Sie sind auf Seite 1von 17

page:1 colour:1 blacktext

We thank Mahzarin Banaji, Vincent Yzerbyt, and an anonymous reviewer for their helpful
comments on an earlier draft.
Address reprint requests to Ap Dijksterhuis, Department of Social Psychology, University of
Nijmegen, POBox 9104, 6500 HE, Nijmegen, The Netherlands. e-mail: U212177 vm.uci.kun.nl.
Arie Kruglanskis work on this project was supported by National Science Foundation Grant
SBR9417422.
oiN:i or rxirixrN1:i soci:i isxcnoiocx 32, 254270 (1996)
:1icir No. 0012
Motivated Social Cognition: Need for Closure Effects on
Memory and Judgment
Ai Dixs1rniis :Nn An v:N KNiiirNnrc
University of Nijmegen, The Netherlands
Air W. Kici:Nsxi
University of Maryland
:Nn
C:ri Scn:ir
University of Nijmegen, The Netherlands
Received: April 18, 1995; revised: October 10, 1995; accepted: December 13, 1995
In two experiments, the relation between need for closure and group perception was
studied. Dispositional need for closure was measured with a scale developed by
Webster and Kruglanski (1994). In Experiment 1, subjects were presented with
behavioral information about a group of soccer hooligans and were asked to form an
impression of this group. Subjects high in need for closure recalled relatively more
stereotype-consistent information, while subjects low in need for closure recalled more
inconsistent information. Furthermore, subjects high in need for closure judged the
target group more stereotypically and perceived the group as more homogeneous
compared to subjects low in need for closure. In Experiment 2, in which the target
group consisted of nurses and in which the stimulus set was more complex, most of the
results of Experiment 1 were replicated. It was concluded that dispositional need for
closure is an important determinant of the way social groups are perceived and
judged. 1996 Academic Press, Inc.
The role that stereotypical expectations play in person perception and
group perception is widely documented. Over the past fteen years, an
254
0022-1031/96 $18.00
Copyright 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.
APJ:JESP96 ART: 1295 21-3-1996
page:2 colour:1 blacktext
accumulating body of research has shown that initial (stereotypical) expecta-
tions play a major role in the way information about persons and group
members is processed (Belmore, 1987; Hemsley & Marmurek, 1982; Stern,
Marrs, Millar & Cole, 1984; Vonk & van Knippenberg, 1995), what informa-
tion is preferentially retrieved from memory (e.g., Hastie & Kumar, 1979;
Rothbart, Evans & Fulero, 1979; for reviews, see Fyock & Stangor,1994;
Rojahn & Pettigrew, 1992; Stangor & McMillan, 1992) and the way these
expectations aect impressions and judgments (e.g., Bodenhausen & Lichten-
stein, 1987; Brewer, 1988; Dijksterhuis & van Knippenberg, 1995; Fiske
& Neuberg, 1990; Macrae, Hewstone & Griths, 1993).
Until recently, the role of motivation in person and group perception was
largely neglected (but see Erber & Fiske, 1984; Kruglanski & Freund, 1983).
However, in the last few years, a shift toward a more pronounced role of
motivation in person and group perception becomes apparent (see, e.g.,
Doosje, Spears &Koomen, 1995; Fiske, 1993; Fiske &Neuberg, 1990; Pendry
& Macrae, 1994; Stangor & Lange, 1994; Yzerbyt, Schadron, Leyens
&Rocher, 1994). For instance, recent researchshows that motivational factors
such as accuracy goals (Kruglanski, 1990), outcome-dependency (Fiske
& Neuberg, 1990; Pendry & Macrae, 1994), the extent to which people feel
entitled to judge a person (Yzerbyt et al., 1994) are all important motivational
factors that aect the way social targets are judged.
NEED FOR CLOSURE
Kruglanski and colleagues contributed to this shift by introducing the
concept of need for closure (e.g., Kruglanski, 1990; Kruglanski, Webster
& Klem, 1993; Webster & Kruglanski, 1994). Our aim is to use the individual
dierence measure of the need for closure to shed light on epistemic motiva-
tional factors in group perception. Specically, the predictive value of peoples
need for closure for several ndings in group perception literature is tested.
Previous relevant research on those issues employed the term need for
structure, a semantic precursor to the needfor closure construct. In an early
paper, Kruglanski and Freund (1983, p. 450) dened the need for structure as
the needtohave some knowledge ona giventopic, any knowledge as opposed
to confusion and ambiguity .... A virtually identical denition was later given
for the need for closure, as [the] desire for a denite answer on some topic,
any answer as opposed to confusion and ambiguity . . . (Kruglanski, 1989,
p.13). Initial attempts to study the need for structure operationalized it
situationally via time pressure (Kruglanski & Freund, 1983) or the instruc-
tions tomake global judgments of a target (Freund, Kruglanski &Shpitzajzen,
1985). The same motivation was subsequently manipulatedvia environmental
noise (e.g., Kruglanski & Webster, 1991; Kruglanski et al. 1993), task attract-
iveness (Webster, 1993), or mental fatigue (Webster, Richter & Kruglanski, in
press).
255 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:3 colour:1 blacktext
Kruglanski (1989, p. 14) recognized that the need for structure/closure may
vary stably across individuals. As he put it, Whereas a need for closure is
likely to vary across topics and situations, stable individual dierences in such
a need may also exist. Thus, persons might reliably vary in the disposition to
apprehend their world in clear-cut, unambiguous terms (Ibid., p. 15). This fact
was duly noted by the original developers of the Personal Need for Structure
Scale, Thompson , Naccarato, and Parker (1992), who specically indicated:
Kruglanski does acknowledge... the possibility of the inuence of chronic
tendencies, with respect to the need for structure or closure. (p. 2). Thus,
Thompson et al. (1992) attest that the need for structure and closure are one
andthe same. Infact, they goso far as toassert that In order todistinguishour
personality measures fromthe situational perspective of Kruglanski, the Need
for Closure will be referred to as Personal Need for Structure (PNS)...
(emphasis ours, p.3).
The PNS scale was further validated and published by Neuberg and
Newsom (1993). The scale was also used in research by Schaller, Boyd,
Yohannes and OBrien (1995), and that of Thompson, Roman, Moscowitz,
Chaiken and Bargh (1994). In parallel, Webster and Kruglanski (1994) devel-
opedtheir ownversion of the need for closure scale whichoverlaps in part with
Thompsons et al. (1992) scale, and which Dutch version was employed in the
present investigation. This scale consists of ve subscales relating to peoples
motivation to come to a quick decision, the unwillingness to carefully elabor-
ate information and the desire to avoid confusion and ambiguity. The scales
are: (a) preference for order, (b) discomfort with ambiguity, (c) decisiveness, (d)
perdictability, and (e) close-mindedness.
To summarize, the PNS and the Need for Closure scales represent two
separate attempts to operationalize the very same construct of Kruglanskis
lay epistemic theory. This construct was referred to as the need for structure in
early publications (e.g., Kruglanski, & Freund, 1983) and as the need for
closure subsequently (e.g., Kruglanski, 1989; Kruglanski &Webster, in press).
The present research studies how need for closure/structure aects the
process of stereotyping. Previous research in this domain has demonstrated
that this need inuences stereotype-formation and may often result in erron-
eous stereotypes because of the less complex information-gathering strategies
it induces (Schaller et al., 1995). Research has also demonstrated initial need
for structure (closure) eects on stereotype-use by increasing the tendency to
draw stereotypic trait-inferences (Neuberg & Newsom, 1993, Experiment 4),
or perform competency evaluations (Jamieson & Zanna, 1989). But need for
closure eects onstereotyping may extendto further, yet unexplored, phenom-
ena considered below.
EMPIRICAL PREDICTIONS
Imagine the case in which subjects are asked to form an impression about
a social group. After subjects are provided with a category label (instigating an
256 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:4 colour:1 blacktext
initial expectancy), they are presented with expectancy-congruent and expect-
ancy-incongruent information (cf. Bodenhausen & Lichtenstein, 1987; Dijk-
sterhuis & van Knippenberg, 1995; Stangor & Ruble, 1989; see Fyock &
Stangor, 1994, for a review of research using this paradigm). First, it may well
be that expectancy-inconsistent information (here stereotype-inconsistent
information) is processed dierently by people high in need for closure versus
those low in this need. High need for closure implies, among other things,
discomfort with ambiguity and closed-mindedness. Inconsistent information
is exactly the kind of information that may instigate ambiguity. While
stereotype-consistent information simply corroborates the (judgmental) im-
plications of a stereotype, inconsistent information may force people to
reconsider their initial judgments. Moreover, since people high in need for
closure rely more on stereotypes during information processing (Kruglanski
&Freund, 1983), we expect themto pay more attention to consistent informa-
tion and show better recall for consistent information. People low in need for
closure, assumed to be less or not at all bothered by the fact that their initial
judgment may have to be reconsidered, are expected to pay more attention to
inconsistent information and to preferentially recall this information.
Furthermore, in line with situationallyinduced needfor closure, the disposi-
tional form of this motivation should aect person judgments as well (cf.
Kruglanski &Freund, 1983). People high in need for closure/structure should
rely more on their stereotypes when processing behavioral information (cf.
Neuberg & Newsom, 1993). Moreover, people high in need for closure are
expected to be reluctant to reconsider their initial expectancies (that is, to
avoid ambiguity). Therefore, we expect people high in need for closure to
make more stereotypical judgments about a social group than people low in
need for closure. In other words, their characterizationof a central tendency in
such groups will be close to the prevalent stereotype about the group in the
society at large. It is also plausible that high need for closure subjects will
perceive lower variability of the target group compared to lowneed for closure
subjects. First, subjects high in need for closure may well pay less attention to
stereotype-inconsistent information . If they do not incorporate inconsistent
information in their impression of the target group (or at least do so to a lesser
degree than their low need for closure counterparts) these subjects should end
up with a more homogeneous impression of the group. More homogeneous
stereotypes among subjects high in need for closure are also appealing from
a functional perspective. One component of high need for closure is the desire
for predictability (Kruglanski et al. 1993; Webster & Kruglanski, 1994).
Obviously, a highly variable stereotype has lower predictive value concerning
group members behaviors or their characteristics than a less variable stereo-
type. In concrete terms, if one is convinced that all soccer hooligans are
aggressive bastards, this stereotype is very informative about a particular
soccer hooligan one may encounter. On the other hand, a more heterogeneous
stereotype, such as the idea that, although hooligans may be quite aggressive
257 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:5 colour:1 blacktext
Extensive validation work shows that the Dutch version of the scale largely resembles the
original version (Cratylus, 1995).
in general, many non-aggressive hooligans do exist as well, provides less
predictive information. Hence, the preference for knowledge (for instance,
a stereotype) that aords predictability may lead to the construction of low
variability stereotypes by high need for closure individuals.
To summarize, we expect subjects high in need for closure to recall more
stereotype-consistent information and less stereotype-inconsistent informa-
tion, as well as to make more stereotypical central-tendency judgments about
a social group, and perceive it as more homogeneous than subjects lowin need
for closure.
EXPERIMENT 1: METHOD
Subjects
Twenty-sevenundergraduate students fromthe Universityof Nijmegen servedas subjects in the
experiment. Subjects received D. 5 (approximately 3 US$) in return for participation.
Pretesting of Stimulus Materials
Before the experiment proper was carried out, thirty-three students (comparable to the
experimental subjects) rated fourteen social groups on fourteen trait dimensions. Nine-point
scales were used withpoles labeledmembers of this groupare not at all...(1) and members of this
group are very...(9). Soccer hooligans were chosen as the target group for Experiment 1. They
were perceived by pretest subjects as aggressive (M:8.75), and unfriendly (M:2.14). Subse-
quently, forty subjects rated 98 behaviors on aggressiveness and friendliness. Nine- point scales
were used with the poles labeled performing this behavior is not at all (aggressive/friendly) (1)
and performing this behavior is very (aggressive/friendly) (9). Out of this pool, twelve behaviors
were selected, 6 of which were aggressive and 6 friendly. Furthermore, 6 irrelevant llers, neutral
with respect to aggressiveness and friendliness, were selected.
Procedure
Subjects attended the experiment in groups of up to 5 persons per session. They were placed in
individual cubicles containing an Apple Macintosh computer, used to provide all information
concerning the experiment. They were asked to read a series of behavioral descriptions about
members of a group of soccer-hooligans and to form an impression of this group. All descriptions
were accompanied by dierent rst names. For example, a behavioral description might read
Hanksometimes starts a ght in a bar. In all conditions, the rst two items as well as the last two
items to appear on the screen, were llers. All other items were presented in random order and
subjects were allowed to read the behaviors at their own pace. Reading latencies were measured.
After completion of the reading task and after a 1-min. break, subjects judged the groupon several
trait dimensions and completed a measure of preceived variability. Subsequently, subjects were
presented with a free-recall task. After its completion, they were thanked and paid.
Upon leaving the laboratory, subjects were asked to ll out the Dutch version of the Need for
Closure scale (Cratylus, 1995). This activity, subjects were told, had nothing to do with the
experiment they had just completed. Specically, subjects were led to believe that an experimen-
ters colleague was developing a scale called Need for Closure, and he needed subjects for work
on improvement of the scale. All subjects complied with the request by completing the scale. At
that point, they were debriefed and dismissed. Debrieng indicated that they indeed perceived the
tasks as unrelated.
258 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:6 colour:1 blacktext
Dependent Variables
First, the reading latencies were measured. After reading the behavioral information there
was a 1-min. break, following which subjects were presented witha judgment task. Subjects had to
rate the target group on four trait dimensions. In a pilot study, these traits were tested for
stereotypicality for soccer hooligans. Two traits were chosen consistent with the stereotype
(intolerant and selsh) and two inconsistent with the stereotype (intelligent and industrious).
These judgments were rated on 9-point scales ranging fromthe members of this group are not at
all...(1) to the members of this group are very... (9). The four traits were presented in random
order.
Subsequently, subjects engaged in a task designed to assess their perceived variability of
stereotypes. They were presented again with the four traits they had seen before. This time, the
traits were accompanied by a 100-point scale, with poles marked the members of this group are
not at all... (1) and the members of this group are very...(100). Subjects were asked to imagine
between which two points the vast majority (90%) of the group could be placed on a certain trait.
The dierence between the two points was taken as a measure of perceived variability. Again, the
traits were presented in random order.
After subjects completed this task, they were asked to write down as many of the original
behavioral descriptions as they could. They were given 6 minutes to complete this task.
RESULTS
First, Need for Closure scores were calculated. On the basis of these scores,
a median split division was made separating subjects high in need for closure
(scores higher than 154, N:13) from those low in need for closure (scores
lower than 154, N:14).
Recall
The number of correctly recalled consistent and inconsistent descriptions
were counted for each subject. An item was counted as correct if the general
meaning of the original item was reected in the description (cf. Gordon
& Wyer, 1987; Hastie & Kumar,1979; Srull, 1981, 1983). Subsequently,
proportions of correct recall were calculated and subjected to a 2 (need for
closure: high vs low) between-subjects;2 (Type of behavioral description:
consistent vs inconsistent) within-subjects ANOVA. First a main eect for
need for closure was found. Subjects high in need for closure performed worse
on the free recall task compared to subjects low in need for closure
(F(1,25):5.03, p:.005). As predicted, this main eect, however, was qualied
by an interaction of need for closure by type of behavioral description
(F(1,25):4.43, p:.05). As Table 1 shows, a memory advantage for consistent
behaviors was obtained for subjects high in need for closure, while the data of
the subjects low in need for closure showed the reverse pattern. It is also of
interest that while subjects with lowneed for closure generally exhibitedbetter
recall than subjects with high need for closure, this dierence was less
pronounced in regard to stereotype-consistent behaviors (F(1,25):1.48, n.s.)
and more pronounced in regard to stereotype-inconsistent behaviors
(F(1,25):15.25, p:.001).
259 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:7 colour:1 blacktext
TABLE 1
Rrc:ii (PrcrN1:crs) :Nn Rr:niNc L:1rNcirs (iN Src) ro CoNsis1rN1 :Nn INcoNsis1rN1
INrox:1ioN
High NFC Low NFC
Recall:
Consistent 39.3 50.0
Inconsistent 28.6 60.3
Reading latencies:
Consistent 5.27 5.49
Inconsistent 4.79 6.35
Reading L atencies
If, as hypotheisized, subjects high in need for closure and subjects low in
need for closure encode the information dierently, reading latencies should
showeects of this dierential elaboration. Meanreading latencies for consist-
ent and inconsistent behaviors were calculated for each subject. A 2 (need for
closure: high vs low ) between-subjects;2 (type of behavioral description)
within-subjects ANOVAshowed a marginally signicant two-way interaction
(F(1,25):3.09, p:.10, see Table 1 for means). The means were in the
predicted direction. Both subjects high and low in need for closure did spend
about the same amount of time on the consistent behaviors. However, subjects
low in need for closure showed longer reading latencies for inconsistent
behaviors, while subjects high in need for closure showed shorter reading
latencies for inconsistent behaviors.
Trait Judgments and Perceived Variability
Scores on the two inconsistent traits were recoded in a way that a high score
represented a stereotypical judgment. Mean scores were calculated for the
consistent trait dimensions as well as for the inconsistent trait dimensions.
A 2 (need for closure:high low) between-subjects;2 (trait type: consistent vs
inconsistent) within-subjects ANOVA revealed a main eect of need for
closure. As predicted, subjects high in need for closure judged the target group
more stereotypically compared to subjects low in need for closure
(F(1,25):4.76, p:.04, see Table 2 for means). The two-way interaction was
not signicant (F(1,25):2.19, p:.16).
To analyze perceived variability, two scores were calculated. One score
represented perceived variability on the consistent traits, the other score
represented perceived variability on the inconsistent traits. These scores were
subjected to a 2 (need for closure: high vs low) between-subjects ;2 (trait type:
consistent vs inconsistent) within-subjects ANOVA. Again, a signicant dier-
ence was found between need for closure conditions. Subjects low in need for
260 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:8 colour:1 blacktext
TABLE 3
Cori:1ioNs or DrirNnrN1 V:i:nirs vi1n Nrrn ro Ciosir
Correlation with NFC p-value
Stereotypicality of judgment .34 .10
Perceived variability 9.49 .02
Recall of consistent information 9.21 n.s.
Recall of inconsistent information 9.52 .007
TABLE 2
JincxrN1s (Hicnr Scors RrirsrN1 S1rro1xiic:i JincxrN1s) :Nn Prcrivrn
V:i:niii1x
High NFC Low NFC
Judgment 5.89 4.94
Perceived variability 27.9 40.7
closure indicated higher perceived variability within the group of soccer
hooligans than subjects high in need for closure (F(1,25):4.69, p:.04, see
Table 2 for means). The two-way interaction was not signicant
(F(1,25):1.29, p:.27).
Correlations
The need-for-closure measure allows us to calculate correlations between
need for closure and the dependent variables. The correlations are listed in
Table 3.
The only reliable (p:.05) correlations are these between need for closure
and perceived variability and between need for closure and recall of inconsist-
ent information. Both correlations are negative. Higher need for closure
convariedwith lower perceivedvariability and deteriotated recall of inconsist-
ent information. Other correlations, though not statistically reliable, may be
informative. Judgments tended to become more stereotypical when need for
closure was high. Furthermore, while higher need for closure may result in
decreased recall overall, the negative correlation between need for closure and
memory for inconsistent information is more pronounced than the negative
correlation between need for closure and memory for consistent information.
Although, as stated earlier, the latter correlations are not statistically signi-
cant, they may suggest that need for closure has a somewhat stronger impact
onthe processing of inconsistent informationcomparedtoconsistent informa-
tion.
261 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:9 colour:1 blacktext
DISCUSSION
The results of Experiment 1 conrm predictions derived from the need-for-
closure construct. First, while subjects low in need for closure tend to recall
more inconsistent behaviors than consistent behaviors, subjects high in need
for closure, apart from performing worse overall, recall consistent behaviors
with a greater probability than inconsistent behaviors. The reading latencies
suggest anencoding dierence. While people lowin need for closure seemtobe
more open to unexpected, disturbing information, subjects high in need for
closure tend to show a preference for information in line with expectations
present during the encoding stage. Furthermore, both trait-judgments and
perceived variability are aected by need for closure. Trait-judgments become
more stereotypical and perceived variability lower when subjects have a high
(vs low) need for closure.
The absence of two-way interactions for judgments and perceived variabil-
ity is important. Giventhe results onthe reading latencies andthe correlations,
one might tentatively conclude that need for closure only inuences the way
inconsistent information is processed (i.e., low need for closure leads to more
individuation) and that need for closure does not inuence the way consistent
information is processed (i.e., high need for closure does not lead to more
stereotyping). However, the absence of two-way interactions (indicative for
dierent-need-for-closure eects on consistent information and inconsistent
information) implies that both processes are at work, that is, need for closure
inuences processing of both consistent and inconsistent information. Of
course, these two related eects are not necessarily of the same magnitude.
EXPERIMENT 2
Though the results of our rst experiment are encouraging, their interpreta-
tion is clouded somewhat by certain features of the present procedure. First,
note that the soccer hooligan stereotype is negative. It is possible that
persons high on the need for closure have better memory for negative versus
positive behaviors, perhaps because the former tend to be more salient, and
need for closure disposes persons to pay particular attention to salient or
accessible features (Ford&Kruglanski, 1995; Thompsonet al., 1994). Thus , to
demonstrate that high (vs low) need for closure individuals have better
memory for stereotype-consistent, and worse memory for stereotype-incon-
sistent information, it would be desirable to replicate Experiment 1 using
a positive rather than a negative stereotype.
Second, it might be argued that dispositional dierences in the need for
closure would aect recall only under conditions of lowcognitive load(present
in Experiment 1). Under high cognitive load, by contrast, all subjects may, for
reasons of cognitive economy, attend primarily to stereotype-consistent infor-
mation. To check this possibility, in Experiment 2 we presented subjects with
30 instead of 18 behavioral descriptions and allowed subjects a relatively short
time to read the information.
262 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:10 colour:1 blacktext
Anal dierence fromExperiment 1 was in terms of the relationbetweenthe
behavioral information presented to subjects and the trait dimensions on
which the target group was judged. Whereas in experiment 1 the trait
dimensions were rather general and not descriptively related to the behavioral
information presented, in Experiment 2 those dimensions were more descrip-
tively relevant to the stimulus information.
METHOD
Subjects
Thirty undergraduate students from the University of Nijmegen served as subjects. Each
received D.5 in return for participation.
Prestesting of Stimulus Materials
Nurses were chosen as the target group for Experiment 2, on the basis of the pilot-study
described in Experiment 1. They were perceived as industrious (M:7.78) and non-aggressive
(M:2.40). Subsequently, forty students rated 98 behaviors on aggressiveness and industrious-
ness. Ninepoint scales were used with poles labeled performing this behavior is not at all (e.g.
industrious) (1) and performing this behavior is very (e.g. industrious) (9). Twenty behaviors
were selected of which 10 were industrious, 5 were lazy and 5 were aggressive. Furthermore, 10
irrelevant llers were selected, neutral with respect to aggressiveness and industriousness.
Procedure
Subjects attended the experiment in groups of up to seven persons per session. Upon entering
the laboratory, they were placed in individual cubicles. Subsequently, subjects were told that they
had to wait for about ten minutes because of a technical problemwith the computer system. They
were asked whether theywould mind lling out a questionnaire while waiting. Subjects were led to
believe that a colleague of the experimenter was developing a scale on the Need for Closure and
that he needed subjects because he was still working on improvement of the scale. All subjects
cooperated by completing the scale.
After having done so, subjects were told that the technical problems were overcome and the
experiment would start. As in Experiment 1, subjects were asked to read a series of behavioral
descriptions about members of a group of nurses and to form an impression of this group. All
descriptions were accompanied by dierent rst names. In all conditions the rst three items as
well as the last three to appear on the screen were llers. For instance, a behavior might read
Anne sometimes works over 50 hours a week. All other items were presented in random order.
Each behavioral description appeared on the screen for four seconds only (rather than allowing
subjects to proceed at their own pace as in Experiment 1).
After completion of the reading task, subjects were given a ve-minute ller task. They were
asked to write down as many capitals of countries as they could come up with. After ve minutes,
they were asked to judge the target group on two trait dimensions and to complete a measure of
perceived variability. Subsequently, they were presented with a free-recall task. After having
nished the latter task, subjects were thanked, paid and debriefed. The debrieng indicated that,
just as intended, the subjects perceived the tasks as unrelated.
Dependent Variables
After reading the behavioral information and completing the ller task, subjects were presented
with a judgment task. Specically, they were asked to rate the target group on two trait
dimensions. Those were specically represented in the behavioral descriptions (i.e., industrious-
263 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:11 colour:1 blacktext
TABLE 4
Rrc:ii or INnis1iois, L:zx :Nn Accrssivr Brn:vios (PrcrN1:crs)
High NFC Low NFC
Behaviors:
Industrious 32.0 42.0
Lazy 38.7 57.4
Aggressive 41.3 69.3
ness and aggressiveness). The judgments were given on nine-point scales ranging from the
members of this group are not at all...(1) to the members of this group are very...(9). The traits
were presented in a random order.
Subsequently, subjects were again presented with the two traits in order to obtain variability
assessments. This time, they were asked to indicate two points ona 100-point scale, between which
the vast majority (90%) of the group could be placed on a certain trait. The poles of the scale were
labeled the members of this group are not at all...(1) and the members of this group are
very...(100). The dierence between the two points was taken as a measure of perceived
variability. Again, the traits were presented in a random order.
After subjects completed this task, they were asked to write down as many of the original
behavioral descriptions as they could. They were given six minutes to complete this task.
RESULTS
Onthe basis of the need-for-closure scores, a mediansplit divisionwas made
separating subjects high in need for closure (higher than 159, N:15) from
those low in need for closure (lower than 159, N:15).
Recall
The number of correctly recalled industrious, lazy and aggressive descrip-
tions were counted for each subject, according to procedure used in Experi-
ment 1. Subsequently, proportions of correct recall were subjectedto a 2 (Need
for closure: high vs low) between-subjects;3 (Type of behavioral description:
industrious vs lazy vs aggressive) within-subjects ANOVA. Replicating Ex-
periment 1, a need-for-closure main eect was found. Subjects high in the need
for closure exhibited worse overall recall than did subjects low in the need for
closure (F(1,28):13.59, p:.002, see Table 4 for means). Furthermore, a main
eect of behavioral description was obtained (F(2,27):8.56, p:.002). Spe-
cically, industrious behaviors were recalledless than lazy behaviors which, in
turn, were recalled less than aggresive behaviors. These two main eects were
qualiedby aninteractionof needfor closure by type of behavioral description
(F(2,27):3.87, p:.04). Simple means revealed no signicant dierences
for subjects high in the need for closure (F:1). By contrast, low in need for
closure subjects recalled the stereotype-inconsistent lazy and aggressive be-
haviors with greater frequency than the stereotype-consistent industrious
behaviors (F(2,27):9.97, p:.001). This eect was borderline for the lazy vs
264 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:12 colour:1 blacktext
TABLE 5
JincxrN1s or INnis1ioisNrss (Hicnr Scors RrirsrN1 Mor S1rro1xiic:i JincxrN1s),
:ccrssivrNrss (Lovr Scors RrirsrN1 Mor S1rro1xiic:i JincxrN1s), :Nn Prcrivrn
V:i:niii1x
High NFC Low NFC
Industriousness 7.13 5.93
Aggressiveness 6.20 6.67
Perceived variability 33.4 45.3
industrious comparison (F(1,27):3.44, p:.08) and signicant for the aggres-
sive vs industrious comparison (F(1,27):24.54, p:.001). The dierence
between the lazy and aggressive behaviors was also marginally signicant
(F(1,27):3.34, p:.08).
It is also noteworthy that, just as in Experiment 1, low need for closure
subjects recalled more inconsistent behaviors than did their high need for
closure counterparts (F(1,27):4.29, p:.05, for lazy) and (F(1,27):15.44,
p:.002, for aggressive). No signicant dierence between high (vs low) need
for closure subjects appearedwith respect to stereotype-consistent industrious
behavior (F(1,27):2.41, p:.11).
Trait Judgments and Perceived Variablity
Judgments were obtained on the two-trait dimension reected in the
behavioral descriptions (i.e., industriousness and aggressiveness). First, these
scores were subjectedto a 2 (Needfor closure: high vs low) between-subjects by
2 (Trait type: industrious vs aggressive) within-subjects ANOVA. This analysis
revealed an signicant interaction of need for closure by trait type
(F(1,28):4.35, p:.05). Subsequently, trait scores for industriousness were
subjected to a 2 (Need for closure: high vs low) between-subjects ANOVA.
A signicant main eect for need for closure was found (F(1,28):4.28,
p:.05). Again, subjects high in need for closure rated the target group in more
stereotypical terms, that is, as more industrious. (see Table 5).
The judgments of aggressiveness were also subjected to a 2 (Need for
closure: high vs low) between-subjects ANOVA. Although cell means were in
the predicted direction (see Table 4), no signicant eect was obtained (F:1).
For every subject, two variability scores (i.e., on industriousness and on
aggressiveness) were calculated. On these scores a 2 (Need for closure: high vs
low) between-subjects;2 (trait type: industriousness vs aggressiveness) with-
in-subjects ANOVA was carried out. A main eect for need for closure
(F(1,60):7.07, p:.02) was again obtained. As in Experiment 1, subjects high
in need for closure perceived less variability in the target group than subjects
lowin the need for closure. The two-way interaction did not reach signicance
(F:1).
265 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:13 colour:1 blacktext
TABLE 6
Cori:1ioNs or DrirNnrN1 V:i:nirs vi1n Nrrn ro Ciosir
Correlation with NFC p-value
Stereotypicality of judgment
1. Industriousness .28 .14
2. Aggressiveness 9.05 n.s.
Perceived variability 9.47 .01
Recall of industrious information 9.09 n.s.
Recall of lazy information 9.20 n.s.
Recall of aggressive information 9.39 .04
Correlations
The correlations between need for closure and the dependent variables are
listed inTable 6. The correlations roughly replicate those of Experiment 1. The
positive relation between need for closure and stereotypicality of judgment as
well as the negative correlation between need for closure and perceived
variability were again obtained. Furthermore, recall of inconsistent informa-
tion seems to be negatively correlated with the need for closure. However, the
latter eect is only signicant for aggressive behaviors (p:.04), but not for
lazy behaviors.
DISCUSSION
Generally, the data of Experiment 2 replicate those of Experiment 1.
However, there is a dierence. Specically, whereas in Experiment 1 high need
for closure subjects recalled more stereotype-consistent than stereotype-
inconsistent information, in Experiment 2 they exhibited a nonsignicant
trend in the other direction. A possible explanation of this nding is the
relatively high saliency of negative, and in particular aggressive behaviors.
Among the aggressive behaviors were statements such as sometimes ghts in
a barand hit a patience once. These extreme behaviors may have grabbed
the subjects attention, and hence may have been encoded eciently and
recalled with ease compared to other, less attended behavioral information.
This explanation is supported by comparing the recall data of low need for
closure subjects across our two experiments. Whereas in both, low need for
closure subjects recalled inconsistent information better than consistent infor-
mation, in Experiment 1 the recall advantage of the former information was
10%whereas inExperiment 2 it was 30%withrespect toaggressive behaviors.
The relative salience of aggressive behaviors may also account for the fact
that the high (vs low) need for closure tendency to judge the target group of
nurses in stereotypic, hence nonaggressive, terms was nonsignicant (though
the means were in the predicted direction). In other words, while it might have
266 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:14 colour:1 blacktext
been relatively easy to downplay the inconsistent behaviors in Experiment
1 and the lazy behaviors in Experiment 2, the saliency of the aggressive
behaviors might have prevented high need for closure subjects from discount-
ing them in forming a judgment. Hence, subjects aggressiveness judgments
may have been more aected by information about aggressive behaviors than
were industriousness judgments by information about lazy behaviors.
GENERAL DISCUSSION
The present two experiments highlight the relevance of dispositional dif-
ferences in need for closure to recall of stereotype-consistent and inconsistent
information, as well as to stereotypical judgments and perceivedheterogeneity
within stereotyped groups. To summarize, high (versus low) need for closure
subjects tended to perceive groups in more stereotypical terms, and viewthem
as more homogeneous with respect to the stereotyped traits. Furthermore,
high (versus low) need for closure subjects tended to recall less stereotype-
inconsistent behaviors, and subjects low on the need for closure tended to
recall more of suchbehaviors than they did of stereotype-consistent behaviors.
It is noteworthy that those ndings were replicated across two highly dierent
stereotypes, one negative (soccer hooligans) and one positive (nurses) and
across procedural dierences including variations in pace of processing and
complexity of the stimulus information.
Whereas in the presence of a negative stereotype high need for closure
subjects recalled more stereotype-consistent versus inconsistent behaviors,
this trend was (nonsignicantly) reversed in the presence of a positive stereo-
type. We attempted to account for this nding by invoking the high saliency of
negative (e.g., highly hostile or aggressive) behaviors that are, therefore, readily
recalled despite being stereotype-inconsistent. Be that as it may, our results
suggest that high need for closure subjects are particularly likely to be aected
by negative stereotypes, perhaps more so than by positive stereotypes. Insofar
as the negative stereotypes are the ones with adverse social consequencesthis
may mean a bias toward negative social attitudes on part of individuals with
a high need for closure.
The present researchis consistent withneed-for-closure theory (Kruglanski,
in press; Kruglanski & Webster, in press) whereby subjects with high degrees
of such need tend to rely more on stereotypes and pay less attention to
individuating information (Jamieson & Zanna, 1989; Kruglanski & Freund,
1983, study 2; see also Neuberg & Newsom, 1993, for comparable ndings for
dispositional need for structure). A major contribution of the present research
is testing this general reasoning with recall as the dependent variable. Our
consistent nding that high need for closure subjects recalled less behavioral
informationthandid lowneedfor closure subjects is consistent withthe notion
that they froze on their initial stereotypes and processed less extensively
available individuating information. Furthermore, high versus low need for
closure appeared particularly biased against the recall of individuating infor-
267 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:15 colour:1 blacktext
mation that was stereotype-inconsistent, a trend that may defend their
frozen conceptions against dissolution.
It is important to ask what mediating mechanism underlies the relation
between need for closure and recall of stereotype-consistent and inconsistent
information. Two alternative accounts are possible, neither of which is deni-
tively addressed by our data. One is that high (versus low) need for closure
subjects simply possess stronger and better developed stereotypes to begin
with. Such stereotypes could then bias recall toward stereotype-consistent
information (Higgins & Bargh, 1987; Pryor, McDaniel & Kott-Russo, 1986;
Ruble & Stangor, 1986). This possibility is consistent with the nding that our
high (versus low) need for closure subjects held generally more extreme
stereotypes and adjudged the target groups as more homogeneous with
respect to the stereotyped traits.
The second possibility is that high (versus low) need for closure subjects are
more motivated to adhere to whatever stereotypes they may possess. To test
this possibility may require equating high and low need for closure groups on
the strength of initial stereotypes and perhaps collecting additional measures
to tap their motivated adherence to those stereotypes (e.g., derogation of
inconsistent information or of sources conveying it). Further research may
protably address those possibilities in an attempt to decide among them.
REFERENCES
Belmore, S. M. (1987). Determinants of attention during impression formation. Journal of
Experimental Psychology: L earning, Memory and Cognition, 13, 480489.
Boldenhausen, G. V., & Lichtenstein, M. (1987). Social stereotypes and information processing
strategies: The impact of task complexity. Journal of Personality and Social Psychology, 52,
871880.
Brewer, M. B. (1988). A dual process model of impression formation. In R.S. Wyer & T.K. Srull
(Eds.), Advances in social cognition (Vol. 1, pp. 136). Hillsdale, NJ: Erlbaum.
Cratylus (1995). De Nederlandse Need for closure schaal. Nederlands tijdschrift voor de
psychologie, 50, 231232.
Dijksterhuis, A., &vanKnippenberg, A. (1995). Memory for stereotype-consistent and stereotype-
inconsistent information as a function of processing pace. European Journal of Social
Psychology, 25, 689694.
Doosje, B., Spears, R., & Koomen, W. (1995). When bad isnt all bad: The strategic use of sample
informationin generalizationand stereotyping. Journal of Personality and Social Psychology,
69, 642655.
Fiske, S. T. (1993). Social cognition and social perception. Annual Reviewof Psychology, 44, 155194.
Fiske, S. T., & Neuberg, S. L. (1990). A continuummodel of impression formation from category-
based to individuating processes: Inuences of information and motivation on attention and
interpretation. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp.
174). San Diego, Academic Press.
Ford, T. E., & Kruglanski, A. W. (1995). Eects of epistemic motivations on the use of accessible
constructs in social judgment. Personality and Social Psychology Bulletin, 21, 950962.
Freund, T., Kruglanski, A. W., & Shpitzajzen, A. (1985). The freezing and unfreezing of impres-
sional primacy: Eects of the need for structure and the fear of invalidity. Personality and
Social Psychology Bulletin, 11, 479487.
268 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996
page:16 colour:1 blacktext
Fyock, J., & Stangor, C. (1994). The role of memory-biases in stereotype maintenance. British
Journal of Social Psychology, 33, 331343.
Gordon, S. E., & Wyer, R. S. (1987). Person memory: Category-set -size eects on the recall of
a persons behaviors. Journal of Personality and Social Psychology, 53, 648662.
Hastie, R., & Kumar, P. A. (1979). Person memory: Personality traits as organizing principles in
memory for behaviors. Journal of Personality and Social Psychology, 37, 2538.
Hemsley, G. D., & Marmurek, H. H. C. (1982). Person memory: The processing of consistent and
inconsistent information. Personality and Social Psychology Bulletin, 8, 433438.
Higgins, E.T.,& Bargh, J. A. (1987). Social cognition and social perception. In M. R. Rozenzweig
and L. W. Porter (Eds.), Annual review of psychology, 38, 369425.
Kruglanski, A. W. (1990). Motivations for judging and knowing: Implications for causal attribu-
tion. In E. T. Higgins and R. M. Sorrentino (Eds.), The handbook of motivation and cognition:
Foundation of social behavior, Vol. 2, New York: Guilford, pp. 333368.
Kruglanski, A. W. (in press). Motivated gatekeeper of our minds: Need for closure eects on
interpersonal and group processes. In E. T. Higgins and R. Sorrentino(Eds.) The handbook of
motivation and cognition: Foundation of social behavior, Vol. 3. New York: Guilford.
Kruglanski, A. W., & Freund, T. (1983). The freezing and unfreezing of lay inferences: Eects of
impressional primacy, ethnic stereotyping and numerical anchoring. Journal of Experimental
Social Psychology, 19, 448468.
Kruglanski, A. W., & Webster, D. M. (1991). Group members reactions to opinion deviates and
conformists at varying degrees of proximity to decision deadline and of environmental noise.
Journal of Personality and Social Psychology, 61, 215225.
Kruglanski, A. W., Webster, D. M., & Klem, A. (1993). Motivated resistance to openness to
persuasion in the presence or absence of prior motivation. Journal of Personality and Social
Psychology, 65, 861876.
Kruglanski, A. W., & Webster, D. M. (in press). Motivated closing of the mind: Seizing and
Freezing. Psychological Review.
Macrae, C. N., Hewstone, M., & Griths, R. J. (1993). Processing load and memory for stereo-
typebased information. European Journal of Social Psychology, 23, 7787.
Neuberg, S. L., & Newsom, J. T. (1993). Personal need for structure: Individual dierences in the
desire for simple structure. Journal of Personality and Social Psychology, 65, 113131.
Pendry, L. F., & Macrae, C. N. (1994). Stereotypes and mental life: The case of the motivated but
thwarted tactician. Journal of Experimental Social Psychology, 30, 303325.
Pryor, J. B., McDaniel, M. A., & KottRusso, T. (1986). The inuence of the level of schema
abstractness upon the processing of social information. Journal of Experimental Social
Psychology, 22, 312327.
Rojahn, K., &Pettigrew, T. F. (1992). Memory for schema-relevant information : Ameta-analytic
resolution. British Journal of Social Psychology, 31, 81109.
Rothbart, M., Evans, M., &Fulero, S. (1979). Recall for conrming events : Memory processes and
the maintenance of social stereotypes. Journal of Experimental Social Psychology, 15, 343
355.
Ruble, D. N., & Stangor, C. (1986). Stalking the elusive schema: Insights formdevelopmental and
social psychological analyses of gender schemas. Social Cognition, 4, 227261.
Schaller, M., Boyd, C., Yohannes, J., & OBrien (1995). The prejudiced personality revisited:
Personal need for structure and formation of erroneous group stereotypes. Journal of
Personality and Social Psychology, 68, 544555.
Srull, T. K. (1981). Some tests of associative storage and retrieval models. Journal of Experimental
Psychology: Human L earning and Memory, 7, 440462.
Srull, T.K. (1983). Organizational and retrieval processes in person memory: An examination of
processing objectives, presentation format, and the possible role of selfgenerated retrieval
cues. Journal of Personality and Social Psychology, 44, 11571170.
Stangor, Ch., & Lange, J.E. (1994). Mental representations of social groups : Advances in
269 MOTIVATED SOCIAL COGNITION
APJ:JESP95 ART: 1295 21-3-1996
page:17 colour:1 blacktext
understanding stereotypes and stereotyping. In M. Zanna (Ed.) Advances in experimental
social psychology, (Vol. 26, pp. 357416).
Stangor, C., & McMillan, D. (1992). Memory for expectancy-congruent and expectancy-incon-
gruent information: A review of the social and social developmental literatures, Psychologi-
cal Bulletin, 111, 4261.
Stangor, C., & Ruble, D. N. (1989). Stereotype development and memory: What we remember
depends on how much we know. Journal of Experimental Social Psychology, 25, 1835.
Stern, L. D., Marrs, S., Millar, M. G., & Cole, E. (1984). Processing time and the recall of
inconsistent and consistent behaviors of individuals and groups. Journal of Personality and
Social Psychology, 47, 253262.
Thompson, E. P., Roman, R. J., Moskowitz, G. B., Chaiken, S., & Bargh, J. A. (1994). Accuracy
motivation attenuates covert priming eects : The systematic reprocessing of social informa-
tion. Journal of Personality and Social Psychology, 66, 474489.
Thompson, M.M., Naccarato, M.E., & Parker, K.H (1992). Measuring cognitive needs: The
development and validation of the Personal Need for Structure (PNS) and Personal Fear of
Invalidity (PFI) measures. Unpublished manuscript, Univ. of Waterloo.
Vonk, R., & van Knippenberg, A. (1995). Processing attitude statements from ingroup and
out-group members: Eects of within-group and within-person inconsistencies on reading
times. Journal of Personality and Social Psychology, 68, 215227.
Webster, D. M. (1993). Motivated augmentation and reduction of the overattribution bias.
Journal of Personality and Social Psychology, 65, 261271.
Webster, D. M., & Kruglanski, A. W. (1994). Individual dierences in Need for cognitive closure.
Journal of Personality and Social Psychology, 67, 10491062.
Webster, D. M., Richter, L., and Kruglanski, A. W. (in press). On leaping to conclusions when
feeling tired: Mental fatigue eects on impressional primacy. Journal of Experimental Social
Psychology.
Yzerbyt, V. Y., Schadron, G., Leyens, J. P., & Rocher, S.(1994). Social judgeability: The impact of
meta-informational cues on the use of stereotypes. Journal of Personality and Social Psychol-
ogy, 66, 4855.
270 DIJKSTERHUIS ET AL.
APJ:JESP95 ART: 1295 21-3-1996

Das könnte Ihnen auch gefallen