The Power of Representation: How Identity Visibility Shapes Educational Trajectories

The Power of Representation: How Identity Visibility Shapes Educational Trajectories

What the research shows about role models, identity-affirming classrooms, and restricted representation, across race and ethnicity, gender, and language

By Lindsey Kundel, Editor in Chief, InGenius Prep

Every student, at some point, scans the room.

Not for exits. For evidence. Who teaches the advanced class. Whose names hang on the wall of past winners. Whose language is spoken in the hallway without apology, and whose is spoken only at home. Long before anyone asks a teenager what they want to study, the room has already offered a quiet answer to a different question: who belongs here?

Families often treat representation as a matter of values, something a school either cares about or doesn’t. The research treats it as something more measurable. Over the past two decades, economists, psychologists, and education researchers have tested three things. What happens when students see people like themselves in positions of expertise. What happens when classrooms affirm, rather than erase, parts of who students are. And what happens when that visibility is restricted by law or custom. Some of the findings are strong. Some are contested. A few are too new to judge.

This paper lays out what the evidence shows, where it is thin, and what families and educators can reasonably do with it.

Representation is not a feeling a school produces. It is a set of conditions a student encounters, and those conditions shape decisions years before an application is written.

A Note on Scope

Identity is broad, and the laws and pressures discussed here touch many identities at once. This paper focuses on three: race and ethnicity, gender, and language. These are the areas where we found the clearest research, including randomized experiments and long-run studies.

Other dimensions of identity matter just as much to students’ lives: sexual orientation and gender identity, disability, religion, first-generation status, socioeconomic background. Covering them responsibly is beyond the scope of this paper until we can build on stronger sources. We intend to return to them.

What Representation Theory Says

“Representation” gets used loosely, so it helps to separate three layers.

Cultural representation is what a curriculum, a library, or a textbook tells students about whose stories count. Academic representation is who students see doing the work: teaching the class, presenting the research, winning the competition. Institutional representation is what a school or system formally recognizes: whom it hires, what it measures, what its policies name.

Three ideas from the research explain why these layers matter.

The first is the oldest. In Albert Bandura’s theory of self-efficacy, people judge what they can do partly through vicarious experience: watching someone similar to themselves succeed. Seeing a peer master something is evidence that it can be mastered.

The second is the stereotype inoculation model, developed by social psychologist Nilanjana Dasgupta. It proposes that contact with same-sex experts in STEM settings strengthens women’s self-concept in STEM, their attitudes toward it, and their motivation to pursue STEM careers. Those experts can be advanced peers, professionals, or professors. Notably, in her studies women’s self-concept benefited from contact with female experts even while negative stereotypes about their gender stayed active. The role model doesn’t erase the stereotype. It gives the student a counterexample to hold against it.

The third comes from economics. Researchers studying teacher race have formally defined role-model effects as the provision of information. A role model tells a student, concretely, that a path is open to someone like them.

The institutional layer is where these ideas meet numbers. In 2020–21, 80 percent of U.S. public school teachers were White, compared with 46 percent of students. Nine percent of teachers were Hispanic, compared with 28 percent of students. That gap says nothing about any single classroom. It does tell us that many students go years without the academic representation the research below examines.

Visible Role Models and Student Self-Efficacy

The strongest evidence on role models comes from randomized studies. Randomization lets researchers separate the effect of the role model from the characteristics of families who seek one out.

Race and ethnicity: the long-run effects of same-race teachers

The most cited study here draws on Tennessee’s STAR class-size experiment, which randomly assigned young children to classrooms. Black students randomly assigned to at least one Black teacher in grades K–3 were 9 percentage points (13 percent) more likely to graduate from high school. They were also 6 percentage points (19 percent) more likely to enroll in college than Black schoolmates who were not. Black teachers had no significant long-run effects on White students. The college effects came mainly through two-year colleges and were concentrated among disadvantaged boys. North Carolina administrative data produced similar findings.

Two details matter for families. First, these were teachers students had between kindergarten and third grade, and the effects surfaced more than a decade later. Second, the authors are careful about why. They could not definitively identify the mechanism. They did find suggestive evidence that the effects were larger when Black teachers were the same sex as their students, which is consistent with role modeling as one of several channels.

Gender: brief exposure, durable choices

Two field experiments show how little exposure it can take.

At a U.S. university, economists Catherine Porter and Danila Serra brought successful women who had majored in economics into introductory economics classes. Female students in those classes became 8 percentage points more likely to major in economics. That is a large effect, given that only 9 percent of women majored in economics at baseline. Each visit was a 15-minute discussion of the role model’s experience.

In France, a larger experiment reached high school students directly. Fifty-six female scientists made randomly assigned classroom visits to about 20,000 students in the Paris region. The share of Grade 12 girls enrolling in selective, male-dominated STEM programs rose from 11 to 14.5 percent. The effects were driven by girls who were already high achievers in mathematics.

The French study also produced one of the most useful findings for anyone designing a program. The most effective visits improved students’ perceptions of STEM careers without overemphasizing women’s under-representation in science. The role model works best as a window onto the work itself, not as a reminder of the gap.

Across these studies, the pattern holds: seeing someone like you succeed in a field changes what you believe is possible, and those beliefs show up later in enrollment and in the choice of major.

The limits deserve equal clarity. The effects are concentrated in particular groups: disadvantaged boys in the teacher study, high-achieving girls in the French study. They are averages, not guarantees. And they speak more to which paths students choose than to how students perform once they are on them.

How Identity-Affirming Environments Affect Academic Performance

Role models are one form of representation. The environment is another. Curriculum, language of instruction, and classroom culture can affirm parts of a student’s identity or ask the student to set them aside.

Ethnic studies: promising, and contested

In San Francisco, researchers studied a ninth-grade ethnic studies course using a regression-discontinuity design, which compares students just above and just below an assignment cutoff. The preregistered study examined the course’s longer-run effects. It was described as the first quantitative evidence of ethnic studies’ longer-run academic impact. The researchers also noted that the course had been developed over several years and was taught by well-trained teachers. Both conditions limit how far the results generalize.

The findings have been challenged. Two critics, one of them a University of Pennsylvania data scientist, argued in a published essay that the study design was too flawed to support any conclusion about the course, positive or negative. Families and educators should hold both the findings and the critique. This is a live methodological debate, not a settled result.

Language: the clearest causal evidence

The strongest evidence on identity-affirming environments comes from language. In Portland, Oregon, seats in dual-language immersion programs are assigned by lottery. In these programs, students learn academic content in English and a partner language. The lottery creates a natural randomized experiment.

Across seven cohorts of lottery applicants, researchers found positive effects on English reading in fifth and eighth grades. The gains were 13 to 22 percent of a standard deviation, equivalent to 7 to 9 months of learning. There was little benefit in math or science, and no harm. By sixth and seventh grade, lottery winners were 3 to 4 percentage points less likely to still be classified as English learners. That effect was stronger for students whose home language matched the partner language.

That last finding matters. For English learners, learning partly in their home language did not slow their progress toward English proficiency. It sped it up.

Māori-medium schooling

New Zealand offers a system-level example, discussed in the international section below. Students in kaupapa Māori schools, which are grounded in te reo Māori and Māori philosophy, have recently posted higher national qualification rates than students in comparable English-medium schools. That evidence is descriptive rather than experimental, a limit addressed below.

The Impact of Suppressed Representation

If representation can shape trajectories, so can its absence, including absence created deliberately by law or by cultural pressure. This section describes what such policies do and what research can and cannot yet say about their effects. It takes no position on the policies themselves.

English-only instruction in California

The clearest historical case involves language. California’s Proposition 227, passed in 1998, required that English learners be taught in English. Students received about one year of intensive English instruction before moving into mainstream classes. Enrollment in a bilingual program required a signed parental waiver. The measure affected roughly 1.4 million English learners.

Research on its academic effects is mixed. A long-term appraisal by Laura McCloskey, Kenji Hakuta, and colleagues found a localized negative impact on English learners, most visible in eighth grade. Sixth graders, by contrast, showed more relative progress. A later analysis from New America concluded that the measure did little to change academic outcomes. It did, however, sharply reduce the number of bilingual teachers and programs. By 2008, only 5 percent of bilingual programs remained. In 2016, California voters approved Proposition 58, which repealed key provisions of Proposition 227 and allowed non-English instruction again without parental waivers.

What this case shows most clearly is not a test-score effect. It is an infrastructure effect. A law can remove the programs and teachers that make a form of representation possible, and rebuilding them takes years after the law changes.

Restrictions on how race and sex are discussed in classrooms

Since 2021, many U.S. states have considered laws limiting how teachers discuss race and sex. An Education Week analysis found that 44 states had introduced bills or taken other steps to restrict how teachers discuss racism and sexism since January 2021. Twenty states were imposing such restrictions through legislation or other means. Trackers differ in what they count. A UCLA School of Law project that also includes changes to DEI policy counted 27 states that had passed measures.

Supporters and opponents disagree sharply about what these laws prohibit and why. What we can say plainly is that we have found no rigorous research measuring their effects on student outcomes. The laws are recent, they vary widely in scope and enforcement, and outcome data take years to accumulate. Claims about their academic effects, in either direction, are currently ahead of the evidence.

Representation that cannot be counted: France

Cultural and legal traditions can limit representation without any classroom rule. France’s republican model holds that the state should deal with individuals rather than communities or groups, to give everyone equal treatment. A 1978 data protection law prohibits collecting or processing personal data that reveals racial or ethnic origin, or religion. A 2007 Constitutional Council decision further limits how studies of diversity can be conducted. French official statistics do record country of birth and nationality.

The consequence for education research is practical. A system that does not record ethnicity cannot easily measure ethnic gaps, or test whether an intervention closes them. It is telling that France’s well-known role-model experiment targeted gender, a category French data do record.

What is not counted is hard to study, and what is not studied is hard to improve. That holds wherever one stands on the policy itself.

International Comparisons of Identity Support Models

No country has solved this. Comparing approaches, though, clarifies the choices.

New Zealand: language as foundation. In 2023, 1,857 students were enrolled in Years 11 to 13 across 64 kaupapa Māori senior secondary settings. Eighty-nine percent of those kura serve students who face more socio-economic barriers to attainment. A New Zealand Qualifications Authority briefing found their NCEA and University Entrance rates higher than those of all students in English-medium schools with similar barriers. The difference was largest at NCEA Level 3 and University Entrance. About 80 percent of the credits these students earned came from the national curriculum, and 20 percent from the Māori-medium curriculum. These settings affirm identity while still preparing students for the national qualification.

The caveat is real. The Ministry of Education has itself noted that the small number of Māori-medium candidates makes comparisons difficult and sometimes misleading. Families who choose these schools may also differ from those who don’t.

Brazil: curriculum mandates and university quotas. Brazil has pursued representation through law at two levels.

In schools, Law 10.639, signed in 2003, made teaching African and Afro-Brazilian history and culture mandatory in primary and secondary education. A 2008 amendment extended the requirement to Indigenous history and culture. Implementation research repeatedly documents a gap between mandate and classroom practice.

In universities, Law 12.711/2012 reserves half of the seats in federal universities and institutes for public school students. A proportion of those seats is designated for Black, brown, Indigenous, and disabled students. The law was revised in 2023. Evidence on outcomes is mixed and mostly comes from single institutions. One study at the Federal University of Pernambuco found that quota students did better on dropout rates and on-time graduation, while non-quota students earned higher GPAs and failed fewer courses. A ten-year review of the research warned that admission policies without retention support produce inclusion in name only.

France: universalism with targeted exceptions. As described above, France avoids ethnic categories in public statistics and policy. It still uses targeted programs, such as the female role-model visits, where a category like gender is recognized.

United States: decentralization. In the U.S., most representation policy is set by states and districts. That is how California could mandate English-only instruction and then reverse it, and why classroom restrictions vary from state to state. The result is wide variation in what a student encounters depending on where they live.

Set side by side, the most promising models pair recognition with rigor and support. New Zealand’s kura still teach mostly to the national curriculum, and Brazil’s own researchers argue that quotas need retention support. The clearest limits appear where mandates arrive without the teachers, training, and follow-through to carry them out.

Where Admissions Fits

Most representation research concerns the years before college. It connects to applications in one concrete place.

On June 29, 2023, the Supreme Court held race-conscious admissions unconstitutional in Students for Fair Admissions v. Harvard. In the same opinion, Chief Justice Roberts wrote that nothing in it prohibits universities from considering an applicant’s discussion of how race affected their life. He added that universities may not use essays to reinstate the system the Court struck down. The discussion must be concretely tied to a quality of character or unique ability the applicant would bring. Several colleges revised their essay questions in response; Harvard, for example, replaced its optional open-topic essay with five required 200-word questions.

For families, the practical consequence is this. Identity now enters selective admissions mainly through the student’s own account of their experience. The reflection those essays ask for — who a student has seen, what they have carried, what they have built from it — is the same reflection this paper describes. Our earlier Journal piece, Crafting Competitive Narratives Post-Harvard v. SFFA, covers how students approach those essays.

Recommendations for Families, Teachers, and Schools

These recommendations stay close to what the evidence supports.

For families

Look for role models in the field, not just in general. In both the U.S. and French experiments, short, field-specific exposure changed what students chose to study. Practical versions include an alumna who works in the field your child is curious about, a university lab open day, or a conversation with a working professional.

Keep the home language. In the Portland lottery study, learning partly in a home language did not cost English learners progress in English, and it helped many. For multilingual families, maintaining the home language is not a trade-off against English.

Ask what your child sees. “Do you feel represented?” is hard to answer. “Who teaches the advanced classes?” and “Who’s on the team?” are easier, and the answers are often revealing.

Talk about the work, not only the gap. The most effective French role models focused on what the careers are actually like, not on under-representation itself.

For teachers

Bring practitioners into the room. Fifteen minutes with someone doing the work moved major choices in the economics study.

Treat implementation as the intervention. San Francisco’s course was built over years and taught by trained teachers. Brazil’s curriculum mandate shows what happens when training doesn’t follow.

Know what your state requires. Where classroom restrictions exist, read the actual text and your district’s guidance before changing instruction. Education Week has described these laws as a source of real uncertainty for teachers.

For schools

Measure what you want to improve. France’s experience shows how hard it is to address gaps a system cannot see.

Treat teacher diversity as a pipeline question. The gap between who teaches and who learns is national, and it will not close classroom by classroom.

Pair access with support. Brazil’s quota research points to retention as the difference between access and outcomes.

Evaluate rigorously. Portland’s lottery made it possible to learn what immersion actually does. Schools that assign seats by lottery, or roll programs out in stages, can learn as well.

The Room, Revisited

The student scanning the room is not asking for flattery. They are gathering evidence about what is possible.

The research does not say representation determines outcomes. It says representation changes the information students act on, and that the effects can last for years: a teacher in third grade, a fifteen-minute visit, a classroom where a home language is an asset rather than an obstacle. It also says that removing these conditions has consequences that take years to undo.

Every room a student enters makes an argument about who belongs. The adults who design those rooms get to decide what argument it makes.

Frequently Asked Questions

What does representation mean in education?

It describes whether students see people and experiences like their own in what they learn, who teaches them, and what their school recognizes. Research separates cultural, academic, and institutional layers.

Does having a same-race teacher affect students?

In a Tennessee experiment, Black students randomly assigned to a Black teacher in K–3 were more likely to graduate from high school and enroll in college. No long-run effects were found for White students.

Do role models really change what students study?

Two randomized studies found that brief visits from women in economics or science increased the share of female students choosing those fields. The effects were concentrated among particular groups, such as high-achieving girls in math.

Does bilingual education slow English learning?

Lottery-based evidence from Portland suggests not. Students in dual-language immersion gained 7 to 9 months in English reading by middle school and were reclassified out of English-learner status sooner.

Can students write about identity in college essays after SFFA?

Yes. The Supreme Court’s 2023 decision allows colleges to consider an applicant’s discussion of how race affected their life, as long as it is tied to the applicant’s character or abilities.

Sources

1. Bandura, A. “Self-Efficacy: Toward a Unifying Theory of Behavioral Change.” Psychological Review, 84(2), 1977. https://doi.org/10.1037/0033-295X.84.2.191
2. Stout, J.G., Dasgupta, N., Hunsinger, M., and McManus, M.A. “STEMing the Tide: Using Ingroup Experts to Inoculate Women’s Self-Concept in Science, Technology, Engineering, and Mathematics (STEM).” Journal of Personality and Social Psychology, 100(2), 2011. https://doi.org/10.1037/a0021385
3. Gershenson, S., Hart, C.M.D., Hyman, J., Lindsay, C.A., and Papageorge, N.W. “The Long-Run Impacts of Same-Race Teachers.” American Economic Journal: Economic Policy, 14(4), 2022. https://www.aeaweb.org/doi/10.1257/pol.20190573
4. Gershenson, S., et al. Data and Code for: The Long-Run Impacts of Same-Race Teachers. openICPSR, 2022. https://www.openicpsr.org/openicpsr/project/145941/version/V1/view
5. National Center for Education Statistics. “Characteristics of Public School Teachers.” Condition of Education. https://nces.ed.gov/programs/coe/indicator/clr
6. Porter, C., and Serra, D. “Gender Differences in the Choice of Major: The Importance of Female Role Models.” American Economic Journal: Applied Economics, 12(3), 2020. https://www.aeaweb.org/doi/10.1257/app.20180426
7. Breda, T., Grenet, J., Monnet, M., and Van Effenterre, C. “How Effective Are Female Role Models in Steering Girls Towards STEM? Evidence from French High Schools.” The Economic Journal, 133(653), 2023. https://doi.org/10.1093/ej/uead019
8. Bonilla, S., Dee, T.S., and Penner, E.K. “Ethnic Studies Increases Longer-Run Academic Engagement and Attainment.” Proceedings of the National Academy of Sciences, 118(37), 2021. https://www.pnas.org/doi/10.1073/pnas.2026386118
9. EdSource. “Research Finds Ethnic Studies in San Francisco Had Enduring Impact.” September 7, 2021. https://edsource.org/2021/research-finds-ethnic-studies-in-san-francisco-had-enduring-impact/660856
10. Tablet. “Studies Fail to Support Claims for New California Ethnic Studies Requirement.” https://www.tabletmag.com/sections/news/articles/studies-fail-to-support-claims-new-california-ethnic-studies-requirement
11. Steele, J.L., Slater, R.O., Zamarro, G., Miller, T., Li, J., Burkhauser, S., and Bacon, M. “Effects of Dual-Language Immersion Programs on Student Achievement: Evidence from Lottery Data.” American Educational Research Journal, 54(1S), 2017. https://doi.org/10.3102/0002831216634463
12. CA Voter Guide. “Proposition 58: English Proficiency (2016).” https://cavotes.org/ballot-measure/2016-prop-58/
13. McCloskey, L., Pellegrin, N., Thompson, K., and Hakuta, K. Proposition 227 in California: A Long-Term Appraisal of Its Impact on Language Minority Student Achievement. The Civil Rights Project, UCLA. https://civilrightsproject.ucla.edu/reports/proposition-227-in-california-a-long-term-appraisal-of-its-impact-on-language-minority-student-achievement
14. New America. “California’s Proposition 227’s Legacy on Bilingual Education.” https://www.newamerica.org/insights/shifting-collective-values/
15. Schwartz, S. “Map: Where Critical Race Theory Is Under Attack.” Education Week, June 11, 2021 (updated 2025). https://www.edweek.org/policy-politics/map-where-critical-race-theory-is-under-attack/2021/06
16. “Over Half of States Have Passed Measures to Stop CRT, Equity Efforts in Schools, Report Says” (reporting on UCLA School of Law’s CRT Forward Tracking Project). https://okcfox.com/news/nation-world/over-half-of-states-have-passed-measures-to-stop-crt-equity-efforts-in-schools-report-says
17. The Conversation. “How French Law Makes Minorities Invisible.” https://theconversation.com/how-french-law-makes-minorities-invisible-66723
18. INSEE. “Ethnic-Based Statistics.” https://www.insee.fr/en/information/2388586
19. New Zealand Qualifications Authority. Aide-Memoire: NCEA in Kaupapa Māori Senior Secondary Settings (OC00816), 2024. https://www2.nzqa.govt.nz/assets/About-us/Official-releases/2024-2025/NCEA-in-Kaupapa-Maori-Senior-Secondary-settings-OC00816.pdf
20. Te Ao Māori News. “Kaupapa Māori Students More Likely to Get NCEA Merit and Excellence Endorsements.” October 16, 2024. https://www.teaonews.co.nz/2024/10/16/kaupapa-maori-students-more-likely-to-get-ncea-merit-and-excellence-endorsements/
21. Murray, S. Achievement at Māori Immersion and Bilingual Schools. New Zealand Ministry of Education, Education Counts. https://www.educationcounts.govt.nz/publications/series/11240/maori_immersion_2004
22. Universidade Estadual Paulista (UNESP). A Educação das Relações Étnico-Raciais na Escola: Análise de Políticas Públicas Educacionais e o Enraizamento da Lei 10.639/03. 2018. https://repositorio.unesp.br/handle/11449/154018
23. Universidade Federal de Uberlândia. As Tramas do Racismo à Brasileira: O Ensino da História e Cultura Afro-Brasileira nas Instituições Escolares de Itapagipe/MG (2003–2006). https://repositorio.ufu.br/handle/123456789/18766
24. Revista Educação e Fronteiras, 12, 2023. Study of Law 12.711/2012 at UFGD. https://ojs.ufgd.edu.br/educacao/article/download/17402/9634
25. Revista Brasileira de Política e Administração da Educação. “Os 10 Anos da Lei de Cotas no Ensino Superior Brasileiro.” https://seer.ufrgs.br/index.php/rbpae/article/download/137592/95346/656139
26. Albuquerque, F.L.C., et al. Performance evaluation of quota and non-quota students at UFPE. IPEA. https://repositorio.ipea.gov.br/bitstreams/b333dc79-6dfe-4ceb-8df5-3759ce04eb16/download
27. Stanford Law School. “Students for Fair Admissions v. Harvard FAQ: Navigating the Evolving Implications of the Court’s Ruling.” December 12, 2023. https://law.stanford.edu/2023/12/12/students-for-fair-admissions-v-harvard-faq-navigating-the-evolving-implications-of-the-courts-ruling/
28. Faegre Drinker. “Supreme Court Decides Students for Fair Admissions, Inc. v. President and Fellows of Harvard College.” June 29, 2023. https://www.faegredrinker.com/en/insights/publications/2023/6/supreme-court-decides-students-for-fair-admissions-inc-v-president-and-fellows-of-harvard-college-and-students-for-fair-admissions-inc-v-university-of-north-carolina-et-al
29. Christian Science Monitor. “Why the College Essay May Never Be the Same.” August 22, 2023. https://www.csmonitor.com/USA/Education/2023/0822/Why-the-college-essay-may-never-be-the-same
30. InGenius Prep. “Crafting Competitive Narratives Post-Harvard v SFFA.” InGenius Journals, August 29, 2025. https://ingeniusprep.com/journal/harvard-affirmative-action-storytelling-admissions/