Study Purposes and Research Questions
“Two experiments showed that framing an athletic task as diagnostic of negative racial stereotypes about Black or White athletes can impede their performance in sports.
In Experiment 1, Black participants performed significantly worse than did control participants when performance on a golf task was framed
as diagnostic of “sports intelligence.” In comparison, White participants performed worse than did control participants when the golf task was framed as diagnostic of “natural athletic ability.”
Experiment 2 observed the effect of stereotype threat on the athletic performance of White participants for whom performance in sports represented a significant measure of their self-worth. The implications of the findings for the theory of stereotype threat (C. M. Steele, 1997) and for participation in sports are discussed.” (Stone, Sjomeling, Lynch & Darley, 1999. p 1213)
Research Methods and Design
“Gordon Allport (1954) observed that being the target of a negative stereotype about an important social identity is distressing and promotes a number of defensive reactions. Specifically, when negative stereotypes are made salient in a situation, Allport predicted that members of a stigmatized group may respond with “obsessive concern” about being labeled and treated in terms of the negative characterization of their group. Their concern could turn to anger, aggressiveness, and militant action, but the potential threat of being stereotyped negatively might also lead targets to become anxious, withdrawn, and even “self-hating.” Indeed, research has shown that to reduce the threat engendered by being the target of a negative stereotype, people may attribute their predicament to the perceiver’s prejudice (Crocker, Voelkl, Testa, ; Major, 1991), engage in self-serving social comparisons (e.g., see Major, Schiaccitano, ; Crocker, 1993), or disengage their self-esteem from the situation (e.g., Major, Spencer, Schmader, Wolfe, ; Crocker, 1998). Whereas many of the specific reactions described by Allport have yet to be investigated empirically, the current literature suggests that being the target of a negative stereotype represents a significant threat to self-regard, one that creates substantial concern and discomfort for individual members of a stigmatized group (Crocker, Major, ; Steele, 1998). Researchers have recently turned their attention to whether negative reactions to being the target of a negative stereotype occur among traditionally nonstigmatized groups (e.g., Aronson et al., 1999; Brown ; Josephs, 1999).
The purpose of the current research was to examine whether majority group members such as White college students also experience concern and suffer distress when they are the target of a negative stereotype about their racial identity. Specifically, we examined the responses of majority group members in the context of sports in which there appear to be negative stereotypes about White athletes. Our research was guided by the hypothesis that many people who participate in sports hold negative racial stereotypes about athletes, and when these stereotypes become salient in a performance context, they may threaten the self-worth of those to whom they apply.” (Stone, et al., 1999) This hypothesis is relevant to traditionally stigmatized groups such as minorities or women, but our central tenet is that it applies equally to White males and females who are not often examined as the target of negative racial stereotypes (Fiske, 1998). Sports and athletics may represent one of the few domains in which Whites are stereotyped negatively and suffer psychologically as a result in proposing this hypothesis, we make at least two theoretical assumptions that require some discussion.
First, there is the assumption that people hold racial stereotypes that are specific to athletes. Racial stereotypes about athletes — particularly Black and White athletes — represent overgeneralized beliefs about the causes of athletic success and failure. The available archival and empirical evidence suggests that people hold very specific positive and negative dispositional beliefs about what lies at the heart of Black and White athletic performance. The current study focuses on the negative beliefs about race and athletic performance that, when salient in a sports performance context, may cause increased concern and distress among the athletes to whom they apply. Our second prediction is that the salience of negative stereotypes can have a negative impact on the athletic performance of Black and White participants in sports.
Although negative stereotypes about athletes can affect performance through interaction with people who hold race-based expectancies (e.g., Darley & Fazio, 1980; Horn, Lox, & Labrador, 1998), the purpose of the current research was to examine a more subtle process by which stereotypes might influence performance in sports: that of stereotype threat. The theory of stereotype threat (Steele, 1997) states that when a negative stereotype about a group becomes salient as the criterion for evaluating performance, individual group members become concerned that their performance may confirm the validity of the negative stereotype. The increased concern created by the threat imposes an additional psychological burden to the task, which, in turn, reduces an individual’s ability to perform to their potential. Thus, in the absence of interpersonal interaction, negative stereotypes, when made salient in a stereotype relevant domain, can have a negative impact on performance in that context.
The primary goal of the current research was to examine how stereotype threat processes might operate in the context of sports for both Black and White participants. Devine and Baker (1991) found that the attributes assigned to the social category of Black athlete included unintelligent and ostentatious, and Biernat and Manis (1994) reported that Black males were perceived to be more athletic than White males. In perhaps the most comprehensive study, Sailes (1996) asked Black and White college students to rate the intelligence, academic preparation, athletic style of play, competitiveness, physical superiority, athletic ability, and mental temperament of Black and White college athletes.
The results showed that White participants rated Black athletes as significantly less intelligent, less academically prepared, and more temperamental, whereas Black participants rated White athletes as significantly less competitive and as exhibiting less “athletic style.”1 Thus, judgments about the characteristics of Black and White athletes tend to reflect the stereotype that Blacks athletes are physically superior but intellectually inferior to White athletes. Recent research suggests that such stereotyped beliefs about Black and White athletes can influence perceptions of an athlete’s performance (e.g., Sapolsky, 1980). In a perceptual confirmation paradigm, Stone, Perry, and Darley (1997) had White participants evaluate a basketball player while listening to a radio broadcast of a college basketball game.
Half the participants were led to believe the target player was White, and half were led to believe the target was a Black athlete. On the broadcast, it was clear that the identified athlete played very well. The results showed that whereas the White male target was perceived as exhibiting less “natural athletic ability” but more “court smarts” and “hustle,” the Black male target player was perceived as exhibiting less court smarts and hustle, but more natural athletic ability. The Black target was also perceived to be a better basketball player even though all participants heard the same running account of the athlete’s basketball performance. Thus, the available empirical evidence tends to mirror what is found in the popular sports literature: People hold racial stereotypes about athletes that are both positive and negative. Specifically, Black athletes are perceived to have natural athletic ability (which is a positive sports attribute) but are thought to be less intelligent, even in a sports context (a negative sports attribute). In contrast, White athletes are perceived to have less natural athletic ability (a negative sports attribute) but are thought to be intelligent and perhaps harder working (positive sports attributes).
These stereotyped beliefs have been shown to influence judgment and perception processes, but it is not currently known whether they can also influence how athletes perceive themselves. According to Steele and his colleagues (e.g., Steele, 1997; Steele & Aronson, 1995), when a negative stereotype about a group becomes salient as a criterion for test evaluation, individual group members can become concerned about confirming the negative stereotype. The concern raised by the salience of the stereotype can subsequently cause individual group members to perform more poorly than they would in a neutral context. For example, Steele and Aronson showed that when the specter of innate intelligence was made salient in the context of a standardized test of academic ability, or when race was simply made salient, White students performed significantly better than did Black students. However, when the test was framed as non diagnostic of innate intelligence, or when race was not made salient, Black and White students performed equally well. Similar performance effects have been observed among women when the gender stereotype concerning math ability is made salient in a testing context (e.g., Spencer, Steele, & Quinn, 1999). Thus, even subtle indications that one is being evaluated on the basis of a negative characterization about one’s group identity can have a debilitating effect on performance in a stereotype relevant Domain.
“The hypothesized effects of stereotype threat on athletic performance were tested in an experiment in which Black and White participants completed a sports-related task that was based on the game of golf. We predicted that if the negative stereotypes about Black athletes (i.e., poor sports intelligence) and White athletes (i.e., poor natural athletic ability) were made salient during an athletic performance, concern over confirming the stereotype would cause each group’s athletic performance to suffer, relative to when each group’s performance on the task was being evaluated on the basis of either the positive stereotype concerning their group (i.e., natural athletic ability for Blacks, sports intelligence for Whites) or a non-stereotype relevant criterion.
In a 2 (race: Black or White) X 4 (test frame: natural athletic ability, sports intelligence, race prime, or no-prime control) experimental design, participants completed a pre-performance questionnaire and then performed a “standardized” measure of athletic performance that was based on the game of golf. Performance on the task, as measured by the number of strokes required to complete the 10-hole golf course, was the primary dependent measure.
Participants were 82 male and female. Black and White undergraduates at Princeton University, who were offered $4 for participation in a study on sports psychology. No one with extensive knowledge of golf (e.g., members of the golf team or students who reportedly played more than once a week) was used in the study. The data from 2 participants (1 Black and 1 White) were excluded before the analysis because they failed to follow instructions during the experiment. The final sample consisted of 40 Black and 40 White
Participants completed the procedures individually. When they arrived at the laboratory, a Black male experimenter explained that they would complete a brief questionnaire, perform a sports test that was based on the game of golf, and then answer questions about their performance after the test was completed. Participants then read a handout that explained the purpose of the study. The athletic test was described as a standardized measure of sports psychology called the Michigan Athletic Aptitude Test (MAAT). Participants were told that the MAAT was developed in 1988 by the exercise and sports psychology department at the University of Michigan. The handout noted that the test was based on the game of golf but had been normalized such that each successive level in the test represented a standard increase in performance difficulty. Ostensibly, performance on the test had been shown to correlate with actual performance on many of the physical and mental activities relevant to most college varsity sports, such as basketball, baseball, and hockey. At this point, the cover story altered its course according to which test frame condition participants had been assigned to randomly.
Test frame manipulation.
Participants in the natural athletic ability condition read that the test was designed to measure personal factors correlated with natural athletic ability. Natural athletic ability was defined as “one’s natural ability to perform complex tasks that require hand—eye coordination, such as shooting, throwing, or hitting a ball or other moving objects.” It was explained that as test difficulty increased, so would the demand on their natural athletic ability or hand-eye coordination. Participants randomly assigned to the sports intelligence condition read that the test was designed to measure “personal factors correlated with the ability to think strategically during an athletic performance.” The handout explained that as test difficulty increased, so would the demand on their ability to use different strategies while performing the athletic test.
Participants in the race-prime and no-prime control conditions read that the test was designed to measure psychological factors correlated with “general sports performance.” The handout explained that as test difficulty increased, so would the demand on the psychological factors that correlate with general sports performance. After participants had read the handout, the experimenter reiterated the instructions and answered questions. Participants then completed a pretest questionnaire.
First, they completed a five-item self-report measure of situational anxiety (Mattsson, 1960). Participants were instructed to indicate their current level of anxiety on 7-point scales with endpoints ranging from uneasy to at ease, comfortable to uncomfortable, upset to peaceful, relaxed to tense, and in control to not in control. To measure self-handicapping processes, participants indicated how many hours of sleep they had the night before, and they rated how focused they felt and how much stress they had been under lately. They also indicated how much bias they perceive in standardized tests on a 7-point scale ranging from not at all (1) to very much (7).
Last, participants completed a brief demographic questionnaire on which they indicated their gender, age, year in school, and racial identity. The racial identity question was used to prime race prior to performance on the athletic test (e.g., Steele ; Aronson, 1995). The demographic questionnaire was presented first in the questionnaire packet in the race-prime condition.
Research Results and Conclusions
Gender did not significantly moderate the effects of race and test frame effects on the performance and self-report measures reported below; consequently, we collapsed across this variable in the analyses.
The number of strokes required to finish the course was submitted to a 2 (race) X 4 (test frame) analysis of variance (ANOVA). The ANOVA revealed only the predicted Race X Test Frame interaction effect, F (3, 72) = 5.70, p ; .001. Orthogonal contrasts and simple effects analyses were then conducted to determine the meaning of the interaction across the eight experimental conditions (see Table 1). First, a planned contrast of the mean differences showed that in the no-prime control condition, White participants (M = 24.6) performed slightly, though not significantly, worse than Black participants (M = 22.1), F (l, 72) = 1.71, p ; .20. The slight difference here suggests that despite White dominance in the professional sport of golf, Black and White participants were on a relatively level playing field in the current golf setting.
Next, the simple interaction effect between race and the natural athletic ability, sports intelligence, and race-prime conditions was significant, F (l, 72) = 7.15, p ; .001. Planned contrasts revealed that among Black participants, performance was significantly better when the test was framed as a measure of natural athletic ability (M = 23.10) compared with when it was framed as a measure of sports intelligence (M = 27.20) or when race was primed prior to performance (M = 27.30), F (l, 72) = 6.27, p ; .01. Performance in the sports intelligence and race-prime conditions did not differ significantly (F ; 1). Also, as predicted, performance by Black participants when the test was framed as diagnostic of sports intelligence or when race was primed was significantly worse compared with the no-prime control group using Dunnett's two-tailed t test for comparison, « (72) = 2.45, p ; .05. Thus, the data supported the prediction that making race or sports intelligence salient would undermine the performance of Black participants on the golf task.
Table I: Effects of Race and the Athletic Test Frame on Performance, Change in Anxiety, and Discounting in Experiment 1
“The present research showed that the salience of stereotype threat in the context of an athletic performance can adversely affect the performance of both Black and White individuals who play sports. Referring back to Allport's (1954) observations, we find that the current data support his assumption that once targets become aware that they are being evaluated in terms of a negative stereotype, they become intensely concerned about being labeled and treated in terms of the negative characterization. The concern they feel appears to overwhelm their ability to think and perform as well as they do when a negative stereotype is not made salient in the situation. Thus, the challenge to the self-imposed by the salience of a negative stereotype about an important social identity appears to consume the very resources that targets require to overcome the threat. The specific conditions of stereotype threat that lead to poorer athletic performance, however, are different for various groups. For example, Blacks suffer when the stereotype concerning their supposed poor sports intelligence is made salient, whereas Whites suffer when the stereotype concerning their supposed poor natural athletic ability is made salient. The problems that follow from stereotype threat salience appear to occur primarily among individuals for whom sports performance is important to their sense of self-worth. It also appears that contextual factors, such as the way in which performance is framed and cues that distract people from thinking about the negative stereotype, can reduce the harmful impact of a salient negative stereotype. The current data suggest that stereotype threat processes stem from the context in which targets perform, and the implication is that changes in the performance context can reduce the negative impact that stereotypes can have when they are brought to mind (e.g., Crocker, 1999; Steele, 1997).” (Stone, et al, 1999 pp 1225-1226).