CU R RE N T D I R E CT I O NS IN P SYC H OL OG I C AL SC I EN C E

Identity, Belonging, and

Achievement
A Model, Interventions, Implications
Geoffrey L. Cohen1 and Julio Garcia2
1
University of Colorado, Boulder, and 2Yale University

ABSTRACT—In this article we discuss how social or group hen, 2007a). A state of mere belonging, the minimal conditions
identities affect achievement. We also present a model of needed to make individuals feel connected to others in an
identity engagement that describes how a salient social achievement domain, was produced in students. Students were
identity can trigger psychological threat and belonging either told that their birthday fell on the same day as someone
concerns and how these can produce persistent perfor- who was a math major, assigned to a small group identified as the
mance decrements, which through feedback loops can ‘‘numbers group,’’ or presented with the possibility of forming
increase over time. The character of such processes may social ties in math. Each of these manipulations, by creating a
be revealed only over time because they are recursive in sense of connectedness to others in math, independently raised
nature and interact with other factors in chronically participants’ achievement motivation, for instance increasing
evaluative social environments. Finally, we address how their persistence on an insoluble math puzzle by 70%.
this model helped in the development of successful inter-
ventions. SITUATIONAL IDENTITY THREAT
KEYWORDS—stigma; stereotype threat; achievement gap;
A major lesson of recent research is that social identities, such as
race; gender; achievement; identity; social identity
one’s race or gender, can interact with other factors in situational
to affect achievement. Steele and colleagues examine situational
What motivates people to achieve is one of psychology’s classic factors that can trigger stereotype threat (Steele, Spencer, &
questions. It is thought that motivation often depends on whether Aronson, 2002), a fear of confirming a negative stereotype about
an endeavor is linked to one’s sense of social identity and feel- one’s group. It can undermine performance by raising stress and
ings of belonging. For instance, one’s family, racial or ethnic increasing mental load (Schmader & Johns, 2003). For instance,
group, and religious affiliation can be important sources of social when told that a difficult standardized test measured intellectual
identity and social belongingness and so provide powerful mo- ability, African American college students performed much
tives for achievement. worse than they did when told the same test did not measure their
Research has begun to shed light on how identity contributes ability (see Steele et al., 2002). White students, by contrast,
to motivation. A seminal study led students to identify with math maintained their performance. That is, African Americans un-
by having their teachers label them as a ‘‘very good arithmetic derperformed when aware that failure could reinforce a negative
student’’ or as ‘‘working hard in arithmetic’’ (Miller, Brickman, & stereotype about their race. Similar effects have been found for
Bolen, 1975). Compared with participants who had not been so women in math, White men in certain athletic domains, and poor
labeled, the students in these two groups earned higher math students in school (see Steele et al., 2002, for a review). Negative
performance, an effect present after 2 weeks. In a series of characterizations of valued social identities—when salient in a
studies with our colleague Gregory Walton, we explored the re- situation—can be threatening and so undermine achievement.
lationship between social identity and motivation by making an Social identities can affect the motivation to achieve through
achievement domain a basis of social belonging (Walton & Co- their interaction with a sense of belonging. Belonging uncer-
tainty, doubt as to whether one will be accepted or rejected by
key figures in the social environment, can prove acute if rejec-
Address correspondence to Geoffrey Cohen, Department of Psychol-
ogy, University of Colorado, Muenzinger Psychology Bldg., Boulder, tion could be based on one’s negatively stereotyped social
CO 80309-0345; e-mail: [email protected]. identity (Walton & Cohen, 2007b; see also Mendoza-Denton,

Volume 17—Number 6 Copyright r 2008 Association for Psychological Science 365

 Identity, Belonging, and Achievement

Downey, Purdie, Davis, & Pietrzak, 2002). For instance, for they can be vulnerable to collective threat, a threat to the image of
nonstereotyped students, a bad day in school or an experimen- their group (Cohen & Garcia, 2005). Because this threat involves
tally created sense of social isolation in an academic domain had other group members and situations in which one is not per-
little effect (Walton & Cohen, 2007b). However, for African sonally evaluated, it can be more pervasive and chronic than
American students, apparently identical hardships undermined other ego threats. For example, merely making African Ameri-
their sense of belonging and motivation to achieve. Moreover, cans aware that somebody else in their group could confirm a
when African Americans’ belonging uncertainty was experi- negative stereotype of their race led to a drop in their self-esteem
mentally resolved, they did not experience the undermining and test performance. These three ideas informed both our
effects of hardships. As discussed below, those who experienced theoretical model of social-identity threat and interventions that
such resolution later earned a higher GPA than their African closed achievement gaps in schools.
American peers.
The above research illustrates three key ideas. First, although
a situation may seem identical for all those in it, it may differ SOCIAL-IDENTITY THREAT IN CHRONICALLY
radically for different groups of people. In classrooms, White EVALUATIVE ENVIRONMENTS
students are in a situation testing their academic skills. African
American and Latino students also confront this prospect. In a social environment, no one performance or psychological
However, they also face the extra burden of knowing that their state occurs in isolation. It interacts with other psychological
skills, and those of others in their group, could be viewed through and environmental factors often through recursive cycles. This
the lens of a stereotype that questions their group’s intellectual notion lies at the core of the Identity Engagement Model pre-
and academic abilities. This concern can occur regardless of the sented in Figure 1. Although we focus on the implications of our
actual level of prejudice. Second, intellectual achievement is model for minorities’ school achievement, similar processes can
malleable (Aronson, Fried, & Good, 2002). Subtle factors can affect performance in a variety of settings, including intergroup
profoundly affect performance. For instance, the effects of ste- conflict and negotiation, and can apply to any number of social
reotype threat on women’s math performance and career aspi- identities (Cohen et al., 2007).
rations were reduced when they were merely asked to generate a As shown in Figure 1, on entering an important social envi-
list of qualities shared by men and women (Rosenthal & Crisp, ronment like a classroom, people will often make a general
2006). Third, because people derive self-worth from their group, assessment by asking, sometimes implicitly, ‘‘Is this is a situa-

Is identity engaged?

Yes

– +
Vigilance

Cues disconfirm Cues confirm

identity threat identity threat

Performance contingent on Threat appraisal:

task- and individual-related factors Ability and desire to cope with threat?
(e.g., feedback quality,
self-efficacy, ability)
Yes No

Sustained or improved performance Lower performance

Fig. 1. The Identity Engagement Model of the effects of social-identity threat on performance.
If people think their identity could be the basis of negative treatment, that identity will be
psychologically engaged, leading them to be vigilant for situational cues related to whether
or not they are in fact being treated negatively. If the cues disconfirm such identity threat,
performance will be relatively more contingent on task- and individual-related factors, often
leading to sustained or improved performance. If the cues confirm identity threat, an appraisal
process will follow, in which people assess whether they have the ability and desire to cope with
the threat. If the answer is yes, sustained or improved performance will tend to follow. If the
answer is no, lower performance will tend to follow. Recursion can occur at several junctures,
as when threat-confirming cues exacerbate vigilance or when low performance exacerbates the
experience of threat, leading in turn to worse performance.

366 Volume 17—Number 6

 Geoffrey L. Cohen and Julio Garcia

tion in which my identity (e.g., as a minority) could be tied to the laboratory may prove only the first stage of a prolonged and
negative outcomes?’’ If they answer yes, their identity will be complex process in the real world. Second, because recursive
psychologically engaged. People tend to become vigilant when processes depend on continual feedback loops, a well-placed
this occurs. They monitor for cues as to whether they are vul- interruption can produce large and long-term effects. This can
nerable to negative treatment based on their identity (Kaiser, prove especially likely if the interruption occurs early enough to
Brooke, & Major, 2006; Purdie-Vaughns, Steele, Davies, Ditl- prevent a downward spiral from emerging or introduces a posi-
mann, & Randall-Crosby, 2008). A minority student, for exam- tive recursive cycle. For instance, interventions may prove
ple, might scrutinize a teacher’s nonverbal behavior for bias. especially effective if they reduce threat, which then improves
Such vigilance is a general and adaptive process. If people people’s performance, further reducing threat, in a self-rein-
discover cues that refute the threat of devaluation, they will tend forcing cycle.
to feel fairly treated and their performance will depend relatively
more on task- and individual-related factors, such as the quality
of instruction and their self-efficacy. This will often lead to INTERVENTION
sustained or improved performance. In one study, relative to
Whites, African Americans receiving critical feedback on an Our model suggests several points of intervention. We discuss
essay saw the feedback as more biased and displayed less mo- two: the vigilance stage and the threat-appraisal stage. During
tivation to revise it (Cohen & Steele, 2002). But when told that the former, the aim of the intervention is to lessen people’s
the feedback was motivated by the instructor’s use of high tendency to interpret experience in light of social identity.
standards and belief in their ability to reach them—that is, when During the latter, it is to buffer people against the emotional
assured that the stereotype was not in play—African Americans impact of such an interpretation. Neither approach, contrary to
responded as positively as White students did. Moreover, when common wisdom, involves directly refuting the stereotype or
told this, stereotyped students complied more with specific propagating a colorblind racial ideology that downplays the
suggestions for improvement and thus their performance bene- importance of ethnicity. Indeed, colorblind messages may un-
fited more from the feedback (Cohen & Steele, 2002). dermine minorities’ trust and belonging, particularly if such
If, however, people discover cues that confirm the existence of messages occur in the absence of actual institutional diversity
a threat, such as racial bias, an appraisal stage will tend to follow. or convey that the distinctive qualities of one’s culture will be
Here individuals assess whether they have the ability and desire ignored or should be suppressed (Purdie-Vaughns et al., 2008).
to deal with the threat. Performance can suffer if people think the The efficacy of theory-driven interventions is illustrated in
threat surpasses their ability or desire to overcome it. However, if randomized, double-blind field experiments conducted by us
individuals think that they can overcome the threat, a challenge and our colleagues focusing on real-world academic outcomes.
response is likely, in which performance is maintained or im- The first intervention, occurring at the vigilance stage, lessened
proves. For instance, identifying with one’s group can be a race-based doubts about belonging. First-year college students
psychological resource that offsets the threat of negative ste- were asked, at the end of the difficult freshman year, to review the
reotypes (Cohen & Garcia, 2005). results of a survey of upperclassmen at their school (Walton &
Psychological processes are recursive and interact with other Cohen, 2007b). This survey conveyed that almost all students,
factors in the social environment. The sense of threat can lower a regardless of race, feel uncertain of their belonging in college in
person’s performance, heightening their sense of threat, which their freshman year and that these feelings wane with time. This
then further lowers performance in a repeating cycle until either led students to view their doubts about belonging in school as
performance stabilizes at a low level or an adaptation occurs. Such common rather than unique to them or their race and as transi-
adaptation could occur when students disidentify or disengage tory rather than fixed (see also Good, Aronson, & Inzlicht, 2003).
from school or when an authority intervenes to interrupt the re- As expected, the intervention had no consistent effect on
peating cycle. Recursion can also occur in the vigilance stage. If Whites. However, it buttressed African Americans’ sense of
people see bias, their vigilance may rise, heightening sensitivity to belonging in school on stressful days. African Americans re-
bias-confirming cues. This repeating cycle could not only initially ceiving the intervention also earned a higher GPA in the sub-
undermine trust but make trust increasingly difficult to regain. sequent semester, significantly closing the racial achievement
Other repeating cycles can occur in interplay with the environ- gap, an effect that our follow-up data indicate persisted into their
ment. For instance, an underperforming student may be viewed by junior year of college. More generally, interventions that en-
teachers as less able, or be assigned to a lower academic track, or courage students to see their academic potential as expandable
affiliate with lower-performing peers, any of which could inhibit rather than fixed, and their difficulties as surmountable with
performance further and in turn increase psychological threat. effort and practice, have proved effective at boosting at-risk
There are at least two implications of our model. First, the students’ academic performance (Aronson et al., 2002; Black-
effects of social-identity threat unfold over time in interaction well, Trzesniewski, & Dweck, 2007; Wilson, Damiani, & Shel-
with other psychological and environmental factors. Effects in ton, 2002). Obviously such strategies are effective only to the

Volume 17—Number 6 367

 Identity, Belonging, and Achievement

degree that the barriers to success in an environment are in fact effective psychological interventions, this one seemed to inter-
surmountable. rupt a recursive cycle (Blackwell, Trzesniewski, & Dweck, 2007;
Our second strategy intervenes at the threat-appraisal stage. It Wilson et al., 2002). While control-group African Americans
increases people’s psychological resources for coping with dropped in their performance and trust in school authorities with
threat, their sense of global adequacy and efficacy. Underpin- time, African Americans doing the intervention did not (Sherman
ning this strategy is the notion that people want and need to see & Cohen, 2006).
themselves in a positive light—to have a sense of self-integrity
(Steele, 1988). Self-affirmation is a process through which
people buttress their self-integrity by asserting or manifesting FURTHER QUESTIONS AND CONCLUDING REMARKS
some value that is important to them (Steele, 1988; see also
Generally people are able to overcome identity threats in spite of
Sherman & Cohen, 2006). For example, when reminded that
having to face them repeatedly. What strategies do they use and
they are the kind of person whose family, profession, or religion
do these differ in effectiveness (Cohen & Garcia, 2005)? What
is important to them, people are better able to tolerate a threat in
role, if any, does a person’s developmental stage play in the
another domain, such as the threat induced by doing poorly in
experience and impact of social identity threat and the strategies
school or work.
they use in response? Additionally, research shows the impact of
In two double-blind field experiments, 7th-grade students
negative stereotypes on those who are subjected to them. How-
completed an in-class self-affirmation exercise early in the
ever, these stereotypes sometimes have a positive impact on the
school year, a stressful time. In a series of structured writing
performance of the nonstereotyped (e.g., Whites and men), who
assignments, they wrote about the importance of a personal va-
may benefit from being on the upside of the negative stereotype
lue, such as religion or relationships with friends, and its role in
(Walton & Cohen, 2003; see also Shih, Pittinsky, & Ambady,
their lives (Cohen, Garcia, Apfel, & Master, 2006; Sherman &
1999). What processes are involved in such effects? More
Cohen, 2006). Compared with a control group completing neu-
pragmatically, how can psychological interventions be con-
tral writing exercises, African Americans in the affirmation
verted into practices that can be implemented throughout a
condition earned a higher fall-term GPA in the course. As shown
school, a district, or the nation? Often the effectiveness of in-
in Figure 2, the intervention cut the percentage of African
terventions decreases due to the lack of control inherent in
Americans earning a D or below by half. This rate was also lower
translating them to a large scale.
than historical norms from three previous cohorts of students in
Granting that many questions still remain, the research dis-
the same course and no different from the rate for White stu-
cussed here underscores the role that factors not directly linked
dents. Follow-up data indicate that the intervention effect on
to the intellect—those related to identity and its effects on
core-course GPA persisted for at least 2 years. As with other
psychological belonging and threat—can play in motivation and
achievement (see also Zigler & Butterfield, 1968). It also points
30% to the importance of examining how social-psychological pro-
African Americans
Percentage of Students Earning

European Americans cesses play out over time in real-world settings. In social envi-
ronments like school and work, psychological processes can feed
a D or Below in Course

25%
off their own consequences and interact with other psychological
20% and environmental factors. Psychological theories should take
into account these interactive and recursive feedback loops.
Perhaps most important, the research also demonstrates the
15%
power of the interplay between rigorous laboratory research and
equally rigorous field research. By informing one another, these
10% two approaches not only sharpened a theoretical understanding
of an important problem, social identity, but also provided the
5% groundwork for a series of pedagogical practices that could
lessen academic inequality and optimize the performance of all
0%
students. This interplay also shows how timely interventions at-
Historical Control Affirmation tuned to social-psychological processes can have long-term effects
Norms Condition Condition that help to solve social problems and better social conditions.
(Three prior cohorts)
Fig. 2. Percentage of students receiving a grade of D or below for African
American students and White students from three previous (prestudy) Recommended Reading
cohorts of students in the course, for African American students and White Aronson, J. (Ed.). (2002). Improving academic achievement: Impact of
students in the control condition, and for African American students and psychological factors on education. San Diego, CA: Academic
White students in the affirmation condition. Press. A well-written and comprehensive edited book addressing

368 Volume 17—Number 6

 Geoffrey L. Cohen and Julio Garcia

the social-psychological processes affecting academic achieve- Improving academic achievement: Impact of psychological factors
ment. on education (pp. 303–328). San Diego, CA: Academic Press.
Crocker, J., Major, B., & Steele, C. (1998). Social stigma. In D.T. Gilbert, Good, C., Aronson, J., & Inzlicht, M. (2003). Improving adolescents’
S.T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology standardized test performance: An intervention to reduce the
(Vol. 2, pp. 504–553). New York: McGraw-Hill. A comprehensive effects of stereotype threat. Journal of Applied Developmental
and thoughtful review of research on the psychological implica- Psychology, 24, 645–662.
tions of stigmatization. Kaiser, C.R., Brooke, V., & Major, B. (2006). Prejudice expectations
Davies, P.G., & Spencer, S.J. (2005). The gender-gap artifact: Women’s moderate preconscious attention to cues that are threatening to
underperformance in quantitative domains through the lens of social identity. Psychological Science, 17, 332–338.
stereotype threat. In A.M. Gallagher & J.C. Kaufman (Eds.), Mendoza-Denton, R., Downey, G., Purdie, V.J., Davis, A., & Pietrzak, J.
Gender differences in mathematics: An integrative psychological (2002). Sensitivity to status-based rejection: Implications for Af-
approach (pp. 172–188). New York: Cambridge University Press. rican American students’ college experience. Journal of Person-
An important analysis of the implications of stereotype threat for ality and Social Psychology, 83, 896–918.
the gender gap in math and science. Miller, R.L., Brickman, P., & Bolen, D. (1975). Attribution versus
Shapiro, J.R., & Neuberg, S.L. (2007). From stereotype threat to ste- persuasion as a means for modifying behavior. Journal of Per-
reotype threats: Implications of a multi-threat framework for sonality and Social Psychology, 31, 430–441.
causes, moderators, mediators, consequences, and interventions. Purdie-Vaughns, V., Steele, C., Davies, P., Ditlmann, R., & Randall-
Personality and Social Psychology Review, 11, 107–130. An in- Crosby, J. (2008). Social identity contingencies: How diversity
tegrative and nuanced review of research on the situational threat cues signal threat or safety for African Americans in mainstream
of being negatively stereotyped. institutions. Journal of Personality and Social Psychology, 94,
Woodhead, M. (1988). When psychology informs public policy: 615–630.
The case of early childhood intervention. American Psychologist, 43, Rosenthal, H., & Crisp, R.J. (2006). Reducing stereotype threat by
443–454. A classic article addressing the complexities involved in blurring intergroup boundaries. Personality and Social Psychol-
testing, refining, and scaling up academic interventions. ogy Bulletin, 32, 501–511.
Schmader, T., & Johns, M. (2003). Converging evidence that stereotype
threat reduces working memory capacity. Journal of Personality
and Social Psychology, 85, 440–452.
Acknowledgments—We thank David Sherman, Gregory Wal- Sherman, D.K., & Cohen, G.L. (2006). The psychology of self-defense:
ton, and Sarah Wert for comments on an earlier draft. Portions of Self-affirmation theory. In M.P. Zanna (Ed.), Advances in experi-
the authors’ research cited in this article were supported by a mental social psychology (Vol. 38, pp. 183–242). San Diego, CA:
grant from the W.T. Grant Foundation, as well as by grants from Academic Press.
Shih, M., Pittinsky, T.L., & Ambady, N. (1999). Stereotype suscept-
the Nellie Mae Education Foundation and the Institute for Social
ibility: Identity salience and shifts in quantitative performance.
and Policy Studies at Yale University. Psychological Science, 10, 80–83.
Steele, C.M. (1988). The psychology of self-affirmation: Sustaining the
REFERENCES integrity of the self. In L. Berkowitz (Ed.), Advances in experimental
social psychology (Vol. 21, pp. 261–302). New York: Academic
Aronson, J., Fried, C.B., & Good, C. (2002). Reducing the effects of Press.
stereotype threat on African American college students by shaping Steele, C.M., Spencer, S.J., & Aronson, J. (2002). Contending with group
theories of intelligence. Journal of Experimental Social Psychol- image: The psychology of stereotype and social identity threat.
ogy, 38, 113–125. In M. Zanna (Ed.), Advances in experimental social psychology
Blackwell, L., Trzesniewski, K., & Dweck, C.S. (2007). Implicit theories (Vol. 34, pp. 379–440). New York: Academic Press.
of intelligence predict achievement across an adolescent transi- Walton, G.M., & Cohen, G.L. (2003). Stereotype lift. Journal of Ex-
tion: A longitudinal study and an intervention. Child Development, perimental Social Psychology, 39, 456–467.
78, 246–263. Walton, G.M., & Cohen, G.L. (2007a). Mere belonging. Unpublished
Cohen, G.L., & Garcia, J. (2005). ‘‘I am us’’: Negative stereotypes as manuscript, Yale University, New Haven, CT.
collective threats. Journal of Personality and Social Psychology, Walton, G.M., & Cohen, G.L. (2007b). A question of belonging: Race,
89, 566–582. social fit, and achievement. Journal of Personality and Social
Cohen, G.L., Garcia, J., Apfel, N., & Master, A. (2006). Reducing the Psychology, 92, 82–96.
racial achievement gap: A social-psychological intervention. Wilson, T.D., Damiani, M., & Shelton, N. (2002). Improving the aca-
Science, 313, 1307–1310. demic performance of college students with brief attributional
Cohen, G.L., Sherman, D.K., Bastardi, A., Hsu, L., McGoey, M., & Ross, interventions. In J. Aronson (Ed.), Improving academic achieve-
L. (2007). Bridging the partisan divide: Self-affirmation reduces ment: Impact of psychological factors on education. San Diego, CA:
ideological closed-mindedness and inflexibility in negotiation. Academic Press.
Journal of Personality and Social Psychology, 93, 415–430. Zigler, E., & Butterfield, E.C. (1968). Motivational aspects of changes in
Cohen, G.L., & Steele, C.M. (2002). A barrier of mistrust: How negative IQ test performance of culturally deprived nursery school chil-
stereotypes affect cross-race mentoring. In J. Aronson (Ed.), dren. Child Development, 39, 1–14.

Volume 17—Number 6 369