Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR

She Swings, She Hits, She’s Great, She’s Benched:


Implications of Gender-Based Shifting Standards
for Judgment and Behavior

Monica Biernat
University of Kansas
Theresa K. Vescio
Pennsylvania State University

The implications of the shifting standards model for under- in terms of whether stereotypes serve as interpretive
standing behavior toward stereotyped groups were examined in frames or comparative standards.
two studies on gender and athleticism. Participants played the Most common in the literature is research and theory
role of co-ed softball team managers, who made team selections, suggesting that stereotypes provide interpretive frame-
position assignments, and judgments about a series of male and works that guide the encoding and processing of infor-
female players. The data supported three hypotheses: (a) Stereo- mation, resulting in assimilative (or stereotype-consis-
types of male superiority as athletes lead to the use of shifting tent) judgments of individual group members (e.g.,
standards to judge athletic performance; (b) zero-sum behaviors Brewer, 1988; Fiske & Neuberg, 1990). But stereotypes
(allocation of limited resources) show evidence of pro-male bias, also may provide perceivers with standards against which
whereas non-zero-sum behaviors (verbal and nonverbal reac- individual group members are compared. This is the
tions) show evidence of pro-female bias; and (c) objective judg- premise of the “shifting standards model,” which sug-
ments are somewhat better predictors of zero-sum behaviors, gests that group members are judged on stereotypic
whereas subjective judgments are better predictors of non-zero- dimensions with reference to the expectations associated
sum behaviors. with their particular category membership (Biernat &
Manis, 1994; Biernat, Manis, & Nelson, 1991). Thus,
women may be judged relative to women and men rela-
Social psychologists have long been interested in the tive to men. The result is that judgments may reveal con-
trast effects (e.g., women judged more financially suc-
influence that stereotypes of social groups have on judg-
ments of individual group members (for reviews, see cessful than men, despite earning less money) or null
Fiske, 1998; Hamilton & Sherman, 1994; Kunda & effects of stereotyping (e.g., men and women judged
Thagard, 1996; Stangor & Lange, 1994; von Hippel, equally tall, despite men’s perceived objective height
Sekaquaptewa, & Vargas, 1995). The persistence of this being taller) (Biernat et al., 1991).
interest stems partly from the assumption that stereo- A number of social judgment researchers have fo-
typic judgment importantly underlies intergroup behav- cused attention on the factors that influence whether
ior and partly from the complex and sometimes assimilative (stereotypic) or contrastive (nonstereotypic)
counterintuitive nature of the influence of stereotypes.
Although it is often the case that stereotypes exert an Authors’ Note: This research was supported by National Institute of
assimilative influence on judgments of individual group Mental Health Grant No. R01MH48844 awarded to the first author. We
members (such that individual targets are evaluated con- are grateful to Janet Swim and Chris Crandall for their helpful com-
ments on an earlier draft of this article. Correspondence concerning
sistently with stereotypes), there is ample evidence that this article should be addressed to Monica Biernat, Department of Psy-
members of negatively stereotyped groups sometimes chology, University of Kansas, 1415 Jayhawk Blvd., Lawrence, KS 66045-
receive equally positive (or more positive) evaluations 7556; e-mail: biernat@ku.edu.
than do members of positively stereotyped groups. PSPB, Vol. 28 No. 1, January 2002 66-77
These conflicting effects have most often been explained © 2002 by the Society for Personality and Social Psychology, Inc.

66
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 67

judgments might result (e.g., target extremity: Manis, We assume, generally, that perceivers behave toward
Nelson, & Shedler, 1988; concreteness vs. abstractness of individual group members in a manner consistent with
stereotypic knowledge: Stapel & Koomen, 1998; Stapel, their judgments. From the shifting standards perspec-
Koomen, & van der Pligt, 1997; stereotype-consistency of tive, however, the critical question is whether perceivers
targets: Bettencourt, Dill, Greathouse, Charlton, & behave in accordance with their subjective or with their
Mulholland, 1997; Jussim, Coleman, & Lerch, 1987). objective judgments. For example, if a given man and
However, the shifting standards model suggests that woman receive equally laudable subjective evaluations
both effects may emerge in the same context but under (e.g., both are “very good”) but the woman receives
different judgment conditions (Biernat & Manis, 1994; more negative objective evaluations than the man (e.g.,
Biernat, Manis, & Kobrynowicz, 1997; Biernat et al., she is predicted to generate less sales revenue), how will
1991). The present research was designed to explore a an employer behave toward these two targets?
provocative behavioral question this suggestion raises. If
Zero-Sum and Non-Zero-Sum Behaviors
both stereotypic and counter- or non-stereotypic judg-
ments of a target emerge in a given context, is behavior To answer this question, we make a distinction
toward that target more likely to follow from one’s rela- between two broad categories or forms of behavior—
tively positive or negative judgment? those that involve allocation of limited resources and
those that do not. Included in the first category are
Judgment Effects actions characterized by the fact that behaving toward
Much research on the shifting standards model has one individual restricts the behavioral options that are
documented that counterstereotypic or nonstereotypic available toward another. Thus, behavioral choices (e.g.,
judgment patterns tend to be documented on subjective who gets promoted, who is assigned to the best tasks or
response scales, whereas stereotype-assimilative patterns positions) and allocation of “valuable” assets (e.g.,
tend to emerge on objective scales. Subjective scales are money) have a zero-sum quality and are highly meaning-
those that rely on numbers and language that are open ful to recipients in that tangible resources are at stake.
The second class of behaviors has a non-zero-sum char-
to interpretation because they have no explicit ties to
acter in that the same or similar actions can be bestowed
external reality (e.g., Likert-type scales or semantic dif-
on a limitless number of targets without resource deple-
ferentials; good/poor, heavy/light, weak/strong). In
tion. Falling under this category are such behaviors as
contrast, objective scales are “common rule” in nature—
nonverbal cues, verbal praise or punishment, and so
they are tied to units of measure that cannot be differen-
forth. Thus, a hiring decision is a zero-sum behavior—by
tially adjusted depending on the category membership
hiring one individual, others are rejected. In contrast, a
of those being evaluated (e.g., dollars earned, inches,
pleasant behavioral interaction style (e.g., smiling, mak-
estimated standardized test score). Because their units
ing eye contact, etc.) is a non-zero-sum behavior that can
have a constant meaning, objective response scales are be displayed to many individuals without being expended.
readily able to capture perceivers’ available stereotyped These behavior types parallel the objective/subjective
perceptions. And because subjective language can shift judgment distinction outlined above. Zero-sum behav-
in meaning depending on who is being described (e.g., iors, similar to objective judgments, require that one
“good” can mean something different when applied to a take a “cross-category” or “absolute” perspective on the
woman than a man), stereotyping effects may be masked set of actors being evaluated—each actor’s standing can
or reversed. be interpreted unambiguously in the context of the oth-
Greater evidence of stereotyping on objective than ers. For example, a favorable hiring decision is similar to
subjective response scales is the signature shifting stan- a high estimated standardized test score in that both can
dards pattern, and it has been documented with regard be understood at face value and interpreted as evidence
to gender in the domains of height, weight, financial sta- of one actor’s higher standing relative to all others in the
tus, verbal ability, job competence, and leadership, and set (regardless of category membership). Because of this
with regard to race in the domains of verbal ability, ath- similarity, we suggest that zero-sum behaviors will gener-
leticism, and job competence (Biernat, Crandall, Young, ally coincide with objective target judgments.
Kobrynowicz, & Halpin, 1998; Biernat & Kobrynowicz, On the other hand, non-zero-sum behaviors are simi-
1997; Biernat & Manis, 1994; Biernat et al., 1991). How- lar to subjective judgments in that both invoke a “within-
ever, all of these findings represent judgment effects; less category” perspective by which actors are compared to
clear are the implications of such judgment patterns for group expectations. For example, non-zero-sum behav-
behavior toward members of stereotyped groups. iors such as cheers or smiles reflect a “wow” effect—they
Exploring these implications is the purpose of the pres- are demonstrations of surprise that an individual has sur-
ent research. passed a low performance standard for his or her group.
68 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

Similarly, a very positive subjective evaluation is based on evaluations. For example, Black writers were told how
the violation of a negative group expectation. For this much “potential” was in the essay; White writers were
reason, we predict that non-zero-sum behaviors will gen- told “when I read college work this bad I just want to lay
erally be consistent with subjective judgments of group my head down on the table and cry” (Harber, 1998, p. 625).
members. In short, we suggest that objective judgments In a more direct demonstration of standard-setting,
will be better predictors of zero-sum behaviors than sub- Biernat and Kobrynowicz (1997) also found lower stan-
jective judgments but that subjective judgments will be dards for women and Blacks to minimally qualify as com-
better predictors of non-zero-sum behaviors than objec- petent for a job (relative to White men). Although these
tive judgments. may seem like positive outcomes for Blacks and women,
Some empirical evidence that is consistent with the Foschi (1992) describes low standards of this sort, and
prediction regarding zero-sum behaviors can be found the reaction to their being surpassed, as “patronizing” in
in a field study that examined gender effects on subordi- nature (p. 195)—the positivity does not reflect sincere
nate and supervisor cross-evaluations (Gupta, Jenkins, & judgment of quality (see also Jackman, 1994). Patron-
Beehr, 1983). Using a sample of employees from a broad izing standards and their resultant favorable subjective
range of work settings, these researchers found an over- evaluations and non-zero-sum behaviors can be con-
all tendency for women (both supervisors and subordi- trasted with zero-sum behaviors: Zero-sum behavioral
nates) to be subjectively evaluated more positively than choices are based not on the target’s success in surpass-
men. At the same time, however, the behavior of employ- ing a lenient standard but on the objective representa-
ers toward the employees seemed more consistent with a tion of devalued groups as being less good, able, compe-
stereotype-consistent objective representation of men as tent, and so forth, than valued or high-status groups.
better than women—male subordinates were awarded
The Present Studies
more promotions and pay raises than the same women
who were subjectively evaluated so positively. Gupta et al. To put these ideas about judgment and behavior to a
(1983) write about employers’ treatment of women as more rigorous test, we turned to the domain of athleti-
follows: “While opinions (evaluations) may be positive, cism and to the stereotype that men have superior ath-
actions (promotions) still follow tradition” (p. 183). letic ability relative to women. Consider a sport in which
Similarly, in the domain of race, Weitz (1972) studied both men and women participate: co-ed softball. In this
the behavioral choices of White participants who were context, a male and female player may both be perceived
led to believe that they would be interacting with either a as “very good.” Nonetheless, we suspect that women
Black or White partner. She found that although some afforded this label are assumed to hit and throw with less
Whites reported extremely positive subjective feelings of power than “very good” men. This occurs, presumably,
“friendliness” toward their Black partners, these same because of shifting standards; looked at in a different
individuals engaged in the most negative behavioral manner, a female player might more readily be labeled
responses to the partner (e.g., they were less likely to “excellent” than a male player who performs the same
choose to wait with the partner during a break in the because she is being evaluated with respect to lower stan-
experiment and they chose to interact with the partner dards. Hypothesis 1 is that judgments of male and female
in only the least intimate of experimental tasks). We con- softball players will show the typical shifting standards
ceptualize these as zero-sum behaviors in that they effect: Men will be judged objectively better than women
involve choices between alternatives: waiting with the (e.g., in expected batting average), but this effect will be
partner (vs. alone) and selecting among seven tasks masked in subjective evaluations (e.g., “good” vs. “poor”).
those three that one would prefer to do alone rather In the domain of softball, zero-sum behaviors include
than with the partner. Again, behavior followed not from such decisions as who is chosen for the team, who plays
the subjective sense of friendliness but rather from the key positions (e.g., shortstop), and the order in which
stereotypical representation of Blacks as less desirable players bat (e.g., top vs. bottom of the batting order).
partners than Whites. Non-zero-sum behaviors include pats on the back,
With regard to non-zero-sum behavior, the extant evi- cheers and praise following a performance, and a variety
dence is less abundant and clear. However, a study by of nonverbal responses. Hypothesis 2 of our research is
Harber (1998) examined the written feedback that that zero-sum behaviors will favor male over female play-
White judges offered to either Black or White authors of ers (e.g., men more likely to be chosen for a team, to play
poorly written essays. Consistent with the notion that key positions, and to bat early in the lineup) but that
positive non-zero-sum behaviors reflect the surpassing of non-zero-sum behaviors will more likely favor female as
low performance standards, comments on essay content opposed to male players (e.g., females more likely than
(a non-zero-sum behavior) were more favorable and sup- males to be praised for the same performance). Finally,
portive for Black than White authors, as were subjective Hypothesis 3 predicts that judgments will be correlated
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 69

with behavior in the manner specified above: Objective credit for their participation and were run through the
judgments will better predict zero-sum behaviors than procedure individually. After arriving at the lab, partici-
will subjective judgments, whereas subjective judgments pants were told to imagine they were managers of a co-ed
will better predict non-zero-sum behaviors than will softball team and that their job was to engage in all the
objective judgments (see also Biernat, Vescio, & Manis, decision making that a real manager would do in putting
1998). together and fielding a team.1
We report two studies designed to test these three pre- Participants were provided with photographs of 18 po-
dictions. In both studies, participants were asked to play tential White players—9 men and 9 women. These pho-
the role of a softball team manager, making such behav- tos were high school yearbook pictures that were
ioral decisions as who should play on the team and what selected for use on the basis of two pretesting sessions.
positions they should play. Potential team members were Specifically, 28 participants in the first pretest were asked
men and women who had been subjectively equated in to sort a series of slides (all drawn from yearbooks) into
terms of their athletic ability in a pretest session (low, categories of high (rating = 4), moderate, (rating = 3), slight
moderate, or high in athletic ability). The inclusion of (rating = 2), and low athleticism (rating = 1). These sorts
targets of different levels of athleticism was done to were done separately for the male and female groups
examine constraints on the processes described above. such that perceived athleticism was defined in a within-
Much research suggests that both stereotypes (e.g., tar- sex fashion. For use in these studies, we chose three pho-
get gender) and individuating information (e.g., tographs of each sex that were consensually indicated to
appearance-based athleticism) influence judgment, and be high, moderate, and low in athleticism (e.g., three
indeed, individuating information often plays the larger highly athletic men, three highly athletic women, three
role (see Kunda & Thagard, 1996). We suspected that for moderately athletic men, etc.). Specifically, the low-ath-
clearly nonathletic and clearly athletic-appearing tar- letic female targets received a subjective athletic rating
gets, the effects of stereotypes would be less strong; of 1.35 (low athletic male M = 1.43), the mean for moder-
instead, most of the “action” of stereotyping should be ately athletic female targets was 2.54 (moderate male M =
observed for the moderately athletic, more ambiguous 2.50), and the mean for highly athletic female targets
targets. As Stapel and Winkielman (1998) put it, a con- was 3.68 (high male M = 3.43). These subjective ratings
struct “will be a guide to interpretation only when there within and across athleticism level were statistically
is something to be interpreted, that is, when the target equivalent for male and female targets.
stimulus is ambiguous rather than unambiguous” (p. 635). Because the first pretest allowed for only within-sex
In short, gender-based judgment and behavioral effects comparisons, we conducted a second, cross-sex compari-
were expected to be strongest for those targets who were son pretest in which the entire set of photos selected
moderately athletic in appearance. from the first pretest was given to a new set of 10 under-
In Study 1, we examined managers’ objective judg- graduate judges (6 women, 4 men). The photos were
ments and zero-sum behavioral choices of these subjec- presented in one random order, and judges were asked
tively equated male and female players. This study to rate “how athletic” they appeared to be on a 1 (not at
allowed for a test of the zero-sum behavioral prediction all athletic) to 7 (very athletic) response scale. Low-athletic
of Hypothesis 2. In Study 2, participants made both sub- women (M = 2.68) were judged equivalent to low-athletic
jective and objective assessments of players’ batting and men (M = 2.85, p > .50), moderately athletic women (M =
fielding abilities, thereby allowing a test of the shifting 4.50) were judged equivalent to moderately athletic men
standards judgment prediction (Hypothesis 1), and also (M = 4.90, p > .20), and high-athletic women (M = 5.47)
engaged in a non-zero-sum behavioroid activity, thereby were judged somewhat more athletic than high-athletic
allowing a fuller test of Hypotheses 2 and 3. men (M = 4.97, p < .05). Although we hoped to equate
the targets in perceived athleticism, this latter pattern
STUDIES 1 AND 2 (which likely reflects use of shifting standards) actually
provides a more conservative test of our hypothesis that
Method
women will be objectively evaluated and treated more
Because the procedures of the two studies were quite negatively than males.
similar, they are described jointly below. The studies In the main study, the set of 18 selected photos was
were conducted during successive semesters, so there presented to participants in a single, randomly ordered
was no overlap in the samples. Participants in Study 1 stack and identified by letters (Persons A through R). As
were 97 undergraduates (57 women, 39 men, 1 unspeci- their first task, the “managers” were asked to sort
fied) and in Study 2 were 67 undergraduates (28 women, through the photos and to select their team. Of the 18
39 men) at the University of Kansas who were enrolled in potential players, they were asked to choose 13 to be on
introductory psychology courses. All received course their team (i.e., to reject 5 players). The unselected pho-
70 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

tos were then removed by the experimenter. After choos- (c) assess our prediction that objective judgments better
ing a team, participants were next asked “to decide who predict zero-sum behaviors and subjective judgments
will play in which positions in your starting lineup.” Par- better predict non-zero-sum behaviors.
ticipants were provided with a list of 10 positions— A final change in Study 2 was a manipulation of the
pitcher, catcher, first base, second base, third base, short- order in which judgments were made. In one condition,
stop, right field, right center field, left field, and left cen- subjective batting and fielding judgments preceded
ter field—and indicated by letter which player they objective judgments, and the zero-sum behaviors pre-
wanted to place in each position. This system required ceded the team selection and assignment measures. In
that 3 players be “benched.” The next task was for partici- the other condition, objective judgments preceded sub-
pants to create a batting order for the 10 starting players; jective judgments, and team selections and assignments
participants again were allowed to sort and handle the preceded the non-zero-sum behavioral measure. As will
photos in any way they desired to make these decisions. be seen below, this order manipulation had no effects on
All of the choices—team selection, benchings, batting any of the measures.
order, and fielding positions—were conceptualized as At the end of both studies, participants also com-
zero-sum behaviors. pleted a background questionnaire that included items
In the next series of judgments, participants were assessing their knowledge of and experience in playing
asked to consider the entire set of photos again, includ- softball/baseball, their own likely batting average and
ing those they had rejected or benched. For each player, fielding percentages were they to play on a co-ed team,
presented on a screen in slide format in the original A to and their judgment of the relative softball skills of col-
R order, participants estimated a seasonal batting aver- lege-age men versus college-age women. As might be
age and fielding error rate (in percentages). These were expected, there were a number of sex differences on
conceptualized as objective judgments. these background factors. In both studies, men
In addition to the above measures (which were described themselves as more knowledgeable about
included in both Studies 1 and 2), Study 2 added several baseball/softball than women, as playing baseball/soft-
other dependent variables and involved some amend- ball more frequently than women, as likely to have a
ments to the general procedure. First, in addition to esti- higher batting average than women, and as likely to have
mating batting averages and error rates of the players in a lower fielding error percentage than women. Men
“objective” units, participants were asked to rate these were also more likely than women to believe that men
abilities on subjective judgment scales (i.e., 1 = very poor have better softball skills than women. However, men
hitter to 7 = very good hitter; 1 = very poor fielder to 7 = very were no more likely than women to have ever played soft-
good fielder). These subjective estimates, along with the ball before, and none of the participants was a college
objective judgments, will allow for a direct test of softball/baseball player.2
Hypothesis 1. Results
Second, a measure of non-zero-sum hypothetical
“behavior” was included. Specifically, participants were ZERO-SUM BEHAVIORAL CHOICES
asked to think about each player in turn, to imagine that Team selections. Participants chose 13 out of 18 possible
this player hit a single, and then to indicate how they players to be on their teams. The first panel of Table 1
would respond (as manager) to the hit. Seven response depicts the number of players chosen of each type (sex by
options were offered and participants were asked to athletic appearance) within each study. Target Sex × Level
check all that would apply to their likely reaction to a of Athleticism × Participant Sex (× Order in Study 2)
player’s single: (a) say “nice job!” to the player, (b) jump mixed-model ANOVAs on these choices indicated main
up and down excitedly, (c) pat the player on the back, effects of both target sex—Study 1: F(1, 94) = 16.64, p <
(d) tell others how pleased you are with the player, (e) .0001; Study 2: F(1, 63) = 3.05, p < .09—and level—Study 1:
nothing in particular, (f) cheer loudly, and (g) smile at F(2, 188) = 374.54; Study 2: F(2, 126) = 253.33, ps <
the player. These responses are not clearly ordinal but .0001—as well as significant Target Sex × Level interac-
they do distinguish between a nonresponse (nothing in tions—Study 1: F(2, 188) = 13.54; Study 2: F(2, 126) =
particular), a mild response (smile at the player), and 16.53, ps < .0001. There were no reliable effects of partici-
more animated, positive responses. The addition of the pant sex, all Fs < 1. Overall, participants chose more men
subjective judgments and non-zero-sum behavioral mea- (6.75, SD = .74 in Study 1; 6.68, SD = .81 in Study 2) than
sure to Study 2 allows us to (a) directly test the signature women (6.25, SD = .72 in Study 1; 6.32, SD = .81 in Study 2)
shifting standards judgment prediction (greater evi- to be on their teams, but the favoring of men was only
dence of stereotyping on objective than subjective reliable in the case of the moderately athletic-appearing
response scales), (b) examine whether non-zero-sum targets, Study 1: t(96) = 6.92; Study 2: t(66) = 8.87, ps <
behaviors favor female as opposed to male players, and .0001.
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 71

TABLE 1: Team Selections, Benching Decisions, Infield Positions, and Top of Batting Order Decisions

Study 1 Study 2
Sex of Player Low Moderate High Low Moderate High

Team selections
Female 1.12 (.69) 2.33a (.70) 2.80 (.45) 1.31 (.97) 2.13a (.69) 2.88 (.33)
Male 1.07 (.61) 2.88a (.39) 2.80 (.49) 0.99 (.89) 2.94a (.24) 2.75 (.47)
Benching decisions
Female 0.50 (.68) 0.85b (.68) 0.43 (.59) 0.60 (.58) 0.96b (.73) 0.40 (.52)
Male 0.51 (.60) 0.29b (.50) 0.42 (.57) 0.46 (.53) 0.25b (.44) 0.31 (.50)
Number in infield positions
Female 0.21 (.41) 0.38c (.60) 1.15 (.69) 0.22 (.45) 0.27c (.45) 1.10 (.80)
Male 0.16 (.36) 1.19c (.74) 0.90 (.64) 0.06 (.24) 1.37c (.79) 0.97 (.74)
Number in top five of batting order
Female 0.24 (.50) 0.55d (.60) 1.19 (.79) 0.45 (.66) 0.39d (.55) 1.24 (.82)
Male 0.15 (.35) 1.08d (.73) 1.42 (.80) 0.16 (.41) 1.43d (.89) 1.33 (.81)

NOTE: Standard deviations appear in parentheses. Within each study, means with matching superscripts differed at p < .0001.

Benching. A similar pattern appeared in choices of corrected for number of players of a given sex on the
which 3 of the 13 team players were benched rather than team, women (MStudy1 = .28, SD = .12; MStudy2 = .25, SD =
placed in the starting lineup (second panel of Table 1). .13) continued to be less frequently assigned to infield
The main effects of target sex—Study 1: F(1, 94) = 13.66; positions than men (MStudy1 = .33, SD = .11; MStudy2 = .36,
Study 2: F(1, 63) = 27.73, ps < .0001—indicated that SD = .12), Study 1: F(1, 94) = 7.09; Study 2: F(1, 63) = 4.77,
women were more likely to be benched than men, but ps < .05. Thus, even correcting for their lesser presence
this effect was qualified by reliable Target Sex × Level on the team, women had less chance of being chosen for
interactions—Study 1: F(2, 188) = 10.10; Study 2: F(2, infield positions.3
126) = 10.43, ps < .0001. As indicated in the table, the ten- As for batting order, the most prestigious top posi-
dency to bench female players was only reliable among tions also were assigned on the basis of target sex. Men
the moderately athletic targets, t(96) = 5.93 in Study 1; were more likely than women to be assigned to the first,
t(66) = 6.23 in Study 2, ps < .0001. In general, not only third, and fourth (clean-up) positions, but women were
were moderately athletic women less frequently selected slightly more likely than men to bat second. To summa-
as team members than men, but even given their lower rize these data we computed the number of men versus
representation, they were the most likely to be among women assigned to the top half (top five) of the batting
the three benched players. Indeed, in analyses of bench- lineup. These means appear in the fourth panel of Ta-
ing likelihood corrected for the number of players of a ble 1. As was the case for the other zero-sum choices,
given sex placed on the team, the main effect of sex ANOVAs indicated main effects of target sex—Study 1:
remained reliable in each study—Study 1: F(1, 94) = F(1, 94) = 20.29; Study 2: F(1, 63) = 10.61—and level of
4.14, p < .05, male M = .18, SD = .10; female M = .28, SD = athleticism—Study 1: F(2, 188) = 118.02; Study 2: F(2,
.11; Study 2: F(1, 63) = 16.73, p < .0001, male M = .15, SD = 126) = 53.40—as well as reliable Target Sex × Level
.09; female M = .31, SD = .11. Participant sex had no interactions—Study 1: F(2, 188) = 9.77; Study 2: F(2,
effects on the benching selections (all ps > .36). 126) = 22.77, ps < .0001. Once again, the tendency for
men to bat sooner than women was only reliable in the
Field positions and batting order. To summarize the field case of moderately athletic targets. Analyses of batting
position data, we computed the number of men versus order assignments correcting for the number of players
women assigned to infield positions (the more presti- of each sex on participants’ teams indicated that women
gious assignments; see third panel of Table 1). In each (MStudy1 = .32, SD = .15; MStudy2 = .33, SD = .14) were signifi-
study, the same ANOVA approach described above cantly less likely to be placed in the top of the batting
revealed main effects of target sex—Study 1: F(1, 94) = order (positions one to five) than men (MStudy1 = .39, SD =
13.54; Study 2: F(1, 63) = 13.57—and athleticism level— .15; MStudy2 = .44, SD = .13); Study 1: F(1, 94) = 11.37;
Study 1: F(2, 188) = 85.70; Study 2: F(2, 126) = 69.96— Study 2: F(1, 63) = 7.79, ps < .0001.
and significant Sex × Level interactions—Study 1: F(2, In sum, these findings indicate a clear zero-sum
188) = 34.63; Study 2: F(2, 126) = 29.40, ps < .0001. behavioral preference for male as opposed to female
Women, and particularly moderately athletic women, softball players: Despite equivalent subjective assess-
were less likely to be selected for infield positions relative ments of athleticism in the pretest, men were more likely
to men. In analyses of fielding position assignment to be chosen for the team, less likely to be benched, more
72 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

likely to be assigned to play infield positions, and more less well, objectively, than comparably labeled male
likely to bat in the top of the batting order than women. players.
The poor quality assignments for women were notable In Study 2, however, we were able to directly test the
even after correcting for the relative lack of women play- shifting standards hypothesis because participants
ers on the team. judged targets’ batting and fielding ability in both objec-
tive and subjective units. To make the subjective and
OBJECTIVE AND SUBJECTIVE BATTING
objective versions of these judgments comparable, and
AND FIELDING JUDGMENTS
to control for individual differences in response scale
In Study 1, participants made only “objective” judg- use, we used within-subjects standardization. That is,
ments of the targets by estimating seasonal batting aver- each judgment (e.g., batting average) was computed as a
ages and fielding error rates. The mean batting average deviation from the participant’s mean batting average
estimates were submitted to a Target Sex × Athleticism estimate across the 18 targets, divided by the standard
Level × Participant Sex ANOVA. As in the earlier analy- deviation for that participant’s batting average estimates.
ses, the target sex, F(1, 92) = 14.54, level, F(2, 184) = We submitted the batting estimates (objective, subjec-
98.38, and Target Sex × Level interaction, F(2, 184) = tive) to a Participant Sex × Order × Target Sex × Athleti-
29.03, ps < .0001, were again significant. At each level of cism Level × Scale repeated-measures ANOVA. A num-
athleticism, men were judged to have higher batting ber of effects were significant, including the main effect
averages than women, and these differences were reli- of target sex, F(1, 63) = 38.30; the main effect of athleti-
able for both moderate (Mmale = .454, SD = .186; Mfemale = cism level, F(2, 126) = 292.09; and the Sex × Level inter-
.287, SD = .145) and highly athletic targets (Mmale = .461, action, F(2, 126) = 92.82, ps < .0001. As in Study 1, the ten-
SD = .188; Mfemale = .365, SD = .165; ts(96) = 12.02 and 6.77, dency for men to be judged better batters than women
respectively, ps < .0001. These data provide good, was reliable for both the moderate (Mmale = .72, SD = .37;
although indirect, evidence of the shifting standards Mfemale = 0.41, SD = .35) and highly athletic targets (Mmale =
effect—male and female targets were judged equiva- .62, SD = .42; Mfemale = .44, SD = .38), ts(66) = 15.73 and
lently in pretested subjective athleticism, but estimates 2.25, respectively, ps < .05. Of most theoretical interest,
of their batting averages nonetheless differed. In addi- however, was the predicted Target Sex × Response Scale
tion, a main effect of participant sex emerged in these interaction, F(1, 63) = 22.91, p < .0001, which was not
judgments, F(1, 92) = 14.54, p < .0001, with women (M = moderated by any higher order interactions. Men were
.390, SD = .155) estimating higher batting averages over- particularly likely to be judged better batters than wo-
all than men (M = .285, SD = .078). men on the objective response scale (Ms = .24 and –.24,
Study 1 fielding error rates were analyzed in a compa- SDs = .24), relative to the subjective response scale (Ms =
rable ANOVA. Reliable main effects and the two-way .12 and –.12, SDs = .25). In unadjusted estimates, objec-
interaction were again documented: target sex: F(1, 92) = tive (batting average) estimates were .363 and .314 for
40.18; level: F(2, 184) = 95.92; and Target Sex × Level: men and women, respectively, and comparable subjec-
F(2, 184) = 20.69, ps < .0001. Male fielders were reliably tive estimates were 4.48 and 4.12.
viewed as less error prone than female players, signifi- A similar analysis was conducted on Study 2 judg-
cantly so at both the moderate (Mmale = 13.1%, SD = 11.1; ments of fielding (error rates were reverse scored such
Mfemale = 27.3%, SD = 17.1) and high athletic levels (Mmale = that positive numbers indicate better play). This analysis
12.9%, SD = 9.5; Mfemale = 17.7%, SD = 13.0; ts(96) = 9.35 also revealed significant effects of target sex: F(1, 63) =
and 4.28, respectively, ps < .0001). Female participants 63.63; athleticism level: F(2, 126) = 369.24; and their
(M = 27.5%, SD = 12.7) also estimated higher overall interaction: F(2, 126) = 57.50, ps < .0001. Overall, men
error rates than male participants (M = 16.4%, SD = 8.3), were judged to field better than women (Mmale = .20,
F(1, 92) = 22.49, p < .0001. Mfemale = –.20, SDs = .20), and comparisons within athleti-
In sum, the objective judgments of targets in Study 1 cism level indicated that this sex difference was reliable
produced a pattern comparable to that for zero-sum only among the moderate targets (Mmale = .71, SD = .29;
behaviors: Male players were predicted to have better Mfemale = –.31, SD = .34), t(66) = 16.38, p < .0001. However,
seasonal batting averages and to field with fewer errors there was no evidence that response scale moderated
than female players. Although this study did not allow these effects, Sex × Scale F < 1. Subject sex and order had
for a direct comparison of subjective and objective target no effects on either the batting or fielding estimates.
judgments made by the same participants, the results are NON-ZERO-SUM BEHAVIOR
consistent with the shifting standards hypothesis that the
same subjective label (e.g., “moderately athletic”) means Only in Study 2 did we include assessments of (hypo-
something different when applied to male and female thetical) non-zero-sum behavior: Participants’ role-
targets: Female players were perceived as performing played behavioral responses to a single hit by each
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 73

TABLE 2: Non-Zero-Sum Behavioral Responses to “Singles” Hit by


TABLE 3: Correlations Among Objective and Subjective Judgments,
Players, Sex × Level Interaction, Study 2
Zero-Sum and Non-Zero-Sum Behaviors, Study 2
Level of Athleticism
Sex of Objective
Player Low Moderate High Team (Zero-Sum) Non-Zero-Sum Batting

Female 2.21 (1.20) [.97] 2.21a (1.21) [.95c] 2.19b (1.18) [.94d] Non-zero-sum –.21
Male 2.19 (1.29) [.94] 1.69a (1.16) [.79c] 1.79b (1.14) [.84d] Objective batting .51*** [.31**] –.21 [–.01]
Subjective batting .46*** [.19] –.31** [–.24*] .66***
NOTE: Standard deviations appear in parentheses. Numbers in brack-
ets indicate the proportion of respondents who responded in any way NOTE: All variables reflect the relative difference in male-female judg-
other than “nothing in particular” to the single. Means with matching ments/behaviors (e.g., team = number of male players – number of fe-
superscripts differ at p < .01. male players). Numbers in brackets indicate partial correlations, with
estimates on the other judgment scale (subjective or objective)
partialed out.
player. As indicated in the Method section, respondents *p < .06. **p < .05. ***p < .0001.
checked off as many of the seven listed reactions that
would characterize their response, as manager, to a
player’s hit. Because these reactions were not clearly ting average (r = .49, p < .0001). However, because rele-
ordinal, we simply calculated the number of responses vant measures were not included in this study, we could
that were checked, excluding the “nothing in particular” not test the relative predictive power of subjective and
response (i.e., if this were the only option checked, it was objective judgments on both zero-sum and non-zero-
scored a “0”; any other response was counted toward a sum behaviors.
“total reaction” score). These figures were submitted to a This was remedied, however, in Study 2. We predicted
Participant Sex × Order × Target Sex × Athleticism Level that the tendency to estimate that men have higher bat-
repeated-measures ANOVA. A main effect of participant ting averages than women should predict team selec-
sex reflected female managers’ (M = 2.82, SD = 1.00) tions (a zero-sum behavior), whereas the male-female
greater likelihood of behaviorally responding to a single difference in subjective batting judgments should pre-
than male managers (M = 1.49, SD = .77), F(1, 63) = dict the non-zero-sum behavior of relatively favorable
37.60, p < .0001. The analysis also revealed main effects reactions to female players’ hits. Table 3 reports the cor-
of target sex, F(1, 63) = 23.36, and athleticism level, F(2, relations among male-female differences in team selec-
126) = 8.20, as well as their interaction, F(2, 126) = 10.43, tion, non-zero-sum behavioral responses, objective bat-
ps < .0001, which is depicted in Table 2. Whereas the ting averages, and subjective batting quality estimates.
zero-sum behaviors reflected preference for men (men For example, the correlation of –.21 between team selec-
more likely to be selected for the team, to play infield tion and non-zero-sum behavior indicates that although
positions, to bat in the top of the lineup), this non-zero- men were preferred on the team over women, men were
sum index indicated favoritism toward female players, responded to less favorably than women when they hit
especially those moderate and high in athleticism. This singles. As can be seen in the table, objective and subjec-
table also depicts the proportion of respondents who tive batting estimates were positively related (.66): The
indicated any kind of non- zero-sum response to the hit tendency to judge men more favorably than women co-
other than doing “nothing in particular” (i.e., the occurred on both scales. Also, both subjective and objec-
seven-item index was reduced to a 0-1 coding system). tive batting judgments were associated with the team
These data indicate the same pattern—managers selection and non-zero-sum responses to hits.
reacted more favorably to a single hit by a woman than a However, the pattern of partial correlations, in which
man. Interestingly, however, in the case of low-athletic the effects of the “other” judgment scale were partialed
men (for whom expectations were presumably low), a out of the analyses, generally supported our predictions
single was greeted with as much behavioral fanfare as (see bracketed entries in Table 3). Although preference
that expressed toward all female players. for men on the team was associated with both objective
and subjective batting judgments, the correlation
CORRELATIONS BETWEEN
between objective judgments and team selections
JUDGMENT AND BEHAVIOR
remained significant after controlling for subjective
The Study 1 data allowed for a test of only one compo- judgments; the converse was not true (see Column 1 of
nent of Hypothesis 3: Objective judgments will predict Table 3). Similarly, although relatively positive reaction
zero-sum behaviors. Consistent with this prediction, the to male as opposed to female players’ hits was negatively
relative preference for male over female players in team related to both objective and subjective batting judg-
selection (zero-sum behavior) was significantly corre- ments, the partial correlations reflected a stronger effect
lated with the male-female difference in predicted bat- of subjective judgments on this behavior.4 In other
74 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

words, the tendency to view men as better batters than literature suggesting that the influence of stereotypes
women was associated with less positive behavioral reac- and other heuristics is strongest when stimuli are ambig-
tions to male relative to female players’ hits. To examine uous in nature (see Kunda & Thagard, 1996, for a review;
whether the key partial correlations were reliably differ- Stapel & Winkielman, 1998). The athletic competence
ent, we used Fisher’s r to z transformation and calculated of moderately athletic targets was presumably rather
z using Fisher’s formula. The difference between the cor- unclear and open to construal, whereas the appearance
relations of .31 and .19 was not reliable (z = 1.03, ns), but of the high- and low-athletic targets (the “jocks” and the
the difference between the correlations of –.01 and –.24 “wimps”) left less room for stereotype-guided interpreta-
was significant (z = 1.88, p < .04, one-tailed). In short, tion (Darley & Gross, 1983; Kunda & Sherman-Williams,
zero-sum behaviors showed only a slight tendency to be 1993).5
better predicted by objective than subjective judgments, The novel aspect of the present research was our
but non-zero-sum behaviors were reliably better pre- focus on behavioral responses to male and female tar-
dicted by subjective than objective judgments. gets. We engaged participants in playing the role of soft-
Discussion ball team manager and took them through the process
of decision making that a real team manager would fol-
The two studies described here were designed to low. Managers were asked to make zero-sum behavioral
extend research on the shifting standards model to choices: deciding who should be on the team, who
incorporate behavioral responses to members of stereo- should play which positions, and who should bat. As pre-
typed groups. In prior work, we have documented that dicted, these behaviors indicated a favoring of men: Men
stereotypes lead perceivers to evaluate members of con- were more likely to be selected for the team, to be
trasting groups against different standards. This ten- assigned to important infield positions, and to bat in the
dency can be seen in the findings that (a) subjective early part of the batting order; they were also less likely
judgments, which allow for meaning shifts, often mask than women to be benched. These patterns were particu-
stereotyping effects at the same time that objective judg- larly true of comparisons between male and female play-
ments reveal them and (b) perceivers “translate” subjec- ers of moderate athleticism, and the position effects
tive language in accordance with group stereotypes (e.g., remained reliable even after correcting for the a priori
“good at math” means higher math grades for Asian than lower likelihood of women being on the team.
for Black students) (Kobrynowicz & Biernat, 1997). However, at the same time that these zero-sum behav-
Consistent with these data, the present studies indi- ioral choices favored male players, Study 2 indicated that
cated that male and female targets who were described non-zero-sum behavior favored female players. By our defi-
as similarly athletic by pretest participants were nonethe-
nition, non-zero-sum behaviors include both verbal and
less judged to have objectively different softball perfor-
nonverbal activities that are unlimited in availability—
mances. Male players (except those who were pretested
behavior toward one individual is not restricted by
to be low in athleticism) were judged to bat and field sig-
behavior toward others. Admittedly, our indicator of
nificantly better than female players. In Study 2, a more
non-zero-sum behavior was indirect and hypothetical—
direct replication of the typical shifting standards effect
managers were asked to indicate how they would
was documented in batting judgments: The tendency to
judge men as better batters than women was stronger on respond if a player hit a single during a game (from
objective (batting average) than subjective (good-poor doing nothing, to smiling, to cheering wildly). Nonethe-
hitter) response scales. That this effect occurred when less, this behavioral measure produced the predicted
scale was manipulated within-subjects adds further cre- pattern of results: A good hitting performance by a
dence to our claim that standard shifts occur readily, per- female player was greeted with more enthusiastic behav-
haps without awareness (see Biernat, 1995). However, a ior than comparable performance by a male player.
similar effect did not emerge for fielding estimates. We We suggest that this pattern occurs precisely because
have no solid explanation for the null effect of response of the low standards perceivers hold regarding female
scale on this dimension, although we suspect that athleticism. A single is better coming from a woman than
judges’ lack of familiarity in estimating error rates (com- a man because it is less expected. The “wow” feeling that
pared to batting averages) may have contributed to the results from a surpassed expectation then translates into
lack of differentiation between objective and subjective favorable non-zero-sum behavior. The data from both
measurement. studies paint a picture in which female athletes receive
It is also worth noting that most of the behavioral and fewer of the good limited resources available in a sports
judgmental differentiation between male and female context but receive more praise and pats on the back
players was due to the targets labeled “moderately ath- when their performance is good. Both patterns,
letic.” This finding is quite consistent with a large body of although apparently contradictory, reflect the operation
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 75

of gender stereotypes and the corresponding perfor- Kawakami, Johnson, Johnson, & Howard, 1997), sug-
mance standards they invoke. gesting that the former are better predicted by implicit
A final question in the present research concerned measures of outgroup attitudes and the latter by more
the relationship between judgment and behavior. To direct self-report measures (see also Dovidio & Fazio,
date, research on the shifting standards model has 1992; Fazio, 1990; Fazio, Jackson, Dunton, & Williams,
focused on judgment effects, revealing that subjective 1995).
and objective response scales often reveal different pat- This spontaneous/deliberative distinction seems to map
terns of evaluation of men and women, or Blacks and fairly well onto our zero-sum/non-zero-sum distinction.
Whites, on stereotyped dimensions. But the implications Spontaneous behaviors are similar to our non-zero-sum
of this differential judgment pattern for intergroup behaviors—they include the kinds of “uncontrollable”
behavior are less clear. Specifically, a logical question nonverbal behaviors that may occur during intergroup
that follows from research on judgment is whether interaction. And most zero-sum behaviors are likely to be
behavior toward target individuals follows from objec- deliberative as well; any choice between alternatives or
tive assessments (which tend to show assimilative stereo- allocation of resources is likely to involve conscious
typing effects) or from subjective assessments (which are deliberation. However, one can imagine a non-zero-sum
less likely to do so). The key prediction of this research behavior that involves deliberation (e.g., Should I smile
was that objective judgments would better predict zero- at this person or not? Should I laugh at this joke?) or a
sum behaviors, whereas subjective judgments would zero-sum behavior that does not (e.g., I happen to spend
better predict non-zero-sum behaviors. more of my available time helping one person rather
This prediction was tested via the correlational analy- than another). The spontaneous/deliberative distinc-
ses reported in Study 2 (see Table 3). Both subjective and tion may be more appropriate for considering the types
objective assessments of batting performance were cor- of issues addressed by Dovidio and Fazio and col-
related with the main zero-sum behavior (team selec- leagues—how self-report versus less reactive measures of
tions) and with the non-zero-sum behavior (verbal and attitudes predict self-report (deliberative) versus less
nonverbal behavioral reactions). However, partial corre- reactive (more implicit and spontaneous) measures of
lations generally supported the predicted pattern of behavior. Indeed, building on the Skinnerian distinction
results: The tendency to choose male as opposed to between “operant” and “respondent” behaviors (see
female players for the team was more strongly related to Skinner, 1938), motivation researchers have long argued
objective than subjective batting judgments (although that Thematic Apperception Test (TAT) measures of
not significantly so), and the tendency to react favorably motives are better predictors of operant or spontaneous
in verbal and nonverbal behavior was more strongly behaviors, whereas self-report measures of values better
related to subjective than objective judgments. That is, predict respondent behaviors (“what people cognitively
choosing men as opposed to women for the team was decide should be done”) (Biernat, 1989, p. 73; see
most strongly associated with judging that male players McClelland, 1980, 1985; McClelland, Koestner, &
have better batting averages than female players, and Weinberger, 1989).
greater behavioral reaction toward male than female For our purposes, however, the zero-sum/non-zero-
players’ hits was most strongly associated with judging sum distinction better captures the fact that in a given sit-
women as subjectively better batters than men. uation, some behaviors are of limited and some of unlim-
Although these findings were not as strong and strik- ited “supply.” This distinction is important in that lim-
ing as we might like, they do suggest that both subjective ited supply (zero-sum) behaviors require consideration
and objective evaluations of members of stereotyped across a number of potential targets, whereas unlimited
groups have meaningful, but different, behavioral impli- supply (non-zero-sum) behaviors do not. As the shifting
cations. Objective judgments, which do not allow for standards model suggests, it is when one is making cross-
within-category meaning shifts, are linked to limited target comparisons that stereotypic between-group dif-
resource zero-sum behaviors; subjective judgments, ferences are likely to be noted and revealed. When
which reflect within-category standard shifts, are linked behavior toward one target is independent of behavior
to the kinds of verbal and nonverbal reactions that are toward another, however, it may follow from intra- rather
not typically exhausted in one’s behavioral repertoire. than intergroup conceptions of targets (“Coming from a
Other researchers in the domain of prejudice and ste- female player, a single makes me cheer!”). This is not to
reotyping have made distinctions between types of deny that non-zero-sum behaviors such as nonverbal
behavioral response to stereotyped group members. For interaction style may reveal stereotypic biases—indeed,
example, Dovidio and his colleagues have distinguished a highly cited study on interracial interactions docu-
between spontaneous and deliberative behaviors ments the less favorable nonverbal behaviors that tend to
(Dovidio, Brigham, Johnson, & Gaertner, 1996; Dovidio, be directed toward Blacks, with negative effect (Word,
76 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN

Zanna, & Cooper, 1974). However, we suggest that just as when non-zero-sum behaviors were regressed on both objective and
subjective judgments, Bs = – .01 (ns) and –.31 (p = .05), respectively.
subjective judgment scales tend to mask stereotyping 5. Nonetheless, it is worth noting that batting average estimates,
effects, non-zero-sum behaviors may mask the full extent fielding error rates, and non-zero-sum behavioral reactions also dif-
of perceivers’ stereotypic bias because they allow for fered for the highly athletic men versus women. Perhaps this reflects
perceivers’ expectations that even the best women players will perform
within- rather than across-group comparison and less competently on the field than the best men players (whereas the
consideration. male and female “wimps” are equally bad) and that because of low
expectations, a single is to be greeted with some flourish. In terms of
More generally, we suggest that both subjective and zero-sum behavior, however, participants appeared willing to treat the
objective judgment, and both zero-sum and non-zero- highly athletic women and men with equal favor. This may reflect a nor-
sum behavior, have important implications for individ- mative or motivational concern that is relevant only when it comes to
zero-sum choices: A manager of a co-ed team is obligated to place some
ual targets. Our research focused on the game of soft- women on the team, in infield positions, and high in the batting order,
ball, where we found that female players were evaluated and for these selections, the most athletic females are chosen. But at
against lower standards than male players, such that they the same time, the manager may still judge those “competent” female
players as less good than their male counterparts.
were less likely to receive valuable resource allocation
but more likely to receive positive nonverbal feedback
for good performance. This is surely likely to be a confus- REFERENCES
ing situation for both male and female athletes, and the Bettencourt, B. A., Dill, K. E., Greathouse, S. A., Charlton, K., &
playing field may only be leveled after a female athlete Mulholland, A. (1997). Evaluations of ingroup and outgroup
members: The role of category-based expectancy violation. Jour-
has shown repeated and extensive high competence on nal of Experimental Social Psychology, 33, 244-275.
the diamond (such that good performance is no longer Biernat, M. (1989). Achievement motives and values: Different con-
structs with different effects. Journal of Personality, 57, 69-95.
unexpected). Further research is needed to test whether Biernat, M. (1995). The shifting standards model: Implications of ste-
the effects reported here are equally strong when actual reotype accuracy for social judgment. In Y. T. Lee, L. J. Jussim, &
rather than hypothetical behavior is considered and to C. R. McCauley (Eds.), Stereotype accuracy: Toward appreciating group
differences (pp. 87-114). Washington, DC: American Psychological
examine the generalizability and boundary conditions Association.
of this pattern of findings. For example, in the work- Biernat, M., Crandall, C. S., Young, L. V., Kobrynowicz, D., & Halpin, S. M.
place, women and Blacks may face a situation in which (1998). All that you can be: Stereotyping of self and others in a mil-
itary context. Journal of Personality and Social Psychology, 75, 301-
scarce resources and choice assignments are not forth- 317.
coming but nonverbal feedback may be overly favorable Biernat, M., & Kobrynowicz, D. (1997). Gender- and race-based stan-
or unrealistic (Harber, 1998; Massey, Scott, & Dorn- dards of competence: Lower minimum standards but higher abil-
ity standards for devalued groups. Journal of Personality and Social
busch, 1975; Pettigrew & Martin, 1987; Steele, 1992). Psychology, 72, 544-557.
How both perceivers and targets negotiate the some- Biernat, M., & Manis, M. (1994). Shifting standards and stereotype-
based judgments. Journal of Personality and Social Psychology, 66, 5-20.
times conflicting judgments and behaviors that stereo- Biernat, M., Manis, M., & Kobrynowicz, D. (1997). Simultaneous
types evoke remains an important research question. assimilation and contrast effects in judgments of self and other.
Journal of Personality and Social Psychology, 73, 254-269.
Biernat, M., Manis, M., & Nelson, T. E. (1991). Stereotypes and stan-
NOTES dards of judgment. Journal of Personality and Social Psychology, 60,
485-499.
1. Study 1 also included a manipulation in which half of the partici- Biernat, M., Vescio, T. K., & Manis, M. (1998). Judging and behaving
pants were told “because this is a co-ed team, your team must have toward members of stereotyped groups: A shifting standards per-
roughly equal numbers of male and female players” and half were not. spective. In C. Sedikides, J. Schopler, & C. A. Insko (Eds.), Inter-
This manipulation had no effects on any of the decisions or judgments group cognition and intergroup behavior (pp. 151-175). Hillsdale, NJ:
and therefore will not be discussed further. In retrospect, we believe Lawrence Erlbaum.
this one-sentence manipulation was too subtle and that by repeatedly Brewer, M. B., (1988). A dual process model of impression formation.
framing the task in terms of co-ed softball, all participants felt some In T. K. Srull & R. S. Wyer (Eds.), Advances in social cognition (Vol. 1,
necessity of being relatively egalitarian in their choices. pp. 1-36). Hillsdale, NJ: Lawrence Erlbaum.
2. In addition to the analyses reported below, we examined whether Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming bias in
any of these measured variables (knowledge, own experience, etc.) labeling effects. Journal of Personality and Social Psychology, 44, 20-33.
moderated the results. Other than a general tendency for softball Dovidio, J. F., Brigham, J. C., Johnson, B. T., & Gaertner, S. L. (1996).
“experts” to estimate lower fielding error rates and lower batting aver- Stereotyping, prejudice, and discrimination: Another look. In
ages overall than nonexperts, and for those who judged themselves as C. N. Macrae, C. Stangor, & M. Hewstone (Eds.), Stereotypes and ste-
worse players to estimate higher error rates than those who judged reotyping (pp. 276-319). New York: Guilford.
themselves as good players, there was no evidence in either study of Dovidio, J. F., & Fazio, R. H. (1992). New technologies for the direct
moderation of the effects reported in the text. Participant sex effects and indirect assessment of attitudes. In J. Tanur (Ed.), Questions
emerged only rarely, as described in the Results section. about questions: Inquiries into the cognitive bases of surveys (pp. 204-
3. These effects were even stronger when we considered only the 237). New York: Russell Sage.
most prestigious infield positions—pitcher, first base, shortstop, third Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A.
base—or when we considered assignment to the more prestigious left (1997). On the nature of prejudice: Automatic and controlled
half of the field—pitcher, shortstop, third base, left and left center processes. Journal of Experimental Social Psychology, 33, 510-540.
field—relative to the right. Fazio, R. H. (1990). Multiple processes by which attitudes guide behav-
4. Of course, results from a multiple regression approach also sup- ior: The MODE Model as an integrative framework. In M. P. Zanna
port this pattern—when team choice was regressed on both objective (Ed.), Advances in experimental social psychology (Vol. 23, pp. 75-
and subjective judgments, Bs = .37 (p < .05) and .21 (ns), respectively; 109). Orlando, FL: Academic Press.
Biernat, Vescio / STANDARDS AND INTERGROUP BEHAVIOR 77

Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). McClelland, D. C. (1980). Motive dispositions: The merits of operant
Variability in automatic activation as an unobtrusive measure of and respondent measures. In L. Wheeler (Ed.), Review of personal-
racial attitudes: A bona fide pipeline? Journal of Personality and ity and social psychology (pp. 10-41). London: Sage Ltd.
Social Psychology, 69, 1013-1027. McClelland, D. C. (1985). How motives, skills, and values determine
Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In what people do. American Psychologist, 40, 812-823.
D. T. Gilbert, S. T. Fiske, & G. Lindzey (Ed.), The handbook of social McClelland, D. C., Koestner, R., & Weinberger, J. (1989). How do self-
psychology (Vol. 2, 4th ed., pp. 357-411). New York: McGraw-Hill. attributed and implicit motives differ? Psychological Review, 96, 690-
Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression for- 702.
mation, from category-based to individuating processes: Influ- Pettigrew, T., & Martin, J. (1987). Shaping the organizational context
ences of information and motivation on attention and interpreta- for Black American inclusion. Journal of Social Issues, 43, 41-78.
tion. In M. Zanna (Ed.), Advances in experimental social psychology Skinner, B. F. (1938). The behavior of organisms. Englewood Cliffs, NJ:
(Vol. 23, pp. 1-74). New York: Academic Press. Prentice Hall.
Foschi, M. (1992). Gender and double standards for competence. In Stangor, C., & Lange, J. E. (1994). Mental representations of social
C. L. Ridgeway (Ed.), Gender, interaction, and inequality (pp. 181- groups: Advances in understanding stereotypes and stereotyping.
207). New York: Springer-Verlag. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol.
Gupta, N., Jenkins, G. D., Jr., & Beehr, T. A. (1983). Employee gender, 26, pp. 357-416). Orlando, FL: Academic Press.
gender similarity, and supervisor-subordinate cross-evaluations. Stapel, D. A., & Koomen, W. (1998). When stereotype activation
Psychology of Women Quarterly, 8, 174-184. results in (counter)stereotypical judgments: Priming stereotype-
Hamilton, D. L., & Sherman, J. W. (1994). Stereotypes. In R. S. Wyer & relevant traits and exemplars. Journal of Experimental Social Psychol-
T. K. Srull (Eds.), Handbook of social cognition (Vol. 2, pp. 1-68). ogy, 34, 136-163.
Hillsdale, NJ: Lawrence Erlbaum. Stapel, D. A., Koomen, W., & van der Pligt, J. (1997). Categories of cat-
Harber, K. D. (1998). Feedback to minorities: Evidence of a positive egory accessibility: The impact of trait concept versus exemplar
bias. Journal of Personality and Social Psychology, 74, 622-628. priming on person judgments. Journal of Experimental Social Psy-
Jackman, M. R. (1994). The velvet glove: Paternalism and conflict in gen- chology, 33, 47-76.
der, class, and race relations. Berkeley: University of California Press. Stapel, D. A., & Winkielman, P. (1998). Assimilation and contrast as a
Jussim, L., Coleman, L. M., & Lerch, L. (1987). The nature of stereo- function of context-target similarity, distinctness, and dimen-
types: A comparison and integration of three theories. Journal of sional relevance. Personality and Social Psychology Bulletin, 24, 634-
Personality and Social Psychology, 52, 536-546. 646.
Kobrynowicz, D., & Biernat, M. (1997). Decoding subjective evalua- Steele, C. (1992, April). Race and the schooling of Black Americans.
tions: How stereotypes provide shifting standards. Journal of Experi- Atlantic Monthly, pp. 68-78.
mental Social Psychology, 33, 579-601. von Hippel, W., Sekaquaptewa, D., & Vargas, P. (1995). On the role of
Kunda, Z., & Sherman-Williams, B. (1993). Stereotypes and the encoding processes in stereotype maintenance. In M. P. Zanna
construal of individuating information. Personality and Social Psy- (Ed.), Advances in experimental social psychology (Vol. 27, pp. 177-
chology Bulletin, 19, 90-99. 254). New York: Academic Press.
Kunda, Z., & Thagard, P. (1996). Forming impressions from stereo- Weitz, S. (1972). Attitude, voice, and behavior: A repressed affect
types, traits, and behaviors: A parallel-constraint-satisfaction the- model of interracial interaction. Journal of Personality and Social
ory. Psychological Review, 103, 284-308. Psychology, 24, 14-21.
Manis, M., Nelson, T. E., & Shedler, J. (1988). Stereotypes and social Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal media-
judgment: Extremity, assimilation, and contrast. Journal of Person- tion of self-fulfilling prophecies in interracial interaction. Journal
ality and Social Psychology, 55, 28-36. of Experimental Social Psychology, 10, 109-120.
Massey, G., Scott, V., & Dornbusch, S. (1975, November). Racism with-
out racists: Institutional racism in urban schools. The Black Scholar, Received July 5, 2000
pp. 10-19. Revision accepted February 17, 2001

You might also like