The Montessori Method: A Century of Evidence from the Casa dei Bambini to the Randomized Trial

The Montessori method, developed by the Italian physician Maria Montessori beginning in 1907, organizes early education around self-directed activity in a prepared environment: mixed-age classrooms, freedom of choice among specially designed self-correcting materials, long uninterrupted work periods, and a teacher who observes and guides rather than instructs. Once evaluated mainly through weak observational designs, Montessori education has in the past two decades accumulated an unusually strong evidence base for a named pedagogy, including lottery-based natural experiments, preregistered randomized trials in France, a national randomized controlled trial in the United States, and a Campbell Collaboration meta-analysis. This review traces the method's origins and diffusion, describes its defining practices, and weighs the evidence on academic achievement, executive function, social development, motivation, and creativity. The overall pattern favors Montessori education, particularly in the early years and at high implementation fidelity, though null findings exist, effects vary across populations and adaptations, and the mechanisms responsible remain genuinely contested. The review closes with practical implications for families and educators.

On January 6, 1907, a physician named Maria Montessori opened a room for some fifty young children of working parents in a tenement block in San Lorenzo, then one of the poorest quarters of Rome. She called it the Casa dei Bambini, the Children's House. Montessori, born in 1870 in Chiaravalle, had earned her medical degree at the University of Rome in 1896 as one of the first women in Italy to do so, and had come to education by an unusual route: clinical work with institutionalized children with developmental disabilities, and through it the writings of the French physicians Jean Itard and Édouard Séguin, whose sensory training materials she adapted and extended (Kramer, 1976). In San Lorenzo she furnished the room with child-sized tables and chairs, gave the children access to her didactic materials, and watched. What she reported observing became the founding narrative of her movement: young children, given freedom to choose real work and repeat it undisturbed, displayed spontaneous concentration, self-discipline, and an appetite for order and mastery that the adults around them had not suspected (Montessori, 1912).

The method spread with extraordinary speed. Her book Il Metodo appeared in English in 1912 as The Montessori Method and ran quickly through multiple printings; the first American Montessori school opened in Tarrytown, New York, in 1911, and enthusiastic magazine coverage briefly made Montessori a household name in the United States. The first American wave then collapsed almost as quickly, undone in part by the influential 1914 critique of William Heard Kilpatrick, a prominent disciple of John Dewey, who judged the method scientifically outdated, and by Montessori's own insistence on controlling training and materials (Whitescarver & Cossentino, 2008). The movement's center of gravity shifted to Europe and, during Montessori's wartime exile in India in the 1940s, to South Asia; she and her son Mario had founded the Association Montessori Internationale (AMI) in 1929 to safeguard the pedagogy, and she died in the Netherlands in 1952 (Kramer, 1976). A durable American revival began in the late 1950s, when Nancy McCormick Rambusch's work led to the Whitby School in Connecticut and, in 1960, to the American Montessori Society, which adapted teacher training to American norms (Whitescarver & Cossentino, 2008). Today Montessori is by a wide margin the most diffused named alternative pedagogy: the 2022 Global Montessori Census documented 15,763 Montessori schools in 154 countries, with the largest numbers in the United States, China, Thailand, Germany, Canada, and Tanzania, and roughly nine percent operating in the public sector (Debs et al., 2022).

What makes a classroom Montessori

Because the name "Montessori" is not legally protected, anyone may attach it to a school, and implementation varies enormously; any serious reading of the research must therefore begin with what the method actually specifies (Marshall, 2017). The canonical Montessori classroom groups children in three-year age bands (most commonly three to six years) and offers a "prepared environment": open shelves of sequenced, largely self-correcting materials for practical life, sensorial discrimination, language, and mathematics, each isolating a particular concept and each demonstrated to a child individually or in a small group when the teacher judges the child ready. Children choose their own work and pursue it through a long uninterrupted work period, ideally around three hours; extrinsic rewards, grades, and whole-class instruction are largely absent; and the trained adult's central task is observation, with instruction delivered as brief individual lessons (Montessori, 1912; Marshall, 2017). The 2022 census research identifies a consistent core of such practices, including multi-age groupings and free choice of activity, that continues to distinguish Montessori classrooms worldwide (Debs et al., 2022).

Montessori arrived at these arrangements empirically, by iterative observation, but justified them with a developmental theory: the young child's "absorbent mind," sensitive periods for particular acquisitions, and the conviction that movement, the hand, and cognition are intertwined. Several of these commitments anticipated positions in contemporary developmental science, including embodied approaches to cognition and the modern emphasis on executive function, and several practices she pioneered, such as child-scaled furnishings and the teacher-as-observer, have long since diffused into mainstream early childhood provision (Marshall, 2017). The question this review addresses is narrower and harder: when the method is implemented as a package, do children fare better than they would have in conventional schooling?

The lottery studies and randomized trials

The central methodological problem in evaluating any chosen pedagogy is self-selection: families who seek out Montessori schools differ from those who do not, in resources, education, and orientation toward schooling, and those differences, not the pedagogy, may drive any advantage their children show. The strongest Montessori research addresses this problem by exploiting oversubscribed public Montessori schools that admit by random lottery, comparing lottery winners with losers, or by outright random assignment.

The study that inaugurated this literature appeared in Science. Lillard and Else-Quest (2006) compared children who won or lost the admissions lottery of an inner-city public Montessori school in Milwaukee serving predominantly minority families. At age five, the end of the Montessori primary cycle, Montessori children outperformed controls on standardized measures of reading and mathematics, showed stronger executive control and more advanced social cognition, engaged more positively with peers on the playground, and appealed more often to justice and fairness in reasoning about social dilemmas. At age twelve the academic gap was no longer evident, but Montessori students wrote more creative and structurally sophisticated essays and showed more positive social skills. The study was small and its two-cohort design limited, as the authors acknowledged, but it demonstrated that the selection problem could be overcome.

A decade later, Lillard and colleagues (2017) followed 141 children, about half admitted by lottery to two public Montessori magnet preschools in a high-poverty American city, across four testing waves from age three to six. The groups were indistinguishable at entry. Over three years the Montessori children pulled ahead on academic achievement, theory of mind, and mastery orientation, reported greater liking of scholastic tasks, and at age four showed stronger executive function. Notably, the program also equalized: the achievement gap between lower-income Montessori children and higher-income conventionally schooled children narrowed across the study until it was no longer statistically significant, whereas income gaps persisted in the comparison group.

The most rigorous evidence to date is a national randomized controlled trial published in the Proceedings of the National Academy of Sciences. Lillard, Loeb, and colleagues (2025) tracked children who applied to 24 oversubscribed public Montessori preschool programs across the United States, nearly 600 children in all, comparing lottery winners and losers through the end of kindergarten. Montessori children ended kindergarten with significantly higher reading, short-term memory, theory of mind, and executive function, with intention-to-treat effect sizes exceeding a fifth of a standard deviation, large by the standards of school research; effects grew rather than faded across the study period, bucking the fade-out pattern common in preschool evaluations. Strikingly, the authors calculated that three years of public Montessori preschool also cost less per child than the traditional programs the control children attended, largely because Montessori's mixed-age design permits higher child-to-teacher ratios at ages three and four; press materials accompanying the study put the saving at roughly thirteen thousand dollars per child.

The French experiments: adaptation, fade-out, and a sleeper effect

A separate randomized literature comes from France, and it complicates the picture in instructive ways. Courtier and colleagues (2021) randomly assigned preschoolers in a public school serving a disadvantaged neighborhood to conventional classrooms or to classrooms using a Montessori curriculum adapted to French public education, with fewer materials, shorter work periods, and comparatively limited Montessori training for teachers. On mathematics, executive function, and social skills, the adapted Montessori classrooms produced outcomes comparable to, not better than, conventional ones. The exception was reading: disadvantaged kindergarteners from the Montessori classrooms substantially outperformed their conventionally schooled peers (d = 0.68), reaching a level comparable to advantaged children attending an accredited private Montessori school.

The follow-up is among the more intriguing results in the field. Le Diagon, Van der Henst, and Prado (2025) recovered 97 of these children in fifth grade, five years after they had all returned to conventional classrooms. The kindergarten reading advantage had disappeared, a classic fade-out. But the Montessori-educated children now outperformed their peers on mathematical problem-solving (d = 0.58), an advantage that had not existed in kindergarten and that appeared specific to conceptual understanding rather than calculation fluency. Such "sleeper effects" are reported elsewhere in early childhood research, but their appearance in a randomized Montessori trial suggests both that early self-directed learning may build capacities whose value surfaces only when later schoolwork demands them, and that single-timepoint evaluations, in either direction, can mislead.

Meta-analysis and the wider comparative literature

Around this randomized core lies a much larger body of quasi-experimental and observational work, now synthesized in a Campbell Collaboration systematic review. Randolph and colleagues (2023), screening thousands of records down to 32 studies meeting inclusion criteria, concluded that Montessori education has meaningful positive effects relative to traditional education on both academic outcomes and non-academic outcomes, including executive function, creativity, and social-emotional measures, with many effects in the moderate range. The review is candid about the limitations of the underlying literature, much of which is non-randomized, and readers should note that its senior author is also the field's most prolific researcher; but as formal syntheses go, it is the most comprehensive available, and its direction is consistent with the lottery studies.

Individual comparative studies fill in texture, including honest complications. Using data on thousands of low-income children in Miami's Title I public prekindergartens, Ansari and Winsler (2014) found that Latino children in Montessori programs made substantially larger school-readiness gains across the pre-K year than Latino children in conventional High/Scope-based programs, while Black children showed strong gains in both program types, with no comparable Montessori advantage, a reminder that pedagogy interacts with population and context. In South Carolina, home to one of the largest public Montessori sectors in the country, a multi-year evaluation of thousands of students by Furman University's Riley Institute found that public Montessori students modestly but consistently outperformed matched non-Montessori peers in English language arts and mathematics, scored higher on a measure of creativity, attended school more regularly, and were suspended less often (Culclasure, Fleming, Riga, & Sprogis, 2018). Looking further down the road, Dohrmann and colleagues (2007) matched Milwaukee high school graduates who had attended public Montessori programs from preschool through fifth grade with demographically similar peers and found significantly higher mathematics and science achievement in high school, with no differences in English, social studies, or grade-point average.

Not every study is favorable, and the exceptions matter. Lopata, Wallace, and Finn (2005), comparing 543 urban fourth and eighth graders across a public Montessori school and three types of conventional schools, found no Montessori advantage in mathematics or language arts at fourth grade, and lower language arts scores for Montessori students at eighth grade. Marshall's (2017) even-handed review in npj Science of Learning concluded that the evidence base, while promising, remained thin relative to the movement's age and reach, and that many individual studies were compromised by self-selection, small samples, and unmeasured implementation quality. Both cautions predate the strongest recent trials, but they remain the correct frame: the question is not whether any study found nothing, some did, but what distinguishes the settings where Montessori education delivers from those where it does not.

Fidelity: when Montessori works and when it does not

The best available answer is implementation fidelity. Lillard (2012) followed preschoolers across classic Montessori programs (full material sets, no conventional supplements), supplemented Montessori programs that mixed Montessori materials with conventional activities, and conventional programs. Children in classic Montessori classrooms showed significantly greater school-year gains in executive function, reading, mathematics, vocabulary, and social problem-solving than children in either of the other settings, whose outcomes did not differ from each other. The finding rationalizes much of the literature's inconsistency at a stroke: because the name is unprotected, "Montessori" schools in outcome studies range from faithful implementations to conventional classrooms with Montessori shelving, and studies that cannot verify fidelity, including, plausibly, some null results, may be evaluating the label rather than the method. It also converges with the French trials, where a deliberately diluted adaptation produced correspondingly selective effects.

Honesty requires naming the remaining limits of this evidence base. Lottery studies compare applicants, so their results generalize to families who seek out Montessori schools, not necessarily to all families; attrition and non-compliance (lottery winners who leave, losers who find other Montessori seats) complicate even randomized designs, though intention-to-treat analyses such as the 2025 national trial's handle this conservatively. Many outcome measures are those conventional schooling values, standardized reading and mathematics, while Montessori education claims broader aims that are harder to test. Much of the research has been conducted by investigators sympathetic to the method, and independent replication, particularly outside the United States and France, is thinner than one would wish. And the mechanism question is unsettled, as the next section describes.

Mechanisms: executive function, motivation, and creativity

Why might the method work? One candidate is executive function, the self-regulatory capacities of working memory, inhibition, and flexible attention that robustly predict school achievement. In a Science review of interventions that aid executive function development, Diamond and Lee (2011) identified school curricula, Montessori among them, as promising vehicles, noting that successful programs embed repeated practice with progressively increasing challenge, a fair description of a child working through the graded Montessori materials, and that children with the weakest executive functions benefit most. Several of the lottery studies, and the national trial, indeed found executive function advantages. Yet the story is not simple: Denervaud, Knebel, Hagmann, and Gentaz (2019), testing 201 Swiss children, found that Montessori students outperformed traditionally schooled peers in academic outcomes and creativity, but not in global executive function apart from working memory; their mediation analysis suggested that creative skills, rather than executive function per se, carried the academic benefit. The relative contributions of self-regulation, motivation, and creative flexibility remain genuinely open.

Motivation supplies a second, well-documented candidate. Rathunde and Csikszentmihalyi (2005) equipped roughly 290 demographically matched Montessori and traditional middle school students with experience-sampling devices that captured their subjective states throughout the school day. During academic work, Montessori students reported greater affect, energy, intrinsic motivation, flow experience, and "undivided interest," the conjunction of high motivation and high importance, than traditional students, whose academic work more often felt important but unenjoyable; outside academics, the two groups' experiences were indistinguishable. Whatever its ultimate mechanism, a pedagogy under which schoolwork feels chosen rather than imposed appears to change how learning is experienced, which is consistent with the greater liking of scholastic tasks observed in Montessori preschoolers (Lillard et al., 2017) and with decades of self-determination research linking autonomy to engagement.

Practical implications

Several implications follow for families and educators. First, the evidence base for Montessori education is now stronger than for virtually any other named early childhood pedagogy: converging results from lottery-based natural experiments, preregistered randomized trials, a national RCT, and a formal meta-analysis point in the same positive direction, most clearly for early literacy, executive function, social understanding, and school-readiness outcomes in the three-to-six age range, including, and perhaps especially, for children from lower-income families. Parents weighing a Montessori preschool are not betting on folklore.

Second, fidelity is the variable a parent can actually inspect. Because the name is unprotected, families should look past the sign to the practices the research identifies: credentialed teachers trained by recognized bodies, full sets of Montessori materials in daily use rather than on display, genuine three-year age groupings, a long uninterrupted morning work period, and the absence of heavy conventional supplementation, since supplemented programs in the best available comparison performed no better than conventional ones (Lillard, 2012). Third, expectations should be calibrated by timescale: some advantages appear early and persist, some fade, and at least one, mathematical problem-solving, emerged only years later; a Montessori preschool is best understood as an investment in capacities, concentration, autonomy, conceptual grounding, whose returns are not all immediate (Le Diagon et al., 2025). Fourth, the motivational findings deserve weight in their own right. A schooling in which children experience academic work with interest and energy, rather than dutiful boredom, is a meaningful outcome even where test scores tie. Finally, families should bring the same honesty to the decision that good researchers bring to the data: Montessori education is well supported, not proven superior for every child in every implementation, and the quality of the particular school in front of you matters more than the word on its door.

A century seldom flatters an educational innovation; most reforms of 1907 are historical curiosities. The Montessori method has instead passed through the most demanding filter modern social science can apply to a pedagogy, randomized comparison, and emerged, at high fidelity, with its central claims substantially intact: that young children given freedom within a prepared structure concentrate deeply, regulate themselves, master academic foundations, and like the work of learning. Maria Montessori called her first book an essay in "scientific pedagogy." The description was premature in 1912, when her evidence was a physician's field notes from a room in San Lorenzo. It is considerably less premature now, when the children of that room's successors, chosen by lottery ticket rather than by fortunate birth, keep arriving at kindergarten ahead of where they would otherwise have been. The research does not say the method is magic. It says, with increasing confidence, that it works, and that is a rarer thing.

References

1. Ansari, A., & Winsler, A. (2014). Montessori public school pre-K programs and the school readiness of low-income Black and Latino children. Journal of Educational Psychology, 106(4), 1066–1079. https://doi.org/10.1037/a0036799
2. Courtier, P., Gardes, M.-L., Van der Henst, J.-B., Noveck, I. A., Croset, M.-C., Epinat-Duclos, J., Léone, J., & Prado, J. (2021). Effects of Montessori education on the academic, cognitive, and social development of disadvantaged preschoolers: A randomized controlled study in the French public-school system. Child Development, 92(5), 2069–2088. https://doi.org/10.1111/cdev.13575
3. Culclasure, B., Fleming, D. J., Riga, G., & Sprogis, A. (2018). An evaluation of Montessori education in South Carolina's public schools. Greenville, SC: The Riley Institute at Furman University. ERIC ED622145. https://files.eric.ed.gov/fulltext/ED622145.pdf
4. Debs, M., de Brouwer, J., Murray, A. K., Lawrence, L., Tyne, M., & von der Wehl, C. (2022). Global diffusion of Montessori schools: A report from the 2022 Global Montessori Census. Journal of Montessori Research, 8(2), 1–15. https://journals.ku.edu/jmr/article/view/18675
5. Denervaud, S., Knebel, J.-F., Hagmann, P., & Gentaz, E. (2019). Beyond executive functions, creativity skills benefit academic outcomes: Insights from Montessori education. PLOS ONE, 14(11), e0225319. https://doi.org/10.1371/journal.pone.0225319
6. Diamond, A., & Lee, K. (2011). Interventions shown to aid executive function development in children 4 to 12 years old. Science, 333(6045), 959–964. https://doi.org/10.1126/science.1204529
7. Dohrmann, K. R., Nishida, T. K., Gartner, A., et al. (2007). High school outcomes for students in a public Montessori program. Journal of Research in Childhood Education, 22(2), 205–217. https://doi.org/10.1080/02568540709594622
8. Kramer, R. (1976). Maria Montessori: A biography. New York: G. P. Putnam's Sons. https://books.google.com/books/about/Maria_Montessori.html?id=2fElAQAAIAAJ
9. Le Diagon, S., Van der Henst, J.-B., & Prado, J. (2025). Early Montessori education shows delayed benefits for mathematical problem-solving in a 5-year longitudinal randomized controlled trial. Scientific Reports, 15, 43961. https://www.nature.com/articles/s41598-025-27687-2
10. Lillard, A. S. (2012). Preschool children's development in classic Montessori, supplemented Montessori, and conventional programs. Journal of School Psychology, 50(3), 379–401. https://doi.org/10.1016/j.jsp.2012.01.001
11. Lillard, A. S., & Else-Quest, N. (2006). Evaluating Montessori education. Science, 313(5795), 1893–1894. https://doi.org/10.1126/science.1132362
12. Lillard, A. S., Heise, M. J., Richey, E. M., Tong, X., Hart, A., & Bray, P. M. (2017). Montessori preschool elevates and equalizes child outcomes: A longitudinal study. Frontiers in Psychology, 8, 1783. https://doi.org/10.3389/fpsyg.2017.01783
13. Lillard, A. S., Loeb, D., Berg, J., Escueta, M., Manship, K., Hauser, A., & Daggett, E. D. (2025). A national randomized controlled trial of the impact of public Montessori preschool at the end of kindergarten. Proceedings of the National Academy of Sciences, 122(43), e2506130122. https://doi.org/10.1073/pnas.2506130122
14. Lopata, C., Wallace, N. V., & Finn, K. V. (2005). Comparison of academic achievement between Montessori and traditional education programs. Journal of Research in Childhood Education, 20(1), 5–13. https://eric.ed.gov/?id=EJ726432
15. Marshall, C. (2017). Montessori education: A review of the evidence base. npj Science of Learning, 2, 11. https://www.nature.com/articles/s41539-017-0012-7
16. Montessori, M. (1912). The Montessori method: Scientific pedagogy as applied to child education in "the Children's Houses" (A. E. George, Trans.). New York: Frederick A. Stokes. https://archive.org/details/montessorimethod00montuoft
17. Randolph, J. J., Bryson, A., Menon, L., Henderson, D. K., Kureethara Manuel, A., Michaels, S., et al. (2023). Montessori education's impact on academic and nonacademic outcomes: A systematic review. Campbell Systematic Reviews, 19(3), e1330. https://doi.org/10.1002/cl2.1330
18. Rathunde, K., & Csikszentmihalyi, M. (2005). Middle school students' motivation and quality of experience: A comparison of Montessori and traditional school environments. American Journal of Education, 111(3), 341–371. https://www.jstor.org/stable/10.1086/428885
19. Whitescarver, K., & Cossentino, J. (2008). Montessori and the mainstream: A century of reform on the margins. Teachers College Record, 110(12), 2571–2600. https://doi.org/10.1177/016146810811001202