HighScope and the Perry Preschool Experiment: Plan, Do, Review, and Six Decades of Evidence

The HighScope curriculum occupies a singular position in early childhood education: it is the pedagogy at the center of the Perry Preschool Project, a randomized experiment begun in 1962 whose participants have now been followed into their mid-fifties. This review traces the approach from its Piagetian origins through its signature plan-do-review routine, and examines the longitudinal evidence: large long-term effects on high school graduation, earnings, and crime; economists' estimates of a 7–10% annual social return; and emerging intergenerational benefits. It also weighs the honest complications — the early fade-out of IQ gains, small single-site samples from the 1960s, reanalyses suggesting educational benefits concentrated among girls, and modern evaluations in which whole-child curricula, HighScope included, show more modest advantages in ordinary classrooms. The review closes with practical implications for families weighing what the Perry evidence does, and does not, promise.

In 1962, David P. Weikart, a young psychologist in the public school district of Ypsilanti, Michigan, confronted a pattern he found intolerable: children from the city's poorest neighborhoods, most of them African American, were failing in school at predictable rates, and the district's response was to wait for the failure and then label it. Weikart's answer was preventive rather than remedial — a preschool operated out of Perry Elementary School for three- and four-year-olds judged at high risk of school failure. The program itself was intensive but not exotic: daily 2½-hour classes taught by certified teachers, small groups, and a weekly home visit to each family (Schweinhart, 2005). What made it historic was its design. Weikart and his colleagues identified 123 children and randomly assigned 58 to the preschool program and 65 to a no-program control group, creating one of the first true experiments in early education — and, through follow-ups at ages 14, 15, 19, 27, 40, and most recently the mid-fifties, one of the longest-running randomized trials in all of social science (Schweinhart et al., 2005; García, Heckman, & Ronda, 2023).

The teaching approach developed inside that experiment was first called the Cognitively Oriented Curriculum, reflecting its debt to Jean Piaget's account of the child as an active constructor of knowledge; the broader tradition also drew on John Dewey's progressive education and, later, Vygotskian ideas about adult scaffolding. In 1970 Weikart left the school district to found the independent HighScope Educational Research Foundation in Ypsilanti, which has developed, disseminated, and studied the curriculum ever since (Schweinhart, 2005). The model's signature is the plan-do-review sequence: each day, children state or draw an intention ("I'm going to build a garage with the big blocks"), carry it out during an extended work time, and then reflect with a teacher and peers on what they did and what happened. Around this routine, teachers organize the classroom into interest areas and plan for "key experiences" — later reformulated as key developmental indicators — spanning personal initiative, social relations, creative representation, movement and music, logic and mathematics, and language and literacy (Hohmann & Weikart, 2002). The adult's role is deliberately double-sided: not a lecturer, but not a bystander either — an observer and conversational partner who extends the child's self-chosen work. Readers familiar with Montessori practice will recognize the family resemblance: a prepared environment, long stretches of self-directed activity, and an adult who follows the child with intention.

The experiment and the famous fade-out

The Perry program's first results looked like a triumph of the most conventional kind: large gains on intelligence tests. At age five, 67% of program children scored an IQ of 90 or above, against 28% of the control group (Schweinhart, 2005). Then the gains eroded. By the early elementary grades the two groups' IQ scores had converged, a pattern soon observed across compensatory preschool programs and enshrined in policy debates as "fade-out." For a generation of skeptics, that convergence was the story: preschool produced a temporary boost on tests and nothing more.

The Perry data ultimately dismantled that conclusion, but the fade-out itself is real and deserves honest treatment. Duncan and Magnuson (2013), reviewing dozens of program evaluations, confirm the general pattern — cognitive test impacts of early education programs commonly shrink or disappear within a few years — while also documenting the paradox that makes Perry famous: in the handful of studies with very long follow-ups, adult outcomes such as educational attainment, earnings, and crime nonetheless show lasting effects. Schweinhart's (2005) own interpretation is that multi-age IQ tests behaved more like achievement tests of what children had recently experienced, so scores tracked the stimulation of the current environment rather than a permanently altered ability. Whatever the mechanism, the honest summary is that Perry did not permanently raise measured intelligence in childhood — and that this turned out to matter far less than anyone expected.

Lifetime effects: the follow-ups at 19, 27, and 40

The project's monograph series traced the two groups with extraordinary tenacity — across all measures and waves, only 6% of data were missing (Schweinhart, 2005). By age 19, the program group showed better scholastic placement and achievement, less delinquency, and higher rates of high school graduation and employment (Berrueta-Clement et al., 1984). By age 27 the differences had extended to earnings, home ownership, and arrests. The age-40 follow-up, which reached 97% of surviving participants, reported the numbers that made the study a fixture of policy debate: 65% of the program group had graduated from regular high school versus 45% of the no-program group — a difference driven overwhelmingly by women (84% vs. 32%); 76% versus 62% were employed; median annual earnings were $20,800 versus $15,300; 36% versus 55% had been arrested five or more times; and 28% versus 52% had ever been sentenced to prison or jail (Schweinhart et al., 2005).

The educational effect for girls appears to have run through a specific institutional channel: program girls were far less likely to be placed in classes for "mental impairment" (8% vs. 36%) or held back a grade, and children not shunted onto those tracks were far more likely to finish school (Schweinhart, 2005). For boys, the program's largest effects appeared instead in employment and, above all, in crime. This asymmetry — schooling effects concentrated in girls, crime reduction concentrated in boys — recurs throughout the literature and, as discussed below, survives the most demanding statistical reanalysis only in part.

The economists' verdict

Perry's second life began when economists took over the ledger. Belfield, Nores, Barnett, and Schweinhart (2006), using the age-40 data, estimated that the program returned $244,812 per participant on a cost of $15,166 (constant 2000 dollars, 3% discount rate) — roughly $16 per dollar invested, of which $12.90 accrued to the general public, 88% of it from crime savings. Those headline ratios deserve their fame and their footnotes in equal measure. When James Heckman and colleagues subjected the data to a stricter accounting — correcting for compromises in the original randomization protocol (siblings assigned together, some children of working mothers swapped between groups), computing proper standard errors, and using conservative assumptions about the social cost of crime — the estimated annual social rate of return fell to roughly 7–10%, substantially below earlier claims but still statistically significant for both males and females and above the historical return on stock market equity (Heckman, Moon, Pinto, Savelyev, & Yavitz, 2010). A 7–10% annual return on a public investment is, by any normal standard, an excellent investment; it is simply not the miracle sometimes advertised.

The more scientifically interesting question is how a two-year, half-day preschool could change adult lives at all, given that the IQ effects faded. Heckman, Pinto, and Savelyev (2013), decomposing the experimental data, found that a sizable portion of the adult treatment effects ran through experimentally induced changes in personality and behavioral skills — reductions in externalizing behavior (aggression, misconduct) and, for girls, improvements in academic motivation — rather than through lasting cognitive gains. On this reading, the Perry classroom's daily regimen of planning, choosing, persisting, and reviewing was quietly training precisely the self-regulatory and social capacities that decades later showed up as fewer arrests and steadier work.

The most recent chapter extends the ledger across generations. García, Heckman, and Ronda (2023), analyzing newly collected data on participants in their mid-fifties and on their children, report long-lasting program effects on cognition at age 54 — a finding the authors explicitly frame against popular fade-out narratives — along with improvements in marital stability, health, and earnings. The children of program participants, by then adults in their late twenties, showed higher education and employment, less criminal activity, and better health than children of controls. The effects of a preschool attended in the Kennedy administration are, on this evidence, measurable in people born decades after it closed.

Does the curriculum itself matter? The Curriculum Comparison study

Perry established that a high-quality participatory preschool could change lives; it could not say whether the HighScope curriculum in particular deserved the credit. To probe that question, Weikart launched a second randomized study in 1967, assigning 68 children living in poverty to one of three well-implemented models: HighScope; a traditional nursery school organized around loosely structured play; and a scripted Direct Instruction program in which teachers drilled academic skills and rewarded correct answers (Schweinhart & Weikart, 1997). For a decade the horse race looked like a tie — all three groups posted an average 27-point IQ rise during the preschool years, and no meaningful differences in school achievement emerged — with, if anything, the Direct Instruction group ending preschool with the highest IQ scores (Schweinhart, 2005).

The groups diverged where few had thought to look: community behavior in adulthood. By age 23, the HighScope and nursery school groups each held significant advantages over the Direct Instruction group, and did not differ significantly from each other on any outcome. Only 6% of either child-initiated group had been treated for emotional impairment or disturbance during schooling, versus 47% of the Direct Instruction group; 10% of the HighScope group had ever been arrested for a felony, versus 39% of the Direct Instruction group (Schweinhart & Weikart, 1997). The authors' conclusion — that scripted academic instruction at ages three and four may purchase a short-lived test-score advantage at the cost of long-term social development — remains a touchstone in debates over play-based versus didactic preschool.

It also demands caveats, which the honest reader should hold firmly. The sample was tiny — roughly two dozen children per group — so single percentage-point contrasts carry wide uncertainty; the Direct Instruction model tested was the 1960s original, which its developers substantially revised in later decades; and the finding that HighScope did not outperform a good traditional nursery school on any outcome is reported far less often than the finding that both outperformed drill. The study is best read as suggestive evidence that how young children are taught has long-run behavioral consequences, not as proof that one brand of child-initiated curriculum surpasses another.

How strong is the evidence? Criticisms and reanalyses

Because so much policy weight rests on Perry, the study has attracted unusually intense methodological scrutiny — and it is a mark of its quality that the data have been opened to independent reanalysis. The results counsel calibrated confidence rather than either dismissal or reverence. Anderson (2008), applying corrections for the multiple statistical comparisons that longitudinal studies inevitably accumulate, reanalyzed Perry alongside the Abecedarian and Early Training projects and found that the statistically robust long-term benefits were concentrated among girls, with little evidence of significant long-term effects for boys once multiplicity was accounted for. Heckman and colleagues (2010), using different methods that model the randomization compromises directly, reached a more favorable verdict — significant returns for both sexes — but at rates well below the popularized figures. The two reanalyses disagree in emphasis, and that disagreement is itself informative: with 123 subjects, conclusions about subgroups sit near the edge of what statistics can resolve.

Generalizability is the second standing question. The Perry sample was drawn from one Midwestern city's low-income African American population in the early 1960s; the program employed certified public school teachers serving five to six children each, ran for two school years, and visited every family weekly. Schweinhart (2005) himself is unusually candid that the findings apply only to "reasonably similar" programs and explicitly do not apply to typical Head Start or state preschool programs as actually staffed and funded. Duncan and Magnuson (2013) add the sobering observation that modern programs evaluated at scale generally show smaller initial effects than Perry did. The strongest legitimate reading of Perry, then, is not "preschool works" but "a specific, expensive, well-implemented preschool worked for a specific disadvantaged population" — a narrower claim, though still among the most consequential in education research.

The modern curriculum in ordinary classrooms

What of HighScope as it exists today, in thousands of classrooms far from Ypsilanti? Here the evidence is thinner and more mixed than the Perry halo suggests, and the contrast deserves plain statement. In the national Head Start Family and Child Experiences Survey of 2000, 20% of teachers reported using HighScope — second only to the Creative Curriculum's 39% — and HighScope classrooms scored significantly higher on observed classroom quality than classrooms using other curricula, with children posting larger fall-to-spring gains in letter-word identification and cooperative behavior (Zill et al., 2003). Those results are encouraging but observational: teachers were not randomly assigned to curricula, so better outcomes may reflect better-resourced or better-trained programs choosing HighScope rather than effects of the curriculum itself.

Rigorous modern comparisons are less flattering. Jenkins and colleagues (2018), analyzing data from a federal evaluation in which curricula were experimentally assigned, found that widely used "whole-child" curricula — the family that includes HighScope and Creative Curriculum — produced higher observed classroom process quality but did not improve children's school readiness relative to the varied curricula teachers assembled on their own, whereas curricula targeting specific literacy or mathematics skills did boost achievement in those domains. A systematic review by Chambers, Cheung, and Slavin (2016) points the same direction: comprehensive approaches that balance child-initiated activity with intentional, skill-focused instruction show the clearest benefits for early literacy and language, outperforming purely developmental-constructivist programs. Notably, there are few methodologically strong, large-scale evaluations of the current HighScope curriculum as a stand-alone intervention — a striking gap for so venerable a model, and one its own research foundation would presumably welcome closing.

Research on the model's central mechanism is more supportive, if softer. Epstein (2003), reviewing studies of planning and reflection in early childhood settings, reports that children given more opportunities to plan and review their own activities scored higher on measures of language, literacy, and social development, with converging findings from investigations in the United Kingdom and the Netherlands — evidence that is correlational, drawn from multiple curriculum approaches, but consistent with the developmental logic that daily intention-setting and recall exercise executive and verbal capacities.

International reach and influence

HighScope's influence long ago outgrew its enrollment figures. By the early 1990s the foundation had certified more than a thousand trainers across 34 U.S. states and ten other countries (Epstein, 1993), and the model today operates through licensed institutes on several continents. When the OECD convened authors of the world's most influential early childhood curricula to present their frameworks side by side, HighScope — presented by Weikart himself — stood alongside Reggio Emilia, New Zealand's Te Whāriki, Belgium's Experiential Education, and the Swedish national curriculum (OECD, 2004). More diffusely, practices now unremarkable in preschools everywhere — child-initiated planning time, interest-area room arrangement, a consistent daily routine, small-group recall conversations — carry HighScope's fingerprints, and the Perry findings remain the single most-cited empirical anchor for public investment in early childhood, from Head Start reauthorizations to state pre-K campaigns (Duncan & Magnuson, 2013; Heckman et al., 2010).

Practical implications

For families and educators, several implications follow from this literature. First, the Perry evidence is best understood as a demonstration of what a whole configuration of quality can do: qualified and well-paid teachers, very small groups, two years of half-day programming, a validated participatory curriculum, and genuine partnership with parents through home visits. A program displaying the HighScope name inherits none of those results automatically; as Schweinhart (2005) puts it, "it's not what you say you do but what you actually do that counts." Parents touring any preschool — HighScope, Montessori, or otherwise — should look for the substance behind the label: children making real choices and talking about them, adults who converse rather than merely manage, and a daily rhythm children can predict and narrate.

Second, the mechanism evidence redirects attention from test scores to dispositions. The lasting Perry effects appear to have traveled through motivation, self-regulation, and social behavior rather than through IQ (Heckman, Pinto, & Savelyev, 2013) — a finding that should reassure families anxious about academic drilling at age four and should equally caution programs against a laissez-faire reading of "play-based." The HighScope adult is intentional and intellectually engaged with children's plans; the evidence favors that engaged middle path, and modern reviews suggest supplementing it with deliberate, playful attention to early literacy and mathematics rather than leaving those to chance (Chambers, Cheung, & Slavin, 2016; Jenkins et al., 2018).

Third, expectations should be honest. Perry served deeply disadvantaged children in the 1960s; its dramatic effect sizes are unlikely to reappear for children already growing up with books, conversation, and stability, for whom a good preschool is one productive environment among several rather than a rescue. The study's returns were calculated against the counterfactual of no preschool at all. What advantaged families can reasonably take from this literature is not a promised earnings premium but a validated picture of what developmentally serious early education looks like — and evidence that scripted academic instruction in the preschool years buys nothing durable and may carry social costs (Schweinhart & Weikart, 1997).

Six decades on, the HighScope story is really two stories that must be told together. One is among the most remarkable in social science: a small experiment, faithfully followed for a lifetime, showing that a few hundred hours of participatory early education altered graduation, earnings, crime, health, and even the fortunes of a second generation — and that it did so by shaping character and habits rather than test scores. The other is quieter and more cautionary: a modern curriculum whose advantage over other thoughtful approaches in ordinary classrooms remains modest and under-studied, and whose most famous numbers describe a program more intensive than most schools bearing its name. Both stories are true. Together they suggest that the active ingredient was never a trademark but a set of commitments — children as planners and agents of their own learning, adults as serious intellectual companions, families as partners — and that those commitments, wherever they are genuinely enacted, rest on some of the strongest long-term evidence early education has ever produced.

References

1. Anderson, M. L. (2008). Multiple inference and gender differences in the effects of early intervention: A reevaluation of the Abecedarian, Perry Preschool, and Early Training Projects. Journal of the American Statistical Association, 103(484), 1481–1495. https://doi.org/10.1198/016214508000000841
2. Belfield, C. R., Nores, M., Barnett, S., & Schweinhart, L. (2006). The High/Scope Perry Preschool Program: Cost–benefit analysis using data from the age-40 followup. Journal of Human Resources, 41(1), 162–190. https://jhr.uwpress.org/content/XLI/1/162
3. Berrueta-Clement, J. R., Schweinhart, L. J., Barnett, W. S., Epstein, A. S., & Weikart, D. P. (1984). Changed lives: The effects of the Perry Preschool Program on youths through age 19. Monographs of the High/Scope Educational Research Foundation, 8. Ypsilanti, MI: High/Scope Press. ERIC ED313128. https://eric.ed.gov/?id=ED313128
4. Chambers, B., Cheung, A. C. K., & Slavin, R. E. (2016). Literacy and language outcomes of comprehensive and developmental-constructivist approaches to early childhood education: A systematic review. Educational Research Review, 18, 88–111. https://www.sciencedirect.com/science/article/abs/pii/S1747938X16300148
5. Duncan, G. J., & Magnuson, K. (2013). Investing in preschool programs. Journal of Economic Perspectives, 27(2), 109–132. https://doi.org/10.1257/jep.27.2.109
6. Epstein, A. S. (1993). Training for quality: Improving early childhood programs through systematic inservice training. Ypsilanti, MI: High/Scope Press.
7. Epstein, A. S. (2003). How planning and reflection develop young children's thinking skills. Young Children, 58(4), 28–36. ERIC EJ679112. https://eric.ed.gov/?id=EJ679112
8. García, J. L., Heckman, J. J., & Ronda, V. (2023). The lasting effects of early-childhood education on promoting the skills and social mobility of disadvantaged African Americans and their children. Journal of Political Economy, 131(6). https://doi.org/10.1086/722936
9. Heckman, J. J., Moon, S. H., Pinto, R., Savelyev, P. A., & Yavitz, A. (2010). The rate of return to the HighScope Perry Preschool Program. Journal of Public Economics, 94(1–2), 114–128. https://www.sciencedirect.com/science/article/abs/pii/S0047272709001418
10. Heckman, J., Pinto, R., & Savelyev, P. (2013). Understanding the mechanisms through which an influential early childhood program boosted adult outcomes. American Economic Review, 103(6), 2052–2086. https://doi.org/10.1257/aer.103.6.2052
11. Hohmann, M., & Weikart, D. P. (2002). Educating young children: Active learning practices for preschool and child care programs (2nd ed.). Ypsilanti, MI: High/Scope Press. ERIC ED465451. https://eric.ed.gov/?id=ED465451
12. Jenkins, J. M., Duncan, G. J., Auger, A., Bitler, M., Domina, T., & Burchinal, M. (2018). Boosting school readiness: Should preschool teachers target skills or the whole child? Economics of Education Review, 65, 107–125. https://doi.org/10.1016/j.econedurev.2018.05.001
13. OECD. (2004). Starting Strong: Curricula and pedagogies in early childhood education and care — Five curriculum outlines. Paris: OECD Directorate for Education. https://childcarecanada.org/documents/research-policy-practice/04/05/starting-strong-curricula-and-pedagogies-early-childhood
14. Schweinhart, L. J. (2005). The High/Scope Perry Preschool Study through age 40: Summary, conclusions, and frequently asked questions. Ypsilanti, MI: High/Scope Educational Research Foundation. https://highscope.org/wp-content/uploads/2024/07/perry-preschool-summary-40.pdf
15. Schweinhart, L. J., Montie, J., Xiang, Z., Barnett, W. S., Belfield, C. R., & Nores, M. (2005). Lifetime effects: The High/Scope Perry Preschool study through age 40. Monographs of the High/Scope Educational Research Foundation, 14. Ypsilanti, MI: High/Scope Press.
16. Schweinhart, L. J., & Weikart, D. P. (1997). The High/Scope Preschool Curriculum Comparison study through age 23. Early Childhood Research Quarterly, 12(2), 117–143. https://www.sciencedirect.com/science/article/pii/S0885200697900090
17. Zill, N., Resnick, G., Kim, K., O'Donnell, K., Sorongon, A., McKey, R. H., et al. (2003). Head Start FACES 2000: A whole-child perspective on program performance — Fourth progress report. Washington, DC: U.S. Administration for Children and Families. ERIC ED478791. https://eric.ed.gov/?id=ED478791