European Journal of Social Psychology

Eur. J. Soc. Psychol. 31, 143±154 (2001)

On the measurement of social orientations

in the minimal group paradigm: norms as
moderators of the expression of intergroup bias

LOWELL GAERTNER1* AND

CHESTER A. INSKO2
1
Texas A&M University, USA
2
University of North Carolina at Chapel Hill, USA

Abstract

Contrary to most other research conducted in the minimal group paradigm tradition, Bornstein, Crum,
Wittenbraker, Harring, Insko and Thibaut (1983a) found little evidence of ingroup favoritism when
they employed a revised measurement system (i.e. the Multiple Alternative Matrices; MAMs). The
current experiment examined whether Bornstein et al.'s effects could be attributed to norms that
prohibit intergroup discrimination, which are made salient by framing the outcome values in the
intergroup allocation task as monetary payment. We manipulated the salience of prohibitive norms by
varying whether participants allocated on the MAMs monetary payment, bonus money or feelings.
Participants more strongly associated `having to be fair and equal' with payment than with bonus or
feelings and category members made fewer allocations that maximized the ingroup's relative and
absolute pro®t and more allocations that minimized intergroup differences when allocating monetary
payment than when allocating bonus money or feelings. Copyright # 2001 John Wiley & Sons, Ltd.

The minimal group paradigm (MGP) experiments (Billig & Tajfel, 1973; Tajfel, 1970; Tajfel & Billig,
1974; Tajfel, Billig, Bundy, & Flament, 1971) suggested that persons assigned to novel social
categories discriminate in favor of their own category. Category members evidenced on a money
allocation task a concern for maximizing the ingroup's relative advantage over the outgroup. The
validity of these ®ndings was challenged when Bornstein et al. (1983a,b) detected with a revised
measurement system little evidence of ingroup favoritism. However, the absence of ethnocentric
behavior ± such as in the results of Bornstein et al. ± does not necessarily re¯ect the absence of
ethnocentric loyalty. Ingroup favoring actions may be suppressed in favor of egalitarian behaviors if
those ingroup favoring actions are deemed inappropriate for the situation. We examine in the current
research norms that prohibit the expression of intergroup discrimination and, consequently, mask the
detection of ingroup favoring motives. In particular, we argue that Bornstein et al. (1983a,b) framed

*Correspondence to: Lowell Gaertner, Department of Psychology, Texas A&M University, College Station, Texas 77843-4235,
USA. E-mail: lag@[Link]

Copyright # 2001 John Wiley & Sons, Ltd. Received 13 January 1999
Accepted 26 June 2000
144 Lowell Gaertner and Chester A. Insko

their intergroup allocation task in a manner that made salient norms of equity and equality that
prohibited discriminatory behavior.

NORMS AND INTERGROUP BIAS

Norms play an important role in the regulation of intergroup relations. Allport (1954), for example,
discussed the importance of social laws as a necessary step to the reduction of intergroup con¯ict.
Franco and Maass (1999) demonstrated that normative sanctions against discrimination moderate the
correlation between implicit and explicit measures of prejudice. Implicit and explicit measures of
prejudice were positively correlated for outgroups towards which the public expression of negative
opinions was deemed acceptable, while the correlations were weaker and non-signi®cant for
outgroups towards which the public expression of negative opinions was deemed unacceptable.
Indeed, Tajfel (1970) initially offered a generic norm of outgroup rejection to explain the discrimina-
tion that arises between minimal groups. This normative account, however, was ultimately abandoned
in favor of a social identity account (Tajfel & Turner, 1979).
Researchers have returned to an examination of the normative in¯uences on members of minimal
groups. For example, members of minimal groups act in accordance with the expected and
demonstrated behavior of other ingroup members. Members of minimal groups expect other ingroup
members to make ingroup favoring allocations (Gaertner & Insko, 2000; Jetten, Spears, & Manstead,
1996) and are more discriminatory when on a previous trial other ingroup members allocated more
resources than fewer resources to the ingroup (Jetten et al., 1996; Locksley, Ortiz, & Hepburn, 1980).
Likewise, the tendency for members of minimal groups to discriminate when allocating positive
outcomes but not negative outcomes has been accounted for, in part, by the normative inappropriate-
ness of differentially exposing ingroup and outgroup members to negative outcomes (Blanz,
Mummendey, & Otten, 1997). Indeed, some perspectives account for the totality of minimal-group
behavior in regard to social norms. Hertel & Kerr (in press) suggest that discrimination between
minimal groups is guided by a norm of group loyalty. Other researchers suggest that the norm of group
loyalty is tempered by self-interest such that members of minimal groups favor their group only when
they can personally pro®t from the allocations of other ingroup members (Gaertner & Insko, 2000;
Rabbie, Schot, & Visser, 1989; Yamagishi, Jin, & Kiyonari, 1999).
Rather than exploring norms that promote discrimination, we examine in the current research
norms that prohibit the expression of discrimination between minimal groups. We study, in particular,
norms that are differentially activated by subtle changes in the representation of resources to be
distributed on intergroup allocation tasks.

REPRESENTATION OF RESOURCES TO BE DISTRIBUTED

That allocation behavior can be signi®cantly altered by subtle changes in the representation of the
resources to be allocated has been acknowledged in previous research. Grzelak, Poppe, Czwartosz, &
Nowak (1988) report that cooperative and altruistic motives are more strongly activated when own and
other's satisfaction with money are represented iconically (i.e. smiling versus frowning faces) than
when represented geometrically (i.e. rectangles of varying length). Likewise, Rabbie and Schot (1990)
suggest that motives for economic self-interest and social competition are activated by allocation tasks
that involve money and points, respectively. Unlike money, points lack absolute value and have
meaning only in relation to the amount of points possessed by other persons or groups; consequently,

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
 Normative sanctions and intergroup bias 145

allocations involving points foster relativistic comparisons. Indeed, category members evidenced a
stronger concern for maximizing the ingroup's relative earnings (as assessed by the Tajfel matrices)
when allocating points than when allocating money (Rabbie & Schot, 1990).

BORNSTEIN ET AL. (1983a) AND NORMATIVE

CONSTRAINTS ON INTERGROUP DISCRIMINATION

Bornstein et al. (1983a) developed the Multiple Alternative Matrices (MAMs) to overcome problems
associated with the measure of intergroup discrimination that traditionally was used in minimal group
research (i.e. the Tajfel matrices1). The Tajfel matrices have been criticized for (a) confounding the
strategies of maximum differentiation and maximum ingroup pro®t (Brewer, 1979), (b) failing to
simultaneously provide participants with the full range of allocation strategies and, therefore, forcing
participants to select a less than optimal preference (Bornstein et al., 1983a), and (c) providing an
ambiguous measurement of the fairness strategy (Bornstein et al., 1983b).
The MAMs present simultaneously seven unconfounded options for distributing resources between
ingroup and outgroup members (see Table 1). Each option represents a distinct social orientation
(Kelley & Thibaut, 1978) or allocation strategy (Tajfel & Turner, 1979). Three of the orientations favor
the ingroup. Max Rel Own represents a concern for maximizing the ingroup's relative gain over the
outgroup. Max Own represents a concern for maximizing the ingroup's absolute gain. Max Joint Own
represents a concern for maximizing the joint earnings of both groups while providing an advantage to
the ingroup. One orientation, Min Dif, represents a concern for minimizing the differences between
groups. The remaining three orientations (Max Rel Other, Max Other, and Max Joint Other) favor the
outgroup in a manner symmetrical to the ingroup favoring orientations. Participants are instructed to
select the single orientation that best re¯ects how they would like the resources to be allocated (for a
criticism see Turner, 1983a,b; for responses see Bornstein et al., 1983b and Insko & Schopler, 1987).
Research with the MAMs did not replicate consistently the ®ndings of research with the Tajfel
Matrices. College-age females tended to minimize intergroup differences (i.e. Min Dif) and college-
age males tended to maximize the joint pro®ts of the ingroup and outgroup (i.e. Max Joint Own;
Bornstein et al., 1983a). On the other hand, Platow, McClintock, and Lieberand (1990; also see Chin &
McClintock, 1993) reported evidence of relative ingroup gain in their MGP research with the MAMs.
This later research, however, involved the allocation of points as opposed to money and, consequently,
may have fostered relativistic motives beyond that which is elicited by the social context of the MGP

Table 1. Ingroup and outgroup outcomes as a function of the multiple alternative matrices
Allocation strategy

Max Max Max Max

Rel Max Joint Min Joint Max Rel
Recipient Own Own Own Dif Other Other Other
Ingroup 42 47 44 34 36 30 22
Outgroup 22 30 36 34 44 47 42

Note: The seven allocation strategies re¯ect distinct options for distributing outcomes (e.g. money) to the ingroup and outgroup
and constitute one set of MAMs. For a mathematical derivation of possible outcome values see Bornstein et al. (1983a).

1
An anonymous reviewer suggested that the Tajfel matrices are misnamed in that the matrices were developed by Tajfel's
colleague Claude Flament. We use the term `Tajfel matrices' for the sake of consistency with previous publications.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
146 Lowell Gaertner and Chester A. Insko

(e.g. Rabbie & Schot, 1990). Bornstein et al. (1983a,b) concluded on the basis of their research with
the MAMs that the ingroup favoritism detected in the initial MGP experiments was due, in most part,
to the confounded system of measurement (i.e. Tajfel Matrices).
It is plausible, however, that Bornstein et al. (1983a,b) inadvertently made salient norms that
prohibited the expression of intergroup discrimination. In particular, Bornstein et al. (1983a)
represented the values in the outcome matrices of the allocation task as monetary payment to be
distributed to ingroup and outgroup members. The term `payment' may convey norms of equity and
equality that censure discrimination in a context in which opposing groups (or persons) are performing
equivalent tasks because payment is typically associated with reimbursement for products or services
rendered. On the other hand, Bornstein et al. also described the outcome values on the Tajfel matrices
as monetary payment and found evidence for maximizing the ingroup's relative and absolute pro®t.
Because the Tajfel Matrices do not present simultaneously all possible unconfounded allocation
strategies the distinct allocation strategies may be less salient to participants on the Tajfel matrices
than on the MAMs. Consequently, the use of the term `payment' in the context of an array of clearly
differentiated response alternatives (i.e. the MAMs) may have constrained participants from expres-
sing discriminatory inclinations. In other words, Bornstein et al.'s MAM research may have created a
situation in which norms of equity and equality were highly salient and the distinction between
discriminatory and egalitarian behavior was strikingly evident ± a situation in which intergroup
discrimination is very unlikely (e.g. Gaertner & Dovidio, 1986).

THE CURRENT RESEARCH

We conducted two experiments to examine whether intergroup discrimination is inhibited by

normative constraints made salient by task characteristics. We framed the MAMs in terms of the
allocation of bonus money, feelings, or payment. We described bonus money as extra money or an
award and informed participants completing the feelings-task that the outcome values represent how
positively they feel about the ingroup and outgroup. We chose bonus and feelings as alternatives to
payment because we assumed that these alternatives are associated less strongly with norms of equity
and equality (an assumption tested in Experiment 1).
Experiment 1 examined whether payment is more strongly associated with norms of equity and
equality than are bonus and feelings. Experiment 2 studied whether variation in the framing of the
MAMs as payment, bonus, or feelings in¯uences levels of intergroup discrimination. If payment
makes salient normative constraints that inhibit intergroup discrimination, payment will be more
strongly associated with norms of equity and equality than will be bonus and feelings and category
members will evidence less intergroup discrimination when allocating payment than when allocating
bonus and feelings.

EXPERIMENT 1

Methods

One hundred and four students (52 males and 52 females) enrolled in upper-level psychology courses
at Texas A&M University participated as part of a classroom demonstration. The experimenter
distributed questionnaire packets and informed the participants that the purpose of the study was to

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
 Normative sanctions and intergroup bias 147

examine the extent to which certain words are associated with the idea of `having to be fair and equal'.
The words `payment', `bonus', and `feelings' appeared on separate pages of the questionnaire and
participants rated on 7-point scales the extent to which they associated each word with `having to be
fair and equal' (1 ˆ not at all; 7 ˆ very much). The order of the words in the questionnaire varied as a
three-level between-subjects factor such that each word appeared ®rst within a given order: (1)
payment, bonus, feelings; (2) bonus, feelings, payment; (3) feelings, pay-ment, bonus. Consequently,
the design was a 3(rating: payment, bonus, feelings)  3(order) mixed factorial and rating served as
the within-subject factor.

Results and Discussion

The ratings were entered into a 2(sex)  3(rating)  3(order) mixed ANOVA. A signi®cant main effect
for rating, F(2, 97) ˆ 22.49, p < 0.0001, indicated that payment, bonus, and feelings were differentially
associated with `having to be fair and equal.' Payment was more strongly associated with having to be
fair and equal (M ˆ 5.01) than was the mean of bonus (M ˆ 3.81) and feelings (M ˆ 3.86), F(1,
98) ˆ 42.97, p < 0.0001, and bonus and feelings did not differ from one another, F(1, 98) ˆ 0.02.2
Consistent with hypotheses, payment more strongly conveys norms of equity and equality than do
bonus and feelings.

EXPERIMENT 2

In Experiment 2 we replicated the MGP procedure of Bornstein et al. (1983a) and varied the
description of the MAMs as involving allocations of payment, bonus, or feelings. Because payment
more strongly conveys norms of equity and equality than do bonus and feelings we expected to ®nd
less discrimination on allocations involving payment than on allocations involving bonus and feelings.

Method

Participants

One hundred and eighty-four students (94 males and 90 females) enrolled in an introductory
psychology course at the University of North Carolina at Chapel Hill (UNC-CH) participated for
partial course credit. Participants were tested in six-person same-sex groups.3

Procedure and Design

The procedure was similar to the procedure of Bornstein et al. (1983a). Participants sat in a main
experimental room facing a large projection screen and were informed that the session was part of a
2
There was a relatively uninteresting Rating  Order effect, F(4, 194) ˆ 4.18, p < 0.003, that indicated that the magnitude of
the difference between payment and the mean of bonus and feelings varied across orders. The direction of the effect,
however, was consistent within each order. Within each order, payment was more strongly associated with having to be
fair and equal than was the mean of bonus and feelings (Fpay-®rst‰1, 30Š ˆ 13.39, p < 0.001, Fbonus-®rst‰1, 36Š ˆ 28.61,
p < 0.0001, Ffeelings-®rst‰1, 35Š ˆ 8.05, p < 0.008) and bonus and feelings did not differ (Fpay-®rst‰1, 30Š ˆ 0.06, p < 0.81,
Fbonus-®rst‰1, 36Š ˆ 2.70, p < 0.11, Ffeelings-®rst ‰1, 35Š ˆ 0.91, p < 0.35).
3
On two occasions, only ®ve participants arrived for the study. A confederate played the role of the sixth participant to maintain
the guise of equivalent ingroup and outgroup size.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
148 Lowell Gaertner and Chester A. Insko

larger project investigating artistic preferences. Participants completed the Klee=Kandinsky painting
preference task in which the experimenter projected onto the screen ®ve pairs of paintings and
mentioned that each pair contained work by the artists Klee and Kandinsky. They recorded privately
their preferred painting from each pair, handed their painting preferences to the experimenter, and
relocated to individual cubicles.
After ostensibly scoring the artistic preferences, the experimenter distributed identi®cation slips
and explained that in the population of college students half prefer the style of Klee and half prefer the
style of Kandinsky and a similar proportion of Klees and Kandinskys were present in the current
session. The identi®cation slip privately informed each participant that he or she was a member of the
group of people who prefer the artistic style of Klee. The experimenter then mentioned that the
purpose of the current session was to address questions pertinent to the larger project that was
investigating artistic preferences. The experimenter explained that participants in the larger experi-
ment will complete an extensive number of psychological tests after completing the Klee/Kandinsky
painting preference task. The experimenter then distributed instruction booklets and the MAMs.
Participants received one of three instruction booklets. That is, the instructions that accompanied
the MAMs varied as a three-level between-subjects factor (instruction: payment, bonus, feelings) and
manipulated the normative constraints on intergroup discrimination.
The instructions in the payment condition were similar to the instructions used by Bornstein et al.
(1983a) and informed participants that the values in the MAMs represented monetary payment. In
particular, the instructions indicated that because the participants in the larger project will complete an
extensive number of psychological tests they might lose interest in the project. The instructions further
mentioned that, to maintain their attention, participants in the larger study will be paid and, to further
maintain their attention, they will determine exactly how much payment other participants receive.
The instructions indicated that the purpose of the current session was to test an allocation procedure
designed to enable persons to determine the amount of payment that other persons receive. In
particular, the instructions indicated that participants would test the allocation procedure by allocating
payment to other members of the Klee and Kandinsky groups who were present in the current
experimental session. The instructions mentioned that although participants could not allocate money
to themselves, they would be allowed to keep the money that others allocated to them.
The instructions in the bonus condition were identical to those of the payment condition, except the
word `payment' was replaced with `bonus money'. Bonus money was de®ned as extra money or an
award (i.e. `money above and beyond payment that is not based upon participation').
The instructions in the feelings condition indicated that the values in the MAMs represented
feelings toward other persons (i.e. higher numbers indicated more positive feelings). Participants were
informed that they would test a new `feelings-scale' that would be used in the larger project. In
particular, the instructions indicated that participants would test the feelings-scale by indicating
their feelings for members of the Klee and Kandinsky groups. Participants were informed that
although they would not evaluate themselves, the other participants present in the session would
evaluate them. The experimenter reminded participants in each of the conditions of the anonymity of
their responses and in the feelings condition an additional precaution was made to emphasize
the privacy of the their responses. Participants were informed that their responses to the feelings-
scale would be placed in a box that contained the responses of all the other participants and would
not be opened until the completion of the experiment. This precaution was initiated to minimize the
extent to which the expression of internal feelings would be constrained by concerns of public
responding.
Participants completed ®ve sets of MAMs after reading the instructions (i.e. participants selected
one of seven social orientations on ®ve consecutive sets of MAMs; see Table 1). The appearance of the
MAMs was identical across conditions and the position of each orientation within a set of MAMs was

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
 Normative sanctions and intergroup bias 149

randomized across the ®ve sets. The experimenter then debriefed and thanked the participants.
Participants in the payment and bonus conditions were awarded one dollar to minimize deception.

Results

Five participants did not complete the MAMs and the responses of four participants who indicated
upon debrie®ng that they were familiar with the work of Klee and Kandinsky and believed that they
were categorized incorrectly were excluded from analyses. The cell sizes for the payment, bonus, and
feelings conditions were for females 29, 29, and 30, respectively, and for males 27, 32, and 28,
respectively.
Our normative account of intergroup discrimination predicts that group members will discriminate
less fervently with allocations involving payment than with allocations involving bonus and feelings.
We entered the proportion of selection of each orientation into separate 2(sex)  3(instruction:
payment, bonus, feelings) analyses of variance.4

Instruction Effects

Signi®cant instruction main effects were found for selection of Max Rel Own, F(2, 169) ˆ 4.44,
p < 0.01, Max Own, F(2, 169) ˆ 3.78, p < 0.02, and Min Dif, F(2, 169) ˆ 4.37, p < 0.01. Table 2
contains the mean proportion of orientation selection as a function of instruction. We decomposed the
instruction effect into two orthogonal contrasts. The ®rst contrast compared the payment condition
with the mean of the bonus and feelings conditions and the second contrast compared the latter two
conditions.
Consistent with the idea that payment conveys norms of equity and equality that inhibit intergroup
discrimination, participants discriminated less when allocating payment than when allocating bonus
and feelings. Relative to participants in the bonus and feelings conditions, participants in the payment
condition made fewer selections of Max Rel Own, F(1, 169) ˆ 8.56, p < 0.004, and Max Own, F(1,
169) ˆ 7.27, p < 0.008, and more selections of Min Dif, F(1, 169) ˆ 8.73, p < 0.004. Participants in the
bonus and feelings conditions did not differ on selections of Max Rel Own, F(1, 169) ˆ 0.36, p < 0.55,
Max Own, F(1, 169) ˆ 0.33, p < 0.57, or Min Dif, F(1, 169) ˆ 0.00, p < 0.96.

Sex Effects

Consistent with the work of Bornstein et al., males and females differed in selection of the MAM
orientations. A main effect for sex on selection of Max Rel Own indicated that males made more
selections of Max Rel Own (Mˆ 0.09) than did females (M ˆ 0.03), F(1, 160) ˆ 5.29, p < 0.02.
Table 2. Proportion of orientation selection as a function of instruction
Max Max Max Max
Rel Max Joint Min Joint Max Rel
Instruction Own Own Own Dif Other Other Other
Payment 0.00 0.05 0.19 0.69 0.05 0.01 0.01
Bonus 0.08 0.16 0.20 0.48 0.06 0.02 0.01
Feelings 0.09 0.18 0.20 0.48 0.01 0.03 0.00

4
p
An analysis of proportions transformed with the arcsin transformation (i.e. Y* ˆ 2  arcsin Y ; Cohen, 1969) produced result
conceptually equivalent to those reported in the text.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
150 Lowell Gaertner and Chester A. Insko

Table 3. Proportion of Max Joint and Min Dif Selection as a function of instruction and sex
Orientation
Max Joint Own Min Dif
instruction instruction
Sex Payment Bonus Feelings Payment Bonus Feelings
Female 0.04 0.13 0.31 0.88 0.64 0.48
Male 0.35 0.26 0.09 0.48 0.33 0.49

Furthermore, Instruction  Sex interactions were found for selections of Max Joint Own, F(2,
169) ˆ 12.84, p < 0.0001, and Min Dif, F(2, 169) ˆ 3.86, p < 0.02. Table 3 contains the proportion
of Max Joint and Max Own selection as a function of sex and instruction.
We explored the interactions by examining separately the orthogonal contrasts of the instruction
effect for males and females. The instruction effect appeared to oppositely affect male and female
selection of Max Joint Own. Females more frequently selected Max Joint Own when allocating bonus
and feelings than when allocating payment, F(1, 169)ˆ 7.79, p < 0.006. Furthermore, females selected
Max Joint Own more frequently when allocating feelings than bonus, F(1, 169) ˆ 5.79, p < 0.02.
Males, on the other hand, more frequently selected Max Joint Own when allocating payment than
when allocating bonus and feelings F(1, 169) ˆ 6.62, p < 0.02. Furthermore, males selected Max Joint
Own more frequently when allocating bonus than feelings, F(1, 169) ˆ 5.88, p < 0.02. The instruction
effect on selection of Min Dif more directly impacted females. Females less frequently selected Min
Dif when allocating bonus and feelings than when allocating payment, F(1, 169) ˆ 11.87, p < 0.001.
Females did not differ on allocations of bonus and feelings, F(1, 169) ˆ 2.26, p < 0.05. Neither
contrast was signi®cant for males (Fs < 2.10). It appears that the instruction effect on selection of Min
Dif applies more directly to females.
An alternative approach to examining the Sex  Instruction interactions that is directly comparable
to the sex differences reported by Bornstein et al. (1983a) is to examine sex differences in each level of
instruction. The simple effect tests of sex on selection of Max Joint Own indicated that males were
more likely to maximize joint outcomes for allocations that involved money (i.e. bonus and payment),
whereas females were more likely to maximize joint outcomes for allocations that involved feelings.
Males made marginally more selections of Max Joint Own than did females in the bonus condition
(Mmales ˆ 0.26 and Mfemales ˆ 0.13), F(1, 169) ˆ 3.67, p < 0.06, and signi®cantly more in the payment
condition (Mmales ˆ 0.35 and Mfemales ˆ 0.04), F(1, 169) ˆ 16.74, p < 0.0001. In the feelings condition,
however, males made signi®cantly fewer selections of Max Joint Own (M ˆ 0.09) than did females
(M ˆ 0.31), F(1, 169) ˆ 8.42, p < 0.004.
Simple effect tests of the Instruction  Sex interaction for selection of Min Dif revealed a pattern
approximately complementary to the pattern of selection of Max Joint Own. Males were less likely
than females to minimize intergroup differences for allocations involving money. However, no sex
differences for minimizing differences occurred for allocations involving feelings. Males made fewer
selections of Min Dif than did females in the bonus condition (Mmales ˆ 0.33 and Mfemales ˆ 0.64),
F(1, 169) ˆ 8.62, p < 0.004, and in the payment condition (Mmales ˆ 0.48 and Mfemales ˆ 0.88), F(1,
169) ˆ 13.26, p < 0.0004. In the feelings condition, however, males (M ˆ 0.49) and females (M ˆ 0.48)
did not differ, F(1, 169)ˆ 0.00, p < 0.96.
The sex differences for selection of Max Joint Own and Min Dif are similar to those reported by
Bornstein et al. (1983a). Bornstein et al. reported that when allocating payment males evidenced a
stronger preference for maximizing the joint earnings of the ingroup and outgroup than did females,
while females evidenced a stronger preference for minimizing intergroup differences than did males.
The current research suggested that males were more concerned than were females with maximizing

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
 Normative sanctions and intergroup bias 151

joint earnings and less concerned with minimizing differences for allocations involving money
(payment and bonus), while females were more concerned than were males with maximizing joint
earnings involving feelings. These patterns suggest that males tend to be more instrumentally focused
while females tend to be more socio-emotionally focused (e.g. Baumeister & Leary, 1995).

Selection of Outgroup Favoring Strategies and Alternation

In addition to the above ®ndings, main effects for instruction, F(2, 169) ˆ 4.23, p < 0.02, and sex, F(1,
169) ˆ 6.82, p < 0.01, occurred for selections of Max Joint Other.5 Selection of Max Joint Other,
however, may have been due to participants who alternated between Max Joint Own and Max Joint
Other in an attempt to minimize the difference between groups while simultaneously maximizing their
joint pro®t. That is, a selection of Max Joint Own in conjunction with a selection of Max Joint Other
provides more money to both groups than does two selections of Min Dif and simultaneously
minimizes intergroup differences (Gaertner & Insko, 2000).
We recoded the MAM selections to control for alternation. We coded a series of MAM selections as
an alternation strategy if a participant selected an equivalent number of the ingroup favoring and
outgroup favoring counterparts (e.g. 2 Max Own selections and 2 Max Other selections). We adjusted
for alternation by (a) adding the frequency of selection of the ingroup and outgroup favoring
counterparts to the frequency of selection of Min Dif and (b) and setting to zero the frequency of
selection of the ingroup and outgroup favoring counterparts. The analysis of the MAMs scored for
alternation produced results identical to those previously reported, except there were no signi®cant
effects for selection of Max Joint Other.

GENERAL DISCUSSION

The current research indicated that intergroup discrimination is in¯uenced by task characteristics that
make salient normative sanctions against discrimination. Persons more strongly associated `having to
be fair and equal' with payment than with bonus or feelings and category members made fewer
allocations that maximized the ingroup's relative and absolute pro®t and more allocations that
minimized intergroup differences when allocating monetary payment than when allocating bonus
money or feelings. It should be noted that participants expressed a strong concern for minimizing
intergroup differences even with the bonus and feelings instructions. Min Dif represented nearly 50%
of category members' allocations (see Table 2). Although the bonus and feelings instructions
minimize external constraints toward fairness, they may not remove totally the constraints.
While the current research focused upon external (or normative) constraints made salient by task
characteristics, internal in¯uences of discrimination may also be affected by task characteristics. In
particular, Rabbie and Schot (1990) suggest that concerns with economic self-interest and social
competition are made salient by allocation tasks that involve money and points, respectively. It appears
that task characteristics in¯uence both normative and motivational factors and, therefore, affect both
the quantity and quality of intergroup discrimination.

5
The instruction main effect indicated that there were more selections of Max Joint Other in the bonus condition (M ˆ 0.06) than
in the feelings condition (M ˆ 0.01), F(2, 169) ˆ 4.23, p < 0.02, while the averaged selection of the bonus and feelings
condition did not differ from the payment condition (M ˆ 0.05), F(1, 169) ˆ 1.61, p < 0.21. The sex main effect indicated
that males made more selections of Max Joint Other ‰M ˆ 0.06Š than did females ‰M ˆ 0.02Š.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
152 Lowell Gaertner and Chester A. Insko

Some readers might be concerned with a seemingly inconsistent sex effect between Experiments 1
and 2. While there were no sex differences in the ratings of the extent to which payment, bonus, and
feelings are associated with `having to be fair and equal' (i.e. Experiment 1), males and females
differentially allocated payment, bonus, and feelings (i.e. Experiment 2). In particular, males were
more concerned than were females with maximizing joint earnings and less concerned with
minimizing differences for allocations involving money (payment and bonus), while females were
more concerned than were males with maximizing joint earnings involving feelings. As suggested
previously, these patterns imply that males tend to be more instrumentally focused while females tend
to be more socio-emotionally focused. The fact that females were, for example, more concerned than
were males with being fair (i.e. minimizing differences) when allocating monetary payment is not
inconsistent with the fact that males and females share congruent perceptions of the applicability of
equity and equality to monetary payment. Research on the detection of salary discrimination, for
example, indicates that while males and females do not differ in the detection of discrimination,
females more strongly judge discrimination as unfair than do males (Rutte & Messick, 1996). Perhaps
sex differences in instrumental versus socio-emotional motivation give rise to behavioral differences
in domains in which perceptual differences are lacking (i.e. perceived applicability of social norms). In
any event, both males and females evidenced more discrimination (i.e. max rel own and max own)
when allocating bonus and feelings than when allocating payment.
Experiments 1 and 2 indicate that the lack of evidence for ingroup favoritism reported by Bornstein
et al. (1983a) can be attributed to the normative sanctions against discrimination made salient by their
payment instruction. Given that the social context of the MGP fosters ingroup favoritism, what aspects
of the context are responsible for the discrimination? Although it has been argued that the MGP
isolates social categorization from other plausible antecedents of discrimination (e.g. prior intergroup
con¯ict, more frequent contact with ingroup members; Tajfel et al., 1971; Tajfel & Turner, 1979),
social categorization is not the only source of discrimination in the MGP. Rabbie and colleagues
(Rabbie & Lodewijkx, 1994; Rabbie et al., 1989) note that because participants in the MGP allocate
money to and receive money from other ingroup and outgroup members, social categorization is
confounded with outcome dependence. Research that unconfounds social categorization and depen-
dence indicates that it is the combination of categorization and dependence that is responsible for
intergroup discrimination. That is, category members favor the ingroup only when their own outcomes
are controlled by other ingroup members (Gaertner & Insko, 2000; Rabbie et al., 1989; Yamagishi
et al., 1999; however, see Gagnon & Bourhis, 1996). Discrimination in the MGP appears to be located
in a process of ingroup reciprocity such that category members favor the ingroup because they expect
ingroup members to return the favorable treatment.
We urge readers not to construe the current ®ndings as justi®cation for the continued use of the
Tajfel matrices. Although the results indicate that discrimination does indeed occur in the MGP, the
above-noted criticisms of the Tajfel matrices still apply. As long as a researcher is interested in
preferences among an array of social orientations it is preferable to allow the participant to make a
choice among all the relevant possibilities. Otherwise the choice is not fully informed and may not be a
valid indicator. The MAMs enable an informed choice.
In conclusion, the current research identi®ed the prohibitive effect of norms of equity and equality
on the overt expression of discrimination between minimal groups. Application of this idea to the
reduction of con¯icts among more than minimal groups, however, probably requires coupling social
norms with other approaches to con¯ict resolution. Motives that accompany membership in more than
minimal groups (e.g. the need to maintain a positively distinct social identity; Tajfel & Turner, 1979)
and the social support for shared self-interest provided by fellow group members (Schopler, Insko,
Graetz, Drigotas, Smith, & Dahe, 1993) may overpower normative sanctions and, consequently, give
rise to the expression of discrimination.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
 Normative sanctions and intergroup bias 153

ACKNOWLEDGMENTS

This research comprised part of the ®rst author's doctoral dissertation and was supported by National
Institute of Mental Health Grant MH53258. The authors thank Bjorn Book-Larson, Anthony
Burrow, David Dombrosky, Gus Hummings, and Sheridan Singleton for their assistance with data
collection and Vaida Thompson and Tim Wildschut for their comments on previous drafts of this
article.

REFERENCES

Allport GW. 1954. The Nature of Prejudice. Addison-Wesley: Reading, MA.

Baumeister RF, Leary MR. 1995. The need to belong: Desire for Interpersonal attachments as a fundamental
human motivation. Psychological Bulletin 117: 497±529.
Billig M, Tajfel H. 1973. Social categorization and similarity in intergroup behavior. European Journal of Social
Psychology 3: 27±51.
Blanz N, Mummendey A, Otten S. 1997. Normative evaluations and frequency expectations regarding positive
versus negative outcome allocations between groups. European Journal of Social Psychology 27: 165±176.
Bornstein G, Crum L, Wittenbraker J, Harring K, Insko CA, Thibaut J. 1983a. On the measurement of social
orientations in the minimal group paradigm. European Journal of Social Psychology 13: 321±350.
Bornstein G, Crum L, Wittenbraker J, Harring K, Insko CA, Thibaut J. 1983b. Reply to Turner's comments.
European Journal of Social Psychology 13: 369±381.
Brewer MB. 1979. In-group bias in the minimal intergroup situation: A cognitive-motivational analysis.
Psychological Bulletin 86: 307±324.
Chin WG, McClintock CG. 1993. The effects of intergroup discrimination and social values on level of self-
esteem in the minimal group paradigm. European Journal of Social Psychology 23: 63±75.
Cohen J. 1969. Statistical Power Analysis for the Behavioral Sciences. Academic Press: New York.
Franco FM, Maass A. 1999. Intentional control over prejudice: When the choice of the measure matters. European
Journal of Social Psychology 29: 469±477.
Gaertner L, Insko CA. 2000. Intergroup discrimination in the minimal group paradigm: Categorization,
reciprocation, or fear? Journal of Personality and Social Psychology 79: 77±94.
Gaertner SL, Dovidio JF. 1986. The aversive form of racism. In Prejudice, Discrimination, and Racism, Dovidio
JF, Gaertner SL (eds). Academic Press: Orlando, FLA; 61±90.
Gagnon A, Bourhis RY. 1996. Discrimination in the minimal group paradigm: Social identity or self-interest?
Personality and Social Psychology Bulletin 22: 1289±1301.
Grzelak JL, Poppe M, Czwartosz Z, Nowak A. 1988. Numerical trap: A new look at outcome representation in
studies on choice behaviour. European Journal of Social Psychology 18: 143±159.
Hertel G, Kerr NL. in press. Priming ingroup favoritism: The impact of normative scripts in the minimal group
paradigm. Journal of Experimental Social Psychology.
Insko CA, Schopler J. 1987. Categorization, competition, and collectivity. In Review of Personality and Social
Psychology, Hendrick C (ed.). Sage: New York; 213±251.
Jetten J, Spears R, Manstead ASR. 1996. Intergroup norms and intergroup discrimination: Distinctive self-
categorization and social identity effects. Journal of Personality and Social Psychology 71: 1222±1233.
Kelley HH, Thibaut JW. 1978. Interpersonal Relations: A Theory of Interdependence. Wiley Interscience: New
York.
Locksley A, Ortiz V, Hepburn C. 1980. Social categorization and discriminatory behavior: Extinguishing the
minimal intergroup discrimination effect. Journal of Personality and Social Psychology 39: 773±783.
Platow MJ, McClintock CG, Liebrand WBG. 1990. Predicting intergroup fairness and ingroup bias in the minimal
group paradigm. European Journal of Social Psychology 20: 221±239.
Rabbie JM, Lodewijkx H. 1994. Con¯ict and aggression: An individual±group continuum. Advances in Group
Processes 2: 139±174.
Rabbie JM, Schot JC. 1990. Group behaviour in the minimal group paradigm: fact or ®ction? In, European
Perspectives in Psychology (Vol. 3), Drenth PJ, Sergeant JA, Takens RJ (eds). Wiley: Chichester; 251±
263.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)
154 Lowell Gaertner and Chester A. Insko

Rabbie JM, Schot JC, Visser L. 1989. Social identity theory: a conceptual and empirical critique from the
perspective of a behavioral interaction model. European Journal of Social Psychology 19: 171±202.
Rutte CG, Messick DM. 1996. Detecting salary discrimination against male and female managers. European
Journal of Social Psychology 26: 727±740.
Schopler J, Insko CA, Graetz KA, Drigotas S, Smith VA, Dahl K. 1993. Individual±group discontinuity: Further
evidence for mediation by fear and greed. Personality and Social Psychology Bulletin 19: 419±431.
Tajfel H. 1970. Experiments in intergroup discrimination. Scienti®c American 223: 96±102.
Tajfel H, Billig M. 1974. Familiarity and categorization in intergroup behavior. Journal of Experimental Social
Psychology 10: 159±170.
Tajfel H, Billig M, Bundy RP, Flament C. 1971. Social categorization and intergroup behaviour. European Journal
of Social Psychology 1: 149±178.
Tajfel H, Turner JC. 1979. An integrative theory of intergroup con¯ict. In Austin WG, Worchel S (eds). The Social
Psychology of Intergroup Relations, Brooks/Cole: Monterey, CA; 33±47.
Turner JC. 1983a. Some comments on. . . `The measurement of social orientations in the minimal group
paradigm,' European Journal of Social Psychology 13: 351±367.
Turner JC. 1983a. A second reply to Bornstein, Crum, Wittenbraker, Harring, Insko, and Thibaut on the
measurement of social orientations. European Journal of Social Psychology 13: 383±387.
Yamagishi T, Jin N, Kiyonari T. 1999. Bounded generalized reciprocity: Ingroup boasting and ingroup favoritism.
Advances in Group Processes 16: 161±197.

Copyright # 2001 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 31, 143±154 (2001)