test the hypotheses.
A sum of the two self-esteem scores was used as the measure of self-
esteem. Attribution ratings included four percentage ratings of
attribution to luck, the difficulty of the task (these two constitute
environmental, external elements), effort and ability (these two con-
stitute internal, self elements), and four seven-point scale ratings
of the same elements in terms of their influence on performance. In
addition, two composite scores were formed by adding the two environ-
mental elements and dividing by the sum of the two self elements,
separately for the two types of ratings (percentage and scale). Four-
teen ratings of the interviewer served as evaluation of the source of
feedback and there were three items rating the original assessment
scale. Four questions addressed evaluation of the particular charac-
teristic. Two social comparison choice measures were included. The
first categorized each subject's choice as above, the same as or below
their own score, while the second indicated whether they chose to look
at a score above 201, between 171 and 201, between 131 and 171 or below
Kolmogorov-Smirnoff tests of normality were conducted for each of
the dependent variables. All but seven violated normality in terms of
skewness, kurtosis or both at a statistically significant level. Con-
sequently a multivariate analysis of variance could not be used as had
originally been planned. Multivariate analyses were necessary due to
the large number of dependent variables. In most cases, each hypothesis
concerned predictions about one or two independent variables and more
than one dependent variable. Such a large number of statistical tests
leads to altered probabilities of Type I and Type II errors. It is
therefore advisable to compute a multivariate statistic before pro-
ceeding to consideration of univariate analyses.
Because of these violations of distribution assumptions, the Sarle
Distribution-Free Multivariate Linear Hypothesis computer program was
used. This package performs a multivariate analysis on a set of vari-
ables by assigning ranks and testing to see if treatments differ in
their mean ranks. (However, it is important to note that we report
mean ratings not mean ranks for each condition in the relevant tables.)
In testing the hypotheses a significance level of .05 was used. Multi-
variate analyses and appropriate univariate analyses are given in the
form of chi-square tests. Pairwise-comparisons testing for between-cell
differences are provided by standard scores. In this study, significant
pairwise-comparisons are those with Z-values > 1.96. A multivariate
analysis including all four independent variables (blocking on sex of
subject and on experimenter team) at once was not possible due to ex-
cessive storage requirements beyond the computer's capacity. Since a
majority of the hypotheses concern an interaction between the feedback
variable and one of the other independent variables, three analyses
paired the feedback variable with nature of the characteristic, impor-
tance of the characteristic and credibility of the source of feedback.
In addition, feedback was tested by itself with the other three inde-
pendent variables (as well as sex of subject and experimenter team) as
blocks. Three of these analyses were significant: credibility/feed-
back, x2 (102) = 140.8 p < .01, nature of the characteristic/feedback,
x (102) = 175.69, p < .0001 and importance of the characteristic/
feedback, x2 (102) = 232.8, p < .0001. The analysis for feedback was
not significant. Results appropriate to each hypothesis will be pre-
sented in order.
Hypothesis I stated that feedback on an important characteristic
has more influence on self-esteem than does feedback on an unimportant
characteristic. This predicts an interaction between feedback and
characteristic importance such that there is no difference between the
two unimportant conditions on self-esteem, but there is a difference
between the two important-characteristic conditions. The chi-square
test was not significant; Hypothesis I was not supported.
Hypothesis II stated that more differentiation between attribution
of causality to the self and environment following positive and negative
feedback occurs when the characteristic is important than when it is
unimportant. Of the ten univariate interactions, three reached statis-
tical significance: luck rating, x2 (3) = 8.09, p < .045, luck per-
centage, x2 (3) = 8.87, p < .05, and effort percentage, x (3) = 12.86,
p < .001. Mean ratings (not mean ranks) can be seen in Table 5. For
all three analyses important-characteristic/positive-feedback differed
from the two unimportant conditions, receiving more attribution to
effort percentage and less to luck. In addition, for effort percentage,
important-characteristic/negative feedback differed from the two unim-
portant conditions. The predicted relationship of a difference between
the two important conditions but not between the two unimportant con-
ditions did not occur; Hypothesis II was not supported.
Mean Ratings of Attribution Items for Hypothesis II
Condition Attribution Items
Luck Luck Percentage Effort Percentage
Positive 1.5ab 9.6cd 27.9ef
Positive 2.0a 16.5c 21.2eg
Negative 1.6 11.1 29.8gh
Negative 2.3b 20.5d 21.5fh
Note: Means with like subscripts are significantly different.
Hypothesis III predicted that subjects would be more likely to
change their evaluation of an unimportant characteristic than of an
important one. The importance-of-the-characteristic by feedback test
was significant for all four items: "importance to me", x2 (3) = 133.3,
p < .0001, "importance to the general public", x2 (3) = 99.0, p < .0001,
"a friend would say I have this characteristic", x2 (3) = 124.1,
p < .0001, and "I would be interested in learning where I stood on this
characteristic", x2 (3) = 73.9, p < .0001. In addition, the three
items rating the assessment scale for each subject's characteristic
reached significance: "valid", x2 (3) = 10.6, p < .01, "accurate",
2 (3) = 18.3, p < .0005, and "useful", x2 (3) = 12.4, p < .006. As
can be seen in Table 6 there was a main effect for importance of the
characteristic on the first 4 items with important characteristics
being rated higher on all items. For each of the last three items the
0 0 0
1 N O
c 4O c3 40 0 )
*0 c H 3-HM 0u 3c )3
0 0) >a.3) > z H i l
u 0 3O E3 C 0
CL, 91 -Z Z
two important conditions differed from the unimportant-characteristic/
negative-feedback condition with the important characteristics re-
ceiving higher ratings. For "accurate" the important-characteristic/
negative-feedback condition differed from the important-characteristic/
positive-feedback condition and for "useful" the important-character-
istic/positive-feedback condition differed from the unimportant-char-
acteristic/negative-feedback condition. The predicted effect of a
significant difference between the two unimportant conditions but not
between the two important conditions did not occur. For "accurate" the
opposite effect occurred. Hypothesis III was not supported.
The next hypothesis was concerned with whether feedback on an
important characteristic more often leads to choice of a self-enhancing
social comparison than does feedback on an unimportant characteristic
(Hypothesis IV). The interaction was not significant for the first
social comparison measure, but it was for the second, x2 (3) = 8.0,
p < .05. Table 7 shows that the two positive-feedback conditions dif-
fered from the unimportant-characteristic/negative-feedback condition.
Mean Ratings for Second Social Comparison Measure (Hypothesis IV)
Condition Mean Rating
Note: Means with like subscripts significantly differed.
This is not the predicted effect of a difference between the two impor-
tant-characteristic conditions but not between the two unimportant-
characteristic conditions. Hypothesis IV was not supported.
Turning from importance of the characteristic to nature of the
characteristic (opinion or ability) the next hypothesis stated that
feedback on opinions would more likely lead to choice of a similar social
comparison under negative feedback than would feedback on abilities
(Hypothesis V). This analysis was significant only for the second social
comparison measure, x (3) = 9.9, p < .02. As can be seen in Table 8,
the two negative-feedback conditions differed from the ability/positive-
feedback condition. This is the opposite of the prediction; Hypothesis
V was not supported.
Mean Ratings for Second Social Comparison Measure (Hypothesis V)
Condition Mean Rating
Note: Means with like subscripts are significantly different. A
higher number signifies choice of a comparison below the subject.
Hypothesis VI was not supported: the nature of the characteristic/
feedback analysis was not significant for self-esteem as had been pre-
The next hypothesis was that differentiation between attribution
of causality to self and environment is greater in the case of feedback
on abilities than opinions (Hypothesis VII). Five of the ten dependent
variables were significantly influenced by nature of the characteristic
and feedback. All environmental elements were significant, as was
ability percentage [task rating, x (3) = 32.7, p < .0001, task per-
centage, x2 (3) = 37.5, p < .0001, luck rating, x2 (3) = 9.5, p < .05,
luck percentage, x (3) = 12.3, p < .01 and ability percentage, x (3) =
9.4, p < .05]. For task rating positive-feedback/ability differed from
negative-feedback/opinion, and both opinion conditions differed from
negative-feedback/ability. For ability percentage the opposite effect
from that predicted occurred in that there was a difference between the
two opinion conditions but not between the two ability conditions. For
luck rating the predicted effect was found in that the two ability
conditions differed but the two opinion conditions did not. In this
instance positive-feedback/ability was attributed less to luck than was
negative-feedback/ability. In addition, positive-feedback/opinion
differed from negative-feedback/ability. For luck percentage positive-
feedback/ability differed from positive feedback/opinion and negative-
feedback/ability differed from the two opinion conditions. Thus
Hypothesis VII was supported for the luck rating item (see Table 9).
Hypotehsis VIII stated that subjects would be more likely to de-
value opinions that abilities following feedback. Only one of the four
items was significant: "friend would say I have this characteristic",
x (3) = 10.8, p < .01. Table 10 shows that ability/negative-feedback
differed from ability/positive-feedback and from opinion/negative-
feedback. In addition, two of the three ratings of the assessment
scale reached statistical significance: "valid", x2 (3) = 8.7, p < .05,
Mean Ratings for Attribution Items (Hypothesis VII)
Task % Luck Luck %
18.75de 2.1h 16.83j
10.8df 1.41 9.3jk
26.4fg 3.2hi 19.2km
9.7eg 1.6 12.5m
subscripts are significantly different.
Mean Ratings for Hypothesis VIII
Condition Evaluation Items
Friend says I have Valid Accurate
Ability 2.5a 4.0c 4.leg
Opinion 2.4 4.1d 4.1fh
Ability 2.lab 3.4cd 3.2ef
Opinion 2.6b 3.4 3.3gh
Note: Means with like subscripts are significantly different.
and "accurate", x2 (3) = 13.8, p < .005. On both items ability/negative-
feedback differed from the two positive-feedback conditions, so this
was the opposite of the predicted effect. For "accurate" ability/
positive-feedback differed from opinion/negative-feedback and the two
opinion conditions differed from each other. There was never a case
where the two opinion conditions differed and the two ability conditions
did not, so Hypothesis VIII was not supported.
The remaining hypotheses deal with the credibility of the source
and the feedback received. It was predicted that feedback from a
credible source would have more effect on self-esteem than would feed-
back from a less credible source (Hypothesis IX). The statistic did
not approach significance, and this hypothesis was thus not supported.
Hypothesis X stated that a less credible source of feedback would
more likely be devalued by subjects than would a more credible source
after failure feedback. Four of the analyses between the credibility/
feedback effect and the fourteen dependent variables were significant:
'how clear was the interviewer in giving you instructions and in ex-
plaining results?", x2 (3) = 9.9, p < .05, "maturity", x2 (3) = 10.8,
p < .05, "experience", x2 (3) = 35.99, p < .0001, and "trust", x2 (3) =
9.97, p < .05. Table 11 shows the differences between means. In the
first item credible-source/negative-feedback differed from the other
three conditions, while credible-source/positive-feedback differed from
all others on the "maturity" item. There was a main effect for source
credibility on the "experience" item and for "trust" credible-source/
negative differed from the two low-credibility conditions. None of the
effects exhibited the predicted difference between the two low-credible-
source conditions with no difference between the credible-source condi-
tions; Hypothesis X was not supported.
Mean Ratings for Evaluation of Source (Hypothesis X)
Condition Evaluation Items
instructions Mature Experienced Trustworthiness
Positive 5.3a 5.0def 4.4gh 4.3
Negative 4.6abc 4.3d 4.4ij 3.9km
Positive 5.4b 4.3e 3.0gi 4.4k
Negative 5.3c 4.4f 3.0hj 4.5m
Note: Means with like subscripts are significantly different.
Hypothesis XI predicted that differential attribution of positive
feedback to self and negative feedback to environment would be more
likely for a high credible source of feedback than for a low credible
source. The analysis was not significant for any of the 10 attribution
variables; Hypothesis XI was not supported.
The overall multivariate analysis with feedback as the independent
variable did not reach significance, which precluded consulting the
appropriate univariate analyses for Hypotheses XII, XIII, XIV, XV, XVI
and XVII. Therefore none of these hypotheses was supported.
Three separate nonparametric multivariate analyses with order,
version of the characteristic and sex of subject each as independent
variable showed no influence of these variables on subject responses.
It did not make a difference what order subjects completed the depen-
dent measure forms, not did the 12 versions of the 4 cells formed by
interaction differentially affect behavior within cells. In addition,
males and females did not show different behavior in this study.
However, the analysis for experimenter team showed that experimenter
team did significantly influence subject behavior, x (34) = 69.54,
p < .0001. Seven of the evaluations of the experimenter yielded sig-
nificantly different ratings for the two interviewers. In all instances
one interviewer was rated more negatively than the other. The variables
that differentiated were: "warmth", x2(1) = 26.6, p < .0001, "sensi-
tivity", x2 (1) = 12.93, p < .0005, "openness", x2 (1) = 6.76, p < .01,
"personableness", x2 (1) = 10.46, p < .005, "how did the interviewer
conduct the session", x2 (1) = 8.7, p < .005, "interviewer's clarity
in giving instructions and in explaining results", x2 (1) = 4.95, p < .05
and "is the interviewer a good person for the job", x2 (1) = 18.55,
p < .0001. In addition, the two social comparison measures were sig-
nificantly different for the two interviewers, first measure, x (1) =
16.1, p < .0001, and second measure, x (1) = 4.0, p < .05. This dif-
ference is partly due to the fact that one interviewer had more people
not make a social comparison choice. Experimenter differences thus
affected subjects responses on one variable administered in the inter-
viewer's presence and on some of the items specifically rating her.
Experimenter differences thus appear to be due to the two interviewers
and not due to the other member of each team. Also, experimenter
differences were not very widespread in influencing subject behavior.
The next set of analyses addressed the issue of the presence of
multiple response modes. In essence subjects had four general self-
enhancement variables provided to them, and a measure of self-esteem.
Analyses have showed that order of presentation had no effect on the
dependent measures. Another question investigated simultaneous versus
alternative use of responses, and asked whether subjects were more
likely to use one response rather than another, or if they used one
were they less likely to use any other. Previous research in other
areas suggests ways to address this question. Pepitone and Wilpizeski
(1960) computed correlations among responses under each condition to
see if responses produced by experimental conditions were related to
each other within the same individual. Steiner and Rogers (1963) stated
that if there are low or negative correlations between response vari-
ables, this is an indication of use of alternative measures. Correla-
tion coefficients were computed between dependent measures for positive
feedback, negative feedback, important characteristic, unimportant
characteristic, abilities, opinions, high credibility and low credi-
bility. There are correlations between 34 dependent variables for each
of these analyses. For each analysis a count was taken of the number
of positive significant correlations between separate dependent measures.
That is, we did not include correlations between related dependent
measures such as the 14 ratings of the source of feedback or between the
10 attribution scores. There are a total of 830 such correlations
between separate dependent measures. Contingency tables were constructed
between pairs of an independent variable and the number of significant-
positive and insignificant-and-significant-negative correlations.
Chi-square tests were then computed. Results of such an analysis must
be viewed with considerable caution given the questionable validity of
this particular set of tests. Feedback, x2 (1) = 20.26, p < .005 and
nature of the characteristic, x2 (1) = 7.95, p < .005 were significant
while the other two analyses were not. For these two independent
variables there was some indication of use of alternative response
measures, specifically under negative feedback conditions and for
opinion characteristics (see Table 12).
Steiner and Rogers (1963) used another analysis in an attempt to
address the question of use of alternative responses. Subject ranks on
one response were compared with the sum of their ranks on the other
responses. This was done in turn for each dependent measure compared
to the sum of all others. A similar analysis was conducted on the
present data, but because it would have been too cumbersome to use all
34 dependent variables, a representative for each of the five major
dependent measures (social comparison, evaluation of the characteristic,
evaluation of the source of feedback, self-esteem and attribution of
causality) was used. The first social comparison measure was used
because it is the more valid of the two measures, and the sum of the
self-esteem scored constituted the self-esteem measure. For evaluation
of the source 2 items with the highest correlations with all others
were summed and used (competent and sensitive). "Importance to me"
correlated best with the items evaluating the characteristic so it was
chosen and the composite attribution scale rating was chosen for the
same reason. After getting the ranks, subjects were dichotomized at
the median on each single variable and also on the sum of all others.
A chi-square test was conducted to determine if people who are above
the median on one measure are below the median on the sum of the other
measures, and vice versa. No chi-square test reached significance,
which means that subjects were not more likely to be below the median
on the sum of the ranks of four measures if they were above the median
Correlations > .36 by Condition and Related Contingency Tables
Condition Items Condition
Ability Clear & Valid .46 Opinion
Clear & Useful .37
Good for job & Valid .37
Insight & Valid .38
Intelligence & Valid .42
Trust & Valid .44
Open & Valid .42
Trust & Friend .38
Trust & Interested Negative
in score .40
Positive Clear & Valid .37
Trust & Accurate .43
Trust & Valid .40
Credible Clear & Valid .42
Clear & Useful .42
Clarity & Valid .42 Low Cred-
Clarity & Accurate .40
Intelligence & Valid .38
Intelligence & Accurate
Intelligence & Useful .39
Trust & Accurate .37
Clarity & Valid .46
Clarity & Accurate .37
Clarity & Useful .40
Useful & Interested in
Intelligence & Valid .40
Intelligence & Useful .40
Valid & Interested in
Useful & Interested in
Clear & Valid .38
Clarity & Valid .41
Valid & Interested in
Useful & Important to
Useful & Interested in
Open & Valid .42
Personableness & Valid .37
Useful & Important to
Useful & Interested in
Table 12 (Continued)
S = Significant positive
N = Not significant, and
Positive 140 690
Credible 134 696
Unimportant 121 709
on the fifth measure (see Table 13 for the contingency tables). This
finding provides no evidence of alternative use of response measures.
Contingency Tables for Median Ranks
Above on single 53
Below on single 38
Above on single 51 44
Below on single 45 52
Above on single 50 55
Below on single 45 40
Above on single 47 41
Below on single 49 55
The central foci of this chapter are discussions of the nature
of the support for self-enhancement theory and social comparison theory.
Then we offer some interpretation of the results and attempt to explain
why the hypotheses in general were not supported.
The first point relevant to self-enhancement theory is that none
of the hypotheses concerned with self-esteem were supported, nor in
fact were there any influences on self-esteem. This could mean at
least two things. First, if self-enhancement had been successfully
achieved then self-esteem would show no effect. This is related to
Hypothesis XVII and stems from the notion that self-enhancement responses
are preferable in an evaluation situation to loss of self-esteem. Given
the lack of support for almost all self-enhancement hypotheses this
notion is unlikely. The second possible reason for no self-esteem
effects is that the feedback variable may have had no effect, or negative
feedback was not enough of a threat to self-concept and thus did not
effect self-esteem. The manipulation check on the feedback variable
suggests that it is unlikely that positive and negative feedback were
not perceived differently. There were significant differences between
pilot subjects' reactions to the two forms of feedback. Perhaps this
difference was not strong enough to carry through into subjects' re-
actions to the feedback. The next place to look is the set of results
relating to evaluation. How did positive and negative feedback
influence subjects? The insignificant multivariate analysis for feed-
back shows that there was no main effect due to that variable. In the
paired-comparison test of the two-factor analyses there were a few,
though not many, influences of feedback on the dependent variables.
Feedback affected social comparison in one case where positive and
negative ability conditions had different means, with positive subjects
more likely to choose a folder with a low number. This effect shows
some self-enhancement attempt, though it is opposite of the prediction
of an effect on opinions but not on abilities.
Turning to attributions of causality, we see that there were
several instances where feedback had effects. Positive feedback on
important and unimportant characteristics had different effects on
ratings of luck, luck percentage, and effort percentage, while negative
feedback only affected unimportant versus characteristics on effort
percentage. This does not suggest self-enhancement; one would expect
negative feedback on an important characteristic to lead to defensive
responses and it did not.
Looking at feedback and the nature of the characteristic, the
clearest effect was on ratings of luck and ability percentage. In the
former, the self-enhancement hypothesis was supported in that positive
and negative feedback on abilities, but not on opinions, significantly
influenced attributions to luck. For ability percentage the opposite
effect occurred, but this merely shows self-enhancing attributions on
opinions rather than abilities. However, these two instances out of
ten possible attribution measures with several interactions is not
On evaluation of the characteristic, negative feedback on important
and unimportant characteristics differentially affected ratings of
validity, accuracy and usefulness of the scale, with unimportant
characteristics receiving a lower rating. That unimportant-character-
istic/positive-feedback did not receive lower ratings may be due to the
phenomenon of raising evaluation of an attribute at which one succeeds.
Negative and positive feedback on important characteristic led to dif-
ferent ratings on accuracy, but this was opposite of the prediction of
a difference for feedback on unimportant and not important character-
istics. The underlying assumption is that it is harder to re-evaluate
important characteristics, but perhaps it is more crucial to re-evaluate
important characteristics after negative feedback than to re-evaluate
unimportant ones after success. This reflects a saving-oneself-from-
implications-of-failure-on-important-characteristics, rather than
The other evaluation-of-the-characteristic results suggest that ability
and opinion operate opposite of their hypothesized effects in that in
two cases positive and negative ability were different but not opinions.
For the accuracy rating both opinions and abilities had different re-
actions to positive and negative feedback. Thus feedback does influence
evaluation of the characteristic.
Finally, there is limited evidence for feedback affecting evalua-
tion of the source. On two items there was a difference between the
two credibility conditions but not between the two low-credibility
conditions. This parallels the case of the greater effect of feedback
on important but not unimportant characteristics in evaluation of the
characteristic. Here again, though it was hypothesized to be harder to
devalue high credible persons, it is more important to do so after
negative feedback than to re-evaluate a low credible person after
Thus there are some cases where feedback led to self-enhancing
responses in that subjects behaved ego-defensively. Most of these
effects were opposite to the present hypotheses, but nevertheless they
do show effects of feedback and are examples of attempts at self-
These self-enhancement instances show some other important things.
First of all, the relationship between perception of feedback on abili-
ties versus opinions is different than hypothesized. Ability feedback
seems to be more important to subjects than opinion feedback. This may
reflect greater concern for standing on abilities than for standing on
opinions. It is also possible that in this particular setting the
ability feedback was more believable than the opinion feedback. How-
ever, in choosing a folder of scales to consult, an analysis of the
nature of the characteristic alone showed that ability and opinion
characteristics did significantly differ, with subjects in ability con-
ditions being less likely to choose to see scores below their own. The
unidirectional drive upwards may lead subjects to the desire to see what
a better performance is like.
Another important point is that some of the results did not support
assumptions that it is cognitively harder to re-evaluate important
characteristics and credible persons. Though these assumptions may
be true, other factors such as the need to minimize failure, may be
We now turn to the evidence relating to Festinger's (1954a) social
comparison theory. Since several response modes were included in this
study, only one of which (social comparison choice) is included in his
theory, only a small part of the results are pertinent. Furthermore,
on the first (and more valid) social comparison measure there was no
evidence for the hypothesized effects. Results on the second, objective
choice measure showed that subjects in positive-feedback conditions
were more likely to choose an objectively high score (not necessarily
higher than their own) than were subjects in the unimportant-charac-
teristic/negative-feedback condition. And ability subjects receiving
positive feedback were more likely to choose an objectively high score
(not necessarily higher than their own score) than were subjects in the
two negative-feedback conditions. This is not evidence for choice of
a similar comparison. But these results may serve as an example of the
desire for social consensus: being unsure, those subjects in opinion
conditions were more likely to look at a score similar to or below their
own, perhaps in an attempt to gain some validation for their own posi-
tion. And subjects in ability conditions may have wanted to see what a
good performance was like.
Turning to more indirect issues, there are other results that bear
on some of Festinger's ideas. The ability and opinion differences
cited above can be seen as support for the unidirectional drive upward
for abilities. Furthermore, negative feedback on abilities seemed to
be of more import to subjects than feedback on opinions. This point
shows that people are not always concerned with opinions. Though
Festinger never states it in his hypotheses, presumably some opinions
are of greater concern to people than others. The importance variable
did not have much influence on social comparison however.
Given that there was evidence for ego-defensive responses the
present study has shown that subjects are not always interested in
accepting accurate appraisal of themselves. This is not to imply that
there is no drive for accuracy (in some circumstances), but it shows
that self-enhancement is another possible response rather than accurate
However, the evidence for self-enhancement was limited, and fur-
thermore only one of the proposed hypotheses received any support. There
are a number of possible explanations for this. It may be that the
independent variables do not operate as hypothesized. There was some
suggestion that present assumptions about the effects of source credi-
bility and ability or opinion importance on the cognitive mechanism
of devaluation were not accurate. This explains some of the failure
to support the experimental hypotheses. In addition there is a series
of possibilities for which there is not much data. This relates to
methodological problems in that subjects may not have perceived the
situation as the experimenter presumed. As previously noted (page 25)
methodological issues may be used to explain failures in other studies
of self-enhancement effects. Symonds (1951, pp. 96-97) notes that
situations apparently similar to an outside observer
may not offer the same intensity of threat of failure
to a given individual. It all depends on how he
values that activity and whether his performance in
it is interpreted by him as important.
Though the pilot test showed that the importance variable was manipulated
successfully, the total situation as experienced by the subject may
not have been significant. This is an age-old problem in experimenta-
tion with humans, especially in reactive settings such as the psychology
laboratory. The experimenter often assumes that the subject in the
laboratory views the situation as the experimenter constructed it.
In fact, the situation is often seen very differently by the subjects
and they may not place the same degree of importance and worth to the
scene as the experimenter does. Receiving feedback on one's perfor-
mance on ability and opinion scales as part of a national research
project may not matter much to students, especially if the nature of
the feedback is somewhat vague and nebulous. In fact, several students
mentioned wanting to see their scale to note which responses had led
to the evaluation. In addition, it is possible that subjects did not
believe the cover story, although we have no evidence bearing on this.
Symonds (1951, p. 176) again makes a valid point concerning this
It is difficult to think of the problem of level of
aspiration wholly from the vantage point of the
experimental studies...because these studies are
defined in terms of the limitations of the experi-
mental laboratory while aspirations penetrate every
corner of living.
This also applied to the process of self-enhancement. For certain
variables and certain realms of experience, the experimental laboratory
may not offer the opportunity to realistically test given principles
of human behavior. For areas as sensitive as self-concept and responses
to evaluation of oneself, the laboratory presents some drawbacks in
being removed from individuals' daily lives and probably not being
perceived as the experimenter assumes it is being perceived. This is
not to say that self-evaluation cannot be studied in the laboratory;
clearly many good investigations have been conducted in such settings.
But experimenters must be particularly aware of the attendant difficulties
in constructing the situation and assuming a certain reaction on the
The present study attempted to preclude some methodological
problems. First, the pilot testing of manipulations prior to experi-
mentation tested for operational differences in independent variables.
Also we included both sexes and more than one set of experimenters in
an attempt to increase generalizability and test for differences. In
the present analyses it was possible to block on both these variables,
and see the effects over and above unintended effects of experimenters
and sex of subject. This notion applied as well to multiple versions
of independent variables, such as the use of 12 characteristics to
represent the cells of the importance and nature of the characteristic
interaction. The fact that there was no effect due to the different
characteristics in the same cell increases confidence that the effects
were not due to a particular operationalization, as may often be the
case in research that does not include multiple representation.
In addition to the point that both sexes, more than one set of
experimenters and multiple versions of independent variables be used
in future, investigators in this area must be particularly careful about
the setting used. It is important for the demonstration of self-
enhancing behaviors that subjects take the situation seriously and
feel a real threat to self-esteem in the negative-feedback conditions.
Perhaps use of existent situations will promote this. For instance,
rather than setting up a cover story for an experiment, a researcher
might use the classroom situation in order to test reactions to feed-
back. The increase in mundane and experimental realism raises questions
about the ethics of such a study and it would be particularly incumbent
on such a researcher that she or he safeguard subject feelings and
avoid long-term negative effects. But within ethical limits such
"natural" situations offer promise in the study of self-enhancement.
PRE-EXPERIMENTAL RATINGS OF THE CHARACTERISTICS
1. For the following list of objects which people might have an opinion
about, please rate how important you think it is for someone to
have an opinion on that object by placing the appropriate number
in the first blank.
1 = Very unimportant 4 = Somewhat important
2 = Unimportant 5 = Important
3 = Somewhat unimportant 6 = Very important
Probation officers //
Physical fitness_ //
Campaign regulations //
Academic freedom //
Birth control //
University of Fla. //
Sex education //
Mental institutions //
Socialized medicine //
Union strikes //
Capital punishment //-
Birthday parties //
Platform shoes //
Malpractice insurance //
Team competition //
Disabled people //
Ethnic groups //
Teaching as a career //
China dishes //
House pets //
Antique cars //
Radial tires //
Rotary engines //
Pollution control //
Food stamps //
Television violence //
Public television //
Wearing t-shirts in public //
2. Now go back over the list and in the second blank rate how interested
you would be in finding out if your opinion on the issue was similar
to that of experts who knew a lot about the issue.
A = Very uninterested D = Somewhat interested
B = Uninterested E = Interested
C = Somewhat uninterested F = Very interested
1. For the following list of abilities please rate how important it is
for you to be good by placing
1 = Very unimportant
2 = Unimportant
3 = Somewhat unimportant
Social skills //
Verbal ability //
Make analogies //
Change flat tire //
Physical defense //
Resist temptation //
Clear thinking_ //
Persuading others //
Reading comprehension //
Perceptual tasks //
Crossword puzzles //
Social interaction //
Music composition //
Musical playing //
Public speaking_ //
the appropriate number in the first
4 = Somewhat important
5 = Important
6 = Very important
Defending beliefs //
Doing laundry //
Driving a car //
Typing fast //
Negotiating traffic //
Arranging flowers //
Building cupboards //
Designing buildings //
Solving interpersonal problems //
Making others feel at ease //
Having a good party_ //
Balancing a checkbook //
Tuning a car //
Sewing clothes //
Doing well under pressure //
Standing up for your rights //
Fine tuning a color TV //
Pursuit rotor ability //
Holding your liquor //
2. Now go back over the list and in the second blank rate how interested
you would be in finding out where you placed on a measure of each
A = Very uninterested D = Somewhat interested
B = Uninterested E = Interested
C = Somewhat uninterested F = Very interested
CONTENTS OF BOOKLET AT SESSION I
Cover letter, background information sheet and
the twelve assessment scales
Institute for Assessment Research
Curtis Foundation Academic Year
New York, New York 10022 1975-1976
You are taking part in a large nation-wide study being conducted
by the Institute for Assessment Research, a department under the aus-
pices of the Curtis Foundation. Our project involves a number of
different research questions, such as:
1. How well can we use demographic data to predict performance
on various ability and opinion assessment procedures?
2. How do various segments of the population differ in their
performance on our scales?
3. What assessment procedures are the most practical for
4. How useful are assessment procedures for the various organiza-
tions and institutions in our society?
These are only a few of the topics we are studying over a period of
several years. I personally would like to thank you for your part in
this research. It is through the cooperation and honest responses of
people like you that we can answer important questions about human
Thank-you again for your part in this project.
With warm regards,
National Executive Director
Student Form B
Please answer the following questions as best you can. This informa-
tion is important to use in the research project. The forms you fill
out will be anonymous, except for the first phase when you meet
individually with one of our interviewers. For that purpose, please
make sure your name is on the first page of this booklet. That page
will be omitted when we conduct the analyses.
1. Sex: Male Female
3. Date of Birth:
5. Grade Level: Freshman Sophomore Junior Senior Graduate Student
6. Course major (if declared):
7. Preference for Course Major (if undeclared):
8. College or University you attend:
9. State of your legal Residence:
10. Place of Birth:
12. Socio-economic Status:
13. Parents' Occupation: Father:
14. Political Affiliation or Preference:
15. Marital Status: Single Married Divorced Separated Widowed
Now turn the page and you will find an assessment scale for one of the
attributes included in our study. Please answer the questions as best
you can. Sometime in the future you will be scheduled to meet with one
of our interviewers who will inform you of your placement on the scale
Please make sure your phone number is near your name on the from page so
that you can be contacted for the individual session.
Thank-you for your cooperation in this project.
Read the paragraphs below and answer the questions as best you can.
1. Advertising has changed remarkably in recent years. Its methods used
to be quite simple and crude as compared with the highly organized
departments and carefully planned advertising campaigns which fea-
ture modern business. For example, I remember the leading merchant
in a little town out West. His store stood on the town's busiest
corner, where his 3 sons helped him sell drygoods. Every Saturday
noon in winter he used to get up on the roof of his store and throw
a new overcoat into the street. The crowd scrambled for it. That
was advertising in those days. Incidentally, one of his sons was
supposed to be quick enough to grab the coat and rush it back to the
shelves to be thrown down again the next Saturday. It was a bad day
for the boys if they let an outsider get away with that decoy coat!
I. The western merchant is portrayed as (1) opposed to advertising,
(2) a miserly money-grabber, (3) a typical advertiser of his day, (4) a
man with modern business ideas, (5) a harsh father
II. Indicate whether these statements are True (+), False (-), or
1. Advertising methods have changed considerably.
2. Modern advertising is highly organized.
3. The western merchant used modern advertising methods.
4. The merchant gave away a coat each Saturday in the year.
5. Giving away merchandise is not good advertising.
6. One of the merchant's sons tried to recover the coat.
2. "I have considered the structure of all volant animals, and find the
folding continuity of the bat's wings the most easily accommodated
to the human form. Upon this model I shall begin my task tomorrow,
and in a year expect to tower into the air beyond the malice or pur-
suit of man. But I will work only on this condition, that the art
shall not be divulged, and that you shall not require me to make
wings for any but ourselves."
"Why," said Rasselas, "should you envy others so great an
advantage? All skill ought to be exerted for the universal good;
every man has owed much to others, and ought to repay the kindness
he has received."
"If men were all virtuous," returned the artist, "I should with
great alacrity teach them all to fly. But what would be the security
of the good, if the bad could at pleasure invade them from the sky?
Against an army sailing through the clouds neither wall nor mountains
nor seas could afford any security. A flight of northern savages
might hover in the wind, and light at once with irresistible violence
upon the capital of a fruitful region that was rolling under them."
1. The author of the passage expresses a philosophy which: (a) is
currently applicable, (b) expresses optimism, (c) was valid only in his
day, (d) reminds one of a military strategist, (e) is similar to existen-
2. In this selection the author is using the literary device of (a) ono-
matopoeia, (b) flashback, (c) symbolism, (d) alliteration, (e) irony.
3. Worldwide peace, according to the passage, could come about by (a)
arming for defense, (b) eliminating evil tendencies, (c) resorting to
strategem, (d) establishing firm controls, (e) letting intellectuals
4. The attitude of the person giving his point of view is one of:
(a) optimism, (b) sprightliness, (c) distrust, (d) innocence, (e) sar-
5. The person whom Rasselas is speaking is: (a) a tailor, (b) a gam-
bler, (c) a bat, (d) an artist, (e) a biologist.
6. The word "volant", according to the context, means: (a) crawling,
(b) flying, (c) violent, (d) carnivorous, (e) ferocious.
Ability to Think Clearly
Read the items carefully and answer as best you can.
1. Given: Either the queen will not punish the prince or if war is not
inevitable, then the prince's lover will not be jealous; but if the
jealousy of the prince's lover implies that war is inevitable, the
king will be angry and if the queen's punishing the prince implies
that the king is angry, then the king is bald, but he's not bald.
(a) Will the queen punish the prince?
(b) Will the king be angry?
2. Given: Albert, Bill, Charlie and David are storekeepers and that
either Albert of Bill must go shopping every day, but never on the
same day, and that when Albert stays home, Charlie always goes
shopping, and when Bill stays home, David always goes shopping.
(a) Can David stay home when Charlie stays home?
(b) When David stays home, does Charlie go shopping?
(c) If Bill goes shopping, can Charlie stay home?
(d) If David does not go shopping, can Albert still go shopping?
3. Given that "Countries receiving U.S. aid are never poor" is true,
what do you know about the following? (True = T, False = F, Un-
related = U). Circle one.
a. Countries failing to receive U.S. aid are never rich. T F U
b. Few countries fail to receive U.S. aid. T F U
c. No rich country receives U.S. aid. T F U
d. Few countries receiving U.S. aid are rich. T F U
e. Countries failing to receive U.S. aid are always rich. T F U
f. No poor country fails to receive U.S. aid. T F U
g. Every rich country receives U.S. aid. T F U
h. A few poor countries receive U.S. aid. T F U
In each of the following sets of propositions, assume A is true; what then
do you know about B?
4. A. Few hedgehogs read the New York Times.
B. A few hedgehogs read the New York Times. T F U
5. A. Every Dodo loves to run a caucus-race.
B. Few Dodos love to run a caucus-race. T F U
6. A. All that glitters is not gold.
B. Some glittering objects are gold. T F U
7. A. All that sparkles is not a diamond.
B. Only diamonds sparkle. T F U
Given the first proposition as true, what can you determine about each
of the following?
7. At least certain underdeveloped nations will eventually become de-
a. At least a few underdeveloped nations will not become
developed. T F U
b. Underdeveloped nations never will become developed. T F U
c. Eventually all nations will become developed. T F U
8. All governments are not neutral.
a. Few governments are neutral. T F U
b. Every government is neutral. T F U
c. No government is natural. T F U
Are the following arguments deductive or inductive?
10. The Republicans are going to win the November election, Mr. Gallup
reported. Of those interviewed who were not undecided 59% said
they plan to vote Republican.
11. Everybody will agree that swimming is the most healthful sport, for
it exercises more muscles of one's body than any other sport and
this is of course more healthful.
12. The Yankees have the odds on the pennant again this year, for they've
won it more than any other club.
13. It's going to rain. Listen to the thunder! Every time I've heard
it thunder like that it rained.
Answer the following questions by picking what you think is the best
1. For some time now in a certain family a child about 11 years old has
been developing poor work habits in regard to school work. The
child leaves homework until the last minute and often does not get
it done or does it carelessly. What should the parents do?
a. Ignore it; do nothing about it.
b. Help the child with the homework.
c. Punish the child if the homework is not completed.
d. Explain the importance of developing good work habits.
e. Set a definite time each day for the child to do homework.
f. Check each day to make sure the child has done the work.
2. If your oven caught on fire what would you do?
a. Close the oven door and wait for the fire to burn out.
b. Pour baking soda on the fire.
c. Call the fire department.
d. Use a fire extinguisher.
3. If alternative A had a 20% probability of giving you a payoff of
$100, alternative B had a 50% probability of giving you a $40 pay-
off, alternative C had a 80% probability of giving you a $25 payoff
and alternative D had a 70% probability of giving you a $30 payoff,
which alternative would you choose?
a. A d. D
b. B e. None of them.
4. When you face an approach-approach conflict what do you do?
a. Pick the alternative that has the surest benefits.
b. List the pros and cons of each alternative.
c. Flip a coin to choose one alternative.
d. Wait for more information.
e. Pick the alternative that has the longterm payoff.
f. Pick the alternative that is best at the moment.
5. A person nears their home and sees that the front door is ajar and
the front room seems to be in disarray. The nearest neighbor is
mile away. What should the person do?
a. Carefully enter the house and get to the phone to call someone.
b. Determine whether anyone is in the house.
c. Run to the neighbor's and phone the police.
d. Call out, "Is anyone there?".
e. Sneak behind the bushes and listen for 5 minutes.
6. For some time now in a certain family a child about 11 years old has
been stealing things from a local department store. No one else
knows about it but the parents. What should they do?
a. Explain the results of crime and point out that "crime doesn't
b. Deprive the child of TV and movies for a month.
c. Threaten the child with telling the police if it happens again.
d. Ignore it; do nothing about the situation.
e. Spend more time with the child and do more things together.
f. Make the child take back the stolen things and pay for what cannot
7. You are having a hard time deciding whether to go to college A or
college B. What would you do?
a. Pick the one where you would get the best education.
b. Pick the one where you would be happiest.
c. Flip a coin to choose one alternative.
d. Go where your parents want you to go.
8. If a friend wanted advice on an interpersonal problem, what would
a. Listen carefully, but make no declarative statement.
b. Tell what you honestly think about the situation.
c. Say that you would rather not be involved.
d. Listen carefully and say what you think the person wants to hear.
e. Listen carefully and help the person sort out his/her thoughts.
f. Try to change the subject.
9. What do you think is the most important principle in decision-
a. Act as fast as the situation warrants.
b. Decide, and don't look back.
c. Act rationally.
d. Imagine all possible outcomes and choose the one that seems right.
e. Consider all the pros and cons.
f. If it feels right, do it.
10. In cohesive groups, which are good ways to insure effective thinking
a. Playing the devil's advocate.
b. Listing the pros and cons.
c. Taking the majority decision.
d. Critically evaluating all ideas.
e. Acting rationally.
f. Consulting outside advisors.
11. A person has been invited to go somewhere with a person they are not
very fond of and they accept. Later a person that they really would
like to know asks them somewhere for the same time. What should
a. Decline the second offer.
b. Accept the second offer and tell the first person they are sick.
c. Accept the second offer and explain the situation to the first
d. Decline the second offer, break the first date and stay home.
e. Decline the second offer and invite the person for another time.
12. As you are driving along the deserted road you come upon a car that
has hit a tree. The nearest phone is 5 miles away. What should
a. Stop the car, get out and try to help any injured persons.
b. Drive ahead to the phone and call the police.
c. Put any injured persons in your car and get them to a doctor.
d. Get out to help, and set up a signal to any cars that might pass.
13. You are a senior in college and have been offered a job that would
provide ample salary and benefits but you really want to study
Egyptian history in graduate school. What should you do?
a. Take the job and put off graduate school until later.
b. Do what your heart desires and go to graduate school.
c. Take the job and concentrate on advancement rather then a useless
d. Tell your employer you want to leave the job in two years in
order to get the graduate degree.
e. Wait to see if you get graduate funding.
Read each item carefully and put the number next to it that expresses
your feeling about the statement. Be sure to answer each item.
1 = Strongly Disagree
2 = Disagree
3 = Undecided
4 = Agree
5 = Strongly Agree
1. A person can learn more by working four years than by going
to high school.
2. The more education a person has the better he or she is able
3. Education is no help in getting a job today.
4. A high school education is worth all the time and effort it
5. Our schools encourage an individual to think for himself or
6. Solution of the world's problems will come through education.
7. Savings spent on education are wisely invested.
8. Parents should not be compelled to send their children to
9. Education is more valuable than most people think.
S10. Education has failed unless it has helped people to understand
and express their own feelings and experiences.
11. How students feel about what they learn is as important as
what they learn.
12. The goals of education should be dictated by children's needs
and interests, as well as by the larger demands of society.
13. Teachers should have academic freedom---freedom to teach what
they think is right and best.
14. Learning is experimental; the child should be taught to test
alternatives before accepting any of them.
15. The true value of education is so arranging learning that
children gradually build up a store house of knowledge that
they can use in the future.
16. Education and educational institutions must be sources of new
social ideas; education must be a social program undergoing
17. Certain facts and knowledge are necessary for the study of
all subjects and these facts do not change very much.
18. Pupils should study only subjects that they feel they will
need in later life.
19. All children should be encouraged to aim at the highest aca-
20. A person's experiences have more to do with his or her learning
ability than does innate intelligence.
21. An activity, to be educationally valuable, should train
reasoning and memory in general.
22. It is more important for children to have faith in themselves
than for them to be obedient.
Attitude toward Self-reliance
This scale contains a list of items such as this:
I think children should be able to brush their teeth daily without
being told to do so by the age of
We want you to read an item, then think of children who are neigher high
nor low in intelligence, physical development, etc., but near the average.
Then mark in the blank the age at which anyone of these "near-average"
children should be able to perform this task.
For example, if you think that the average child should be able to brush
his or her teeth daily without being told to do so by the age of five
years, place the figure five in the blank, like this:
I think children should be able to brush their teeth daily without
being told to do so by the age of 5
In some cases you may wish to put down the age in months rather than
years. In such cases write "mos." after the figure, like this: 2 mos.
We want your own opinion on each item.
1. I believe that children can be taught to make a good selection of
their own meal when dining in a hotel or restaurant by the age of
2. I believe that children should be able to wash and dry the dishes
after a family meal by the age of
3. I feel that children should be able to prepare a simple family
dinner without assistance after it has been planned by a parent by
the age of
4. (Disregard legal age when marking this item.)
I feel that children should be able to drive the family car when
adults are in the car by the age of
5. I think that children should be able to take responsibility for
keeping at least one room clean throughout the week (assuming they
have help in one regular weekly cleaning) by the age of
6. I think that children can be left in charge of the household with
younger members of the family (not more than three children and none
younger than three years and assuming at least one parent is home at
night) by the age of
7. I believe that children should be taught to buy their own clothing
without help from an adult by the age of
8. I think that children should be taught how to obtain their own
answers to their questions such as "How is silk made?" and "Where
does rubber come from?" by the age of
9. I believe that children should be able to take the entire respon-
sibility for their homework (i.e., complete it without parental
reminder or supervision) by the age of
10. I think children are capable of taking entire care of their hair
(i.e., wash, comb, decide how it is to be worn, when it is to be
cut, etc.) by the age of
11. I feel that children should be able to make their own bed daily
(excepting when sheets are changed) without help by the age of
12. I feel that children should be able to plan and prepare a simple
family dinner (e.g., meat, potatoes, vegetable and dessert) by the
13. I think that children should be included in some family discussions
(e.g. on expenditures of money), their opinions being considered
along with the opinions of older members of the family, by the
14. I think children should be able to choose the movies they wish to
attend by the age of
15. I think children should be able to partake in extra-curricular
activities, with some parental guidance in the choice of activities
by the age of
16. I think that children should be able to choose their own reading
material, including books, magazines and newspapers, by the age of
17. I think a child should be allowed to leave the parental home for a
period of time (e.g. to a six weeks camp) by the age of
18. I think that children should be able to make desirable choice from
their wardrobe of clothing to be worn, without suggestions from
an adult, by the age of
19. I think that children should be able to plan and give a simple
afternoon party for immediate friends by the age of
20. I think that a child should be able to prepare and serve lunch to
the younger members of the family when the parents are away for
lunch by the age of
21. I think that children should be able to take the responsibility
for paying one monthly bill for the family (e.g. obtain the money
from the parents, pay the bill of the date it is due or before,
get receipt for payment--this refers to bills such as light, gas
or water bills) by the age of
22. I think that children should be able to make their own arrangements
about going to an evening party, when it means taking a street car
or bus a distance of three miles or less, by the age of (this
refers to a party which begins at seven o'clock and ends at ten
o'clock and includes making decisions as to what time to go, whom
to go with, how to meet companions, etc.)
Attitudes toward Law and Justice
Mark the statements which you agree with in the first column, those you
disagree with in the second column and those you have no opinion about
in the third column.
Agree Disagree No opinion
1. On the whole judges are honest.
2. Police often carry a grudge against
people who get in trouble with the law
and treat them cruelly.
3. For the most part, justice gets done by
the police and the courts.
4. Many of the people in prisons are actually
innocent of the crimes they were con-
5. Most policement are honest.
6. Any jury can be fixed and most of them
7. We would have less crime if our laws were
8. The big-time crooks never get arrested in
this country. It's just the little guy
that gets caught.
9. Most judges are honest and kind-hearted.
10. Juries seldom understand a case well
enough to make a really just decision.
11. The sentences of judges in court are
determined by their prejudices.
12. Court decisions are almost always just.
13. Justice does not have cross-cultural
14. A hungry person has the right to steal.
15. It is all right to break the law if he
or she does not get caught.
16. Justice is nothing more than certain
acts, beliefs and attitudes of the major-
ity of individuals in their daily rela-
tions with others.
17. Law is essential to the enjoyment by
citizens of their inalienable rights.
18. Justice is something over and above
19. Justice is a rarely achieved goal.
20. Law is to secure justice and order among
21. There is no absolute standard of justice.
22. Justice is the maintenance and adminis-
tration of the norms of the people.
1. What is the traditional bridal flower?
a. Orange blossoms c. White calla lily
b. White gladioli d. White rose
2. Generally, when is the best time to arrage flowers?
a. Several hours before the function
b. Immediately before the function
c. Anytime before the function
d. Day before the function
3. Which is true about water in flower containers?
a. Completely change water daily
b. Add cold water every day
c. Room temperature is best
d. Refill water only in the evening
4. What is the general principle for proportion of arrangement?
a. Flowers 1- times height of upright container
b. Depends on texture of flowers
c. Height of flowers should be that of the vase
d. There is no general principle
5. Which of the following influence visual balance?
a. Color intensity c. Distance of focus from center
b. Texture d. Number of colors in arrangement
5. Where is the focal point of an arrangement?
a. Where the brightest color is
b. Two-thirds of the way from the top
c. In the middle of the design
d. Where the arranger prefers
7. In a miniature, what determines the height of the container?
a. Length of the stems c. Diameter of the flowers
b. Size of the flowers d. None of the above
8. What is the effect of candlelight?
a. Makes green more yellow c. Darkens blue
b. Lightens blue d. Deepens purple
9. What are the major principles of flower arranging?
a. composition, unity, focus, balance
b. balance, focus, harmony
c. shape, focus, balance
d. unity, composition, harmony
10. What are the types of color harmony?
a. monochromatic, polychromatic, complementary
b. monochromatic, analogous, complementary
c. hue, tint, pastel
d. chromatic, semi-chromatic, analogoug
11. Which is most true about table arrangements?
a. Include no fruit as guests may eat them
b. Make low enough so as not to obstruct conversation
c. High arrangements are permissible if there is "window" space
d. Should be equally interesting from all sides
12. What is most important in a dried arrangement?
a. Keep away from sun c. Place it high in the room
b. Keep away from draught d. Have it fairly low in height
13. Which are ways to keep flowers longer?
a. Two aspirin per quart water
b. Pounding stems
c. Cut stems straight across
d. Squeeze out stem juice
14. When is the best time to "condition" flowers?
a. Anytime before placement c. At least 24 hours before placement
b. Night before placement d. At night
15. For flower arranging, which are the forms of color?
a. Pure color and shade c. Pure color, white and hue
b. Hue and pure color d. Pure color, white, hue and tone
16. What does "miniature" refer to?
a. Using small flowers c. Representation on reduced scale
b. Bonzai d. Specially grown flowers
Cross-word Puzzle Ability
Given below are items from cross-word puzzles in daily newspapers. You
are to enter in the space provided the best answer you can think of.
The number of blanks indicates the number of letters in each answer.
Weight of India
Ems or Evian
"The strife is "
Nerve of vision
Rorschach, for one
Kind of hygiene or surgery
Part of Mao's name
Alliance of West
Gary of golf
Range of central U.S.
1. Which wood is most commonly used for cupboards?
a. Hardwoods c. Hardwoods and some softwoods
b. Pine and maple d. Softwoods
2. Which is true of plywood?
a. It is three layers glued together
b. It comes in 3, 5, or 7 ply
c. It is between " and 1" thick
d. Different types of wood are glued cross-grain
3. Which wood most easily dents?
a. Pine c. Cherry
b. Redwood d. Maple
4. What is the thickness of lumber for cupboards?
a. 1" c. 3/4" or 13/16"
b. " to 1" d. "
5. Which can be used to shape wood?
a. Lathe c. Drill-press
b. Shaper-press d. English shaper
6. What are the kinds of stress at a joint?
a. direct, indirect
b. shear, compression, tension
c. tension, shear, rabbet
d. full, tension, shear
7. Which are standard glued joints?
a. Cross-lapped c. Dovetail
b. Mortise and tenon d. Mitre
8. What are drawer bottoms made of?
a. i" plywood c. " any wood
b. " pine d. 3/8" pine
9. What is true of the order of assembly of a drawer?
a. Front and one side first c. Back and one side first
b. Chamfer first d. Bottom is last
10. Which is true of sanding?
a. Belt sander substitutes for hand
b. Belt sander runs in both directions
c. Belt sander not appropriate
d. Finish with hand sanding
11. What are the types of glue?
a. Hide, plastic resin, flake c. Plastic resin, hide, white
b. Plastic resin, white d. Resin, celage, white
12. Which is the final grade of sandpaper used?
a. 3/0 c. Smooth
b. 6/0 d. 5/0
13. Which is the most difficult to apply?
a. Varnish c. French polish
b. Lacquer d. Lorton finish
14. Which joint is used for drawers?
a. Glue-joint c. Dowel pin
b. Mortise and tenon d. Dovetail
15. Why are dowel joints grooved spirally?
a. To fit tighter c. They are not grooved spirally
b. For escape of air d. For better resistance
16. What is a dado?
a. Ornamental piece
b. Has nothing to do with carpentry
d. Type of finial
17. What is the most important structural detail in a free-sliding
a. Rigidity c. Ease of slide
b. Balance d. Angle of dovetail joints
18. What is a chamfer?
a. A saw cut c. A machine to cut lumber
b. A type of finial d. A 45 degree cut removing a corner