RESUMEN
Background: Montessori education is the oldest and most widely implemented alternative education in the world, yet its effectiveness has not been clearly established. Objectives: The primary objective of this review was to examine the effectiveness of Montessori education in improving academic and nonacademic outcomes compared to traditional education. The secondary objectives were to determine the degree to which grade level, Montessori setting (public Montessori vs. private Montessori), random assignment, treatment duration, and length of follow-up measurements moderate the magnitude of Montessori effects. Search Methods: We searched for relevant studies in 19 academic databases, in a variety of sources known to publish gray literature, in Montessori-related journals, and in the references of studies retrieved through these searches. Our search included studies published during or before February 2020. The initial search was performed in March 2014 with a follow-up search in February 2020. Selection Criteria: We included articles that compared Montessori education to traditional education, contributed at least one effect size to an academic or nonacademic outcome, provided sufficient data to compute an effect size and its variance, and showed sufficient evidence of baseline equivalency-through random assignment or statistical adjustment-of Montessori and traditional education groups. Data Collection and Analysis: To synthesize the data, we used a cluster-robust variance estimation procedure, which takes into account statistical dependencies in the data. Otherwise, we used standard methodological procedures as specified in the Campbell Collaboration reporting and conduct standards. Main Results: Initial searches yielded 2012 articles, of which 173 were considered in detail to determine whether they met inclusion/exclusion criteria. Of these, 141 were excluded and 32 were included. These 32 studies yielded 204 effect sizes (113 academic and 91 nonacademic) across 132,249 data points. In the 32 studies that met minimum standards for inclusion, including evidence of baseline equivalence, there was evidence that Montessori education outperformed traditional education on a wide variety of academic and nonacademic outcomes. For academic outcomes, Hedges' g effect sizes, where positive values favor Montessori, ranged from 0.26 for general academic ability (with high quality evidence) to 0.06 for social studies. The quality of evidence for language (g = 0.17) and mathematics (g = 0.22) was also high. The effect size for a composite of all academic outcomes was 0.24. Science was the only academic outcome that was deemed to have low quality of evidence according to the GRADE approach. Effect sizes for nonacademic outcomes ranged from 0.41 for students' inner experience of school to 0.23 for social skills. Both of these outcomes were deemed as having low quality of evidence. Executive function (g = 0.36) and creativity (g = 0.26) had moderate quality of evidence. The effect size for a composite of all nonacademic outcomes was 0.33. Moderator analyses of the composite academic and nonacademic outcomes showed that Montessori education resulted in larger effect sizes for randomized studies compared to nonrandomized studies, for preschool and elementary settings compared to middle school or high school settings, and for private Montessori compared to public Montessori. Moderator analyses for treatment duration and duration from intervention to follow-up data collection were inconclusive. There was some evidence for a lack of small sample-size studies in favor of traditional education, which could be an indicator of publication bias. However, a sensitivity analysis indicated that the findings in favor of Montessori education were nonetheless robust. Authors' Conclusions: Montessori education has a meaningful and positive impact on child outcomes, both academic and nonacademic, relative to outcomes seen when using traditional educational methods.
