Ingroup Favoritism in Positive and Negative Domains

Document Sample
Ingroup Favoritism in Positive and Negative Domains Powered By Docstoc
					      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

Ingroup Favoritism in Positive and
Negative Domains
Prof.Dr., Political Science and Public Administration,
Eastern Mediterranean University

Assoc.Prof.Dr., Psychological Counseling Guidance and
Research Center, Eastern Mediterranean University

       Earlier research employing minimal group paradigm reported
ingroup favoritism in the positive domain but no bias in the
negative domain (Mummendey, 1992; Wenzel & Mummendey,
1996). This study tested the asymmetry in ingroup favoritism
between the positive and the negative domains employing a
different strategy; unlike earlier research, naturally formed “real
groups” were employed. One-hundred and sixty Turkish Cypriot
university students rated the degree of applicability of a list of
positive and negative attributes, with different levels of relevance
to categorization, to ingroup and outgroup drivers who were
involved in traffic accidents, and assigned them penalty scores as a
measure of responsibility for the accident. The ingroup was a
citizen and the outgroup was a foreigner with temporary residence
in North Cyprus. The results were as predicted, indicating ingroup
favoritism in both domains and higher responsibility for the
outgroup target.

Key words: Ingroup favoritism, symmetricity in ingroup
           favoritism, intergroup perceptions.

     According to the Social Identity Theory (Tajfel, 1982; Tajfel
& Turner, 1986) categorizing the social world into groups often

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

leads to intergroup bias. In order to achieve and maintain positive
social identities, persons favor their own groups over others.
Although evidence to the contrary is available, most research
demonstrated that the intergroup bias manifests itself as ingroup
favoritism (see Diehl, 1990 for an overview). However, the bulk of
research in this area has been concerned with positive evaluations
and reward allocations with respect to in- and outgroups. The
results of the limited number of studies dealing with negative
evaluations do not permit us to make a conclusive statement about
the nature of intergroup bias. For instance, a series of studies, with
a focus on negative evaluations, Mummendey and her colleagues
(Mummendey, Simon, Dietze, Grünert, Haeger, Kessler, Lettgen &
Schaferhoff, 1992; Mummendey, Otten & Blanz, 1994;
Mummendey, 1995; Wenzel & Mummendey, 1996) were not able
to demonstrate clear ingroup favoritism or only a weaker form as
compared to positive evaluations. The author concluded that there
exists an asymmetry between positive and negative domains and
argued that a norm of fairness precludes the discriminatory process
in negative evaluations and negative outcome allocations.
Reynolds, Turner and Haslam (2000), on the other hand, suggested
that it is not the positivity or the negativity of evaluative
dimensions that accounts in the process, but rather the “fit” of the
positive and negative evaluations to the ingroup and/or the
outgroup (p.68). They reported ingoup bias when the traits were
typical of either one of the groups but no bias when they were not
typical of any group. More specifically there was ingroup bias
when positive traits were typical of the ingroup and when the
negative traits were typical of the outgroup.
       Most of the data on ingroup favoritism came from studies
that employed the minimal group paradigm. Although minimal
group conditions ensure control of relevant variables, group
categorization in real life has more meaningful and significant
implications with respect to intergroup relations and bias.
Perceived group identities are not independent of intergroup
relations and expectations (Turner, Hogg, Oakes, Reicher &
Wetherell, 1987). Since categorization often serves to enhance and
maintain a positive self-esteem, claiming or disclaiming a given
group identity has significance in respect to persons’ motives,

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

goals, and expectations. Minimal group conditions involve groups
that are constructed randomly or on the bases of trivial criteria.
Such groupings may be perceived as irrelevant especially when
negative evaluations or resource allocations are to be made without
adequate reason for discriminating against the outgroup. As
Reynolds et al. (2000) suggested, membership to a minimal group
under negative conditions may “not provide an appropriate or
meaningful basis for self-definition … and therefore, group
identification and intergroup discrimination are minimized” (italics
added, p. 68).
        In real life conditions, groups are formed naturally through
social processes and such 'real groups’ (Mullen, Brown & Smith,
1992) are meaningful and contain perceptual and evaluative
connotations (Crocker & Luhtanen, 1990). Memberships of such
groups are inferred from physical appearance, properties,
belongings, posts and positions, signs and symbols. For example,
physical characteristics constitute cues to a target’s sex group and
skin color to the individual's racial group. The categorization
process triggered by signs and symbols or daily events cannot be
context-independent. An intergroup situation in real life, as
compared to an artificially created situation, bears more
significance for self-other definition. Because such groupings
necessitate memories and images of intergroup relations in the past
and encompass expectations in the future. In the present study, an
attempt was made to examine social categorization as it occurs in
real life.
        A real life situation calls for context-relevant perceptions
and evaluations. Certain aspects of the situation or of the
categorization context are more relevant than some others or not
relevant at all. A high degree of correspondence between salient
aspects of the situation and the evaluative dimension is expected to
accentuate between-group discrimination. Therefore, the relevance
of the stimuli and the evaluative dimension are important within
the context of categorization and should be significant variables in
the study of intergroup bias. The meaning of the context-related
term “relevance” as used here is somewhat different from the
meaning of the term "relevance" in earlier studies. Usually, the
term relevance has been used to point out the perceived

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

significance of attributes or evaluative dimensions by participants
who are making intergroup comparisons (Mullen, Brown & Smith,
1992). Attributes stereotypical of the ingroup or of the outgroup
have been found to increase intergroup bias (Spears & Manstead,
1989; Mullen et al., 1992). Similarly, Reynolds et al.’s (2000) use
of the terms “fit” or “normative fit,” indicating whether an
evaluative item was typical of the ingroup or of the outgroup, is
also related to the extent that an attribute characterizes a
stereotype. A context-related definition of the term “relevance”, on
the other hand, refers to the degree to which an evaluative item or
dimension relates to the context of categorization. No a priori
reference is made to any stereotype or attribute of a social group.
For instance, to evaluate an individual's achievement outcome (i.e.,
either success or failure), it is more relevant to focus on issues such
as hard work and ability, than on issues such as the individual
being sincere and polite. Similarly, arrogance and aggressiveness
may be more relevant in the evaluation of an interpersonal conflict
or dispute than charm and sincerity. In neither example is any
reference made to group membership of the individual. Items that
are highly relevant in the particular situation should be meaningful
for an observer and affect his/her evaluation of the actor. In short,
the argument here is that intergroup comparisons under natural
conditions cannot be independent of the context or social situation
that leads to categorization. In addition, the relevance of the
evaluative items for the categorization outcomes should increase
intergroup bias.
       In the current study, ingroup favoritism in the positive and
the negative domains was examined by employing social
categorization as it occurs in real life. For this purpose a traffic
accident is taken as the important event. Traffic accidents are
significant events that are perceptually highly salient since they
involve potential threats to individuals’ life and property as well as
to public order and security. Exposure to such events are expected
to have significance for the self and, therefore, to motivate
intergroup bias. In establishing a link between causal attributions
and social identity, Hewstone (1990) suggested that attributional
biases relate to the maintenance and enhancement of a positive
social identity. More specifically, a positive identity can be

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

achieved (a) by attributing ingroup success or outgroup failure to
internal, stable, global, and personally controllable factors, and (b)
by attributing ingroup failure or outgroup success to external,
unstable, and uncontrollable factors. Similarly, rating an ingroup
actor higher on positive attributes and lower on negative ones than
an outgroup actor will serve the need to achieve positive
distinctiveness. It also follows that when a comparison is made
between the ingoup and the outgroup, the former is perceived as
less responsible for a negative outcome than the latter. In a list of
positive and negative attributes, positive attributes will be judged
more characteristic of the ingroup than the outgroup, and the
reverse would be the case with the negative attributes. An ingroup
driver who was involved in a car accident, on the other hand,
would be penalized less severely than an outgroup driver. In this
study, we expected that ingroup favoritism would occur with the
penalty assignments as well as the evaluative responses in the
positive and negative domains. In addition, we expected ingroup
favoritism to be more salient in the negative domain because
negative attributes are more relevant to categorization than positive
ones. With respect to the degree of relevance of an attribute to the
particular behavior or event used in the categorization process, it
was predicted that attributes with a higher relevance have a
stronger impact on ingroup bias than attributes with a lower
relevance. It also follows that ingroup favoritism is strongest with
the high-relevant negative attributes and weakest with the low-
relevant positive attributes.

1. Method
1.1 Pilot studies
       Prior to the presentation of the event of a traffic accident, the
evaluative valence (i.e., domain) and relevance of 20 attributes had
to be determined. The majority of these attributes (i.e., words)
were taken from a list used in an earlier study (Rustemli, Mertan &
Ciftci, 2000). Two pilot studies were conducted to determine the
evaluative valence and relevance of the attributes to categorization.
In the first pilot study, 56 university students (i.e., judges),
classified the attributes into either the positive or negative domain.

       Review of Social, Economic & Business Studies, Vol.7/8, 175-190

To determine the relevance of the attributes to an actor (i.e., in this
case a driver who was involved in a car accident) one half of the
judges selected half of the attributes from each evaluative domain
that described (a) a good driver who has never been involved in
any car accident (i.e., the positive, high-relevance attributes), and
(b) a bad driver who has been involved in car accidents (i.e., the
negative, high-relevance attributes). The other half of the judges
did the same but this time with the request to select half of the
attributes from each evaluative domain that did not describe or
were less descriptive of the two types of drivers (i.e., the low-
relevance attributes).
       There was complete agreement among the judges in
classifying attributes as positive or negative on the evaluative
dimension and, although somewhat lower, high degrees of
agreement on the relevance dimensions.            In Table 1, the
classification of the attributes along the evaluative and relevance
dimensions is presented. The numbers in parentheses indicate the
percentages of judges who placed the attributes in the respective
relevance categories. As shown in the table, the percentages of
judges classifying the attributes along the relevance dimension
varied between 67.9 (helping) and 100.0 (responsible, cold, and

  Table1. Classification of Attributes by Evaluative Domain and Relevance
  Evaluative domain      High                      Low

  Positive               Careful (91.1)            Friendly (98.2)
                         Courageous (69.6)         Good (75.0)
                         Reliable (75.0)           Helpful (67.9)
                         Responsible (100.0)       Honest (92.9)
                         Tolerant (73.2)           Intelligent (78.6)
                         Aggressive (96.4)         Cold (100.0)
  Negative               Dangerous (100.0)         Lazy (96.4)
                         Disrespectful (89.3)      Rude (76.8)
                         Irritable (89.3)          Unfair (78.6)
                         Selfish (69.6)            Unmannerly (83.9)
Note: The numbers in parentheses indicate the percentages of judges who placed
the attributes in the respective relevance category.

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

       The second pilot study was conducted to determine whether
there existed clear valence differences between the positive and the
negative attributes. This was essential for the examination of
differences in ingroup favoritism between the negative and positive
domains. Another sample of 24 judges, again all university
students, rated the attributes as being either socially positive or
negative. No reference was made to a “good driver” or a “bad
driver”. A 7-point scale was used ranging from “very positive”
(3), “positive” (2), “a little positive” (1), “neither positive nor
negative” (0), “a little negative (-1), “negative” (-2) to “very
negative”(-3). A comparison of the absolute averages of the
positive and negative attributes revealed no significant valence
difference between the two categories, t (23) = 1.24, p > .05. The
averages for the positive and the negative attributes were 2.36 and -
2.24, respectively.

1.2 Participants and Design
       The participants were 160 Turkish Cypriot students (i.e.,
equal numbers of males and females) taken from two universities
in North Cyprus. The ages of the participants varied between 17
and 27 (Mean age = 19.80; sd = 1.97). The study involved a 2
(target group: ingroup versus outgroup) by 2 (attribute domain:
positive versus negative), by 2 (attribute relevance: high versus
low) factorial design with repeated measures on all factors.

1.3 Materials
        Photographs of collided cars were used as stimulus material.
Two cars were        photographed twice; once with a DP… license
plate and once with a ZZ… license plate. By law, the former plate
(i.e., DP) can be owned by citizens of North Cyprus only and the
latter (i.e., the ZZ license plate) can be owned by a foreigner who
resides in North Cyprus on a temporary basis. Hence, we assumed
that the driver of the DP licensed car was perceived as a member of
the ingroup and the driver of the ZZ plated car as a member of an
outgroup. No human being appeared in the photographs. The
photographs were scanned into a computer and cleared of all kinds

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

of cues except the cars and the license plates. A questionnaire was
prepared by using two versions of the photographs. Each version
had both the DP- and the ZZ-plated cars but the make of cars was
alternated to balance a potential effect arising from this factor.
Thus, the ingroup-outgroup manipulation was achieved by the
different license plates.
        The first part of the questionnaire included the instructions
and an explanation of the study. In order to motivate interest in the
study and to disguise its purpose, the explanation read as “… a
study aiming to investigate how people perceive and react to
drivers who get involved in car accidents”. Questions asking for
personal information about age, sex, place of birth, etc. followed
the explanation. The second page of the questionnaire started as
follows: “Now, you will see photographs of cars taken after traffic
accidents. What we want you to do is to inspect these pictures very
carefully, and then, indicate the degree to which the driver was
responsible for the accident, and make judgments about the
driver’s personality. Since there is no right or wrong answers to
these questions, please indicate your immediate responses”. The
second page of the questionnaire had either the DP plated car or the
ZZ plated car followed by a set of questions about the accident and
the driver. The degree of responsibility of driver (i.e., actor) for the
accident was measured by asking participants to impose a penalty
by indicating a score out of hundred points, with the instruction
that 100 meant complete responsibility and 0 (zero) no
responsibility at all. To assess the personality of the driver, the
attributes were presented as a list and the participants were asked
to indicate the extent of applicability on a four-point scale ranging
from “not applicable at all” (1) to “most applicable” (4). The
attributes in the list were presented in the same random order.
       The next page was identical to the second with the exception
that if the former page had the DP plated car, the next would have
the ZZ plated car and reversely. Thus, each participant was
presented with both pictures but with different plates.

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

1.4 Procedure
      The questionnaires were administered individually or in
small groups on the campus of each university.

1.4.1 Results
       The penalty scores imposed to the drivers constituted the
responsibility measure. As predicted, the average score for
responsibility attributed to the ingoup (M = 54.75, SD = 29.52)
was significantly lower than responsibility score for the outgroup
(M = 67.62, SD = 26.86); t (158) = 5.59, p < .001.
       The alpha coefficients for the categories of the attribute
relevance-domains varied between .45 (i.e., the high-relevance-
positive category) and .72 (i.e., the low-relevance-positive
category). Although a coefficient of .45 is relatively low, it is not
rare in the literature. The coefficient for high-relevance-negative
attributes was .70, and for low-relevance-negative attributes. 65.
       Ingroup favoritism scores for each of the attribute domain
and relevance categories were computed. To do so, the outgroup
evaluation was subtracted from the ingroup evaluation for the
positive attributes, and the ingroup evaluation was subtracted from
the outgroup evaluation for the negative attributes, next the average
for each domain-relevance category was computed. Thus, a
positive score would mean ingroup and a negative score outgroup
favoritism, and a value of zero or close to zero would mean no
discrimination between the groups. As Table 2 shows, all the
average favoritism scores were positive, the largest being 0.44 for
high-relevance-negative attributes and the lowest being 0.18 for
low-relevance–positive attributes. To examine whether these
averages implied ingroup favoritism, we tested for significant
difference from zero by means of single sample t-tests. The t-
values varied from 3.26 to 6.87, with degrees of freedom equal or
larger than 142 indicating that ingroup favoritism was present in all
of the attribute domain-relevance categories.
       An initial analysis of variance on favoritism scores by sex,
make of car, order of presentation, attribute domain and attribute
relevance indicated no significant interaction and main effects for
the first three factors. Consequently, the data were pooled for

       Review of Social, Economic & Business Studies, Vol.7/8, 175-190

these factors and subjected to a two-factor ANOVA with repeated
measures. The analysis revealed significant main effects both for
the evaluative domain ( F (1, 132) = 6.33, p < .02), and relevance
(F (1, 132) = 15.52, p < .001). Average favoritism for the negative
attributes (i.e., 0.39) was significantly larger than average
favoritism for the positive attributes (i.e., 0.21). Similarly, the
high-relevance attributes produced more bias than the low-
relevance attributes, the averages being 0.34 and 0.26, respectively.
In Table 2, the average ingroup favoritism scores and the standard
deviations are presented for each of the attribute-domain-relevance
categories. Favoritism was the highest for the high-relevance–
negative attributes (i.e., 0.44) and the lowest for the low-relevance
– positive attributes (i.e., 0.18). Paired comparisons of the averages
indicated that ingroup favoritism was significantly higher in the
high-relevance-negative attributes than in the other attribute
categories. Degree of relevance did not result in a significant
difference in ingroup bias when the attributes were positive.

Table2. Average Ingroup Favoritism Scores and Standard Deviations (italics)
             by Attribute Domain and Relevance Categories.
                              Attribute domain
      Attribute relevance        Positive Negative          t-test

      High        M              0.24        0.44           6.43***
                  SD             0.58        0.77
      Low         M              0.18        0.34           5.16***
                  SD             0.69        0.73
      t-test                     1.31        2.34*

Note: * alpha = .05. *** alpha = .001.

       The favoritism scores, as computed in the present study offer
information about the size and direction of the bias, but do not
offer information about the applicability (i.e., in absolute terms) of
the attributes in the domain-relevance categories to members of the
in- and outgroup. At the risk of redundancy, a reanalysis of the data
by a 2x2x2 (target group x evaluative domain x relevance)

       Review of Social, Economic & Business Studies, Vol.7/8, 175-190

ANOVA with repeated measures, is presented very briefly. The
main effects for target group (F (1, 132) = 15.52, p ≤.001), domain
(F (1, 132) = 27.06, p ≤ .001), and relevance (F (1,132) = 6.14, p ≤
.02), as well as the interactions between target group by domain (F
(1,132) = 35.18, p ≤ .001) and the domain by relevance (F (1, 132)
= 70.75, p ≤. 001) were significant. Also the three-way interaction
among the factors was significant (F (1, 132) = 6.33, p ≤ .02). In
Table 3, the average values involved in these analyses are

  Table 3.Cell Averages and Related t-values by Target Group by Attribute
                           Domain by Relevance
                                 Target group
  Attribute                      Ingroup      Outgroup               t
  Positive                        2.36          2.15           3.93***
  High relevance   M              2.30          2.07           4.64***
                   SD             0.53          0.50
  Low relevance    M              2.42          2.24           3.26***
                   SD             0.58          0.61
  Negative                        2.41          2.80           7.06***
  High relevance   M              2.50          2.95           6.87***
                   SD             0.63          0.60
  Low relevance    M              2.32          2.66           5.73***
                   SD             0.63          0.64
Note: All tests are one-tailed; *** p < .001.

        The data in Table 3, suggest that the ingroup was favored
over the outgroup by assigning a higher standing on positive
attributes and a lower standing on the negative ones. Attribute
relevance increased discrimination both between the ingroup and
the outgroup and the positive and negative attributes within target
groups. The significant three-way interaction among the factors
originated from the magnifying effect of attribute-relevance That
is, in line with our prediction the strongest discrimination against
the outgroup occurred with the high-relevance negative attributes
(with averages of 2.95 and 2.50 for the outgroup and the ingroup,

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

respectively), and the weakest discrimination with the low-
relevance positive attributes (with averages 2.42 and 2.24 for the
outgroup and the ingroup, respectively).

1.4.2 Discussion
        In the present study, ingroup favoritism was examined by
positive and negative attributes. Ingroup-outgroup manipulation
was achieved by changing the license plate of a wrecked car - a
license plate that could be owned by a citizen of North Cyprus only
(i.e., the ingroup), and another plate that could be owned by a
foreigner with temporary residence in North Cyprus only (i.e., the
outgroup). The data demonstrated the presence of ingroup
favoritism both in the positive and the negative evaluative
domains. The ingroup was favored more positively than the
outgroup when positive attributes were involved and less when
negative attributed were involved. Favoritism was also present in
the attributions of responsibility for the accident which was
operationalized by a penalty score to the driver.
        These results provide support for the Social Identity Theory
that claimed that positive discrimination of the ingroup would
occur not only in the positive but also in the negative evaluative
domain. The participants favored the ingroup by employing a “we
are better than them” strategy with the positive attributes, and
“they are worse than us” strategy with the negative ones. The
observed symmetry in ingroup favoritism in the positive and the
negative domains replicates an earlier finding by Rustemli et al.
(2000) showing ingroup favoritism by native Turkish Cypriots
(i.e., the ingroup) and denigration of the outgroup involving
immigrants from Turkey who constituted a low status group in
respect to a variety of measures.
        The present findings, however, cannot be explained by status
differences between the ingroup and the outgroup, because the
outgroup involved foreigners in general with no reference to any
social, religious, ethnic, etc. group. The only thing that was clear
was that the outgroup consisted of foreigners. Our findings that
foreigners were perceived as a less favored outgroup was of itself
surprising, since from a historical perspective there was no reason

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

to assume any widespread negative or prejudiced attitudes toward
foreigners in the population of North Cyprus. In history, the island
of Cyprus has hosted peoples of different ethnic backgrounds,
nationalities, religions and languages. Familiarity and contact with
foreigners usually reduce prejudices rather than increase them. No
doubt, the recent conflict between the Turkish and Greek Cypriots
communities on the island might have lead to increased prejudice
and hostilities between both communities and toward foreigners in
general. However, even this reasoning cannot account for the
observed ingroup bias, because the foreigners in the present study
could not be members of the Greek Cypriot outgroup. By the time
the data were collected, intercommunal visits were strictly banned.
        In the present study, the ingroup and the outgroup were “real
groups”. Unlike minimal group conditions, membership to
naturally formed groups is salient and has strong effects on social
identities (Mullen et al., 1992). If minimal group conditions are
adequate to produce intergroup bias, one would expect the mere
existence of real groups to lead to a stronger bias, let aside their
behavior. In the present research, the participants were presented
with negative outcomes involving an outgroup as well as an
ingroup actor. Social comparisons between the ingroup and the
outgroup in such a situation was “probably perceived as
sufficiently threatening to evoke ingroup favoritism and/or
outgroup derogation as a means of defending that, the social,
identity” (Branscombe, Ellemers, Spears & Doosje, 1999, p.46;
italics added). Therefore, the present findings suggest that ingroup
favoritism is not limited to the positive evaluative domain, but also
occurs in the negative domain. The asymmetry in ingroup bias
reported in some earlier studies was based on minimal group
conditions (Mummendey et al., 1992: Wenzel & Mummendey,
1996). Membership of minimal groups may not be sufficiently to
evoke defensive strategies to protect one's social identity.
        When the relative strength of ingroup favoritism in the
positive and the negative domains were compared, the latter was
significantly stronger. When also the absence of any difference
between the average valence of the positive and the negative
attributes is taken into account, the stronger ingroup bias in the
negative domain can be attributed only to the nature of the event

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

used in the categorization process. The negative outcome of the
actor's behavior lead participants to weight the negative attributes
more heavily than the positive attributes. In other words, the
negative attributes possessed a higher relevance for the
categorization process than the positive attributes. High-relevance
attributes such as behaving carefully, aggressively, and
dangerously resulted in stronger effects than the low-relevance
attributes such as behaving honestly, lazily, and coldly. However,
the interaction between evaluative domain and attribute relevance
indicated that the relevance effect was valid in the negative
domain. In line with our predictions, the strongest ingroup
favoritism was observed with the high-relevance, negative
attributes and weakest ingroup favoritism with the low-relevance,
positive attributes. Although a difference in the expected direction
was observed for high- and low-relevance, positive attributes, this
difference did not reach significance. It could be that the positive
attributes are perceived as less relevant in the case of a negative
event than the negative attributes. In general, little differentiation
was observed between high- and low-relevant items in the positive
        In summary, the data from naturally formed “real groups”
that were obtained in the present study by using a new research
methodology demonstrated ingroup favoritism in evaluations in
both the positive and the negative domains and the attribution of
responsibility. Intergroup bias was magnified by the relevance of
the evaluative attribute for the categorization process, and
especially so with negative attributes that possess higher relevance.
However, it is important to note that the present study was limited
in the nature of employed manipulation that consisted of a negative
event - a car accident involving ingroup and outgroup actors. The
stronger ingoup favoritism that was observed in the negative
evaluative domain might be due to the higher relevance of the
domain to the categorization. Additional studies employing
negative as well as positive events that mimic categorization in real
life are needed to examine these effects in greater detail.

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

    DOOSJE, B. (1999) The context and content of social
    identity threat. Malden, Massachusetts; Blackwell. In N.
    Ellemers, R. Spears, & B. Doosje (Eds.), Social identity:
    Context, commitment, content. pp. 46.
CROCKER, J., & LUHTANEN, R. (1990) Collective self-esteem
    and in-group bias. Journal of Personality and Social
    Psychology, 58, pp. 60 – 67.
DIEHL, M. (1990) The minimal group paradigm: Theoretical
    explanations and empirical findings.European Review of
    Social Psychology, Vol.1, pp. 263-92.
HEWSTONE, M. (1990) The ‘ultimate attribution error’ ? A
    review of the literature on intergroup causal attribution.
    European Journal of Social Psychology, 20, pp. 311-335.
MULLEN, B., BROWN, R., & SMITH, C. (1992) In-group bias as
    a function of salience, relevance, and status: An integration.
    European Journal of Social Psychology, 22, pp. 103-122.
MUMMENDEY, A. (1995) Positive distinctiveness and social
    discrimination: An old couple living in divorce. European
    Journal of Social Psychology, 25, pp. 657-670.
MUMMENDEY, A., OTTEN, S., & BLANZ, M. (1994) Social
    categorization and intergroup discrimination. The asymmetry
    in positive and negative outcome allocations. Revue
    Internationale de Psychologie Sociale, 1, pp. 15-30.
    SCHAFERHOFF, S. (1992) Categorization is not enough:
    Intergroup discrimination in negative outcome allocation.
    Journal of Experimental Social Psychology, 28, pp. 125-144.
REYNOLDS, K. J., TURNER, J. C., & HASLAM, S. A. (2000)
    When are we better than them and they worse than us? A
    closer look at social discrimination in positive and negative
    domain. Journal of Personality and Social Psychology, 78,
    pp. 64-80.
RÜSTEMLİ, A., MERTAN, B., & CİFTCİ, O. (2000) Ingroup
    favoritism among native and immigrant Turkish Cypriots:

      Review of Social, Economic & Business Studies, Vol.7/8, 175-190

    Trait evaluations of ingroup and outgroup targets. The
    Journal of Social Psychology, 140, 1, pp. 26-34.
SPEARS, R., & MANSTEAD, A. S. R. (1989) The social context
    of stereotyping and differentiation. European Journal of
    Social Psychology, 19, pp. 01-21.
TAJFEL, H. (1982) Social psychology of intergroup relations.
    Annual Review of Psychology, 33., pp. 1-39.
TAJFEL, H., & TURNER, J. C. (1986) The social identity theory
    of intergroup behavior. In: S. Worchel, & W. G. Austin
    (Eds.), Psychology of Intergroup Relations. Chicago:
    & WETHERELL, M. S. (1987) Rediscovering the social
    group. A self-categorization theory. Oxford. England: Basil
WENZEL, M., & MUMMENDEY, A. (1996) Positive-negative
    asymmetry of social discrimination: A normative analysis of
    differential evaluations of in-group and out-group on positive
    and negative attributes. British Journal of Social Psychology,
    35, pp. 493-507.


Shared By: