Lay Dispositionism and Implicit Theories of Personality

Transcription

Lay Dispositionism and Implicit Theories of Personality
Copyright 1997 by the American Psychological Association, Inc.
0022-3514/97/S3.00
Journal of Personality and Social Psychology
1997, Vol.73, No. I, 19-30
Lay Dispositionism and Implicit Theories of Personality
Chi-yue Chiu
Ying-yi Hong
University of Hong Kong
Hong Kong University of Science and Technology
Carol S. Dweck
Columbia University
Lay dispositionism refers to lay people's tendency to use traits as the basic unit of analysis in social
perception (L. Ross & R. E. Nisbett, 1991). Five studies explored the relation between the practices
indicative of lay dispositionism and people's implicit theories about the nature of personal attributes.
As predicted, compared with those who believed that personal attributes are malleable {incremental
theorists), those who believed in fixed traits (entity theorists) used traits or trait-relevant information
to make stronger future behavioral predictions (Studies 1 and 2) and made stronger trait inferences
from behavior (Study 3). Moreover, the relation between implicit theories and lay dispositionism
was found in both the United States (a more individualistic culture) and Hong Kong (a more
collectivistic culture), suggesting this relation to be generau'zable across cultures (Study 4). Finally,
an experiment in which implicit theories were manipulated provided preliminary evidence for the
possible causal role of implicit theories in lay dispositionism (Study 5),
Lay dispositionism refers to the use of traits as the unit of
analysis in social perception (L. Ross & Nisbett, 1991). The
inferential practices that have been shown to be associated with
lay dispositionism include (a) viewing behavior as reflecting
its correspondent disposition (Gilbert & Jones, 1986; Jones,
1990; Jones & Davis, 1965; Nisbett & Ross, 1980), (b) the
belief that behavior in a particular situation can be predicted
from knowledge of the relevant trait (Kunda & Nisbett, 1986;
L. Ross & Nisbett, 1991), and (c) the expectancy that behavior
will be consistent across situations (Kunda & Nisbett, 1986; cf.
Wright & Mischel, 1987).
Some writers (e.g., Jellison & Green, 1981) believe that in
the United States, most individuals are brought up to assume
that behaviors reflect personal dispositions. Indeed, the fact that
people may spontaneously infer traits from behavior has led
some social psychologists to believe that dispositional inference
is an automatic outgrowth of perceptual experience (Baron &
Misovich, 1993). However, research has indicated that the likelihood and the meaning of dispositional inferences may vary with
the perceiver's inferential goals (Krull, 1993; Krull & Dill,
1996; Lee, Hallahan, & Herzog, 1996), cultural experiences
(Cousins, 1989; Lee et al., 1996; Miller, 1984), and the perceiver's construal of the person-situation relation (Shoda &
Mischel, 1993). Such situational variability and individual differences argue against the view that trait inference is an inevitable consequence of perceptual experience. In the present research we sought to further understand the psychological processes that may underlie individual differences in lay
dispositionism, focusing on the perceiver's conceptions of the
nature of personal attributes.
Recently, there has been increasing recognition among psychologists that an individual's judgments are influenced by the
implicit theories the individual has about the self and other
people (Dweck, Chiu, & Hong, 1995a; Dweck & Leggett, 1988;
M. Ross, 1989; see also Heider, 1958; Kelly, 1955). These theories are referred to as implicit theories because unlike most
scientific theories, these theories are rarely explicitly articulated.
Nonetheless, they may set up an interpretive frame within which
information is processed.
In this article, we propose that lay people have implicit theories about human nature and that the kind of implicit theory
a person holds is related to the tendency to subscribe to lay
dispositionism. More specifically, we predicted that the tendency
to use traits as the basic unit of analysis would be greater among
those who believe that an individual's personality consists of
fixed, static traits {entity theorists) than among those who believe that it consists of dynamic personal qualities that can be
changed and developed (incremental theorists).
Chi-yue Chiu, Department of Psychology, University of Hong Kong;
Ying-yi Hong, Department of Psychology, Hong Kong University of
Science and Technology; Carol S. Dweck, Department of Psychology,
Columbia University.
Preparation of this article and the research reported in it were supported in part by University of Hong Kong Research Grants 337/014/
0010 and 337/014/0015.
We thank Nino Sidari and Catherine Parker for their assistance in
data collection. We also thank Sheri Levy for the information on the
expanded measure.
Correspondence concerning this article should be addressed to Chiyue Chiu, Department of Psychology, University of Hong Kong, Hong
Kong or to Carol S. Dweck, Department of Psychology, Columbia University, New "Vbrk, New \brk 10027. Electronic mail may be sent via
the Internet to [email protected],hk or to [email protected].
The idea that perceivers' implicit theories may guide the way
they make social inferences was first articulated by Jones and
Thibaut (1958), who proposed that perception of a target' 'will
tend to be imposed on the stimulus person by the perceiver; the
missing link in the inference chain will be supplied by the
19
20
CHIU, HONG, AND DWECK
perceiver's own 'theory' of personality" (p. 166). In the present
research we focused on how implicit theories about some structural properties (fixedness or malleability) of personal attributes
may provide an epistemic context for trait inferences and the
meaning of traits.
As an overview, the present research consisted of a series of
five studies designed lo test the predicted relation of implicit
theories and lay dispositionism. First, the proposed relation between implicit theories and the inferential practices indicative
of lay dispositionism was systematically investigated in Studies
1, 2, and 3. In Study 4, we attempted to establish the crosscultural generalizability of this relation. Finally, in Study 5,
participants' implicit theories were manipulated to explore possible causal relations between implicit theories and lay
dispositionism.
Background: Implicit Theories and Self-Judgments
Previous research by Dweck and her associates (see Dweck &
Leggett, 1988; Hong, Chiu, & Dweck, 1995) has shown that
individuals' beliefs about the malleability of self-attributes predict judgments of the self (see also Robins & Pals, 1996). For
example, in the domain of academic achievement, those who
believe their intelligence to be a fixed entity (entity theorists)
tend to focus on judging their level of intelligence. That is, they
are concerned with how much of this fixed commodity they
possess. They also tend to see academic failures as indications of
their intellectual ability, ascribing lo themselves global, stable,
negative ability traits on the basis of a limited number of failure
experiences. By contrast, those who view their intelligence as
a malleable trait (incremental theorists) tend to place the emphasis on developing versus judging their ability. In the face of
obstacles, they tend to focus on the factors that mediate improved performance and increased ability (e.g., effort, problemsolving strategies) rather than on self-judgment.
This relation appears to hold in the domain of social interaction as well (Erdley, Cain, Loomis, Dumas-Hines, & Dweck,
1997; see also Benenson & Dweck, 1986; Chiu, Hong, &
Dweck, 1994; Goetz & Dweck, 1980). Holding an entity versus
incremental theory of one's personality predicts a focus on selfjudgment and a tendency to attribute global, stable negative
traits to oneself in the face of rejection. In short, a belief in
fixed self-attributes appears to be associated wilh a greater tendency to process information in terms of traits and trait evaluation. By contrast, a belief in malleable self-attributes is associated with a greater tendency to process information in terms of
process variables that mediate performance.
ality is viewed as a set of fixed traits, then understanding their
personality may involve assessing their fixed underlying trails.
Thus, entity theorists may view the task of person perception
as being to judge or diagnose these underlying traits from the
available behavioral information. Moreover, because these (raits
are believed to be fixed, a diagnosis made at one time should
be predictive of behavior at a later time.
In contrast, incremental theorists, who believe in more dynamic, malleable qualities that can vary systematically over time
and perhaps across contexts, may view the task of person perception as being to understand the dynamics of behavior, rather
than simply to judge dispositional traits. In doing so, they may
be more likely to take account of the actor's states (e.g., goals,
intentions) as well as relevant aspects of the situation (Hong,
1994). If this is true, then they should display weaker trait
inferences and be more sensitive to the effects of the situation
on behavior.
Past research has provided support for the relation between
entity versus incremental theory and one of the inferential practices associated with lay dispositionism; namely, the tendency
to draw global trait inferences from behaviors even when there
is a plausible situational explanation for the behaviors. Specifically, Erdley and Dweck (1993, Experiment 2) assessed late
grade school children's entity versus incremental theory of personality and later showed them a narrated slide show of a ' "new
boy in school," who, in an attempt to make a good impression
and to hide his ignorance, performed a variety of transgressions
(e.g., trying to cheat on some work, appropriating another
child's leftover materials). None of these transgressions hurt
others, and the situational pressures were emphasized throughout. After the slide show, children were asked to make a number
of judgments about the boy. The results revealed that entity
theorists were significantly more likely to make global trait inferences ("bad," "mean," and "nasty") from these behaviors
than were incremental theorists.
The five studies reported in this article further explored the
relation between implicit theories and the inferential practices
associated with lay dispositionism. Studies 1, 2, and 3 were
designed to provide further evidence for the link between implicit theories and judgments of others by addressing the inferential links that characterize the lay oppositionist model of behavior depicted in Figure 1. As shown in the figure, in the lay
dispositionist model, behaviors across situations are assumed to
be mediated by enduring dispositional traits. According to L.
Ross and Nisbett (1991), if perceivers believe that behaviors
across situations are manifestations of the same underlying trait.
Implicit Theories and Social Perception
This research sought to extend previous work on implicit
theories and judgment, focusing on the relation between implicit
theories and the inferential practices associated with lay dispositionism. Tn the domain of social perception, research by Dweck
and her associates (Erdley & Dweck, 1993; see also Dweck,
Hong, & Chiu, 1993) suggests that just as a fixed view of the
self is associated with a greater emphasis on evaluating selfattributes, a fixed view of others may be associated with a
greater emphasis on judging others' traits. When others' person-
implies
(Study 3),
Dispositional
Trait
Behavior in
One Situation
Figure I.
predicts
(Study 1)
predicts
JStudy 2)
Behavior in a
New Situation
A lay dispositionist model of behavior.
21
LAY DTSPOSITIONISM
perceivers will expect a relatively high level of cross-situational
consistency in these behaviors. For example, knowing that Jack
is more aggressive than John in one situation, a lay dispositionist
would predict Jack to be more aggressive than John in another
situation as well. Thus, if entity theorists believe more strongly
than incremental theorists in a trait analysis of behavior, they
should expect greater cross-situational consistency in trait-relevant behaviors. This hypothesis was tested in Study 1.
In addition, when presented with information about a person's
trait, a lay dispositionist should be fairly confident that this
information can reliably predict this person's behavior in a new
situation (L. Ross & Nisbett, 1991). For example, if Jack is
known to be more aggressive than John, on average, a lay dispositionist will predict with relatively high confidence that Jack
will act more aggressively than John in any particular situation.
In Study 2 we addressed the relation between subscribing to
an entity versus incremental theory and the tendency to make
confident predictions of specific behavior from knowledge of a
person's traits. If entity theorists ascribe a more important role
to traits than do incremental theorists, they should also be more
confident in making these predictions.
Study 3 focused on making confident judgments of aperson's
traits on the basis of knowledge of the person's behavior (e.g.,
if Jack hits John in a particular situation, to what extent can it
be concluded with confidence that Jack is an aggressive person?). As Figure 1 illustrates, when presented with a person's
behavior, a lay dispositionist should be quite confident that this
behavior is indicative of the person's traits (L. Ross & Nisbett,
1991). Accordingly, in this study, we predicted that relative to
incremental theorists, entity theorists would be more ready to
make trait inferences from behaviors.
Study 4 examined the cross-cultural generality of the relation
between implicit theories and lay dispositionism. In relatively
individualistic cultures (e.g., American culture), people are socialized to identify themselves as individuals who can leave
groups at will and are expected to behave according to personal
motivation. Thus, these cultures may focus more on the individual as the origin or agent of social actions (Miller, 1984; Morris & Peng, 1994). By contrast, in relatively collectivistic cultures (e.g., Indian and Chinese cultures), people are expected
to identify themselves as group members and behave according
to such things as role expectations and group norms (Chiu &
Hong, in press; Ho & Chiu, 1994; Hofstede, 1980). These cultures may thus focus on role expectations, group norms, or other
contextual constraints as the impetus for social action (Morris &
Peng, 1994). Will the relation between entity versus incremental
theory and lay dispositionism hold in individualistic cultures
only, or will it hold in both individualistic and collectivistic
cultures? In other words, are entity and incremental theories
indigenous cultural constructions developed to give meaning to
social behavior in individualistic cultures, in which the emphasis
is on personal agency (Miller, 1984; Shweder & Sullivan,
1990)? Or are they more general worldviews that individuals
in all cultures tend to develop to make sense of the stability and
variability in humatfactions (Dweck, Chiu, & Hong, 1995b)?
Study 4 was designed to address this issue.
Finally, to establish the causal relation between implicit theories and the inferential practices associated with lay dispositionism, in Study 5, we manipulated participants' implicit theo-
ries by asking them to read a (fictitious) Psychology Todaytype article that presented persuasive evidence for an entity or
incremental theory. We predicted that, compared with those who
read an incremental theory article, participants who read an
entity theory article would be led to make stronger global trait
inferences from behaviors.
In summary, a number of phenomena in person perception
are associated with lay dispositionism. The present research
tested the hypothesis that the tendency to exhibit some of these
phenomena is linked to a specific theory of personality. Specifically, having a static, entity theory was predicted to orient an
individual toward using traits as the unit of analysis and toward
subscribing to the inferential practices associated with lay dispositionism. In contrast, holding a dynamic, incremental view was
expected to attenuate the strength of lay dispositionism and
make its allied inferential practices less likely. In addition, the
cross-cultural generalizability and causal nature of these relations were explored. The idea that implicit theories may provide
an interpretive framework for social inferences dates back to
Jones and Thibaut (1958). In line with this research tradition,
in the present research we further explored the role of perceiverinduced constraints in social perception by placing the perceiver's implicit theories of personality at the heart of person
perception.
Study I
One important index of lay dispositionism is the expectancy
that trait-relevant behaviors will be consistent across situations
(the behavior-to-behavior link depicted in Figure 1). As L. Ross
and Nisbett (1991) posited, if perceivers believe that behaviors
are mediated by traits, they will expect a person who possesses
a particular trait to act in a manner consistent with this trait
across even very different situational settings. Thus, a lay dispositionist is inclined to make confident predictions of a person's
behavior in new situations on the basis of knowledge of this
person's trait-relevant behavior in a single prior situation.' In
this study, consistent with the view that entity theorists are more
likely to be lay dispositionists, we predicted that relative to
incremental theorists, entity theorists would make more confident predictions of a person's behavior on the basis of knowledge of that person's behavior on one prior occasion.
Method
Participants
The participants were 40 undergraduates (11 men and 29 women) at
Columbia University who participated in the study to obtain course
1
In this study, we were interested more in entity versus incremental
theorists' relative tendency to make confident predictions about crosssituational behavioral consistency than in how accurate their predictions
were. Note, however, that in Kunda and Nisbett (1986), a strong expectancy of cross-si tuational consistency was regarded as a prediction error,
given the consistent finding in personality research that the correlation
of trait-relevant behavior across different situations is typically around
.10 and seldom exceeds .20 (see Mischel, 1968;Mischel & Peake, 1982;
Peterson, 1968).
22
CHIU, HONG, AND DWECK
requirement credit. Because no gender effects were found in this or any
of the other studies, gender is not included in the analyses reported.
Measures
Measure of implicit person theories. Although implicit theories can
be domain specific (e.g., a person can have an entity theory about moral
character and an incremental theory of intelligence; see Dweck et al.,
1995a), and although it is most desirable to make domain-specific assessments and predictions, the stimulus materials we adapted from
Kunda and Nisbett (1986, see below) cut across the domains of morality,
sociability, and intelligence. Therefore, we used a more domain-general
measure of implicit theories. This implicit person theory measure was
embedded in a battery of other measures designed to assess styles of
coping with negative events and consists of three items: "The kind of
person someone is is something very basic about them and it can't be
changed very much;" "People can do things differently, but the important parts of who they are can't really be changed;'1 and "Everyone
is a certain kind of person and there is not much that can be done to
really change that." Respondents indicated their extent of agreement
with each item on a 6-point scale with responses ranging from 1 {very
strongly agree), 2 (agree), 3 (mostly agree), to 6 (very strongly disagree). Unlike individual-difference measures that tap generalized needs
or cognitive styles, our measure taps one simple, unitary belief. The
items are intended to have the same meaning, and continued repetition
of the same idea becomes bizarre and tedious to the respondents. Thus,
only three items were included. Respondents' implicit person theory
was indexed by their mean score on these three items.
Several issues pertaining to the format of the theory measure should
be addressed. First, items depicting an incremental theory were not
included in this measure because previous studies (Boyum, 1988; Leggett, 1985) and our pilot studies found that, even for respondents who
endorsed items depicting entity theories, there was a strong tendency to
endorse items depicting the opposite, incremental, theory, as well as a
tendency to drift toward incremental choices over items. This suggests
that the incremental items are highly compelling and perhaps more socially desirable as well. However, to ensure that disagreement with the
entity theory statements can be taken to represent agreement with the
incremental theory, in one study (see Dweck et al., 1995b), respondents
were given the theory measure and asked to explain their choices. Those
who disagreed with the entity statements gave clear incremental theory
justifications for their responses. Also, in past studies in which implicit
theories have been experimentally manipulated (Bergen, 1991; Dweck,
Tenney, & Dinces, reported in Dweck & Leggett, 1988), participants
given an incremental theory have behaved in ways that are similar to
those who disagreed with the entity theory statements on our scale (see
also Study 5 in the present article). Further, in the domain of intelligence,
in which the theory measure has the same format, an incremental theory
(i.e., disagreement with the entity theory) has been linked to a whole
network of theoretically related variables, such as goals, attributions,
affect, and behavior (see Dweck, 1991, for a review).
Finally, Levy and Dweck (1996) have just obtained direct evidence,
with an expanded implicit theory measure, that those who disagree with
an entity theory on the present measure are the ones who agree with
the incremental theory on the expanded new measure. The expanded
new measure has avoided the overly compelling nature of past incremental items by presenting a very strong form of the incremental theory
(e.g., "Everyone, no matter who they are, can significantly change their
basic characteristics" and "All people can change even their most basic
qualities" ). On the expanded measure, the correlation between the entity
items (which are basically the same as the ones on the present measure)
and the incremental items was - . 7 3 . with a sample size of 652. In a
sample of 101 participants who took both the present measure and the
expanded measure, with a delay of 1 week or less between the two
measures, the correlation between the present measure and the expanded
measure was .83. Of those who were classified as incremental theorists
on the present measure, 88% were classified as incremental theorists on
the expanded measure and only 9% were classified as entity theorists.
Of those who were classified as entity theorists on the present measure,
90% were classified as entity theorists, and none were classified as
incremental theorists on the new measure.
In short, with the awareness that our present implicit theory measure
does not provide a direct assessment of incremental theory, we have
obtained evidence to show that disagreement with an entity theory (the
belief that attributes cannot be changed) can be taken to represent agreement with an incremental theory (the belief that attributes can be
changed). For these reasons, and for consistency with our past work, in
the present article, we refer to those who disagreed with the entity items
as "incremental theorists," but with appropriate caveats.
Second, in the current format, although endorsement of an entity
theory entails agreement with the items, previous studies have shown
that agreement with these statements does not represent an acquiescence
set. For example, MacGyvers (1992) found that there was no relation
between respondents' endorsement of an entity theory and the tendency
to agree with arbitrary items presented in a forceful and compelling
way. In a similar vein, in a number of studies (e.g., Chiu, Dweck,
Tong, & Fu, in press; see also Study 3 in the present article), we included
in the measure or in the study control items that entailed agreement or
disagreement. Entity theorists never showed a greater tendency to agree
with items that are not theoretically relevant. Finally, Dweck et al.
(1995a, 1995b) found that even though the implicit person theory measure has the same format as measures of implicit theories about nonhuman attributes (e.g., implicit theory about the malleability of the world),
they are statistically independent.
Previous studies (Dweck et al., 1995a) have shown that this scale is
a reliable measure, with a high internal reliability (alphas ranging from
.90 to .96 in our previous studies and from .73 to .90 in the present
research). Test-retest reliability for a 2-week interval was .82. As far
as the construct validity of the measure is concerned, the measure does
not correlate with academic aptitude (verbal and quantitative SAT
' scores) or with standard measures of socially desirable responding (the
Paulhus, 1984, Social Desirability Scale) and self-presentation (the Snyder, 1974, Self-Monitoring Scale). This indicates that the measure is
not confounded with intellectual ability or self-presentation concerns.
Also, it does not correlate with optimism about human nature, the Coopersmith ( 1967) measure of self-esteem, or with the Altemeyer (1981)
Right-Wing Authoritarianism Scale and the Kerlinger (1984) measures
of conservatism and liberalism (see Dweck et a!., 1995a). Thus, it is
not confounded with positivity or negativity about the self and others
or with the respondents' ideological rigidity or political stance. Finally,
recall that the defining dimension of implicit theories and their assessment is whether personal attributes are believed to be fixed or malleable.
The theories reflect one simple belief and do not contain in their definition a processing style component or a motivational component. They are
thus distinct from other process-oriented individual differences variables
such as personal need for structure (Neuberg & New som, 1993), atlributional complexity (Fletcher, Danilovics, Fernandez, Peterson, & Reeder,
1986), and need for cognition (Cacioppo & Petty, 1982). In a justcompleted study by Levy and Dweck (1997) the correlations between
implicit theories and these process-oriented individual differences variables were only between .17 and .24.
In the present study, respondents' ratings on the three items were
averaged to yield an overall theory score. To ensure that only participants
who had a clear theory were included, data from 10 participants (25%)
who scored between 3.0 and 4.0 were excluded from the theory group
comparison analyses but were retained in the correlational analyses, in
which we treated endorsement of entity theory as a continuous variable.
Participants who scored at 3.0 or below (n = 15) were classified as
23
LAY DISPOSITIONISM
entity theorists, and those who scored at 4.0 or above were classified as
incremental theorists (n = 15).
The behavioral prediction measure. The Kunda and Nisbett (1986)
questionnaire, consisting of four questions, was used to assess participants' behavioral predictions. Each question asks for a prediction of a
target's behavior in a particular situation on the basis of knowledge of
the target's behavior in one completely different prior situation. Two of
the four questions ask for behavioral predictions in the social domain
(honesty and friendliness), and two ask for behavioral predictions in
ability domains (academic performance and performance in a basketball
game). An example is,
Suppose you observed Jack and Joe in one particular situation and
found that Jack was more friendly than Joe. What do you suppose
is the probability that in a completely different situation, you would
also find Jack to be more friendly than Joe?
Respondents are asked to indicate their prediction on a probability scale
from 0.00 to 1.00. Clear instructions were included to ensure that the
participants understood how to use the probability scale.
The theory measure was given to the participants together with a
battery of measures designed for another study. Then the Kunda and
Nisbett (1986) questionnaire was presented to the participants as an
unrelated study.
Results and Discussion
Participants' predictions within each domain (social traits vs.
ability) were averaged. Agreement with entity theory (treated
as a continuous variable) correlated positively with cross-situational consistency predictions in the social domain (r — .61, p
< .001), in the ability domains (r = .45, p = .01), and in the
two domains combined (r — .49, p < .01). We also subjected
the average predictions in each domain to a 2 X 2 (Implicit
Theory X Domain) analysis of variance (ANOVA). As expected, the implicit theory main effect was significant, F( 1, 28)
= 12.96, p - .001. Entity theorists made stronger predictions
than incremental theorists both in the social domain (M ~ .57
v s . M = .42),?(28) = 4.03,/?< .001, and in the ability domains
(M = .55 vs. M = .44), r(28) = 2.70, p = .01. Interestingly,
the overall prediction of incremental theorists was significantly
below .50 (95% confidence interval [CI] = .4275 ±.0687). This
indicates that incremental theorists, for whom an individual's
personality does not consist of fixed traits, did not believe that
exhibiting more trait-relevant behavior than another person in
one situation meant that the individual would exhibit that kind
of behavior more than the other person in another situation.
Indeed, for them, if a person was found to be friendlier than
another person in one situation, this relation would more likely
than not be reversed in a different situation. In contrast, entity
theorists' overall prediction was significantly greater than .50
(95% CI = .5583 ± .0348). For them, if one person was found
to be friendlier than another person in a particular situation, this
relation would more likely than not generalize to a completely
different situation.
It thus appears that entity theorists expect trait-relevant behavior to be consistent across situations. It also appears that
incremental theorists, when given no information other than the
relative occurrence of a certain kind of behavior in a single
situation, tend not to predict the same relative likelihood of
displaying that kind of behavior in a different situation. On the
contrary, lacking other information, they may expect this relative
likelihood to "even out" across situations (cf. Tversky & Kahneman, 1982).
Study 2
Another inferential practice associated with lay dispositionism is the expectation that knowledge of a person's traits is
a highly reliable basis for making behavioral predictions (the
trait-to-behavior link in Figure I). 2 In this study, we thus predicted that relative to incremental theorists, entity theorists
would make stronger predictions of a target's behavior in a
particular situation on the basis of information regarding the
target's traits.
Method
Participants
The participants were 50 undergraduates (18 men and 32 women) at
the State University of New %rk at Purchase who were paid $4 for their
participation in the study. They all completed the implicit person theory
measure described in Study 1. On the basis of the same criteria as in
Study 1,21 participants were classified as entity theorists; 11 participants
were classified as incremental theorists; and data from 18 participants,
who did not have a clear theory, were excluded from the group comparison analyses but were retained in the correlational analyses in which
endorsement of entity theory was treated as a continuous variable.
Measures
In addition to the implicit person theory measure, the participants
were given a 10-item questionnaire. On each item, trait information
about a target was presented, and participants were asked to make a
prediction about the target's behavior in a particular situation. One
example is, "Henry is more aggressive than Edward on average. What
do you suppose is the probability that Henry would act more aggressively
than Edward in a particular situation?" Participants were given the same
probability scale as in Study 1 (0.00 to 1.00) to indicate their predictions.
The 10 items covered positive and negative traits, as well as traits with
unclear valence: defensiveness, talkativeness, aggressiveness, dishonesty,
helpfulness, shyness, xtudiousness, conscientiousness, boastfulness, and
spontaneity.
Results
For each participant, the 10 predictions were averaged to
give an overall prediction score. Agreement with entity theory
(treated as a continuous variable) was positively correlated with
the overall prediction score (r — .38, p < .05). Theory group
2
Again, we were not interested in whether entity theorists were more
accurate in their predictions than were incremental theorists, or vice
versa, although it should be noted that personality research has found
knowledge about a person's trait to be a rather unreliable predictor of
a person's behavior in a particular situation (see Mischel, 1968, 1973;
L. Ross & Nisbett, 1991). Specifically, the correlation between a person's trait measured by a personality inventory and that person's traitrelevant behavior in a particular situation seldom exceeds .30 (see Mischel, 1968, 1973). Indeed, the correlation coefficient of .30 has been
referred to in the literature as the personality coefficient (see Mischel,
1973).
24
CHIU, HONG, AND DWECK
comparisons performed on this overall score revealed a significant implicit theory effect, /(30) = 2.30, p < .05. As predicted,
relative to incremental theorists, entity theorists made stronger
predictions about the targets' behavior in a particular situation
on the basis of information about the targets' traits (M = .78
vs. M = .64).
Study 3
Tn summary, Studies 1 and 2 provided support for the hypothesis that entity theorists would exhibit a greater tendency than
incremental theorists to display the inferential practices of the
lay dispositionist model depicted in Figure 1. Study 1 showed
that if a trait-relevant behavior is known to have taken place on
one occasion, entity theorists tend to expect the same type of
behavior to recur on other occasions in different situations (the
behavior-to-behavior link in Figure 1). In Study 2 we found
that compared with incremental theorists, entity theorists tended
to make stronger predictions of a person's behavior in a particular situation on the basis of knowledge of this person's traits in
the relevant domain (the trait-to-behavior link). In Study 3 we
tested entity and incremental theorists' relative tendency to make
a confident inference about a person's trait on the basis of
knowledge of behavior in a single situation (the inferential practice specified by the behavior-to-trait link in Figure 1). If this
hypothesis is confirmed, this would further indicate that for
entity theorists more than for incremental theorists, dispositional
traits mediate behaviors and thereby lend consistency to behaviors across situations.
Method
Participants
The participants were 37 college students (23 men and 14 women)
at the State University of New YorY. at Purchase who volunteered to take
part in the study during a class meeting. None of them had participated
in Study 2.
Measures
Measure of implicit theory of character. Recall that implicit theories
are domain specific (see Dweck et al., 1993) and that judgments in a
particular domain are best predicted by implicit theories of attributes in
the same domain. In Studies 1 and 2, we were interested in the participants' predictions of behavior across domains (e.g., morality, intelligence, and sociability). We therefore measured the participants' general
theory about the fixedness or malleability of basic human attributes. In
this study, we were interested in how participants made judgments of a
person's moral character on the basis of a single morally relevant behavior. Therefore, instead of assessing participants' general theory as we
did in Studies 1 and 2, we assessed their implicit theory in a more
specific domain—moral character.
The measure of implicit theory of character consists of three items,
each depicting an entity theory of character: (a) "A person's moral
character is something very basic about them and it can't be changed
much,'" (b) "Whether a person is responsible and sincere or not is
deeply ingrained in their personality. It cannot be changed very much,"
and (c) "There is not much that can be done to change a person's moral
traits (e.g., conscientiousness, uprightness, and honesty)." Participants
were required to indicate on a 6-point scale, with responses ranging
from 1 (strongly agree) to 6 (strongly disagree), how much they agreed
with each statement.
Evidence attesting to the reliability and validity of the measure is
reviewed in Dweck et al. (1995a). Very briefly, the internal reliability
coefficients of the measure obtained from past studies ranged from .85
to .94. Test-retest reliability for a 2-week interval was .80. In addition,
the measure did not correlate with a pessimistic or negative view of
human nature, social presentational tendencies as measured by the Paulhus ( 1984) Social Desirability Scale and Snyder's (1974) Self-Monitoring Scale, or with intellectual competence (SAT scores). Finally, the
measure did not correlate with locus of control factors (Levenson,
1974), ideological rigidity such as authoritarianism (Altemeyer, 1981),
or conservatism and liberalism (Kerlinger, 1984). Among the 37 participants in this study, 10 scored 3.0 or below on this measure and were
classified as entity theorists; 20 scored 4.0 or above and were classified
as incremental theorists. The remaining 7 participants did not have a
clear theory, and, as in Studies 1 and 2, their data were excluded from
the theory group comparisons but were retained in the correlational
analyses.
Measure nf trait judgments. To assess participants' trait judgments,
a list of 35 behaviors was compiled. These behaviors ranged from clearly
positive behaviors (e.g., "risking one's life for another") to mildly
positive behaviors (e.g., "making one's bed in the morning"'), mildly
negative behaviors (e.g., "interrupting someone who is speaking"), and
clearly negative behaviors (e.g., "stealing a car"). The participants
were asked to indicate on a rating scale the extent to which each of these
behaviors was indicative of the goodness or badness (moral character) of
the person who displayed the behavior. The scale ranged from 1 (not a!
all indicative) to 5 (very indicative). The degree of rated indicativeness
was taken as an index of trait judgments.
In addition to the indicativeness judgments, participants were also
asked to judge the positivity and negativity of each behavior on a 9point scale with responses ranging from 1 (verypositive), to 5 (neutral),
to 9 (very negative). This rating allowed us to test whether any difference between entity theorists and incremental theorists on the indicativeness judgment was due to a difference in the perceived positivity or
negativity of the behavior. For example, if entity theorists made more
extreme indicativeness judgments than did incremental theorists, it could
be because entity theorists perceived the behavior to be more extreme
than did incremental theorists.
Results and Discussion
As predicted, entity theorists perceived the behaviors to be
more indicative of the actors' moral character than did incremental theorists. Moreover, as analyses of the evaluation of the
behaviors revealed, this was not because entity theorists perceived the behaviors as being more extreme than did incremental
theorists.
To construct an index of the perceived extremity of the behaviors, we recoded the participants' evaluation of each behavior's
negativity versus positivity from the original scale (from 1 to
9) to a scale that ranged from —4 to 4. The absolute value on
the transformed scale was then used to index how extremely
positive or negative the behavior was perceived to be. We then
took the average of the extremity scores for all 35 behaviors.
The correlation between the extremity score and the implicit
theory measure (treated as a continuous variable) was not significant (r = .01, ns). Theory group comparison analyses performed on the average extremity score revealed no significant
differences between entity and incremental theorists, ?(28) =
0.07, p = .95. The average extremity scores for entity theorists
(M = 2.58) and incremental theorists (M = 2.56) were virtually
25
LAY DISPOSITIONISM
identical. Indeed, the two theory groups did not differ in their
evaluation of the positivity or negativity on any of the 35 behavioral items. This finding also suggests that incremental theorists
were not less inclined than entity theorists to make evaluative
judgments as long as these judgments were attached only to the
behaviors and not to the performers of the behaviors (cf. Bassili,
1989; Trope, 1986, 1989; Trope & Liberman, 1993).
Despite the fact that entity theorists and incremental theorists
made identical evaluations of the overall extremity of the behaviors, they differed significantly in how much they perceived
these behaviors to be indicative of a person's character. Agreement with entity theory and the average of the 35 trait indicativeness judgments was significant ( r = — . 5 0 , p < .01).Theory
group comparisons performed on the average trait indicativeness
judgments revealed a clear difference between entity and incremental theorists, r(28) = 3.09, p < .005. Entity theorists, to a
significantly greater extent than incremental theorists, believed
that a person's character could be revealed in a single behavior
this person performed (M = 3.62 vs. M = 2.96 on a 1 - 5 scale).
Was this effect independent of the valence of the behaviors,
or was it localized in judgments of either negatively or positively
valenced behaviors? To answer this question, we divided the 35
behaviors into three categories on the basis of the participants'
evaluations of the behavior on the original 9-point scale. Specifically, among the 35 behaviors, 15 had an average score between 1 and 3.5 and were classified as positive behavior, and
16 had an average score between 6.5 and 9 and were classified
as negative behavior. The remaining 4 behaviors had an average
negativity score between 3.5 and 6.5 and were classified as
behaviors with weak valence. The average trait indicativeness
rating for each category of behavior was then computed, and a
2 X 3 (Implicit Theory x Valence) ANOVA, with the second
factor as a within-subject variable, was performed on the average trait indicativeness score. The results revealed a clear implicit theory main effect, F ( l , 28) = 12.19, p < .005. The
Implicit Theory X Valence interaction was not significant, F(2,
56) — 1.20, p = .30. The insignificant interaction effect suggested that the difference between entity theorists and incremental theorists in the perceived trait indicativeness of a behavior
was independent of the valence of the behavior. Indeed, as Table
1 shows, relative to incremental theorists, entity theorists judged
a behavior to be more indicative of a person's character regardless of whether the behavior was clearly positive, clearly negative, or only weakly valenced. Thus, entity theorists' more extreme trait indicativeness judgments were not localized in posi-
Table 1
Trait Indicativeness Judgment Scores: The Effect
of Implicit Theories
Entity
theorists
tive or negative behaviors but may reflect the general framework
for social inference their theory has set up.3
In summary, together with the findings from Studies 1 and
2, the findings from the present study suggest that relative to
incremental theorists, entity theorists believe more strongly in
the lay dispositionist model of behavior. First, they are more
inclined to make confident predictions of a person's behavior in
a particular situation on the basis of knowledge of this person's
relevant traits (Study 2). In addition, they believe more strongly
that even a single behavior is indicative of a person's character
(Study 3) and expect a higher level of cross-situational consistency in behavior (Study 1).
Study 4
Studies 1, 2, and 3 showed that, at least among American
participants, an entity theory of personal attributes is systematically linked to the tendency to subscribe to lay dispositionism.
However, different cultures may construct different models of
social inference (Cousins, 1989; Miller, 1984; Morris & Peng,
1994). Whereas the individual is seen as the primary agent of
social actions in individualistic cultures (e.g., American culture), role expectations, situational norms, and contextual constraints are seen as primary regulators of social actions in collectivistic cultures (e.g., Chinese culture; Hong, Chiu, & Kung, in
press; Markus & Kitayama, 1991, 1994). Will the relation between an entity theory of human attributes and the tendency to
subscribe to lay dispositionism be found in cultures that put
less emphasis on the person in understanding social events? To
address this question, Study 3 was replicated with a sample of
Hong Kong Chinese students and a sample of American students.
Method
The participants were 196 (54 men and 142 women) Chinese students
from the University of Hong Kong and 124 students (47 men and 77
women) from Columbia University who had not participated in any of
the previous studies. One week after the Chinese students filled out the
implicit theories, they were asked to fill out the trait judgment measure
in an "unrelated" study. This was the same measure as the one used
in Study 3, in which descriptions of behaviors were presented and participants were asked to rate on a 9-point scale (from 1 to 9) the extent
to which each of these behaviors was indicative of the moral character
of the person who displayed the behavior. The American participants
filled out the implicit theories measure and the trait judgment measure
in the same experimental session, but the two measures were separated
by a number of filler questionnaires. The implicit theories measure, the
trait judgment measure, and the filler questionnaires were presented
to the American participants as a battery of measures put together by
investigators from different research projects for convenience of
administration.
To ensure that the relation between implicit theories and lay dispositionism was not mediated by participants' confidence in other people's
morality, we also assessed the degree to which participants viewed peo-
Incremental
theorists
3
Behavior
Positive behavior
Weakly valenced behavior
Negative behavior
M
SD
M
SD
3.82
3.15
3.54
0.61
0.77
0.86
3.05
2.14
3.08
0.85
0.77
0.69
There was a significant valence main effect, F(2, 56) = 11.55, p
< .001. Independent of the participants' implicit theory, behaviors with
clear valence were judged to be more indicative of a person's character
dian were behaviors with unclear valence (M = 3.31 for positive behaviors, M = 3.23 for negative behaviors, and M — 2AS for behaviors with
weak valence), r(29) = 6.19, p < .0001.
26
CHIU, HONG, AND DWECK
pie as basically moral or immoral. The confidence measure consists of
three pairs of items (e.g., ' 'I believe that most people will take advantage
of others if they can" vs. "Ibelieve that most people are trustworthy").
Participants were asked to choose one statement from each pair and rate
it using a 3-point scale with responses ranging from 1 (very True for
me) to 3 (sort of true for me). Responses to each item were recoded
into a 6-point ( 1 - 6 ) scale with a higher score indicating higher confidence. The confidence measure was formed by taking the average of the
three items.
Finally, both the Chinese and the American participants were asked
to fill out the 16-item Brown Collectivism Scale {Brown et al., 1992)
after they had completed the trait indicativeness measure. Examples of
items from the Brown Collectivism Scale are "It is more important that
my group does well than that I do well as an individual" and "Working
with others is usually more trouble than it's worth" (reverse scoring).
Participants indicated the extent of their agreement with each of these
statements on a 6-point scale with responses ranging from 1 (strongly
disagree) to 6 (strongly agree). The collectivism score was the average
agreement with the 16 items.
Results
Compared with the American participants, the Chinese participants scored higher on collectivism (M = 3.71 vs. M - 2.19
on a scale from 1 to 6), ?(316) = 27.04,/? < .0001, and had
greater confidence in others' morality (M = 3.37 vs. M = 2.94),
r(318) = 6.76, p < .001. The two samples did not differ in
their tendency to endorse an entity theory of morality (M =
3.15 and M = 3.12, respectively), *(316) = 0.33, ns.
To test the effects of cultural groups and implicit theories on
trait indicativeness judgments, we performed a multiple regression analysis. In this analysis, average trait indicativeness judgment was regressed on cultural groups (American vs. Chinese),
implicit theories, and the Cultural Groups x Implicit Theories
interaction. In addition, morality confidence and collectivism
were included as covariates. The implicit theories main effect
was significant, f (1, 312) = 14.98,/? < .001. In line with our
hypothesis, for both American and Chinese students, endorsement of an entity theory was positively related to average trail
indicativeness judgments (r — .22, p < .01 for American students, r = .24, p < .01 for Chinese students, and r — .23, p <
.001 for American and Chinese students combined). The Cultural Groups X Implicit Theories interaction was not significant,
F( 1, 312) = 0.70, ns, indicating that the relation between implicit theories and trait indicativeness was not moderated by
the participants' cultural group membership. The cultural group
effect was also significant, F ( l , 312) = 5.97, p < .05. The
least squares means indicated that after the effects of implicit
theories, confidence in others' morality, and collectivism had
been partialled out, American students still exhibited a slightly
greater tendency to make trait indicativeness judgments than did
Hong Kong Chinese students (M = 3.14 vs. M = 2.84).
Confidence in others' morality was not related to trait indicativeness judgments, F ( l , 312) = 0.03, m (r = .06 for Chinese
students and r = —.14 for American students). Recall that implicit theories predicted trait indicativeness judgments even after
the effect of confidence was partialled out in the multiple regression analyses. Together, the results suggest that confidence in
others' morality did not mediate the relation between implicit
theories and trait judgment.4
In summary, there was support for the cross-cultural general-
ity of the association between implicit theories of morality and
the tendency to make moral trait inferences from behavior.
Study 5
Thus far, we have shown a consistent association between
implicit theories and dispositional judgments. However, our underlying model posits a causal relation, with the theories setting
up a framework within which different types of inferences are
plausible. It is therefore important to test the hypothesized
causal relation. In two past studies (Dweck, Tenney, & Dinces,
1982, cited in Dweck & Leggett, 1988; Bergen, 1991), theories
of intelligence have successfully been manipulated by means of
reading passages. In the second case, with adults, Bergen provided participants with a Psychology Today -type article that
compellingly presented either the entity or the incremental theory. The article vividly described and quoted from extensive
research purporting to document the theory. In both studies
the theory manipulation produced significant effects on major
dependent measures that were in no way touched on in the
passages.
In the present study, theories of personality and character
were manipulated. Each participant was given a short article tn
read that presented either the entity or the incremental theory.
Although participants might arrive with strong predispositions
toward one theory or the other, both theories may represent
basic modes of thought that are at some level familiar to most
individuals, and the manipulations were expected to lead participants to adopt that particular mode of thought. The entity article
informed participants that human nature used to be thought of
as a bundle of potentialities, each of which could be developed,
but that scientists, through years of rigorous research, had arrived at the view that people possessed a finite set of rather
4
In this multiple regression analysis, collectivism was positively related to trait indicativeness judgments, F(l, 312) = 14.54, p < .001.
More collectivistically oriented individuals tended to make stronger trait
judgments from behavior (r = .21, p < .01 for American students, r =
.13,/J = .06 for Chinese students, and r = .16,/> < .001 for both groups
combined). However, as noted, past studies (Lee et al., 1996; Miller,
1984; Morris & Peng, 1994) have found that, compared with individualists, collectivists, who are more sensitive to interpersonal influences on
behavior, are less likely to attribute behavior to internal causes. The
positive relation between collectivism and dispositional judgment suggests that although both dispositional inference and internal attribution
are important aspects of lay dispositionism, they may be different cognitive processes; that is, one can infer a trait from behavior but not see it
as the cause of the behavior (Hamilton, 1981). It is possible that, relative
to individualists, collectivists, keeping group welfare in mind, place
more responsibility on individuals to regulate their social behavior for
the common good and are thus more likely to judge people's morality
on the basis of whether they do or do not do so. That is, their character
judgments will be based on the degree to which individuals perform
socially positive versus socially negative actions. However, collectivists
may still place more emphasis than individualists on social-situational
factors as causes of behavior. For example, to collectivists, a morally
weak person may simply be someone who is more susceptible to negative
situational forces. Thus, although they may judge this person as immoral,
they may still view the causes of the actor's behavior as residing more
in social factors than do individualists (Miller & Bersoff. 1994). This
issue would be an interesting one for future research.
27
LAY DISPOSITIONISM
fixed traits. In contrast, the incremental article informed participants that the view of human nature as a set of fixed traits was
now outmoded and had been replaced by the view of dynamic
human potentialities that could be cultivated and developed over
a lifetime. Each view was buttressed with vivid descriptions of
research and case studies. No mention was made in the passage
of anything directly relevant to any of the dependent measures.
For example, one paragraph of the entity article said,
In his talk at the American Psychological Association's annual convention held at Washington D.C. in August, Dr. George Medin argued that 'in most of us, by the age of ten, our character has set
like plaster and will never soften again.' He reported numerous large
longitudinal studies which show that people 'age and develop, but
they do so on the foundation of enduring dispositions.'
The same paragraph in the incremental article read,
In his talk at the American Psychological Association's annual convention held at Washington D.C. in August, Dr. George Medin argued that 'no one's character is hard like a rock that cannot be
changed. Only for some, greater effort and determination are needed
to effect changes.' He reported numerous large longitudinal studies
which show that people can mature and can change their character.
He also reported research findings showing that people's personality
characteristics can be changed even in their late sixties.
Method
Forty-six students (14 men and 32 women) from Columbia University,
who had not participated in any of the previous studies, were recruited
as participants. They were told that they were participating in three
studies that were grouped together for efficiency. Each part was given
in a separate booklet, with a separate study name, and a separate consent
form. The first part, the theory manipulation, was called a study of social
comprehension and memory (comprehension of and memory for personrelevant information). The second part, a filler task, was called a lexicaldecision task, and the third part was called a social judgment task.
Participants were told after they finished reading the article that the
comprehension and memory tests on the article would come later in the
session. Then, they were presented with a filler task, which was a 15min lexical-decision task, and were asked to decide whether a string of
letters presented on a computer screen was a word. Then, they filled out
a four-item short form of the behavioral prediction questionnaire used
in Study 2 (e.g., "Tina is more defensive than Mara on average, what
do you suppose is the probability that Tina would act more defensively
than Mara in a particular situation?") and a six-item short form of the
trait judgment measure used in Study 3. Finally, as manipulation checks,
four questions were included to assess the degree to which the participants found the article difficult to understand, credible, and persuasive
and the degree to which they agreed with the views espoused in the
article. Participants were asked to indicate their responses to each of
these questions on an 8-point scale with responses ranging from 1 (not
at all) to 8 (very much). To probe for suspiciousness, we had participants write down what they thought the research hypotheses were in
this research. None of them saw a connection between the reading
comprehension task and the judgment task. Thus, the findings from this
study could not be due to the participants' effort to apply what they
had read from the passage to the judgment task.
Finally, previous research (Lord, Lepper, & Preston, 1984; see also
McGuire, 1964) has shown that exposing people to arguments from both
sides reduces the biases produced by a one-sided persuasive message. We
therefore showed participants in the debriefing session both articles and
discussed with them both viewpoints. They were told that, actually, both
views were widely held, and they were given references for further
reading.
It was hypothesized that participants who were led to believe in fixed
traits would tend to view behaviors as indicative of such traits and would
tend toward dispositional judgments more than those who received the
incremental passage.
Results and Discussion
In general, participants found the article easy to understand
(Af - 1.42, SD = 0.70), credible (M = 5.98, SD = 1.22), and
persuasive (M = 5.67, SD - 1.34). They also tended to agree
with the views espoused in the article (Af = 6.71, SD — 0.72).
The theory manipulation (entity vs. incremental) did not have
a differential effect on these manipulation checks (all fs <
1). Thus, the two articles were equally effective in leading the
participants to agree to the implicit theory presented in the
article.
As predicted, participants who were led to believe in an entity
theory of personality scored higher on the behavioral prediction
questionnaire than did participants who were led to believe in
an incremental theory (Af = 77% vs. M = 69%), F ( l , 44) =
4.27, p < .05. Thus, compared with incremental participants,
entity participants made stronger predictions about the targets'
behavior in a particular situation on the basis of information
about the targets' traits. Entity participants also scored significantly higher on the trait indicativeness judgment measure than
did incremental participants (Af = 3.72 vs. M = 3.27), F ( l ,
44) = 5.18, p < .05, indicating that entity participants made
stronger trait judgments from trait-relevant behaviors than did
incremental participants.
General Discussion
Although lay dispositionism is often thought of as a tendency
that is shared by those in American culture, the present research
shows that it is more characteristic of some people than others.
It appears that for entity theorists—those who believe in fixed
traits—more than for incremental theorists, traits are the basic
unit with which social events are analyzed and understood. Specifically, in comparison to incremental theorists, entity theorists
seem to be more confident that trait-relevant behaviors will be
consistent across situations (Study 1). They also appear to believe more strongly that (a) knowing a person's traits allows
one to make confident predictions about that person's behaviors
in a new situation (Studies 2 and 5) and (b) knowing how a
person behaved in a particular situation allows one to make
confident inferences about this person's traits in the relevant
domain (Studies 3, 4, and 5). Furthermore, the relation between
implicit theories and dispositional inferences was found in cultures characterized by markedly different patterns of social understanding (Study 4 ) . This suggests that entity and incremental
theories are not simply indigenous cultural constructions people
in individualistic societies develop to understand social behavior.
Instead, they may be more general worldviews individuals in all
cultures can develop to give meaning to their social experiences.
Finally, there is initial evidence that implicit theories can partially explain the tendency to subscribe to lay dispositionism
(Study 5).
What might mediate the relation between an entity theory
28
CHIU, HONG, AND DWECK
and the tendency to subscribe to lay disposicionism? Could it
be the theory group differences in pessimism, response biases,
or cognitive astuteness? Or is it the different interpretive frames
the two theories set up to understand social events, as we
proposed?
It is evident from Study 4 that confidence in others' morality
does not mediate the relation between implicit theories and trait
judgments because confidence in others did not predict trait
judgments, and implicit theories predicted trait judgments even
after the effect of confidence was statistically controlled. Moreover, the results from Studies 1-3 show clearly that compared
with incremental theorists, entity theorists tended to make more
positive, as well as more negative, disposition-related inferences.
If the findings in the present research were due to response
biases on measures of social perception, then we should expect
entity theorists to display more extreme responding than incremental theorists on all of the social perception measures, but
that was not the case. In Study 3, although entity theorists made
more extreme trait indicativeness judgments than did incremental theorists, their evaluations of the behaviors in the same study
were identical to those of incremental theorists. In short, entity
theorists displayed more extreme responding than incremental
theorists only when our theory predicted that they would do so.
Furthermore, as previously noted, our past research (see
Dweck et al., 1995a, 1995b) showed that endorsing an incremental theory is not associated with social desirability or selfmonitoring as measured by the Paulhus Social Desirability Scale
(Paulhus, 1984) and the Synder Self-Monitoring Scale (Snyder,
1974). Thus, it is not the case that incremental theorists are
more concerned with self-presentation than are entity theorists
and, therefore, tend not to admit making extreme trait inferences.
Endorsing an incremental theory is also not related to the respondents' performance on standardized tests of conceptual
ability (e.g., the Academic Promise Tests; Bennett et al., 1965)
or tests of academic aptitude (SAT scores) or to their selfperceived intellectual ability. This shows that incremental theorists are not more cognitively competent and hence more able
to deal with social information in a complex way than entity
theorists are.
Future research will continue to search for specific mediators
of the implicit theories effects. Our proposal is that an individual's implicit theory about the fixedness or malleability of personality sets up an interpretive framework for understanding the
social world. As the present research suggests, along with entity
theorists' belief in fixed traits, there appears to be a belief in
traits as the unit of social understanding and a belief in a close
correspondence between traits and behaviors. These beliefs
allow entity theorists to make the three kinds of inferences illustrated in Figure 1.
In contrast, incremental theorists, for whom traits are malleable and personality is, thus, dynamic, tend not to weigh static
traits as heavily in understanding people and their behaviors.
Hence, they are less likely than entity theorists to display the
inferential practices associated with lay dispositionism. Instead,
as previous research (Chiu, 1994; Hong, 1994) has shown,
incremental theorists tend to put relatively greater emphasis on
the dynamic psychological factors that mediate behaviors (e.g.,
goals, intentions, expectancies, and affect). For example, Chiu
(1994, Study 1) found that when trait and goal information
were explicitly presented, entity theorists used traits, whereas
incremental theorists used goals as the major basis for organizing impressions of people. Similarly, Hong found that when
college students were presented with behavioral descriptions
(e.g., "Alexis stole the bread from the bakery shop") and were
asked to generate causal explanations for the behaviors, entity
theorists offered appreciably and significantly more trait responses (e.g., "Alexis stole the bread from the bakery shop
probably because he was a thief or a dishonest person") than
did incremental theorists. Incremental theorists, by contrast, had
a greater tendency to explain the behaviors by referring to the
actor's internal psychological states (e.g., "Alexis stole the
bread from the bakery shop probably because he was hungry
or desperate"). In short, there appear to be clear differences
between entity and incremental theorists in their relative emphasis on traits versus psychological states as explanations.
An entity versus incremental theory, as a general epistemic principle, may have wide implications for various inferential practices.
Fur example, entity and incremental theorists may be differentially
liable to trait-based stereotyping. Given entity theorists' greater
readiness to assign traits to people, will they be more ready to
assign traits to social groups as well? In two recently completed
studies, Levy and Dweck (1996) found that entity theorists endorsed existing stereotypes to a significantly greater degree than
did incremental theorists. Moreover, Levy, Stroessner, Dweck,
Hong, and Chiu (1997) found that entity theorists assigned significantly stronger traits to novel groups, whether the group was a
largely positive or a largely negative one.
To conclude, our findings are anticipated by many early theories in personality and social psychology. Kelly (1955) posited
that how individuals interpret and predict future events is to a
large extent guided by their implicit constructions of the social
world. Heider (1958) also proposed that individuals' implicit
theories of personality set up a framework for understanding
social events (see also Trope & Higgins, 1993). Consistent with
these broad theoretical postulates, our research suggests that
one implicit belief that may guide social inferences is the belief
about the fixedness versus malleability of human attributes. In
our view, having an entity theory sets up an emphasis on traits
and hence creates a greater propensity for the inferential practices indicative of lay dispositionism. In contrast, having an
incremental theory increases the relevance of situational and
psychological mediators in social understanding and hence moderates the propensity for lay dispositionism (Dweck et al.,
1993). Indeed, the idea that at least some perceivers may adopt
a dynamic approach to understanding people and their behavior
has received serious attention in recent models of social cognition (see Shoda & Mischel, 1993; cf. L. Ross & Nisbett, 1991).
In line with this growing emphasis and the longstanding tradition
represented by Heider and Kelly, our analysis and our findings
suggest that (a) having an entity versus incremental theory of
personality may set up a different processing framework in the
realm of social perception and (b) lay dispositionism, a construct that has been associated with a rich set of empirical
phenomena, may be differentially fostered by the perceiver's
implicit theories.
References
Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg, Manitoba, Canada: The University of Manitoba Press.
LAY DISPOSITION1SM
Baron, R. M., & Misovich, S.J. (1993). Dispositional knowing from
an ecological perspective. Personality and Social Psychology Bulletin,
79, 541-552.
Bassili, J. N. (1989). Traits as action categories versus traits as person
attributes in social cognition. In J. N. Bassili (Eds.), On-line cognition
in person perception (pp. 61-89). Hillsdale, NJ: Erlbaum.
Benenson, J., &Dweck, C. S. (1986). The development of trait explanations and self-evaluations in the academic and social domains. Child
Development, 57, 1179-1189.
Bennett, G. K., Bennett, M. G., Glendenen, D. M , Deppelt, J. E., Ricks,
J. H., Jr., Seashore, H. G., & Wesman, A. G. (1965). Academic promise tests. New "Vbrk: Psychological Corporation.
Bergen, R. (1991). Beliefs about intelligence and achievement-related
behaviors. Unpublished doctoral dissertation, University of Illinois,
Urban a - Ch ampai gn.
Boyum, L. A. (1988). Students' conceptions of their intelligence: Impact on academic course choice. Unpublished master's thesis. University of Illinois, Urbana-Champaign.
Brown, R., Hinkle, S., Ely, P. G., Fox-Cardamone, L., Maras. P., &
Taylor, L. A. (1992). Recognizing group diversity: Individualist-collectivist and autonomous-relational social orientations and their implications for intergroup processes. British Journal of Social Psychology, 31, 327-342.
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal
of Personality and Social Psychology, 42, 116-131.
Chiu, C. (1994). Bases of categorization and person cognition. Unpublished doctoral dissertation, Columbia University, New York.
Chiu, C , Dweck, C. S., Tong, Y. Y., & Fu, H. Y. (in press). Implicit
theories and conceptions of morality. Journal of Personality and Social Psychology.
Chiu, Y., & Hong, Y. (in press). Justice from a Chinese perspective. In
H. S. R. Kao & D. Sinha (Eds.), Asian perspectives on psychology.
New Delhi, India: Sage.
Chiu, C , Hong, Y, & Dweck, C. S. (1994). Toward an integrative
model of personality and intelligence: A general framework and some
preliminary steps. In R. J. Sternberg & P. Ruzgis (Eds.), Personality
and intelligence (pp. 104-134). Cambridge, England: Cambridge
University Press.
Coopersmith, S. (1967). The antecedents of self-esteem. New York:
Freeman.
Cousins, S. D. (1989). Culture and self-perception in Japan and the
United States. Journal of Personality and Social Psychology, 56,
124-131.
Dweck, C. S. (1991). Self-theories and goals: Their role in motivation,
personality, and development. In R. Dienstbier (Ed.), Nebraska Symposium on Motivation, 1990 (pp. 199-255). Lincoln: University of
Nebraska Press.
Dweck, C. S., Chiu, C , & Hong, Y. (1995a). Implicit theories and
their role in judgments and reactions: A world from two perspectives.
Psychological Inquiry, 6. 267-285.
Dweck, C. S., Chiu, C , & Hong, Y. (1995b). Implicit theories: Elaboration and extension of the model. Psychological Inquiry, 6, 322-333.
Dweck, C. S., Hong, Y, & Chiu, C. (1993). Implicit theories: Individual
differences in the likelihood and meaning of dispositional inference.
Personality and Social Psychology Bulletin, 19, 644-656.
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to
motivation and personality. Psychological Review, 25, 109-116.
Erdley, C , Cain, K., Luomis, C , Dumas-Hines, E, & Dweck, C. S.
( 1997). The relations among children's social goals, implicit personality theories and responses to social failure. Developmental Psychology, 33, 263-272.
Erdley, C. A., & Dweck, C. S. (1993). Children's implicit personality
theories as predictors of their social judgments. Child Development,
64, 863-878.
29
Fletcher, G. J. O., Danilovics, P., Fernandez, G., Peterson. D., & Reeder,
G. D. (1986). Attributioual complexity: An individual differences
measure. Journal of Personality and Social Psychology, 51, 8 7 5 884.
Gilbert, D. T, & Jones, E. E. (1986). Perceiver-induced constraints: Interpretation of self-generated reality. Journal of Personality and Social Psychology, 50, 269-280.
Goetz, T. E., & Dweck, C. S. (1980). Learned helplessness in social
situations. Journal of Personality and Social Psychology, 39, 2 4 6 255.
Hamilton, D. L. (1981). Cognitive representations of persons. In E. T.
Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The
Ontario Symposium (Vol. 1, pp. 135-159). Hillsdale, NJ: Erlbaum.
Heider, F. (1958). The psychology of interpersonal relations. New York:
Wiley.
Ho, D. Y. E, & Chiu, C. (1994). Component ideas of individualism,
collectivism, and social organization: An application in the study of
Chinese culture. In U. Kim, H. C. Triandis, C, Kagitcibasi, G. Choi, &
G. Yoon (Eds.), Individualism and collectivism: Theory* method and
applications (pp. 137-156). Thousand Oaks, CA: Sage.
Hofstede, G. (1980). Cultures consequences: International differences
in work-related values. Beverly Hills, CA; Sage.
Hong, Y. (1994). Predicting trait versus process inferences: The role
of implicit theories. Unpublished doctoral dissertation, Columbia University, New York.
Hong, Y, Chiu, C . & Dweck. C. S. (1995). Implicit theories of intelligence: Reconsidering the role of confidence in achievement motivation. In M. Kernis (Ed.), Efficacy, agency and self-esteem (pp. 197216). New York: Plenum.
Hong, Y., Chiu, C , & Kung, M. (in press). Bringing culture out in front:
Effects of cultural meaning system activation on social cognition. In
K. Leung, Y. Kashima, U. Kim, & S. Yamaguchi (Eds.), Progress in
Asian social psychology (Vol. 1). Singapore: Wiley.
Jellison, J. M., & Green, J. (1981). A self-presentation approach to the
fundamental attribution error: The norm of internality. Journal of
Personality and Social Psychology, 40, 643-649.
Jones, E. E. (1990). Interpersonal perception. New "fork: Freeman.
Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The
attribution process in person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 219-266).
New York: Academic Press.
Jones, E. E., & Thibaut. J. W. (1958), Interaction goals as bases of
inference in interpersonal perception. In R. Tagiuri &. L. Petrullo
(Eds.), Person perception and interpersonal behavior (pp. 151 —
178). Palo Alto, CA: Stanford University Press.
Kelly, G. A. (1955). The psychology of personal constructs. New York:
Norton.
Kerlinger, F. N. (1984). Liberalism and conservatism: The nature and
structure of social attitudes. Hillsdale, NJ: Erlbaum.
Krull, D.S. (1993). Does the grist change the mill? The effect of the
perceiver's inferential goal on the process of social inference. Personality and Social Psychology Bulletin, 19, 340-349.
Krull, D. S., & Dill, J. C. (1996). On thinking first and responding
fast: Flexibility in social inference processes. Personality and Social
Psychology Bulletin, 22, 949-959.
Kunda, Z., & Nisbett, R. E. (1986). The psychometrics of everyday
life. Cognitive Psychology, 18, 195-224.
Lee, F., Hallahan, M., & Herzog, T (1996). Explaining real-life events:
How culture and domain shape attributions. Personality and Social
Psychology Bulletin, 22, 732-741.
Leggett, E. L. (1985, March). Children's entity and incremental theories
of intelligence: Relationships to achievement behavior. Paper presented at the annual meeting of the Eastern Psychological Association,
Boston.
30
CHIU, HONG, AND DWECK
Levenson, H. (1974). Activism and powerful others: Distinction within
the concept of internal-external control. Journal of Personality Assessment, 38, 377-383.
Levy, S. R., & Dweck, C. S. (1996, June). The relation between implicit
person theories and beliefs in stereotypes. Paper presented at the
Eighth Annual Convention of the American Psychological Society,
San Francisco.
Levy, S., & Dweck, C, S. (1997). Implicit theory measures: Reliability
and validity data for adults and children. Unpublished manuscript,
Columbia University, New York.
Levy, S. R., Stroessner, S. J., Dweck, C. S., Hong, Y. Y, & Chiu, C. Y.
(1997, April). Implicit person theories and stereotype formation. Paper presented at the 68th Annual Meeting of the Eastern Psychological
Association, Washington, DC.
Lord, C. G., Lepper, M. R., & Preston, E. (1984). Considering the opposite: A corrective strategy for social judgment. Journal of Personality
and Social Psychology, 47, 1231-1243.
MacGyvers, V. (1992). Implicit theories and real world outcomes. Unpublished doctoral dissertation, University of Illinois, UrbanaChampaign.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review,
98, 224-253.
Markus, H. R., & Kitayama, S. (1994). A collective fear of the collective: Implications for selves and theories of selves. Personality and
Social Psychology Bulletin, 20, 568-579.
McGuire, W. J. (1964). Inducing resistance to persuasion: Some contemporary approaches. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 1, pp. 191-229). New Tfark: Academic
Press.
Miller, J. G. (1984). Culture and the development of everyday social
explanation. Journal of Personality and Social Psychology, 46, 961 —
978.
Miller, J. G.( & Bersoff, D. M. (1994). Cultural influences on the moral
status of reciprocity and the discounting of endogenous motivation.
Personality and Social Psychology Bulletin, 20, 592-602.
Mischel, W. (1968). Personality and assessment. New York: Wiley.
Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252-283.
Mischel, W., & Peake, P. K. (1982). Beyond deja vu in the search for
cross-situational consistency. Psychological Review, 89, 730-755.
Morris, M. W, & Peng, K. (1994). Culture and cause: American and
Chinese attributions for social and physical events. Journal of Personality and Social Psychology, 67, 949-971.
Neuberg, S. L,, & Newsom, J. T. (1993). Personal need for structure;
Individual differences in the desire for simple structure. Journal of
Personality and Social Psychology, 65, 113-131.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and
shortcomings of social judgment. Englewood Cliffs, NJ: Prentice Hall.
Paulhus, D. L. (1984). Two-component models of socially desirable
responding. Journal of Personality and Social Psychology, 46, 598609.
Peterson, D. R. (1968). The clinical study of social behavior. New York:
Appleton-Century-Crofts.
Robins, R. W., & Pals, J. L. (1996, June). Implicit self-theories in the
academic domain: A longitudinal study. Poster presented at the Eighth
Annual Convention of the American Psychological Society, San
Francisco.
Ross, L., & Nisbett, R. E. (1991). The personal and the situation. New
York: McGraw-Hill.
Ross, M. (1989). Relation of implicit theories to the construction of
personal histories. Psychological Review, 96, 341-357.
Shoda, Y, & Mischel, W. (1993). Cognitive social approach to dispositional inferences: What if the perceiver is a cognitive-social theorist?
Personality and Social Psychology Bulletin, 19, 574-585.
Shweder, R. A., & Sullivan, M. A. (1990). The semiotic subject of
cultural psychology. In L. A. Pervin (Ed.), Handbook of personality:
Theory and research (pp. 399-4)6). New "fork: Guilford Press.
Snyder, M. (1974). The self-monitoring of expressive behavior. Journal
of Personality and Social Psychology, 30, 526-537.
Trope, Y.*( 1986). Identification and inferential processes in dispositional
attribution. Psychological Review, 93, 239-257.
Trope, Y. (1989). Levels of inference in dispositional judgment. Social
Cognition, 7, 296-314.
Trope, Y, &. Higgins, E. T. (1993). The what, when, and how of dispositional inference: New answers and new questions. Personality and
Social Psychology Bulletin, 19, 493-500.
Trope, Y, & Liberman, A. (1993). The use of trait conceptions to
identify other people's behavior and to draw inferences about their
personalities. Personality and Social Psychology Bulletin, 19, 5 5 3 562.
Tversky, A., & Kahneman, D. (1982). Judgment under uncertainty:
Heuristics and biases. In D. Kahneman, P. Slovic, & A. Tversky
(Eds.), Judgment under uncertainty: Heuristics and biases (pp. 3 22). Cambridge, England: Cambridge University Press.
Wright, J. C , & Mischel, W. (1987). A conditional approach to dispositional constructs: The local predictability of social behavior. Journal
of Personality and Social Psychology, 55, 454-469.
Received March 10, 1994
Revision received October 9, 1996
Accepted October 25, 1996 •