00001.tif

PDF

Download

Open document

Flip pages

Contact Us

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
T he quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back o f the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600

THE ROLE OF AFFECT IN PERSONALIZED PERCEPTIONS OF INDIVIDUALS
by
Brian Paul Cotton
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
August, 1995
Copyright 1995 Brian Paul Cotton
UMI Number: 9617095
UMI Microform 9617095
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90007
This dissertation, written by
Brian Paul Cotton
under the direction of h.X?........ Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
—. C . . „
Dean of Graduate Studies
D a te May.. I?..?.?.
DISSERTATION COMMITTEE
wryerson
ii
Table of Contents
Introduction 1
Experiment 1: Method 12
Experiment 1: Results 18
Experiment 1: Discussion 27
Experiment 2: Introduction 29
Experiment 2: Method 31
Experiment 2: Results 39
Experiment 2: Discussion 49
General Discussion 53
Bibliography 62
Appendix A 67
Appendix B 74
Appendix C 81
Appendix D 88
List of Tables
Table 1: Stimulus Configurations and Associated Valences
Table 2: Unstandardized Attraction Means and Standard Deviations, Experiment 1
Table 3: Liking Means and Standard Deviations, Experiment 1
Table 4: Unstandardized Attraction Means and Standard Deviations, Experiment 2
Table 5: Liking Means and Standard Deviations, Experiment 2
Abstract
The differential effects of categorization and personalization on perceptions of
individuals were investigated in two experiments using an idiographic paradigm of
impression formation. The decategorization model and accentuation theory were
used to generate two main hypotheses. First, during personalized perception,
impressions of an individual will be based on the affective valence of an individual's
personal attributes, whereas during category-based perception attribute based
impressions will be assimilated to impressions of the group in which an individual is
a member. Second, the assimilation effect will cause category-based impressions to
become distorted away from the personalized baseline levels and toward the level of
the category valence. In Experiment 1, participants were asked to either form
impressions of 6 individuals or to place them into categories of academic majors.
Half of the target individuals were described primarily by positive attributes, and half
by primarily negative attributes. The first hypothesis was supported by a significant
interaction between task instructions and the affective valence of the attributes. The
second hypothesis was not supported. The procedure was changed in Experiment 2 to
increase the influence of the categories on impressions of the targets. A marginally
significant interaction between task instructions and the valence of personal attributes
was found in Experiment 2. The second hypothesis was again not supported.
Implications for decategorization-based interventions in intergroup conflicts and
directions for future research on the role of affect in personalized perceptions of
individuals are discussed.
The contact hypothesis proposes that contact between members of groups in
conflict will reduce prejudice and discrimination by providing an opportunity to
disconfirm mutually destructive stereotypes (Allport, 1954; Amir, 1969; Cook, 1978;
Stephan, 1987). As intergroup contact proceeds, group members come to perceive
each other less as ingroup and outgroup members and more as unique individuals.
Contact reduces prejudice and discrimination by creating a perception of commonly
held interests and humanity between members of the two groups (Allport, 1954). The
decategorization model (Brewer & Miller, 1984; Miller & Brewer, 1986; Miller &
Harrington, 1990) speculates that the mechanism of contact involves changing group
members' perceptions of themselves and of each other. Individuals in an intergroup
contact situation typically discriminate others as ingroup and outgroup members by
focusing on the distinctions that separate people into easily discernible groups. Such
a focus on highly salient group boundaries inhibits the perception of interpersonal
bonds based on common attributes and attitudes between ingroup and outgroup
members. Instead, outgroups and outgroup members are associated with negative
affect (Messick & Mackie, 1989) and threat (Turner, 1975), perpetuating prejudice
and discrimination. If, however, individuals are decategorized as members of
ingroups and outgroups, then they have an opportunity to identify similarities and
become attracted to each other, thereby reducing prejudice and discrimination.
The shift in perceptions is accomplished by shifting attention from group
boundaries to personal attributes. When group boundaries are highly salient during
contact, perceptions of and responses to others are category-based: individuals are
viewed as interchangeable exemplars of their group, exhibiting the physical or
psychological qualities that define it (e.g., a certain age, skin color, or political
ideology). When category salience is low or not salient, perceptions and responses
are based on each individual's personal attributes, which allow one to discover
similarities and commonalities with another on an interpersonal basis. One mode of
perception is category-based, whereas the other personalized (Brewer & Miller, 1984;
Miller & Brewer, 1986). Contact that improves intergroup relations promotes
personalized perceptions between ingroup and outgroup members by shifting
attention from intergroup to interpersonal bases.
The model predicts that personalized interaction will decrease negative
perceptions of the outgroup and increase positive perceptions of it. This occurs when
the personalized perceptual style used during an encounter with individual outgroup
members is applied to other outgroup members (Brewer & Miller, 1988). At the
intergroup level the cumulative effect of personalized perceptions breaks down the
salience of ingroup-outgroup boundaries, removing the basis of negative perceptions
and behaviors against the outgroup. The majority of empirical investigation into the
decategorization model has been concerned with the effects of personalized
perceptions of outgroup members after a decategorization intervention (e.g.,
Bettencourt, Brewer, Croak, & Miller, 1992; Miller, Brewer, & Edwards, 1985;
Marcus-Newhall, Miller, Holtz, & Brewer, 1993). This research measured
perceptions of individuals for whom a decategorization intervention presumably had
reduced the salience of their group membership. One aspect of this research is that
personalized perceptions may have been influenced by individuals' group affiliation.
That is, reactions to individuals may have been based in part on reactions to the group
with which they were affiliated. The work addressed the effects of decategorization
on perceptions of individuals with varying degrees of group affiliation, and is
probably a fair approximation of the state of affairs in actual intergroup contact
situations. However, little empirical work has been done concerning the effects of
3
decategorization on perceptions of individuals in the absence of any group
affiliations. The main purpose of the present research is to study perceptions of
individuals that are more purely personalized or category-based.
One way to study this is to consider the effects of category-based perceptions
as instances of perceptual accentuation due to social categorization, and personalized
perceptions as the absence of a categorization. Accentuation theory (e.g., Campbell,
1956; Deschamps, 1984; Krueger, 1992; Tajfel, 1959) proposes that the perceptual
consequences of organizing a set of elements, either objects or people, into discrete
categories cause a person to view members of each category as being more similar to
each other, and members of different categories as being more dissimilar to each
other, than they objectively are. That is, the process of categorization inherent in
intergroup situations accentuates intra-group similarity and inter-group dissimilarity
(Turner, 1987). The accentuation effect occurs when a discrete categorization scheme
such as ingroup-outgroup is imposed on a series of people who vary continuously
along a judgment dimension. In addition, there must be some systematic relation
between the discrete and continuous dimensions such that certain values on the
continuous dimension predict membership in one of the discrete categories. Skin
colors, for instance, range continuously from very light to very dark. Overlaid on
skin color is membership in various racial groups (e.g., Caucasian, Indian). The
accentuation effect occurs when two individuals who may be relatively similar in skin
color, but who are in opposing racial groups, are seen as more dissimilar to each other
and more similar to the other members of their racial groups, despite objective
similarities between the two people (Tajfel, Sheikh, & Gardner, 1964).
Accentuation theory can be used to extend the decategorization model to
individual group members in the contact situation. In terms of the model, the
accentuation of inter-category dissimilarity and intra-category similarity is important
for two reasons. First, categorizing individuals into ingroups and outgroups promotes
perceptions of dissimilarity, which hinder the discovery of common interests and
humanity between ingroup and outgroup members that is critical to improving
intergroup relations (Allport, 1954). In addition, categorization deindividuates
individual group members, easing normative restrictions on acts of prejudice and
discrimination against individuals (Maslach, 1974; Ziller, 1964; Zimbardo, 1969).
Personalization can squelch the accentuation effect by shifting attention away from
the ingroup-outgroup boundary and onto other dimensions unrelated to ingroup or
outgroup membership. Second, accentuation theory specifically addresses how
individuals are affected by categorization. This enables one to generate hypotheses
about the individual-level effects of category-based and personalized perceptions of
affiliation between individuals.
One effect of categorization is on the degree of perceptual distinction among
individuals when they viewed as members of a category rather than as unique
individuals. According to the decategorization model, category-based interaction
with a collection of individuals usually leads to the perception that they are all very
similar on a number of stereotypic dimensions associated with outgroup membership
(i.e., the outgroup homogeneity effect, see Linville, Salovey, & Fischer, 1986; Mullen
& Hu, 1989; Ostrom & Sedikides, 1992; Park & Judd, 1990 for reviews). When a
collection of individuals is categorized as an outgroup, often there is decreased
perceptual discrimination among them (Hensley & Duval, 1976; Park, Judd, & Ryan,
1991), and increased prejudice by ingroup members toward them (Wilder &
Thompson, 1988). Any new information learned about individual outgroup members
is remembered as stereotype-relevant attributes rather than as attributes unique to
each person (Carpenter, 1993; Ostrom, Carpenter, Sedikides, & Li, 1993). As a
result, there is little individual differentiation among outgroup members.
As an accentuation effect, homogeneity results when the perceptions of
individuals are assimilated to the stereotypical perception of their group (Deschamps,
1984; Tajfel et al., 1964). When applied to an intergroup situation, this suggests that
when individuals are placed into ingroup and outgroup categories, they will be
perceived as more similar to other ingroup members and less similar to outgroup
members, than when they are not categorized. The consequence of this distortion is
that individuals in different groups seem dissimilar to each other, dislike each other
(Marks, Miller, & Maruyama, 1981), and are interpersonally unattractive to each
other (Rosenbaum, 1986; Singh & Tan, 1992), thereby decreasing their likelihood of
forming an interpersonal bond.
During personalized perception, on the other hand, there should be more
differentiation among outgroup individuals because the perceptual basis of the
distortion is de-emphasized (Krueger & Rothbart, 1988). This allows a group
member to be evaluated according to his or her undistorted personal attributes.
Personalized interactions with a number of outgroup members breaks down the
perception of outgroup homogeneity by shifting attention away from group
boundaries and onto individuals' personal attributes (Brewer, 1988; Fiske & Neuberg,
1990). This focus on individuating information weakens the perception of
homogeneity among outgroup members (Judd & Park, 1988; Linville, Fischer, &
Salovey, 1989), enabling individual differences among group members to be noticed
and more individuated perceptions of them to be formed (Ruscher & Fiske, 1990;
Ruscher, Fiske, Miki, & Van Manen, 1991). When individuals are perceived
according to their personalities and in a manner that is not distorted by their category
memberships, there is more opportunity to discover similarities upon which to build
interpersonal bonds, ultimately improving perceptions of the outgroup as a whole
(Miller & Harrington, 1990). Unlike perceptions of the outgroup, which tend to
become more positive after personalization (Bettencourt et al., 1992; Miller et al.,
1985), personalized perceptions of individual outgroup members may not always be
positive. Some can become more negative and some remain unchanged from
category-based perception. Therefore, during category-based perception, the valence
of the outgroup interacts with the valence of an individual's attributes to determine
one's perception of another.
One mechanism underlying this interaction may be a distortion of the
objective perception toward the outgroup stereotype. As an objective baseline, a
personalized impression is based primarily upon a person's unique attributes, with
little or no influence of category memberships (Brewer, 1988; Fiske & Neuberg,
1990). A category-based impression, on the other hand, is dominated by the
individual's group membership, and his or her personal attributes are used to confirm
the membership (Brewer, 1988; Yzerbyt & Leyens, 1991). Compared to the baseline,
the category-based perception is distorted toward the stereotypical perception of the
outgroup. Perceptions may be distorted in a negative direction, or in a positive
direction, depending on the valence of the outgroup perception. Category-based
distortion can be viewed as the result of a process of person perception whereby
people form impressions of others by computing a weighted average of the affective
valence of all of the information one has about another (e.g., Anderson, 1981).
When a perception is based solely on a person's unique attributes, this average value
can be used as a rough approximation of a personalized impression. With the
addition of a group membership as another piece of weighted information, the
average value may be used as an approximation of a category-based impression. That
is, if a group membership is highly salient in one's perception of another, the group
membership will be averaged with the personal attributes in the perception
(Anderson, 1981). The information averaging approach is useful in this research
because it quantifies the components o f perception and allows the distortion
hypothesis to be tested in a series o f directional predictions about people's evaluations
o f categorized and personalized individuals .
The Idiographic Paradigm
In this dissertation, decategorization is discussed in terms of a perceptual shift.
It can also be conceptualized as a model describing a continuum of information
processing modes that differentially evaluates individuals according to their unique
attributes (Brewer, 1988; Fiske & Neuberg, 1990). Movement along the continuum
depends on the amount of attention focused on category-linked attributes relative to
the amount focused on personal attributes (Brewer & Miller, 1988). During category-
based processing of individuals, category-defining attributes are highly salient and
personal attributes are subordinated to them. By definition, all the members of the
category share at least one common attribute, which makes them relatively
undifferentiated from each other. As perceptions become more differentiated,
category-based attributes become less salient and personal attributes more prominent.
Individual differences among group members are recognized, but individuals are still
perceived in terms o f group membership. During personalized processing individual
attributes are highly salient such that individuals are recognized as being unique
according to their individual attributes. They are still associated with the group, but
8
group membership is a relatively unimportant attribute in individuals' perceptions of
each other. Thus the process of personalization is a process that directs attention
away from categorical attributes and instead directs it toward personal attributes.
When category-based attributes are highly salient in an interaction, such as skin color
in a fact-to-face meeting, processing is shallow. When personal attributes are salient
processing is deeper because it involves a larger number of less readily apparent
attributes that require more attention when forming an evaluation (Brewer & Miller,
1984). The result is a perception of another that is complex, providing many
dimensions on which individuals may become attracted to each other (Brewer &
Miller, 1984).
In terms of accentuation theory, decategorizing individuals reduces the
accentuation effect by breaking the correspondence between the more continuous
idiosyncratic classification dimensions, such as the affective valence of set of
personal attributes, and the discrete classification dimension, such as in- or outgroup
membership. Personalization shifts attention away from category boundaries, which
reduces their salience in one person's perception of another and weakens the link
between the values along the evaluation dimension and category membership. As
the link is weakened, individuals are perceived more according to their personal
attributes than according to the category attribute. Group members become more
differentiated from each other as one's perceptions of them become more varied. As
the personal attributes come to dominate the perceptions, individuals are better able to
find similarities between each other on the basis of shared personal attributes.
The category-based distortion of personalized perceptions was examined in
this dissertation using an idiographic paradigm of category-based and personalized
perception (Pavelchak, 1989). In the two experiments reported in this dissertation,
participants view sets of attribute information using either a category-based or a
personalized mode. To induce category-based perceptions, participants are asked to
examine a series of descriptions of people and to guess the academic major of each
person. To induce personalized perceptions, participants are asked to examine the
descriptions in order to form an impression of each person. The paradigm is
idiographic in that rather than arbitrarily imposing a category or affective meaning on
a set of attributes, subjects are allowed to ascribe a unique meaning to each target
person's attributes. This feature of the idiographic paradigm is desirable because it
allows participants' liking and attraction to target individuals to be measured directly
rather than assumed (Pavelchak, 1989). Moreover, the idiographic manipulations
have been shown to reliably induce category-based and personalized perceptions by
making the category salient in the category-based condition, and by making an
individual's personal attributes salient in the personalized condition (Pavelchak,
1989).
It is important to note, however, that the operational definitions of category-
based and personalized processing used in the paradigm differ from the processing
characteristics implied by the decategorization model or that typically occur in
naturalistic intergroup settings. In the decategorization model category-based
processing is shallow, attentionally efficient, and rapid, whereas by contrast deeper,
effortful, and more deliberate processing characterize personalized processing (Fiske
& Neuberg, 1990). The reason for this is that the decategorization model assumes
that category identity is manifestly available and thereby eliminates any need for
further processing, because one has a readily available stereotype of that category.
The idiographic paradigm disrupts this ordinary state of affairs by requiring that one
induce the category from personal elements. Thus, the ordinary processing
10
characteristics are reversed such that category-based processing is deeper and
personalized processing is comparatively more shallow. The purpose of this reversal
is to control for effort justification, longer attention, familiarity, and other features
correlated with deep processing which act to produce more favorable ratings of
individuals under personalized than category-based processing. Therefore, if induced
categories have effects, it is not due to inadequate processing of attribute information.
In other words, this paradigm addresses the effects of salient categories on
perceptions of individuals without relying on the shallow thought and minimal
attention to attribute information that are typical of category-based perception.
Experiment 1
According to the decategorization model, perceptions of outgroup members
become more favorable after personalization (Bettencourt et al., 1992; Marcus-
Newhall et al., 1993; Miller et al., 1985). However, when personalized perceptions
are formed in the absence of any influence from a group membership, personalization
may not uniformly result in more favorable perceptions. The affective valence of an
individual's attributes should be much more influential in perceptions of others after
personalization than when the perceptions are category-based, in which case the
category should be more influential. The overall affective evaluations of individuals
under category-based perception will be dependent upon the affective valence of the
group into which they are placed. If an individual is categorized into a negatively-
perceived group, then they will be viewed negatively. This is typically what happens
when the group is an outgroup (Brewer & Kramer, 1985; Messick & Mackie, 1989).
If the group is viewed positively, then the category-based perception of an individual
will be positive. If the group is viewed more or less neutrally, then category-based
perceptions of group members will be neutral. In the paradigm used in this research,
11
the mean affective value of the academic majors is slightly positive, although some
are negative and some are positive.
The main prediction for Experiment 1 is that there will be an interaction
between perceptual orientation (category-based or personalized) and the valence of
the attribute configurations. Overall, category-based perceptions will produce
evaluations that are less positive for individuals described by primarily positive
attributes than they would be under personalized perception. This is because the
overall mean valence of the categories, according to normative ratings, used in this
idiographic paradigm is only .75 standard deviations above the scale midpoint,
whereas the overall mean valence of the positive attribute configurations are 2
standard deviations above the midpoint. Thus the ratings for categorized positive
targets will be depressed relative to ratings for personalized positive targets.
Similarly, the overall mean valence of the negative attribute configurations are .50
standard deviations below the midpoint, therefore category-based perceptions of
individuals who are described by primarily negative attributes will produce
evaluations that are more positive than personalized perceptions of the same
individuals. The interaction will be driven by a distortion of affective evaluations
away from the mean objective values of the configurations, represented by the liking
ratings in the personalized conditions, toward the category values in the category-
based conditions. In other words, the valence of the induced categories will distort
the effect of the valence of the attributes when considered by themselves and
influence evaluations toward the category value. In the personalized condition there
will be no distortion and the evaluations will approximate the levels predicted from
the normative values of the attributes that describe an individual.
12
In the idiographic paradigm, the target individuals were constructed of
primarily positive or negative attributes so that the positively and negatively
described targets would be easily placed into positive and negative categories in the
category-based condition (Pavelchak, 1989). In the absence of a categorization
manipulation, the targets would be easily differentiated according to the valence of
their attributes. Therefore a main effect for attribute valence on evaluations of
individuals is predicted. Moreover, in the category-based conditions, there should be
a main effect for the valence of the induced categories on the evaluations of the
targets such that targets placed into positive categories should be given more
favorable evaluations than those placed into negative categories. However, the
valence main effects should be evident in only one of the two dependent measures. It
should not appear in the interpersonal attraction measure because it is a composite of
three different scales, two on 7-point and one on a 6-point scale. In order to form this
composite measure of interpersonal attraction, the data for these measures were
standardized within each participant prior to analysis. The result of this
transformation will nullify the pre-existing valence differences in the attributes and
categories. Thus the valence main effects will only be found for the direct measure of
affect (i.e., liking) in the evaluations of the target individuals. Finally, because
evaluations of individuals are based on their attributes and/or category memberships
and not on the perceptual mode alone, there will be no main effect for perceptual
mode on evaluations.
Method
Participants. Thirty-eight members (11 males, 27 females) of the introductory
psychology subject pool at the University of Southern California volunteered for
participation in this experiment in exchange for extra course credit. The median age
13
of the participants was 19 years, with ages ranging from 18 to 26. Participants were
randomly assigned to one of the two between-subjects conditions in the design. They
were run in groups ranging from 1 to 4, and all participants in a given experimental
session received the same perceptual manipulation.
Stimulus Materials. The stimulus materials for Experiment 1 consisted of six
attribute configurations, each of which was presented to participants as an individual
who was identified by their initials. The configuration describing each stimulus
person was composed of four attributes, each taken from Anderson's (1968) listing of
valenced attributes. In Anderson's original paper, the affective scale was a 7-point
scale anchored by 0 = very dislikable and 6 = very likable. In order to conform to the
1-7 scales used in Experiments 1 and 2, 1 was added to each of the original scale
points. The four attributes for each configuration were chosen so that the mean
likability ratings for each configuration was either negatively or positively valenced,
falling about .5 and 1.5 scale points on either side of the neutral midpoint of the scale.
Attributes with extremely negative or extremely positive values were avoided, as
were attributes likely to be unfamiliar to the participants (e.g., congenial). The sets of
four attributes were then randomly ordered within each configuration to produce three
positive and three negative configurations. Finally, six 3-letter initial groups were
generated and randomly assigned to a configuration. Initials were used instead of
proper names to label the configurations to avoid implying a sex category for any of
the configurations. The labeled configurations along with the Anderson values and
the mean valence of each configuration are shown in Table 1.
14
Table 1
Stimulus Configurations and Associated Valences
Positive Configurations
ev
imaginative 5.92
cultured 5.50
ambitious 5.84
emotional 3.83
M = 5.27(0.98)
FEA
efficient 5.82
well-read 5.86
methodical 4.25
reliable 6.27
M = 5.55 (0.89)
RTL
idealistic 4.84
attentive 5.50
open-minded 6.30
individualistic 5.67
M = 5.58 (0.60)
Negative Configurations
JMJ
irritating 2.18
well-read 5.86
argumentative 3.27
authoritative 3.74
M = 3.74(1.57)
NAL
talkative 4.52
argumentative 3.27
opinionated 3.57
careless 2.40
M = 3.44 (0.87)
NMM
emotional 3.83
neurotic 2.52
sensitive 4.58
unlucky 3.80
M = 3.68 (0.85)
Note. The original affective values in Anderson (1968) were on a 0-6 scale, with 0 =
very dislikable and 6 = very likable. The values were shifted 1 point to conform to
the 0-7 scales used in Experiments 1 and 2. Standard deviations are in parentheses.
15
The mean affective value of academic majors was obtained from a measure
inserted into a premeasures packet given out to the introductory psychology subject
pool at the beginning of the semester in which Experiment 1 was conducted. This
data was collected 4 weeks prior to the start of data collection for Experiment 1. The
measure asked respondents to rate the affective valence of 49 academic majors
offered at USC on a 7-point fully labeled Likert scale, with 1 = very negative and 7 =
very positive. Data from 246 respondents indicate that the mean affective valence of
the 49 majors was 4.65 ( SD = 0.53), with means ranging from 3.51 to 6.02. This
suggests that the population from which the participants in Experiment 1 were drawn
perceive academic majors to be, on average, slightly positive. In the category-based
conditions, participants were expected to idiographically induce positive or negative
categories for the targets based upon their personal attributes. In the idiographic
procedure, individuals described by primarily positive attributes should be assigned
primarily majors, whereas those who were described by primarily negative attributes
should be assigned primarily negative majors. Therefore, the idiographic induction of
a positive or a negative major served as a within-subjects manipulation of category
valence in the category-based condition.
Participants in both between-subjects conditions were presented with the three
negatively valenced and three positively valenced configurations, constituting the two
levels of the attribute valence within-subjects variable. The stimulus configurations
were presented to subjects in a column at the top of each page in the response packet,
with the three initials heading the column and the four attributes below. (See
Appendix A for the complete task instructions and dependent measures packet for
participants in the personalized processing condition, and Appendix B for the
complete task instructions and dependent measures packet for participants in the
16
category-based processing condition.) Four dependent measures were arranged below
the configuration with liking appearing first, perceived closeness second, perceived
similarity third, and perceived commonality last. This ordering remained constant
across all conditions. In the category-based processing condition, a blank line was
inserted between the configuration and the first measure for participants to write what
they judged each target's major was and if that major was the same as their own. This
information was intended to serve two purposes. First, it would be the means by
which the valence of the induced categories was recorded. Second, it would bolster
the cover story. The pages were then randomly ordered in the response packet for
each experimental session, with the restriction that the first page of the packet
contained the processing instructions for that condition, and that the last page
contained a series of demographic items.
Dependent measures. The first dependent measure used in this experiment
was a general measure of interpersonal attraction, composed of measures of perceived
closeness, perceived similarity, and perceived commonality. The second measure,
liking, was included as a gauge of the affective valence of a participant's perception of
a target. It was included as a test of the distortion hypothesis because it is analogous
to the favorability measure used to obtain normative ratings of Anderson (1968) trait
list. The perceived liking, similarity, and commonality measures were answered on
fully labeled Likert-type scales, with 1 = "very dissimilar/very little in common" and
7 = "very similar/very much in common." The liking scale was reversed such that 1 =
"very likable" and 7 = "very unlikable." The perceived closeness measure was
adapted from a pictorial ordinal-level scale designed by Aron, Aron, & Smollen
(1992) for use in evaluating interpersonal relationships. The scale featured six figures
numbered 1-6. Each figure consisted of a pair of circles with equivalent diameters,
17
enclosed within a rectangle. Each pair of circles was labeled, with the leftmost of the
pair labeled "You" and the rightmost labeled with the initials of a target (e.g., "EJV.").
The geometric centers of the pairs of circles were various distances apart (3.0, 2.5.
2.0, 1.5, 1.0, and 0.5 cm) and were ordered such that the pair with the most distance
between centers (3.0 cm) corresponded to 1 on the scale and the pair with the least
distance (0.5 cm) corresponded to 6 on the scale. The instructions on this item asked
"How close do you feel to this person?" and defined closeness as "a sense of sharing
or interconnectedness between you and this person...the closer the two figures are to
each other, the closer or more interconnected they are. On the other hand, the further
they are apart, the less close or interconnected they are" (Aron et al., 1992).
Procedure. Participants were met at the door of the laboratory as they arrived,
ushered to a seat, and if there were multiple participants in a session, were asked not
to speak among themselves prior to the start of the session. To begin the session, the
experimenter introduced herself and informed the participants of their rights.
Participants were told that this experiment was concerned with measuring how people
respond to other people, based on minimal amounts of information given about them.
Additional verbal instructions indicated that: a) they would be given a packet
containing descriptions of six individuals along with four types of items for indicating
how they responded to each individual; b) each person would be identified only by
their initials and be described by four traits; and c) each of the traits assigned to a
target person had been provided by someone who knew them well and each trait was
equally important in describing that person.
In the category-based perceptual condition, participants were asked to read
through each person's traits and to try to guess their academic major. They were
informed that a list of majors offered at USC would be printed on the front page of
18
the packet to help them, but that they were not restricted to that list in guessing a
major. Participants were instructed that they could use a major more than once, but
that they were to assign only one major per person. In the personalized perceptual
condition, participants were asked to read through the traits that described each
person and to try to form a detailed impression of them. Participants were instructed
to imagine being in a conversation with each person, learning four traits about them,
and then forming an impression of them. All participants were then told that after
they had completed the task for a person, they were to complete the four responding
styles questions about that person, and then to go onto the next person until they had
finished responding to all six people, then to complete the demographic information
requested on the last page of the packet.
The experimenter then asked the participants if they had any questions about
the procedure, and distributed the packets to them. Participants were instructed to
turn their packets over when they had finished and to wait quietly until all other
participants were finished. When all participants in the session were finished, the
experimenter collected their completed packets, probed them for any suspicions about
the procedure and what they suspected the purpose of the experiment was, then
thoroughly debriefed them as to the nature of the manipulations and the expected
results. Finally, the experimenter elicited participants' promises that they would not
reveal the nature of the manipulations or hypotheses of the experiment were, thanked
them, and assigned them credit for their participation.
Results
The idiographic manipulation o f category valence was intended to allow
participants in the category-based conditions to induce positively valenced majors for
targets described by positive attributes and negatively valenced majors for negatively
19
described targets. The majors that participants assigned to each of the 3 positive and
3 negative targets were examined and the mean valence for the majors was calculated
as a check on the idiographic manipulation. For the 3 positive majors the means are
positive: EJV (M = 4.96, SD = 0.33), FEA (M = 4.59, SD = 1.15), RTL (M =4.81,
SD = 0.43). Contrary to predictions, however, the mean valences assigned to
negative targets were also positive: JMJ (M = 4.67, SD = 0.32), NAL (M = 4.77, SD
= 0.30), NMM (jM = 4.90, SD = 0.31). Therefore, the idiographic manipulation of
category failed to produce negative categorizations for the negative targets.
Before performing the main statistical analyses, the liking item was reverse
scored and the data was examined for outliers and extreme cases. None were found.
Interpersonal attraction. The data for the interpersonal attraction measures
(closeness, similarity, commonality) were first standardized, then added together for
each target person to form a composite score for each individual (alpha = 0.81). The
composites from the three negative targets (JMJ, NAL, NMM) were then added, as
were the three positive composites (EJV, FEA, RTL) to yield a composite
identification score for the negative and the positive targets.
The composite scores were then entered into a 2 X 2 X 2 MANOVA with sex
of participant and perceptual condition as the between-subjects variables and valence
of the trait configurations as the within-subjects variable. Sex showed no main effect
(p < .77), no interaction with perceptual condition (p < .11), no interaction with
valence (p < .87), and no three-way interaction (p < .23). As predicted, perceptual
condition had no main effect on participants' evaluations of interpersonal attraction to
the targets (F (1,36) = 0.00, ns.) and there was no main effect for the valence of the
configurations (F (1, 36) = 0.09, ns). More importantly, a significant interaction
between valence and processing condition was obtained (F (1,36) = 6.75, p < .01).
20
The means and standard deviations for the unstandardized interpersonal attraction
composite pooled across sex of participant are presented in Table 2. Personalized
targets who were described by primarily positive attributes were more attractive (M=
43.95, SD = 5.13) than categorized targets described by positive attributes (M =
38.90, SD = 4.31), t(18) = 2.27, g < .025. Personalized targets who were described by
primarily negative attributes (M = 25.21, SD = 5.51) were less attractive than
categorized targets described by negative attributes (M = 27.05, SD = 8.58), but not
significantly so, t(18) = -0.55, ns.
Table 2
Unstandardized Attraction Means and Standard Deviations. Experiment 1
Valence of Target
Perceptual Condition Negative Positive
a a
Category-based 27.05 (8.58) 38.90 (4.31)
a b
Personalized 25.21 (5.51) 43.95 (5.13)
Note. Standard deviations are in parentheses. Higher scores mean higher levels of
attraction. Means in the same column that have different superscripts differ from
each other at g < .025.
Liking. The distortion hypothesis was examined with the liking data. First,
the liking data for the three negative targets (JMJ, NAL, NMM), and the data for the
three positive composites (EJV, FEA, RTL) were added to form a composite liking
score for the negative and the positive individuals. The composite liking scores were
21
then entered into a 2 X 2 X 2 MANOVA with sex of participant and perceptual
condition as the between-subjects variables and valence of the trait configurations as
the within-subjects variable. Sex showed no main effect (g < .28), no interaction with
perceptual condition (g < .25), no interaction with valence (g < .81), and no three-way
interaction (g < .29). As predicted, the manipulation of perceptual orientation
produced no main effect on liking (F (1,36) = 0.30, ns). There was a significant main
effect for valence of configurations (F (1,36) = 136.02, g < .001), and a significant
interaction between perceptual condition and valence (F (1,36) = 7.92, g < .01). The
means and standard deviations for the liking measure pooled across sex of participant
are presented in Table 3.
Table 3
Liking Means and Standard Deviations. Experiment 1
Valence of Target
Perceptual Condition Negative Positive
Category-based 10.95 (2.78)3 16.32 (2.11)
Personalized 9.58 (2.91) 17.68 (2.58)
Baseline values 10.60 (2.75) 16.40 (2.47)
Note. Standard deviations are in parentheses. Higher scores mean higher levels of
liking. Means in the same column that have different superscripts differ from each
other at g < .0125.
22
Accentuation theory predicts that category-based perceptions of individuals
will be distorted toward the valence of one's affective reaction to a category and away
from the valence of the reaction to an objectively perceived, uncategorized individual.
Personalized perceptions, on the other hand, should approximate the levels derived
from the ratings of the mean of the attributes calculated from the Anderson (1968)
data. The personalized predictions were tested in the liking data with a Bonferroni t
-test. The mean Anderson scores of the positive targets is 16.4 (SD = 2.47), and is
10.6 (SD = 2.75) for the negative targets. Two comparisons were made, and the
significance level for each test was set at .025. As predicted, in the personalized
condition, positive targets (M = 17.68, SD = 2.60) were not perceived significantly
differently from the composite objective value, t (18) = 1.27, ns, nor were negative
targets (M = 9.58, SD =2.91) significantly different, t (18) = .93, ns.
The full test of the category-based distortion hypothesis was designed as a 2
(valence of attribute configurations) X 2 (valence of category) X 2 (perceptual mode)
mixed model MANOVA. It is predicted that a significant main effect of category
valence and an interaction between valence of induced major and perceptual
orientation will support the distortion hypothesis. The data for this analysis were
organized by first examining the valence of the major each participant induced for
each target in the negative attribute and positive attribute category-based conditions.
The responses for a given target in the positive attribute and negative attribute
conditions would then be paired with the identical targets for a randomly selected
participant in the corresponding personalized conditions. For example, in the
23
category-based, positive attribute condition, the target with the most positive major,
according to the normative data collected on the majors, was paired with the identical
target in the personalized, positive attribute condition.
The full test was not possible, however, because no participant induced a
negative major for any positive or negative attribute targets. Only positive majors
were induced for all targets for all participants. Therefore, the only comparison
possible was between negative targets in the category-based and personalized
conditions. The results of the single comparison did not, however, support the
category-based distortion hypothesis. Targets described by negative attributes and
assigned positive majors (M = 10.95, SD = 2.78) were not liked more than negatively
described, personalized targets (M = 9.58, SD = 2.91), t(18) = 1.01, ns. Therefore,
although the direction of the means is consistent with the prediction of category-based
distortion, the prediction is not strongly supported.
Given that the participants in this experiment did not assign negative majors to
negative targets, thereby preventing the planned test of the distortion hypothesis, a
series of post-hoc analyses were conducted to detect a category-based distortion effect
in the liking data. The first of these analyses involved examining the induced
category choices each participant made for the positive and negative targets in the
category-based condition. The design of this analysis was to enter the data for the
targets that most closely approximated the hypothesized effects of the idiographic
categorization manipulation. Thus, of the induced majors, the most and least positive
were selected and matched with the identical targets in the personalized condition.
For the 3 positive targets, the liking data was retained for the target who was assigned
the most positive major (according to the normative ratings of the majors), and for
the target assigned the most negative major. For the negative targets the data for the
24
target assigned the most positive major and the one assigned the most negative major
were retained. These targets were then matched with the identical targets from a
randomly selected participant in the personalized condition. This data was then
entered into a 2 X 2 X 2 MANOVA.
If the induced categories distort liking evaluations away from personalized
levels, then positive targets assigned the most positive majors and positive targets
assigned the least positive majors will be liked more and less, respectively, than the
personalized counterparts. Likewise, negative targets assigned the most positive
majors and negative targets assigned the least positive majors will be liked more and
less, respectively, than the personalized counterparts. This will be indicated by an
interaction between the valence of the categories and perceptual orientation. In
addition, there will be a main effect for category valence such that the targets placed
into the most positive majors will be liked more than those placed into the least
positive majors, and a main effect for target valence such that positive targets will be
liked more than negative targets. There will be no main effect for perceptual
orientation, no interaction between target valence and perceptual orientation, and no
3-way interaction.
The target main effect was significant such that positive targets (M = 5.60, SD
= 1.39) were liked more than negative targets (M = 3.44, SD = 1.58), F (1,38) =
147.12, p < .001. The 2-way interaction between target valence and perceptual
orientation was also significant, F (1,38) = 14.53, p < .001, with personalized positive
(M= 5.84, SD = 1.11) and categorized positive (M = 5.70, SD = 1.45) targets being
liked more than their personalized negative (M = 3.67, SD = 1.68) and categorized
negative (M = 4.07, SD = 1.63) counterparts. No other effects were significant.
25
A second post-hoc analysis involved examining both positive and negative
targets in the category-based condition and selecting the positive target and the
negative target who were assigned the most positive major of the 3 targets in each
target valence condition. The 2 targets were then matched with their counterparts
from a randomly selected participant in the personalized condition. The liking data
for these targets were then entered into a 2 (target valence) X 2 (perceptual
orientation) MANOVA. If placing targets into a positive major distorts liking away
from the personalized levels, then there will be a main effect for perceptual
orientation such that categorized targets will be liked more than personalized targets.
There will not be a main effect for target valence, nor will there be an interaction.
Categorized targets (M = 4.48, SD = 1.57) were not liked more than
personalized targets (M = 4.65, SD = 1.50), F < 1.0, nor was there an interaction, F <
1.0. There was, however, a significant main effect for target valence such that
positive targets (M = 5.55, SD = 1.45) were liked more than negative targets (M =
3.60, SD = 1.58), F (1,38) = 56.58, p < .001.
A third post-hoc analysis involved examining the standard deviations of the
categorized and personalized targets from the mean level of liking for each attribute
valence condition. If the addition of category information to an individual's
description distorts liking away from the personalized levels, in which the category
information is absent, then there will be a larger mean standard deviation in the
category-based conditions than in the personalized conditions. This is because
positive categories will distort positive personalized targets upward (more liking) and
negative categories will distort positive personalized targets downward (less liking),
relative to the uncategorized targets. A similar effect will occur with negative targets
26
such that positive categories will distort negative personalized targets upward and
negative categories will distort negative personalized downward. In other words,
positive and negative categories will cause the mean liking ratings for the targets to
be more variable relative to those for uncategorized, personalized targets.
The data for this analysis were generated by computing the standard deviation
of liking from the levels predicted by the baseline Anderson ratings for targets in the
category-based and personalized conditions. The positive target assigned the most
positive major and the positive target assigned the most negative major were selected
and matched with the identical targets for a randomly selected participant in the
personalized condition, and the standard deviations from the baseline levels of liking
were computed for them. Similarly, the negative target assigned the most positive
major and the negative target assigned the most negative major were matched with
the identical targets from a randomly selected participant in the personalized
condition and the standard deviations from the baseline levels were computed. The
matched sets of standard deviations were then compared for each target valence
condition.
The predictions were not supported. Directionally, the standard deviations of
the liking ratings for categorized positive targets (M = 1-08, SD = 1.00) were more
variable than the standard deviations for personalized positive targets (M = 0.83, SD
= 0.71), but this difference was not significant t( 19) = 0.96, ns. For the negative
targets, again, the standard deviations for the categorized targets (M = 1.36, SD =
1.16) were more variable than the personalized targets (M = 1.16, SD = 0.88), but
again this difference was not significant t(19) = 0.53, ns.
2 7
Discussion
The evidence from the interpersonal attraction data support the main
prediction that the affective valence of a target individual's attributes interacts with
the mode of perception to determine the level of interpersonal attraction. The
perceptions can change for the better when the mean valence of the personal attributes
are more positive than the mean valence of the induced categories, but they can
change for the worse when the mean valence of the attributes are less positive than
the mean valence of the induced categories. The pattern of the unstandardized means
clearly shows that the effects of personalization on perceptions of individual group
members differ from the pattern found in past research in which personalized
perceptions of individuals may have been influenced by their group membership. The
data from this experiment, then, show that personalization does not always result in
more favorable perceptions of individuals.
The available evidence from the liking data do not, however, provide strong
support for the distortion hypothesis. According to the decategorization model,
personalization shifts the basis of one's perception of another individual from group
membership to personal attributes. Personalized perceptions are primarily based on
the affective valence of the individual’ s personal attributes. Category-based
perceptions of an individual are primarily based on the affective valence of the group
in which the individual is a member. The addition of the group membership
information in effect distorts the personalized perception toward the group value. The
obtained means for negatively described targets who were given positive majors do
show some shift away from the levels predicted for personalized targets and toward
the predicted category-based levels, but this shift is not significant. In addition, three
28
different post-hoc analyses failed to furnish strong support for the distortion
hypothesis.
One explanation for the lack of support for the distortion hypothesis may be
that the categories into which targets were placed did not exert a strong influence on
the perception of the attributes. Participants in the category-based conditions may
have considered the category label as a piece of information describing a target
individual on par with the attribute information. When a person is perceived as a
member of a social group, his or her group membership is the characteristic that
dominates perceptions of him or her and all other attributes are seen as supporting the
valence of the category (Brewer, 1988; Fiske & Neuberg, 1990). Thus in a category-
based perception, the category label is the piece of information that has the most
weight among all relevant pieces of information that describes the individual and
exerts the most weight in the perception of them (Anderson, 1981).
In Experiment 1, participants may not have weighted the category label any
more than the four attributes in their perceptions of the target individuals. That is, the
idiographic manipulation of category-based perception may have influenced
participants to place an equivalent weight on the academic major of a target, relative
to the attribute information. In the procedure participants are given the attribute
information for each target individual at the top of each page in the packet of
measures, whereas the list of majors appeared on a separate page at the beginning of
the packet. The arrangement of category and attribute information may have caused
participants to focus more attention on the attribute information than on the category
information. This would mean that the category information was of equal or lower
salience than the attribute information in participant's perceptions of the targets. The
lowered salience of the category information may therefore have caused it to be
29
weighted less than the attribute information, reducing its ability to distort perceptions
away from the objective levels.
A second experiment was conducted in order to correct the potentially weak
manipulation of category salience in the category-based perception conditions. The
idiographic paradigm was again employed, but it was altered to induce participants to
place greater weight on the academic major categories in the category-based
perception conditions.
Experiment 2
In Experiment 1, participants in the category-based processing condition may
not have placed more weight on the category membership than on the other attributes
of each target they viewed. This may have been due to the low amount of thought or
attention participants had to devote to the categorization task, relative to the
personalization task. In the category-based conditions, participants were asked to
categorize each of 6 target individuals into an academic major based on the each
target's attributes. This may have been relatively easy given that the participants,
primarily freshmen and sophomores, interact with other students from different
majors in introductory-level courses every day. They likely had a number of
prototypical majors readily available in memory into which they could place the
targets. Furthermore, they were provided with a list of 24 majors to refer to if they
were somehow not able recall any majors. The focus of their attention, then, was on
the attributes o f each target, which they used as the basis for categorization, rather
than on the categories themselves. Compared to the participants in the
personalization conditions, who were instructed to focus exclusively on the attributes,
participants in the category-based conditions may have weighted the category
information equally or even less than the attribute information. Therefore, the non-
30
significant distortion effect in the category-based conditions may be the result of a
weak manipulation of category salience.
Experiment 2 was designed to strengthen the salience of the category
information in the perceptions of the target individuals in the category-based
conditions. The design of Experiment 2 is a conceptual replication of Experiment 1.
Participants were instructed to read a series of 6 attribute configurations, 3 negative
and 3 positive, and to either guess the likely academic major of each or to form an
impression of each. The procedure was changed, however, to require participants
engaging in the categorization task exert more attention than participants in the
personalization task. The increased thought or salience in the direction of the
category membership of the targets was intended to make the weight of the category
information higher than the attribute information in the perception of a target. Two
features of Experiment 2 were designed to increase the salience of the academic
major categories in the category-based perceptions. First, participants were not
provided with a list of majors to which they could refer as they categorized the
targets. Second, the attribute information for each target was removed from
participants' sight before they completed the interpersonal attraction and liking
measures. This experiment also differed from Experiment 1 in that two measures
were added to the procedure to serve as manipulation checks on the relative amount
of thought participants exerted in the category-based and personalized conditions.
These measures are the latency at which participants complete the task for each target
(guess a major or form an impression) and the number of emergent attributes
generated for each target individual. These measures will be described in more detail
in the method section.
31
The predictions for this experiment parallel those made for Experiment 1.
Namely, there will be an interaction between perceptual mode (category-based or
personalized) and the valence of the attribute configurations. Category-based
perceptions will produce evaluations that are less positive for individuals described by
primarily positive attributes than they will be under personalized perception.
Category-based perceptions of individuals who are described by primarily negative
attributes will produce evaluations that are more positive than personalized
perceptions of the same individuals. For the interpersonal attraction measure, there
will be no main effect for perceptual condition or valence of the configurations, but
there will be a significant interaction between the two. For the liking measure, there
will be a significant distortion of evaluations away from the mean objective values of
the configurations toward the mean category value in the category-based condition.
In the personalized condition there will be no such distortion. In addition, the
manipulation checks will indicate that participants in the category-based perceptual
conditions exert more thought placing the target individuals into academic majors
than participants in the personalized conditions exert forming impressions of the
targets.
Method
Participants. Forty-seven members (16 males, 31 females) of the introductory
psychology subject pool at the University of Southern California volunteered for
participation in this experiment in exchange for extra course credit. The median age
for the participants was 19 years, with ages ranging from 18 to 22 years. Participants
were randomly assigned to one of the two between-subjects conditions in the design
and were run individually.
32
Materials. The stimulus materials for this experiment were identical to the six
attribute configurations used in Experiment 1. However, the presentation of the
stimulus materials was changed from Experiment 1 to allow response latency data to
be collected. Rather than presenting the participants with the configurations on the
same page as the dependent measures, they were presented on a computer monitor in
the same format as in Experiment 1 (i.e., the four attributes were arranged in a column
headed by a 3-letter set of initials).
The presentation of the six stimulus configurations was controlled by a
computer. Before each experimental session, the computer randomly generated the
presentation order for the six configurations. Participants were able to control when
the computer presented each configuration on the monitor with two buttons, marked
"7" and "enter", on an 8-button keypad. The latency data was collected via this
keypad, the interpersonal attraction and liking data was collected with a 6-page
response packet, and the recall data was collected on a single sheet of paper described
in more detail below. Demographic data was collected at the end of the filler task.
Dependent measures. Three types of dependent measures were used in this
experiment. The first was the interpersonal attraction and liking measures used in
Experiment 1. These measures were presented in a packet of pages, with each page
headed by one of the six sets o f initials. The four measures were in the same order on
each page as in Experiment 1: liking, perceived closeness, perceived similarity,
perceived commonality. The pages in the packet were arranged in the same order as
the corresponding attribute configurations were presented.
The second type of measure used in this experiment was a participant's
response latency for each of the six attribute configurations. In order to check the
relative amount of thought participants exerted as they guessed majors or formed
33
impressions, a measure of cognitive effort was needed. One index of the relative
amount of cognitive effort involved in a given task is the amount of time it takes a
person to complete the task (Fiske, 1980). All else constant, a task that takes longer
to complete should also require more cognitive effort to complete than one that takes
a shorter amount of time. Moreover, the more thought directed at a set of information
during person perception, the more salient it is likely to be in that perception (Fiske &
Taylor, 1984). In this experiment a participant's response latency either to guess a
target's likely academic major or form an impression of them was taken as a measure
of the amount of thought required to complete the experimental task. Response
latency was defined as the amount of time in milliseconds between stimulus onset and
when the participant stopped the presentation of that target by pressing the
appropriate button on the keypad. The data collection routine of the computer
program had a default setting to terminate the presentation of the stimulus
configurations after 45000 ms. If any participants did not stop the presentation by
pressing the appropriate button on the keypad and they did not begin filling out the
identification measures for that target, then they were assigned a 45000 ms latency for
that trial. The exception to this rule was that the data for participants who were
obviously not engaged in the task for a given target (e.g., filling out the measures)
was dropped and replaced with the mean latency for the rest of the targets.
The third type of measure used in this experiment was designed as a post-hoc
manipulation check on amount of thought expended in a perceiving a person. The
more thought expended in perceiving a person, the more well-developed or complex
that perception will be (Linville et al., 1989; Ostrom et al., 1993). The measure of the
complexity of perceptions involved the number of emergent attributes derived from a
free-recall protocol collected from each participant near the end of the experimental
34
session. An emergent attribute was defined as an adjective or phrase used to describe
a particular target that was not contained in the original description of that target.
Emergent attributes have been used in cognitive psychology to explain how a concept
that results from the combination of two or more parent concepts contains not only
attributes that are inherited from the parent terms, but also novel attributes that
emerge from their conjunction to make it coherent (Cohen & Murphy, 1984;
Hampton, 1987). In social cognition research emergent attributes are thought to make
a set of attributes describing a person cohere into something sensible, almost by
filling in the gaps between attributes inherited from disparate social categories (Asch
& Zukier, 1984; Hastie, Schroeder, & Weber, 1990; Kunda, Miller, & Claire, 1990).
For example, if a person is a member of two seemingly disparate social categories
such as "Communist" and "investment banker", the attribute "greedy" may be invoked
into the representation to account for the leftist’s capitalistic tendencies. Applied to
the idiographic paradigm used in this research, the notion of emergent attributes may
be used as a gauge to measure the degree to which the perception of an individual is
based on high or low amounts of thought. In the idiographic paradigm, emergent
attributes are more likely to be contained in a participant's category-based perception
of a target individual than in their personalized of the same target. This is expected
because the category-based perceptions are designed to require more thought than
personalized perceptions. Note that in the world outside of the laboratory, category-
based perception is likely to require very little thought relative to personalized
perception. In an intergroup contact setting, the group memberships of individuals
are signaled by readily apparent cues such as skin color, gender, language spoken, and
uniform color. Ordinarily, personalization requires that attention be shifted away
from the readily apparent cues to more subtle personal attributes (Brewer, 1988;
35
Miller & Harrington, 1990). This shift is likely to require much more thought relative
to category-based perception of individuals. However, in Experiment 2, category-
based perception was deliberately made to require more thought than personalized
perception. Thus, more emergent attributes will be found in recall protocols from
subjects in the category-based conditions than in those from subjects in the
personalized conditions.
(See Appendix C for the affiliation and recall dependent measures packet for
participants in the personalized processing condition, and Appendix D for the
dependent measures packet for participants in the category-based processing
condition.)
Procedure. To begin the session, participants were met at the door of the
laboratory by the experimenter, ushered in, and seated. After introducing himself and
informing the participants of their rights, the experimenter explained that the
experiment was concerned with the processes that underlie how people perceive and
then respond to each other. The experiment would examine three different aspects of
these processes, and consequently would have three parts. The first part of the
experiment would focus on how people perceive and respond to people whom they
had never before and would be computer controlled. The experimenter then handed
the participant an introductory information sheet to read while he left, ostensibly to
start the computer program. The sheet of paper contained the perceptual condition
instructions and instructions about how to use the responding apparatus.
For participants in the category-based condition the instructions read:
In this first part of the experiment, you will be asked to view 6
individuals' descriptions on the screen in front of you. You will see a
series of 4 traits describing a single person who is identified only by
their initials. Each one of these characteristics has been assigned to the
person by someone who knows them well, and each o f the
characteristics is equally important in describing the person. Your task
is to read through the traits and to try to guess each person's major.
You can assign a major to more than one person, but please assign
only one major per person.
For participants in the personalized condition, the instructions were the same except
that the task asked them:
...to read through the traits and to form an impression of them [the
target individual]. You should think in terms o f being in a
conversation with the person and learning four traits about them. After
learning the traits, what is your impression o f them? What is this
person like? Try to form a detailed impression of them in your mind.
Participants in both conditions were given identical instructions on the use of the
responding apparatus. The instructions read:
In order to ensure that you are able to consider the people at your own
pace, you will control the presentation o f the individuals. The
computer will prompt you with an instruction to press the "enter"
button on the keypad when you are ready to view an individual's
description. When you are ready, press that button and the person's
description will come on the screen. As soon as you have guessed that
person's major [or formed an impression of them], press the "7" button
to halt the presentation. The computer will then instruct you to fill out
the responding styles items for that person. After you have finished
the items for that person, you will press the "enter" button and the
computer will generate the description for the next person. When the
description is ready, the computer will ask you to press the "enter"
button to see the next person's description. There will be 6 individuals
whom you will be responding to. After you have finished the sixth
person, the experimenter will return to begin the second part of the
experiment.
While the participants were reading the instructions, the experimenter turned
on the video equipment used to monitor the participant during the session and turned
on the presentation and data collection program. After the participant finished
reading the instructions, the experimenter summarized the main points of the
instructions. Participants in either condition were treated the same except that in the
category-based condition, the experimenter emphasized that participants were to "try
to guess each person's likely major, even though there is no real right or wrong
37
responses in this part task" and that the main interest of the experiment was "in how
you come about your answer, rather than in the answer itself." In the personalized
condition, the experimenter's instructions instead emphasized the suggestion that
participants may wish to pretend that they had engaged in a conversation with each
person and learned four attributes about them. It should be noted that the
experimenter was not blind to conditions. However, it was necessary for the
experimenter to know which condition each participant was in order to set up the
computer program and to assemble the response materials correctly according to
condition.
Next, all participants were given the keypad and a clipboard with the
identification measures for each target. Finally, all participants were reminded that
they were to stop the presentation of an individual's description as soon as they had
guessed that person's major (or fully formed the impression of them). After the
experimenter was satisfied that the participant understood all of the instructions, he
left the room and returned to the main computer and monitoring station adjoining, but
not visible to the participant room.
While the participant was completing the first part of the experiment, the
experimenter monitored their progress on a video screen. If the experimenter
observed a participant failing to follow directions on a trial (e.g., filling out the
identification measures for that target before stopping the presentation), he noted it
and stopped the participant before the next trial occurred to remind them to press the
"7" button before filling out the measures.
After the participant had finished the first part of the experiment, the
experimenter returned to introduce the second part of the experiment. He explained
that whereas the first part of the experiment was concerned with how people perceive
38
and respond to people they had never met before, the second part was concerned with
how people perceive and respond to other people in general. The participant was then
given the Attributional Complexity Scale (Fletcher, Danilovics, Fernandez, Peterson,
& Reeder, 1986), which was described as a person perception questionnaire with a
few demographic items at the end. This scale was meant only to serve as a distracter
task to prevent any incidental rehearsal of the attributes presented in the first part of
the experiment, and was not included in the analyses. The experimenter then left the
room, telling the participant that he would return in a few minutes to see if they had
finished. After the participant had finished the questionnaire, the experimenter
returned to introduce the final part of the experiment.
The experimenter collected the questionnaire and gave the participant a sheet
of paper on which to write their protocols. He explained that whereas the first two
parts of the experiment were concerned with how people actively perceived and
responded to others, the last part was concerned with the role one's memory plays in
how people perceive and respond to each other. Participants were instructed to write
a description for each of the people they had viewed in the first part of the
experiment. They were asked to begin each description with the person's initials, but
that if they could not remember the initials, they were begin each with a number (e.g.,
"Person 1"). They were also instructed not to be concerned with spelling or using
proper grammar in the descriptions. The response sheets were headed with a
summary of the experimenter's instructions, followed by a series of lines on which
participants were to write their descriptions. The experimenter left the room and
announced that he would return in 6 or 7 minutes to see if the participant had finished.
After 7 minutes the experimenter returned and asked if the participant needed
more time to finish. If they did, he left and returned after 4 or 5 minutes and asked
39
again. If they did not need any more time, he collected their protocols, thoroughly
debriefed the participants, gave them credit for their participation, and dismissed
them.
Results
The idiographic manipulation of category valence employed in Experiment 2
was intended to allow participants in the category-based conditions to induce
positively valenced majors for targets described by positive attributes and negatively
valenced majors for negatively described targets. The majors participants assigned to
each of the 3 positive and 3 negative targets were examined and the mean valence for
the majors was calculated as a check on the idiographic manipulation. For the 3
positive majors the means are positive: EIV (M = 5.06, SD = 0.41), FEA (M =5.15,
SD = 0.99), RTL (M = 4.99, SD = 1.13). Contrary to predictions, however, the mean
valences assigned to negative targets were also positive: JMJ (M = 5.01, SD = 0.52),
NAL (M = 4.98, SD = 0.31), NMM (M = 5.22, SD = 1.21). As is Experiment l,th e
idiographic manipulation of category failed to produce negative categorizations for
the negative targets.
Prior to the main performing the main statistical analyses, the liking item was
reverse scored and all of the data were screened. Two participants' data were dropped
due to a pattern of extreme responses. Both participants had to be reminded more
than two times to press the "7" button on the keypad before filling out the measures
for the targets. On the identification measures, both gave responses that were 3
standard deviations higher than the mean for 2 of the 3 negative targets, and gave
responses that were 2 standard deviations below the mean for 1 of the 3 positive
targets. One additional participant's data was dropped because he repeatedly failed to
40
follow the experimenter's instructions during the first part of the session. The final
data set contained the responses from a total of 44 participants.
Latency data. The latency data were collected in the first part of the
experiment. The data consisted of the participant's response latency from stimulus
onset to offset for each of the six targets that were presented. Subjects were closely
monitored by the experimenter during this part of the procedure. If a participant did
not follow the directions for this part of the procedure (i.e., they began filling out the
measures for a particular target without first stopping the presentation of the
stimulus), the latency for that trial was replaced by the mean latency for the entire
series of trials, excluding the faulty trial. Eleven participants (6 in the category-based
condition, 5 in the personalized condition) had to be reminded to press the "7" button
on the keypad before they began filling out the measures for a target. If a participant
produced more than two faulty trials, then the entire set of latency data for that
participant was dropped from analysis. Two participants' data from the personalized
processing condition were discarded from the final data set by this criterion. The
distributions of the target latencies for the remaining data were then examined. The
majority of the distributions were positively skewed, so all latencies were log
transformed to remove the positive skew (cf. Fazio, 1990; Kirk 1968).
The transformed data were first entered into a 2 X 2 X 2 MANOVA with
perceptual condition and sex of participant as between subjects factors and valence of
configuration as a within subjects factor. Sex did not have a main effect on response
latencies (g < .78), did not interact with perceptual condition (g < .33) or
configuration valence (g < .71), nor did it participate in a 3-way interaction (g < .18).
There was a significant main effect for perceptual condition, F (1,40) = 16.79, g <
.0001, no significant main effect for valence, F (1,40) = 0.20, ns, and no significant
41
interaction, F (1,40) = 1.40, ns. On average, participants took substantially longer to
guess targets’ likely academic majors (transformed M = 12.90, SD = .51) than did
participants who formed impression of them (transformed M = 12.38, SD = .41).
The pattern of effects was consistent with the prediction that the category-based
manipulation will require participants to exert more thought than will the personalized
perceptual manipulations.
Recall protocol data. Recall protocols were collected in the last part of the
experiment. The protocols were then coded for number of emergent attributes.
Emergent attributes were defined as attributes that appear in a participant's description
of a target that were not part o f the original presentation of attributes (Hastie et al.,
1990; Kunda et al., 1990). Moreover, emergent attributes could not be category -
labels, e.g., "lawyer", that described a category instead of the individual (Fiske &
Neuberg, 1990). A subset of 21 randomly selected protocols were independently
coded by two raters, both blind to condition. Interrater reliability was moderately
high (Cronbach's alpha = .84), indicating that the coding criteria were sufficiently
reliable to distinguish emergent attributes from the rest of the material in the
protocols. One of the original raters, again blind to condition, then coded the entire
set of 43 protocols.
Overall, there were 165 emergent attributes coded from protocols in the
category-based perceptual condition (M = 7.17) and 132 in the personalized
condition (M = 6.28). The total number of emergent attributes coded from each
participant were then subjected to a Mann-Whitney U-test. The difference between
personalized (M rank = 21.2) and category-based (M rank = 23.7) processing
conditions was not significant (U = 214, z = -.65, ns). Thus there were not
42
significantly more emergent attributes in the category-based protocols than in the
personalized protocols.
Although the number of emergent attributes in category-based perceptions
was not significantly higher than the number in personalized perceptions, the
direction of the difference in the mean number of emergent attributes is consistent
with the expectation that category-based perceptions will be more complex than
personalized perceptions of the target individuals as a result of greater thought. A
second, albeit indirect, piece of evidence supporting this conclusion is the relative
number of targets forgotten by perceptual condition in the protocols. The hypothesis
is that recall for the targets will be better for the category-based perceptions than for
the personalized perceptions. To test this I reanalyzed the protocols according to the
number of targets each participant forgot. Participants in the category-based
condition forgot a total of 8 targets (M = -35, M rank = 26.52), compared to a total of
17 (M = .81, M rank = 18.83) in the personalized condition. This difference was
significant, U = 157, z = - 2.28, p < .02, indicating that participants' recall of the
target individuals was better after category-based perception than personalized
perception.
Taken together, the latency and recall data provide convergent evidence that
the category-based perception manipulation caused participants to expend more
thought on the target individuals than did the personalized perception manipulation.
Interpersonal attraction. The data for the interpersonal attraction measures
(closeness, similarity, commonality) were first standardized then added together for
each target person to form a composite score for each individual (alpha = 0.86). The
composites from the three negative targets (JMJ, NAL, NMM) were then added, as
43
were the three positive composites (EJV, FEA, RTL) to yield a composite
identification score for the negative and the positive targets.
The composite identification scores were then entered into a 2 X 2 X 2
MANOVA with sex of participant and perceptual condition as the between-subjects
variables and valence of the trait configurations as the within-subjects variable. Sex
of participant did have a significant main effect on interpersonal attraction (F (1,40)
= 9.08, p < .004), with males (M = 37.66, SD = 8.20) giving higher interpersonal
attraction scores to the targets than did females (M = 33.22, SD = 10.16). Sex did
not, however, show an interaction with perceptual condition (p < .27) or valence of
configuration (p < .13), nor did it show a three-way interaction (p < .83).
The main effect of sex of participant on measures of interpersonal attraction
was not predicted. One explanation for it may be that male participants felt more
comfortable and competent at the computer-based task than female participants
(Badagliacco, 1990; Krendl & Broihier, 1992; Okebukola & Woda, 1993; Reisman,
1990). The greater positive affect experienced by males, relative to females, may
have influenced perceptions of the targets so that males tended to perceive all targets
more positively than female participants. Sex of participant did not combine with
either of the other factors to differentially affect perceptions of interpersonal
attraction, nor did it produce a significant main effect in either the latency or recall
measures. As predicted, perceptual condition had no main effect on participants'
evaluations of interpersonal attraction to the targets (F (1,42) = 0.02, ns.), and there
was no main effect for the valence of the configurations (F (1,42) = 0.12, ns). The
interaction between valence and processing condition, however was only marginally
significant (F (1,42) = 2.41, p < .13). A power analysis indicated that at least 25
additional participants would be needed for this interaction to reach a conventional
44
level of significance (i.e., g < .05). The means and standard deviations for the
unstandardized interpersonal attraction composites pooled across sex of participant
are presented in Table 4.
Table 4
Unstandardized Attraction Means and Standard Deviations. Experiment 2
Valence of Target
Perceptual Condition Negative Positive
Category-based 28.78 (9.06)3 42.13 (6.26)a
Personalized 24.57 (7.64)a 43.38 (5.70)a
Note. Standard deviations are in parentheses. Higher scores mean higher levels of
attraction. Means in the same column that have different superscripts differ from
each other at g < .025.
The interpersonal attraction data for Experiment 2 show a different pattern
then the data from Experiment 1. Recall that in Experiment 1, target individuals
described by positive attributes were perceived more positively after personalization,
whereas negative targets were perceived only slightly more negatively after
personalization. In Experiment 2, positive individuals were not perceived more
positively when they were personalized (M = 43.38) than when they were perceived
as group members (M = 42.13), t(20) = 0.69, ns.. Individuals described by negative
attributes, on the other hand, were perceived more negatively when they were
personalized (M = 24.57) than when they were viewed as group members (M =
45
28.78), but this difference was only marginally significant, t (20) = 1.66, g < .06.
Although not predicted, this pattern was interesting, particularly because the trait
attribute stimuli were identical to those used in Experiment 1. The difference will be
discussed in more detail below.
Liking. First, the liking data for the three negative targets (JMJ, NAL,
NMM), and the data for the three positive composites (EJV, FEA, RTL) were added
to form a composite liking score for the negative and the positive configurations. The
composite liking scores were then entered into a 2 X 2 X 2 MANOVA with sex of
participant and perceptual condition as the between-subjects variables and valence of
the trait configurations as the within-subjects variable. Sex showed no main effect (p
< .87), no interaction with perceptual condition (g < .50), no interaction with valence
(g < .94), and no three-way interaction (g < .52). As predicted, perceptual condition
produced no main effect on liking scores (F (1,42) = 0.52, ns). There was a
significant main effect for valence of configurations (F (1,42) = 119.21, g < .001).
The predicted interaction between perceptual condition and valence was not
significant (F (1,42) = 0.71, g < ns). The means and standard deviations for the
liking measure pooled across sex of participant are presented in Table 5.
The distortion hypothesis was tested in the liking data As predicted in the
personalized condition, positive targets (M = 17.56, SD =2.18) were not significantly
different from the mean Anderson value, t (22) = 1.30, ns, nor were negative targets
(M = 10.24, SD = 3.24), t (22) = .26, ns. As in Experiment 1, a 2 X 2 X 2 mixed-
model analysis was used to test the distortion predictions in the . The identical
procedure was used to organize the liking data for this analysis. However, no
participant in the category-based condition assigned a negative major to either a
negatively described or a positively described target. Therefore the only comparison
46
relevant to the distortion hypothesis that was able to be tested was the negatively
described targets who were assigned a positive major versus the negatively described
targets were personalized. The results of this comparison were similar to those found
in Experiment 1. Negative targets who were placed into positive categories (M =
11.52, SD= 3.36) were not liked significantly more than negative targets who were
personalized (M = 10.24, SD - 3.24), t(20) = 0.91, ns.
Table 5
Liking Means and Standard Deviations. Experiment 2
Valence of Target
Perceptual Condition Negative Positive
Category-based 11.52 (3.36)3 17.57 (2.25)
Personalized 10.24 (3.24) 17.57 (2.18)
Baseline levels 10.60 (2.75) 16.40 (2.47)
Note. Standard deviations are in parentheses. Higher scores mean higher levels of
liking. Means in the same column that have different superscripts differ from each
other at p < .0125.
Given the lack of negative majors assigned to negative targets, a series of
post-hoc analyses identical to the ones performed in Experiment 1 were performed on
the liking data. The first of these analyses involved examining the induced category
choices each participant made for the positive and negative targets in the category-
based condition. The design of this analysis was to enter the data for the targets that
most closely approximated the hypothesized effects of the idiographic categorization
47
manipulation. Thus, of the induced majors, the most and least positive were selected
and matched with the identical targets in the personalized condition. For the 3
positive targets, the liking data was retained for the target who was assigned the most
positive major (according to the normative ratings of the majors), and for the target
assigned the most negative major. For the negative targets the data for the target
assigned the most positive major and the one assigned the most negative major were
retained. These targets were then matched with the identical targets from a randomly
selected participant in the personalized condition. This data was then entered into a 2
X 2 X 2 MANOVA.
If the induced categories distort liking evaluations away from personalized
levels, then positive targets assigned the most positive majors and positive targets
assigned the least positive majors will be liked more and less, respectively, than the
personalized counterparts. Likewise, negative targets assigned the most positive
majors and negative targets assigned the least positive majors should be liked more
and less, respectively, than the personalized counterparts. This will be indicated by
an interaction between the valence of the categories and perceptual orientation. In
addition, there will be a main effect for category valence such that the targets placed
into the most positive majors will be liked more than those placed into the least
positive majors, and a main effect for target valence such that positive targets will be
liked more than negative targets. There will be no main effect for perceptual
orientation, no interaction between target valence and perceptual orientation, and no
3-way interaction.
The target main effect was significant such that positive targets (M = 5.78, SD
= 1.30) were liked more than negative targets (M = 3.88, SD = 1.66), F (1,42) =
82.12, g < .001. No other effects were significant.
48
A second post-hoc analysis involved examining both positive and negative
targets in the category-based condition and selecting the positive target and the
negative target who were assigned the most positive major of the 3 targets in each
target valence condition. The 2 targets were then matched with their counterparts
from a randomly selected participant in the personalized condition. The liking data
for these targets were then entered into a 2 (target valence) X 2 (perceptual
orientation) MANOVA. If placing targets into a positive major distorts liking away
from the personalized levels, then there will be a main effect for perceptual
orientation such that categorized targets will be liked more than personalized targets.
There will not be a main effect for target valence, nor will there be an interaction.
Categorized targets (M = 4.94, SD = 1.51) were not liked more than
personalized targets (M = 4.91, SD = 1.22), F < 1.0, nor was there an interaction, F <
1.0. There was, however, a significant main effect for target valence such that
positive targets (M = 5.78, SD= 1.29) were liked more than negative targets (M =
4.05, SD = 1.49), F (1,42) = 39.41, p < .001.
A third post-hoc analysis involved examining the standard deviations of the
categorized and personalized targets from the mean level of liking for each attribute
valence condition. If the addition of category information to an individual’ s
description distorts liking away from the personalized levels, in which the category
information is absent, then there will be a larger mean standard deviation in the
category-based conditions than in the personalized conditions. This is because
positive categories will distort positive personalized targets upward (more liking) and
negative categories will distort positive personalized targets downward (less liking).
A similar effect will occur with negative targets such that positive categories will
distort negative personalized targets upward and negative categories will distort
49
negative personalized downward. In other words, positive and negative categories
will cause the mean liking ratings for the targets to be more variable relative to those
for uncategorized, personalized targets.
As in Experiment 1, the data for this analysis were generated by computing
the standard deviation of liking from the levels predicted by the baseline Anderson
ratings for targets in the category-based and personalized conditions. The positive
target assigned the most positive major and the positive target assigned the most
negative major were selected and matched with the identical targets for a randomly
selected participant in the personalized condition, and the standard deviations from
the baseline levels of liking were computed for them. Similarly, the negative target
assigned the most positive major and the negative target assigned the most negative
major were matched with the identical targets from a randomly selected participant in
the personalized condition and the standard deviations from the baseline levels were
computed. The matched sets of standard deviations were then compared for each
target valence condition.
The predictions were not supported. Directionally, the standard deviations of
the liking ratings for categorized positive targets (M = 1-03, SD = 1.00) were more
variable than the standard deviations for personalized positive targets (M = 0.78, SD
= 0.64), but this difference was not significant t(20) = 0.68, ns. Again, for the
negative targets the standard deviations for the categorized targets (M _ = 1.41, SD =
0.58), were not more variable than the standard deviations for the personalized
negative targets (M = 1-53, SD = 0.52), t(20) = - 0.38, ns.
Discussion
The main finding of Experiment 1 was that responses to individuals after
category-based and personalized perceptions showed an interaction between
50
perceptual mode and the affective valence of the attributes that describe the
individuals. This finding was partially replicated in Experiment 2, although the
interaction was only marginally significant. Additionally, the pattern of mean
attraction scores was different from that found in Experiment 1. In Experiment 1,
positive individuals were perceived more positively when they were personalized (M
= 43.95, standardized M = 1-59) than when they were regarded as category members
(M = 38.90, standardized M = -1-85), t (18) = -2.41, p < .025. Negative individuals,
on the other hand, were not perceived more negatively when they were personalized
(M = 25.21, standardized M = -84) than when they were regarded as category
members (M = 27.05, standardized M = --45), t (18) = .90, ns. In Experiment 2
attraction toward positive individuals under personalized (M = 43.38, standardized M
.47) and category-based (M = 42.13, standardized M = --43) perceptions, t (20) = -.56,
ns, did not differ. However, for negatively described individuals, the difference
between personalized (M = 24.57, standardized M = 1-47) and category-based (M =
28.78, standardized M = -1.61) perceptions was marginally significant, t (20) = -1.49,
p < 06.
The overall pattern of effects for the interpersonal attraction data when
Experiment 1 is compared to Experiment 2 indicates that participants' feelings of
attraction to both positive and negative individuals were relatively stable for
personalized perceptions, whereas attraction increased for both positive and negative
individuals under category-based perception when the categories were made highly
salient in Experiment 2. The induced majors in Experiment 2 (M = 5.07, SD = 0.43)
were also more positive than the induced majors in Experiment 1 (M = 4.78, SD =
0.47), t(l 19) = 2.45, p < .01. This suggests that increasing the amount of thought
participants expend on placing targets into positive categories increased the weight of
51
group membership information on participants' evaluations of targets. Recall that the
mean valence of academic majors was 4.65, corresponding to slightly positive affect
felt about the majors. Increasing that weight over the attribute information would
tend to make perceptions of targets, especially negative targets, more positive than
when it was equally weighted. The change in the procedures between the
categorization conditions in the two experiments was designed to induce participants
to place more weight on group membership information, relative to attribute
information. Everything else about the stimulus individuals was identical to
Experiment 1. Thus, although this explanation was not tested directly (i.e., in the
same experimental design), it seems a plausible account for the increased positivity
for all targets in the category-based condition.
An equally plausible explanation for the increase in the positivity of
perceptions in the category-based conditions is related to the mere exposure
phenomenon in which increased familiarization with an object increases a person's
liking for it (Zajonc, 1968). That is, if participants in the category-based conditions in
Experiment 2 became more familiar with the targets, relative to category-based
participants in Experiment 1. This in turn may have caused their perceptions of the
targets to be more favorable than participants in Experiment 1. However, because the
amount of thought required during the categorization task was not manipulated in the
same design, it is not possible to tease apart the two explanations for the increase in
positivity in category-based perceptions. This issue should be addressed in future
research.
If the increase of salience of category information caused participants to
weight that information more heavily than the attribute information, the increased
salience did not cause category-based perceptions to be significantly distorted from
52
baseline levels. The distortion hypothesis was tested in 4 post-hoc analyses, and all
failed to furnish strong support for the category-based distortion of liking evaluations
away from personalized levels. One potential explanation for this is that the academic
major categories did not exert a strong enough influence on perceptions of
individuals' attributes to distort them in the direction of the category value. This
would suggest that participants gave an equal weight to both types of information,
category membership and personal attributes, when forming their perception of the
targets. This would explain the essentially equivalent liking scores for positive
targets in the personalized and category-based conditions. However, this cannot
account for the positive shift in liking scores for negative targets in the category-
based condition, relative to the personalized conditions. Another explanation is that
the category information was weighted more than personal information, but that the
distortion was simply not strong enough to be significant. In other words, although
the weight of the group membership information in Experiment 2 may have been
higher than in Experiment 1, it was not high compared with group membership
information in field settings. In the field, group membership information is often
readily apparent. This information may be cued by such things as skin color, manner
of dress, accented speech, and occupation, which are likely weighted more in one's
initial encounter with another than a person's academic major. Therefore, in
Experiments 1 and 2, the group membership information available to participants was
not sufficiently weighted to cause distorted category-based perceptions. A third
explanation is that the idiographic paradigm failed to provide a strong manipulation
of highly positive and negative categories, which in turn prevented them from
distorting evaluations away from the baseline personalized level. A fourth
explanation is that no ingroup/outgroup competition was salient between the
53
participants and the targets that they evaluated. This may have made participants pay
less attention to category boundaries than typically happens in intergroup contact
situations (Brewer & Miller, 1984). These issues will be discussed further in the
General Discussion.
General Discussion
This dissertation investigated the effects of personalization on perceptions of
individuals in two main hypotheses. Both the decategorization model and
accentuation theory were used to generate the hypothesis that during personalized
interactions, perceptions of individuals will be based on the affective valence of the
attributes that characterize each individual. During category-based interactions,
perceptions of individual group members will be assimilated to one's perceptions of
the group itself. In the idiographic paradigm used in this dissertation, participants
viewed descriptions of 6 target individuals, half described positively and half
negatively. Participants were instructed to either form an impression of each
(personalized perception) or to place each into an academic major category (category-
based perception). The mean affective valence of the categories was slightly positive,
whereas the mean affective valence of the positive individuals was moderately
positive, and the mean affective valence of the negative individuals was moderately
negative. Therefore, an interaction between the affective valence of individual
targets' attributes and the perceptual mode used by participants was predicted.
In Experiment 1, participants' perceptions of negative individuals were more
negative when they were personalized than when they were category-based, and
perceptions of positive individuals were more positive when they were personalized
than when they were category-based. In Experiment 2, perceptions of negative
individuals were again more negative when they were personalized than when they
54
were category-based, but perceptions of positive individuals were slightly more
positive when they were personalized than when they were category-based. The
different patterns for perceptions of category-based targets in the two experiments
was attributed to the increased salience of the categories in Experiment 2.
These results are congruent with other work demonstrating accentuation
effects in interpersonal attraction between members of different groups (e.g., Doise,
Deschamps, & Meyer, 1978; Hensley & Duval, 1976; Krueger & Rothbart, 1988;
Wilder & Thompson, 1988), which show that individual group members are
perceived according to the group tendency or stereotype when they are categorized.
The results from the experiments presented in this dissertation add to these findings
by showing that one basis for the accentuation effect may lie in the differential weight
a person places on other attribute information when it is considered in the context of
category information. In Experiment 1, the mean valence of the induced categories
for positive targets (M = 4.59, SD = 0.64) was identical to the mean valence of the
induced categories for negative targets (M = 4.59, SD = 0.31). However, both
interpersonal attraction (M = 38.90, SD =4.31) and liking (M = 16.32, SD = 2.11)
were higher for positive targets than attraction (M = 27.05, SD = 8.58) and liking (M
= 10.95, SD = 2.78 ) for negative targets. In Experiment 2 the mean valence for the
induced categories for positive targets (M = 5.07, SD = 0.84) was again identical to
the mean valence of the induced categories for the negative targets (M = 5.07, SD =
0.68), but interpersonal attraction (M = 42.13, SD= 6.26) and liking (M = 17.57, SD =
2.25)) were higher for positive targets than attraction (M = 28.78, SD = 9.06) and
liking (M = 11.52, SD = 3.36) for negative targets. This suggests that although the
induced categories were identical in valence, the valence of the personal attributes
did differentially influence attraction to and liking for the targets. Therefore, in the
55
context of categories that have low salience, relative to situations in which category
salience is much higher, personal attributes may influence evaluations of individual
group members. In the context of highly salient categories, such as during
ingroup/outgroup competition, personal attributes may exert less influence. This may
further our understanding of how the cognitive processing of affectively valenced
information can affect perceptions of outgroup members during an intergroup
interaction.
A second hypothesis derived from accentuation theory predicted that
category-based perceptions would be distorted away from the personalized baseline
levels predicted by an averaging model toward the level of the category valence. This
hypothesis was tested in both experiments by comparing participants' scores on a
liking measure in both perceptual conditions. This hypothesis was not supported.
Although perceptions in the personalized conditions for positive and negative targets
did not differ significantly from baseline predictions, they were not significantly
distorted away from the baseline toward the category level. However, the distortion
hypothesis was not strongly tested in that there was no powerful manipulation of the
valence of the induced majors. A series of 4 post-hoc analyses were performed on the
liking data from the negative targets in the category-based and personalized
conditions in Experiment 1 and Experiment 2 as an attempt to find evidence for the
distortion effect. The first of these compared the liking for categorized and
personalized negative targets, with the prediction that liking should be higher for the
categorized than for the personalized targets. The second analysis compared liking
for the most positive and least positive induced majors for positive and negative
targets against liking for the identical targets in the personalized condition. The
prediction for this analysis was for a category valence X perceptual orientation
56
interaction such that positive targets assigned the most positive majors and positive
targets assigned the least positive majors will be liked more and less, respectively,
than the personalized counterparts. Likewise, negative targets assigned the most
positive majors and negative targets assigned the least positive majors should be liked
more and less, respectively, than the personalized counterparts A second prediction
was for a main effect of the valence of the induced majors such that targets assigned
positive majors will be liked more than targets assigned negative majors. The third
analysis compared the most positive induced major for the positive and negative
targets against liking for the identical targets in the personalized condition. The
prediction for this analysis was that the addition of category information to an
individual's description distorts liking away from the personalized levels, in which the
category information is absent, so that there will be a larger mean standard deviation
in the category-based conditions than in the personalized conditions. None of the
results from these analyses provided support for the hypothesis. Consequently, it is
not clear if the lack of a distortion effect is due to a weak manipulation of category
salience or a failure of the idiographic paradigm to induce participants to assign the
targets to strongly valenced positive and negative categories. In order to redress this
ambiguity, a design that explicitly manipulates the valence of the categories into
which targets may be placed should be employed in future research.
Taken together, the interactions found in the interpersonal attraction data from
both experiments provide a first indication that the contact-based interventions
prescribed by the decategorization model may have different effects when individuals
are perceived as being affiliated with a group, versus when a their group affiliation is
not salient. Previous research has demonstrated that perceptions of outgroup
members typically become more positive after personalization (Bettencourt et al.,
57
1992; Miller et al., 1985). This research demonstrates that perceptions after
personalization can become more negative for a certain subset of group members.
That is, if an individual is described by primarily negative attributes and he or she
belongs to an affectively-neutral group, then personalization can remove the
ameliorating effects of the group membership and allow him or her to be perceived
according to his or her true negative nature.
The implication for decategorization-based interventions is that even if
personalization of the outgroup can improve relations between groups, it may degrade
relations between individuals. Of course, the terms "improve" and "degrade" are
contingent on the perceived affective valence of an individuals' attributes and group
of which they are a member. If the group is highly negative and the individual is also
highly negative, then perceptions after personalization are unlikely to change much.
Likewise, if the group is highly positive and the individual is also highly positive,
then personalization is unlikely to make any significant difference in perceptions.
This was probably the case with positive individuals in Experiment 2, wherein the
increased salience of the academic majors also increased their positive weight. This
resulted in category-based perceptions that were as positive as personalized
perceptions. However, it is unlikely that individuals who are evaluated negatively
after personalization will be perceived much more negatively than a negative
outgroup because individuals tend to be perceived more positively, at a baseline level,
than groups (Sears, 1983). Therefore, if relations between individuals are degraded
after personalization, they may not be as bad or hostile as the relations would be on
the group level.
Despite the discouraging results of this research, conclusions about the nature
of individual-level perceptions during category-based and personalized interactions
58
must remain preliminary until several issues are addressed in subsequent research.
One issue concerns the use of the idiographic paradigm. The idiographic paradigm
was used in this dissertation because it allowed participants to place target individuals
into a wide range of categories rather than into arbitrarily determined categories. This
was desirable because it did not intentionally confound the affective valence of a
category with category-based perceptions, per se (Pavelchak, 1989). However, this
manipulation of category-based perception did not seem to arouse strong negative
categorizations. In two experiments, 42 participants categorizing a total of 252
targets did not place a single one in a major that was negative. This may have
occurred because undergraduates in the introductory psychology subject pool do not
see the many academic majors negatively. Indeed, in the normative ratings of the
positivity or negativity of the 49 majors collected prior to Experiment 1, only 5 were
rated negatively. This suggests that undergraduates see the majority of majors at least
slightly positive. In order to generate category-based distortion, strongly valenced
positive and negative categories, such as an ingroup and a hated outgroup, may have
to be used instead of relying on idiographically induced academic majors.
In addition, the idiographic paradigm is highly artificial because participants
must induce categorizations from a limited set of information that may be of little
actual use in determining a person's academic major. In the real world,
categorizations are often made on the basis of unambiguous cues to group
membership such as race, gender, and age (Zebrowitz, 1990). Category information
in the real world may be much more salient in determining things like perceptions of
interpersonal affiliation than in the idiographic laboratory. In this case, category
information may be weighted much more heavily than attribute information in
forming perceptions of individuals. This suggests that using a paradigm that enables
59
participants to place targets into specific, highly valenced categories, may make the
distortion of personalized perceptions much stronger than was found in this research.
A second issue concerns the nature o f the perceptual task. The contact
hypothesis and the decategorization model were intended to be used in intergroup
situations where the boundaries discriminating between ingroup and outgroup were
clear, and personalized individuals were still recognized to some degree as group
members. In this research the intergroup setting was absent. Participants simply
viewed individuals described by four traits, separate from any ingroup and outgroup
affiliations. Group membership was intended to be weighted equally with attribute
information in Experiment 1. The weight of group membership was increased in
Experiment 2, with the effect that both negative and positive individuals were seen
more positively, relative to Experiment 1. This is congruent with the work on the
mere exposure effect which shows that the more a person is exposed to an object, the
more positively he or she feels about it (Zajonc, 1968). However, in an intergroup
situation, group membership may be weighted much higher than attribute
information, making personalization much less effective than if group membership
were assigned less weight in the perception. One implication of this is that positively
described outgroup individuals are still seen negatively after personalization. Indeed,
some evidence suggests that people attend more to negative information than positive
information when forming impressions of others (Yzerbyt & Leyens, 1991). This
suggests that the effects of personalization on the decategorization of individual group
members may be limited in intergroup settings.
A third issue concerns the valence of the academic majors used in the
experiments. In this research the mean valence of the group membership information
was slightly positive. This produced category-based perceptions that tended to
60
become more moderate than personalized perceptions. A different pattern of results
would be expected if the group membership of a target were highly positive or highly
negative. In the case of a highly positive outgroup, category-based perceptions would
be more positive than personalized perceptions only for negative individuals. In the
case of a highly negative outgroup, category-based perceptions for positive targets
would be more moderate than personalized perceptions of them. The implication for
intergroup contact is that positive individuals from positive outgroups and negative
individuals from negative outgroups who are personalized may subsequently be
viewed as confirming the positive or negative stereotype of the outgroup.
A final issue concerns the reciprocal effects of personalizing group members.
That is, after individual group members are personalized, how is the group perceived?
In an ongoing relationship between members of different groups, the group
membership of individuals is unlikely to be forgotten after personalization. At certain
times individuals may be regarded more as group members than individuals, while at
other times the reverse may hold. It may be that personalized group members
contribute to subsequent perceptions of the group (Miller & Brewer, 1986; Miller &
Harrington, 1990). If the first outgroup members that are personalized are found to be
negative, this may maintain the perception of a negative outgroup and worsen the
perception of a positive or neutral outgroup and may undercut the beneficial effects of
positive individuals who are personalized later. This suggests that personalization be
investigated in a dynamic setting in which the intergroup and interpersonal aspects of
perception and interaction are allowed to intermingle through time (e.g., Brown &
Turner, 1981).
61
Despite the limitations of this research, one implication is clear: practitioners
and social engineers who use decategorization-based interventions to improve
intergroup relations should take into account the effects such manipulations can have
on inter-individual relations. Outgroup members who are personalized and found to
be negative may reflect badly on the outgroup, potentially making intergroup
relations more bleak than before the intervention. Therefore, it would be prudent to
take both the intergroup and interpersonal effects of personalization into account
when designing decategorization-based interventions to encourage the beneficial
effects of personalizing individual group members, and to guard against the
potentially deleterious effects.
62
Bibliography
Allport, G. W. (1954). The nature of prejudice. Cambridge, MA: Addison
Wesley.
Amir, Y. (1969). The contact hypothesis in ethnic relations. Psychological
Bulletin. 71. 319-342.
Anderson, N. H. (1968). Likableness ratings of 555 personality-trait words.
Journal of Personality and Social Psychology. 9. 272-279.
Anderson, N. H. (1981). Foundations of information integration theory.
New York: Academic.
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of other in the self
scale and the structure of interpersonal closeness. Journal of Personality and Social
Psychology. 63. 596-612.
Asch, S. E. &Zukier, H. (1984). Thinking about persons. Journal of
Personality and Social Psychology. 4 6 .1230-1240.
Badagliacco, J. (1990). Gender and race differences in computing attitudes
and experience. Social Science Computer Review. 8. 42-63.
Bettencourt, B. A., Brewer, M. B., Croak, M. R., & Miller, N. (1992).
Cooperation and the reduction of intergroup bias: The role of reward structure and
social orientation. Journal of Experimental Social Psychology. 28. 27-52.
Brewer, M. B. (1988). A dual process model of impression formation. In T.
K. Srull& R . S. Wver (Eds.). Advances in Social Cognition .Vol. 1. (pp. 1-36).
Hillsdale, NJ: Erlbaum.
Brewer, M. B. & Kramer, R. M. (1985). The psychology of intergroup
attitudes and behaviors. Annual Review of Psychology. 36. 219-243.
Brewer, M. B. & Miller, N. (1984). Beyond the contact hypothesis:
Theoretical perspectives on desegregation. In N. Miller & M. B. Brewer (Eds.),
Groups in contact: The psychology of desegregation, (pp. 281-302). Orlando, FL:
Academic.
Brewer, M. B. & Miller, N. (1988). Contact and cooperation: When do
they work? In P. A. Katz & D. A. Taylor (Eds.), Eliminating racism: Profiles in
controversy, (pp. 315-326). New York: Plenum.
Brown, R. J. & Turner, J. C. (1981). Interpersonal and intergroup behavior.
In J. C. Turner & H. Giles (Eds.), Intergroup behavior (pp. 33-65). Oxford: Basil
Blackwell.
Campbell, D. T. (1956). Enhancement of contrast as composite habit.
Journal of Abnormal and Social Psychology. 53. 350-355.
63
Carpenter, S. (1993). Organization of in-group and out-group information:
The influence of gender-role orientation. Social Cognition. 11. 70-91.
Cook, S. W. (1978). Interpersonal and attitudinal outcomes of cooperating
interracial groups. Journal of Research and Development in Education. 12. 97-113.
Cohen, B. & Murphy, G. L. (1984). Models of concepts. Cognitive
Science. 8. 27-58.
Deschamps, J. C. (1984). Intergroup relations and categorical identification.
In H. Tajfel (Ed.), The social dimension: European developments in social
psychology (Vol. 2). (pp. 541-559). Cambridge: Cambridge University Press.
Doise, W., Deschamps, J. C., & Meyer, G. (1978). The accentuation of intra
category similarities. In H. Tajfel (Ed.), Differentiation between social groups:
Studies in the social psychology of intergroup relations (pp. 159-168). London:
Academic Press.
Fazio, R. H. (1990). A practical guide to the use of response latency in
social psychological research. In C. Hendrick & M. S. Clark (Eds.), Research
methods in personality and social psychology fVol. 11). (pp. 74-97). Newbury Park,
CA: Sage.
Fiske, S. T. (1980). Attention and weight in person perception: The impact
of negative and extreme behavior. Journal of Personality and Social Psychology.
889-906.
Fiske, S. T. &Neuberg, S. L. (1990). A continuum of impression
formation, from category-based to individuating processes: Influences of information
and motivation on attention and interpretation. Advances in Experimental Social
Psychology (Vol. 23). 1-74.
Fiske, S. T. & Taylor, S. E. (1984). Social Cognition. Reading, MA:
Addison-Wesley.
Fletcher, G. O., Danilovics, P., Fernandez, G., Peterson, D., & Reeder, G. D.
(1986). Attributional complexity: An individual differences measure. Journal of
Personality and Social Psychology. 51. 875-884.
Hampton, J. A. (1987). Inheritance of attributes in natural concept
conjunctions. Memory & Cognition. 15. 55-71.
Hastie, R., Schroeder, C., & Weber, R. (1990). Creating complex social
conjunction categories from simple categories. Bulletin of the Psvchonomic Society.
28,242-247.
Hensley, V. & Duval, S. (1976). Some perceptual determinants of perceived
similarity, liking, and correctness. Journal of Personality and Social Psychology. 34.
159-168.
64
Judd, C. M. & Park, B. (1988). Out-group homogeneity: Judgments of
variability at the individual and group levels. Journal of Personality and Social
Psychology. 54. 778-788.
Kirk, R. E. (1968). Experimental design: Procedures fro the behavioral
sciences. Belmont, CA: Brooks-Cole.
Krendl, K. & Broihier, M. (1992). Student responses to computers: A
longitudinal study. Journal of Educational Computing Research. 8. 215-227.
Krueger, J. (1992). On the overestimation of between-group differences. In
W. Stroebe & M. Hewstone (Eds.), European Review of Social Psychology. Vol. 3
(pp. 31-56). Chichester, UK: J. Wiley & Sons.
Krueger, J. & Rothbart, M. (1988). Use of categorical and individuating
information in making inferences about personality. Journal of Personality and Social
Psychology. 55. 187-195.
Kunda, Z., Miller, D. T., & Claire, T. (1990). Combining social concepts:
The role of causal reasoning. Cognitive Science. 14. 551-577.
Linville, P. W., Fischer, G. W., & Salovey, P. (1989). Perceived
distributions of the characteristics of in-group and out-group members: Empirical
evidence and a computer simulation. Journal of Personality and Social Psychology.
57,165-188.
Linville, P. W., Salovey, P., & Fischer, G. W. (1986). Stereotyping and
perceived distributions of social characteristics: An application to ingroup-outgroup
perception. In J. Dovido & S. L. Gaertner (Eds.), Prejudice, discrimination, and
racism (pp. 165-208). New York: Academic.
Marcus-Newhall, A., Miller, N., Holtz, R., & Brewer, M. B. (1993). Cross
cutting category membership with role assignment: A means of reducing intergroup
bias. British Journal of Social Psychology. 32. 125-146.
Marks, G., Miller, N., & Maruyama, G. (1981). The effect of physical
attractiveness on assumptions of similarity. Journal of Personality and Social
Psychology. 41. 198-212.
Maslach, C. (1974). Social and personal bases of individuation. Journal of
Personality and Social Psychology. 2 9 .411-425.
Messick, D. M. &M ackie, D. M. (1989). Intergroup relations. Annual
Review of Psychology. 40. 45-81.
Miller, N. & Brewer, M. B. (1986). Categorization effects on ingroup and
outgroup perception. In J. F. Dovido & S. L. Gaertner (Eds.), Prejudice.
Discrimination, and Racism, (pp. 209-231). San Diego, CA: Academic.
65
Miller, N. Brewer, M. B., & Edwards, K. (1985). Cooperative interaction in
desegregated settings: A laboratory analog. Journal of Social Issues. 4 1 .63-79.
Miller, N. & Harrington, H. J. (1990). A model of social category salience
for intergroup relations: Empirical tests of relevant variables. In P. J. Drenth, J. A.
Sergeant, & R. J. Takens (Eds.), European Perspectives in Psychology (Vol. 31. (pp.
205-220). Chichester, UK: J. Wiley & Sons.
Mullen, B. & H u, L. (1989). Perceptions of ingroup and outgroup
variability: A meta-analytic integration. Basic and Applied Social Psychology. 10.
233-252.
Okebukola, P. & Woda, B. (1993). The gender factor in computer anxiety
and interest among some Australian high school students. Educational Research. 35.
181-189.
Ostrom, T. M., Carpenter, S.L., Sedikides, C. & Li, F. (1993). Differential
processing of in-group and out-group information. Journal of Personality and Social
Psychology. 64. 21-34.
Ostrom, T. M. & Sedikides, C. (1992). Out-group homogeneity effects in
natural and minimal groups. Psychological Bulletin. 112. 536-552.
P ark,B . & Judd, C. M. (1990). Measures and models of perceived group
variability. Journal of Personality and Social Psychology. 59. 173-191.
Park, B., Judd, C. M., & Ryan, C. S. (1991). Social categorization and the
representation of variability information. In W. Stroebe & M. Hewstone (Eds.),
European Review of Social Psychology. Vol. 2 (pp. 211-245). Chichester, UK: J.
Wiley & Sons.
Pavelchak, M. A. (1989). Piecemeal and category-based evaluation: An
idiographic analysis. Journal of Personality and Social Psychology. 56.354-363.
Reisman, J. (1990). Gender inequality in computing. Computers in Human
Services. 7. 45-63.
Rosenbaum, M. E. (1986). The repulsion hypothesis: On the non
development of a relationship. Journal of Personality and Social Psychology. 51.
1156-1166.
Ruscher, J. B. & Fiske, S. T. (1990). Interpersonal competition can cause
individuating processes. Journal of Personality and Social Psychology. 58. 832-843.
Ruscher, J. B., Fiske, S. T., Miki, H., & van Manen, S. (1991). Individuating
processes in competition: Interpersonal versus intergroup. Personality and Social
Psychology Bulletin. 17. 595-605.
Sears, D. O. (1983). The person positivity bias. Journal of Personality and
Social Psychology. 44.233-250.
66
Singh, R. & Tan, L. S. C. (1992). Attitudes and attraction: A test of the
similarity-attraction and dissimilarity-repulsion hypotheses. British Journal of Social
Psychology. 31. 227-238.
Stephan, W. G. (1987). The contact hypothesis in intergroup relations. In C.
Hendrick (Ed.), Review of Personality and Social Psychology (Vol. 91. (pp. 13-35).
Newbury Park, CA: Sage.
Tajfel, H. (1959). Quantitative judgment in social perception. British Journal
of Psychology. 50. 16-29.
Tajfel, H., Sheikh, A. A., & Gardner, R. C. (1964). Content of stereotypes
and the inferences of similarities between members of stereotyped groups. Acta
Psvchologica. 22. 191-201.
Turner, J. C. (1975). Social comparison and social identity: Some prospects
for intergroup behavior. European Journal of Social Psychology. 5 .5-34.
Turner, J. C. (1987). A self-categorization theory. InJ. C. Turner, M. A.
Hogg, P. J. Oakes, S. D. Reicher, & M. S. Wetherell (Eds.), Rediscovering the
social group: A self-categorization theory, (pp. 42-67 Oxford: Blackwell.
Wilder, D. A. & Thompson, J. E. (1988). Assimilation and contrast effects
in judgments of groups. Journal of Personality and Social Psychology. 54. 62-73.
Yzerbyt, V. Y. & Leyens, J-P. (1991). Requesting information to form an
impression: The influence of valence and confirmatory status. Journal of
Experimental Social Psychology. 27. 337-356.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of
Personality and Social Psychology Monograph Supplement. 9 .1-27.
Zebrowitz, L. A. (1990). Social Perception. Pacific Grove, CA:
Brooks/Cole.
Ziller, R. C. (1964). Individuation and socialization. Human Relations. 17.
341-360.
Zimbardo, P. G. (1969). The human choice: Individuation, reason, and
order versus deindividuation, impulse, and chaos. In W. J. Arnold & D. Levine
(Eds.), Nebraska Symposium on Motivation. Vol. 17. (pp. 237-307). Lincoln, NE:
University of Nebraska Press.
Appendix A
On the following pages, you will see a series of four traits describing a single person who
is identified only by their initials. Each one of these characteristics has been assigned to the
person by someone who knows them well, and each of the characteristics is equally
important in describing the person. Please read through the traits that describe each person
and try to form an impression of this person. You should think in terms of being in a
conversation with them and learning four traits about them. After learning the traits, what
is your impression of them? What is this person like? Try to form a detailed impression of
them in your mind. After you have formed your impression, please complete the
responding styles questions for each of the persons. For all of the questions, please circle
the number on the scales that best corresponds to how you feel.
g.J. V ,
Imaginative
Cultured
Ambitious
Emotional
1. How likable is this person? (How much would you like this person?)
I ___________ I __________ I __________I __________ I __________ I __________ l
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable norunlikable unlikable unlikable unlikable
68
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
EJV
4.
o o o o o o
Yog EJV You EJV
You EJV
You EJV
You EJV
3. How similar is this person to you?
J. 1 1 J
1 2 3 4 5
very moderately slightly neither dissimilar slightly
dissimilar dissimilar dissimilar nor similar similar
6 7
moderately very
similar similar
4. How much do you feel you have in common with this person?
I ___________ I ___________ I ___________ I ___________ I _____ 1
1
very l i t t l e
a moderate amount
6 7
very much
69
F. E, A ,
Efficient
Well-read
Methodical
Reliable
1. How likable is this person? (How much would you like this person?)
I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
p p
You FEA
a.
OO
You FEA
OO
You FEA
4.
G D
You FEA
(O
Y ou FEA
6.
o
You FEA
3. How similar is this person to you?
I.
1 I
I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I.
1 1
1
very l i t t l e
a moderate amount
7
very much
J.M .J.
Irritating
Well-read
Argumentative
Authoritative
1. How likable is this person? (How much would you like this person?)
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
o o O O OO
You JM J You JM J You JM J
Y ou JM J
You JM J You JM J
3. How similar is this person to you?
I -------------------1 ___________ I ____________I ___________ I ___________ I ---------------- 1
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I ----------------- 1 ___________ I ___________l _________ I _____________I __________ l
1 2 3 4 5 6 7
very l i t t l e a moderate amount very much
N. A. L.
Talkative
Argumentative
Opinionated
Careless
1. How likable is this person? (How much would you like this person?)
I _ _ _ _ _ _ _ l _ _ _ _ _ _ I _ _ _ _ _ _ I - - - - - - I - - - - - - 1- - - - - - J
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
NAL You NAL
5.
You
NAL You NAL
OO
You NAL
6.
You NAL
3. How similar is this person to you?
1 I ________
1 2 3 4 5
very moderately slightly neither dissimilar slightly
dissimilar dissimilar dissimilar nor similar similar
6 7
moderately very
similar similar
4. How much do you feel you have in common with this person?
I ______________ I _____________ I ____________ l______________ I _____
1 2 3 4 5
very l i t t l e a moderate amount
6 7
very much
72
N.M.M.
Emotional
Neurotic
Sensitive
Unlucky
1. How likable is this person? (How much would you like this person?)
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
z.
O O OO 0 0
You NMM You NMM Y ou NMM
You NMM You NMM You NMM
3. How similar is this person to you?
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I_ _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _ _ J ------------ J
1 2 3 4 5 6 7
very l i t t l e a moderate amount very much
I
73
R. I .-L
Idealistic
Attentive
Open-minded
Individualistic
1. How likable is this person? (How much would you like this person?)
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
o o OO OO
Y ou HU- You RTL Y ou RTL
a D liC D
You RTL
You RTL You RTL
3. How similar is this person to you?
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 2 3 4 5 6 7
very l i t t l e a moderate amount very much
Appendix B
On the following pages, you will see a series of four traits describing a single person who
is identified only by their initials. Each one of these characteristics has been assigned to the
person by someone who knows them well, and each of the characteristics is equally
important in describing the person. Please read through the traits that describe each person
and try to guess that person's academic major. Use the following list of majors to help
you. You may use any major more than once, but please assign only one major per
person. NOTE: You are not restricted to this list of majors. If you think that a person has
a major that is not on the list, then you may use it instead on one on the list.
Accounting Anthropology Architecture
Biology Business Chemistry
Cinema-TV Communications Economics
English Fine Arts Geography
Geology History International Relations
Mathematics Music Nursing
Philosophy Political Science Psychology
Religion Sociology Theater
After you have made your guess, please complete the responding styles questions for each
of the persons. For all of the questions, please circle the number on the scales that best
corresponds to how you feel.
75
E.J.V.
Imaginative
Cultured
Ambitious
Emotional
Major: _______________________
Is this person's major the sam e as yours? (please circle one) yes
1. How likable is this person? (How much would you like this person?)
no
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
EJV EJV You
5 .
You EJV You EJV
You EJV
You EJV
3. How similar is this person to you?
1
1
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 2
very little a moderate amount
6 7
very much
76
EJLA
Efficient
Well-read
Methodical
Reliable
Major: _______________________
Is this person's major the sam e as yours? (please circle one) yes
1. How likable is this person? (How much would you like this person?)
no
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor uniikabte unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You FEA
FEA You
S .
You FEA
You
OO
You FEA
You FEA
3. How similar is this person to you?
I _ _ _ _ _ _ _ I _ _ _ _ _ _ _ I _ _ _ _ _ _ _ I _ _ _ _ _ _ _ I _ _ _ _ _ _ I - - - - - - J
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I - - - - - - - - - - - - - - - 1 _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ I ------------- J
1 2 3 4 5 6 7
very l i t t l e a moderate amount very much
77
J-IVU.
Irritating
Well-read
Argumentative
Authoritative
Major: _______________________
Is this person's major the sam e as yours? (please circle one) yes no
1. How likable is this person? (How much would you like this person?)
I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ l_ _ _ _ _ _ _ l_ _ _ _ _ _ _ _ l
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
JM J
4
You JMJ
CO
You JMJ
S.
You JMJ You JMJ You JM J
3. How similar is this person to you?
I
1 2 3 4 5
very moderately slightly neither dissimilar slightly
dissimilar dissimilar dissimilar nor similar similar
4. How much do you feel you have in common with this person?
6 7
moderately very
similar similar
I ________
1
very l i t t l e a moderate amount
7
very much
i
78
N. A. L
Talkative
Argumentative
Opinionated
Careless
Major: _______________________
Is this person’s major the sam e as yours? (please circle one) yes
1. How likable is this person? (How much would you like this person?)
no
I ___________ I __________ I _ _
1 2 3
very moderately slightly
likable likable likable
I
4 5 6 7
neither likable slightly moderately very
nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
Eire. Please indicate how close you feel to this person.
You
NAL
You NAL
5 .
You
NAL You NAL
OO
You NAL
You NAL
3. How similar is this person to you?
1 I ________
1 2 3
very moderately slightly
dissimilar dissimilar dissimilar
1 1
4 5
neither dissimilar slightly
nor similar similar
4. How much do you feel you have in common with this person?
6 7
moderately very
similar similar
J
1
very l i t t l e a moderate amount
very much
79
N.M. M .
Emotional
Neurotic
Sensitive
Unlucky
Major: _______________________
Is this person’s major the sam e as yours? (please circle one) yes
1. How likable is this person? (How much would you like this person?)
1
no
1 2 3 4 5
very moderately slightly neither likable slightly
likable likable likable nor unlikable unlikable
6 7
moderately very
unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
NMM You NMM
S.
You NMM You NMM
NMM You
6 .
NMM You
3. How similar is this person to you?
1
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I
1
very little
3 4
a moderate amount
6 7
very much
BO
R- T. L.
Idealistic
Attentive
Open-minded
Individualistic
Major: _______________________
Is this person's major the sam e as yours? (please circle one) yes
1. How likable is this person? (How much would you like this person?)
1
no
1 2 3 4 5
very moderately slightly neither likable slightly
likable likable likable nor unlikable unlikable
6 7
moderately very
unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing
or interconnectedness between you and this person. On the scale below, you
will notice that the pairs of figures are various distances apart. The closer the
two figures are to each other, the closer, or more interconnected they are. On
the other hand, the further they are apart, the less close, or interconnected they
are. Please indicate how close you feel to this person.
You
an. You an.
5 .
You
an. You an.
an. You
S.
an. You
3. How similar is this person to you?
1 1
1 2 3 4 5
very moderately slightly neither dissimilar slightly
dissimilar dissimilar dissimilar nor similar similar
4. How much do you feel you have in common with this person?
6 7
moderately very
similar similar
1 2
very little a moderate amount
very much
8!
Appendix C
1. How likable is this person? (How much would you like this person?)
I __________I __________ I _________I __________ I __________I _________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
K _______________ 2 . ____________ _ _ 3 .
o o
Y w EJV
OO
Ym EJV
CO
You EJV
G D
Yoa EJV
G D
Vm EJV
6.
o
Y ia EJV
3. How similar is this person to you?
I __________I __________I ________ II__________I _________ I _________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I ___________ I ___________ I ___________ l - I ___________ I ___________J
1 2 3 4 5 6 7
very little a moderate amount very much
F.E.A.
1. How likable is this person? (How much would you like this person?)
1 __________ I __________ 1 __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o o
Von FEA
OO
Y ou. FEA
OO
Yh FEA
G D
Von FEA
G D
Y aa FEA
S.
O
Y»» FEA
3. How similar is this person to you?
1 ____________ I ____________ I ___________J ____________ I ____________ I ____________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 2 3 4 5 6 7
very little a moderate amount very much
■I.M. .1 .
83
1. How likable is this person? (How much would you like this person?)
I __________ I __________ I __________ I __________ I __________ 1 __________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
I.__________________ 2 . 3 .
o o
You JM J
O O
You JM J
OO
Ym j m j
QD
Y m JMJ
SD
Ym JMJ
S.
o
Ym JM J
3. How similar is this person to you?
I ____________ I ____________ I ___________J ____________ I ____________ I ____________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 2 3 4 5 6 7
very little a moderate amount very much
84
iL A JL
1. How likable is this person? (How much would you like this person?)
1 __________ I _________ I ___________I ___________ I __________ I __________I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikab le unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharmg or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o o
You NAL
OO
Yoa NAL
OO
Yoa NAL
«• 5. S.
G D
You NAL
G D
Y aa NAL
o
Yoa NAL
3. How similar is this person to you?
1 2 3 - 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 _ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ _ I_ _ _ _ _ _ _ J
1 2 3 4 5 6 7
very little a moderate amount very much
85
1. How likable is this person? (How much would you like this person?)
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
p p
Ym NMM
2.
OO
Yo* NMM
OO
Ym NMM
G D
Ym NMM
5.
G D
Ym NMM
G.
o
Ym NMM
3. How similar is this person to you?
1 2 3 - 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar sim ila r sim ilar similar
4. How much do you feel you have in common with this person?
1 2 3 4 5 6 7
very little a moderate amount very much
86
RiT.L.
1. How likable is this person? (How much would you like this person?)
I __________ I __________ I __________ I __________ I __________ 1 __________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o d
Y «« RTl.
OO
Yoa R1L
GO
Yea Hit.
G D
Yoa HtL
G D
Yea mrt.
s.
o
Yea R T L
3. How similar is this person to you?
I
1 2 3 ' 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I ___________I _ _ ________ I ___________I ___________ I ___________L _________ 1
1 2 3 4 5 6 7
very little a moderate amount very much
In this part of the experiment your task is to write a description of each of the people who
you saw in the first part of the experiment. When you write your descriptions, please
include anything that comes to mind about each person, even if it was not part of the
original description in the first part of the experiment. You may write your descriptions on
the lines provided below, and you may use the back of this page if you need more room for
the descriptions. Please begin each description with the person's initials if you can recall
them. If you cannot, then please identify each person at the beginning of the description
(for example, "Person 1"). Please do not worry about correct spelling or complete
sentences.
Appendix D
E.J.V .
Major: _____________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
1 2 3 4 5 6 7
very moderately slighdy neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o o
You EJV
2.
o o
Ym EJV
OO
You EJV
G D
Yoa EJV
G D
Yaa EJV
G .
o
Yo« EJV
3. How similar is this person to you?
I _________ I _________ I _________ I __________I _________ I _________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I _________ I __________I _________ I _________ I _________ I _________ I
1 2 3 4 5 6 7
very little a moderate amount very much
89
F .E .A .
Major: ______________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
I __________ I __________ I __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
1 ___________________ 2._____________________ 3.
o o
You FEA
OO
Ym FEA
OO
Ym FEA
G D
You FEA
( 2 D
You FEA
6.
o
Ym FEA
3. How similar is this person to you?
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have
1 1 1
in common with this person?
1 1 1 1
1 2 3 4 5 6 7
very little a moderate amount very much
9 0
Major: _______________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o o
You j m j
2.
OO
You JMJ
3.
OO
Yd* JMJ
*■ 5. 6.
G D
You JMJ
G D
You JMJ
o
Yoa JMJ
3. How similar is this person to you?
I __________ I __________ I __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 2 3 4 5 6 7
very little a moderate amount very much
91
N.A.L.
Major. _______________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
1 __________ I ___________I _________ 1 ___________I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
J.________________ 2. ________________ 3.
o o
You NAL
OO
You NAL
OO
You NAL
G O
You NAL
C D
You NAL
6 .
o
You NAL
3. How similar is this person to you?
I __________ I __________ I __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
1 __________ I __________ I __________ 1 __________ 1 __________ 1 __________ I
1 2 3 4 5 6 7
very little a moderate amount very much
92
N.M.M.
Major: _______________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable nor unlikable unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hand, the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
! ____________________ z._________ 3.
o o
You NMM
OO
You NMM
OO
You NMM
GD
You NMM
G D
You NMM
S.
o
You NMM
3. How similar is this person to you?
I __________ I __________ I __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I ____________I ____________ I ____________ I ____________ I ____________ I ____________I
1 2 3 4 5 6 7
very little a moderate amount very much
93
T » L.
Major: _______________________
Is this person's major the same as yours? yes no
1. How likable is this person? (How much would you like this person?)
1 ___________ I _________ I __________ I __________ I __________ I __________I
1 2 3 4 5 6 7
very moderately slightly neither likable slightly moderately very
likable likable likable norunlikabie unlikable unlikable unlikable
2. How close do you feel to this person? By close, we mean a sense of sharing or
interconnectedness between you and this person. On the scale below, you will notice that
the pairs of figures are various distances apart. The closer the two figures are to each
other, the closer, or more interconnected they are. On the other hantC the further they are
apart, the less close, or interconnected they are. Please indicate how close you feel to this
person.
o o
You RTL
2.
OO
Ym HTl.
OO
Ym RTL
OD
y™ on.
d X >
You RTL
6.
o
You RTL
3. How similar is this person to you?
I __________ I __________ I __________ I __________ I __________ I __________ I
1 2 3 4 5 6 7
very moderately slightly neither dissimilar slightly moderately very
dissimilar dissimilar dissimilar nor similar similar similar similar
4. How much do you feel you have in common with this person?
I __________ I __________ I __________ I __________ I __________ l__________I
1 2 3 4 5 6 7
very little a moderate amount very much
94
In this part o f the experiment your task is to write a description of each of the people who
you saw in the first part of the experiment. When you write your descriptions, please
include anything that comes to mind about each person, even if it was not part of the
original description in the first part of the experiment. You may write your descriptions on
the lines provided below, and you may use the back of this page if you need more room for
the descriptions. Please begin each description with the person's initials if you can recall
them. If you cannot, then please identify each person at the beginning of the description
(for example, "Person 1"). Please do not worry about correct spelling or complete
sentences.

Linked assets