1979
Meta-Analysis

Glass & Smith Meta-Analysis

The first major quantitative synthesis of class-size research. Analyzed 77 studies and established that student achievement increases as class size decreases — particularly below 20 students. This landmark review set the stage for the experimental era to follow.

1985
Experiment

Project STAR Begins

Tennessee launches the Student/Teacher Achievement Ratio experiment — the largest randomized class-size trial ever conducted. 11,571 students across 79 schools are randomly assigned to small classes (13–17), regular classes (22–25), or regular classes with a full-time aide.

1989
Experiment

Project STAR Concludes

After four years, results show small-class students outperform peers by 0.20–0.34 SD in reading and math. Effects are roughly double for Black students and those on free lunch. The aide-only intervention shows no significant benefit — establishing that it is class size itself, not just extra adults, that matters.

1996
Policy

SAGE Program & California CSR Launch

Wisconsin’s SAGE program begins targeting K–3 class sizes of 15. California simultaneously launches the nation’s largest class-size reduction initiative, dropping average K–3 classes from 28.5 to 19.5 students. Two major policy experiments are now underway.

1999
Reanalysis

Krueger Reanalysis of STAR

Alan Krueger publishes a rigorous reanalysis of STAR correcting for non-random attrition and school fixed effects. Confirms and strengthens the original findings: effect sizes of 0.19–0.28 SD for small classes vs. regular. Establishes STAR as one of the most robust experiments in education research.

2000
Counterpoint

Hoxby Natural Experiment

Caroline Hoxby uses natural population variation in Connecticut to argue that class-size effects may be near zero once selection bias is removed. This counterpoint sparks a major debate in the field, though subsequent critiques question the external validity and power of the approach.

2001
Follow-Up

Krueger & Whitmore — Long-Run STAR Effects

The first long-run follow-up of STAR students shows small-class students are significantly more likely to take the ACT/SAT college entrance exams, with Black students seeing the largest gains. The race gap in test-taking narrows by 60% for students who were in small K–3 classes.

2003
Socioemotional

UK — Attentiveness & Peer Relations

Blatchford, Edmonds & Martin publish a longitudinal study of English Reception and KS1 classes. Systematic observations of 235 children confirm more off-task behavior in larger classes. Teacher ratings of 5,000+ children reveal mixed peer-relations patterns — an early signal that socioemotional outcomes need nuanced measurement.

2011
Landmark

Chetty et al. — STAR Adult Outcomes

Raj Chetty and colleagues link STAR students to tax records and find that small-class students earn significantly more as adults, are more likely to attend college, save for retirement, and live in higher-quality neighborhoods. Demonstrates that early class-size investments have returns that compound over a lifetime.

2011
Socioemotional

Dee & West — Non-Cognitive STAR Returns

Reanalysis of STAR follow-up data finds K–3 small-class assignment improves teacher-rated student initiative by 0.13–0.18 SD at Grade 4. The authors’ broader NELS:88 analysis yields an implied internal rate of return of 4.6% overall, 7.9% in urban schools. Adds the socioemotional dimension to the STAR evidence base.

2011
Synthesis

Brookings — What Research Says

Whitehurst & Chingos publish an influential Brookings synthesis concluding that large, sustained class-size reductions (7+ students) show the clearest benefits. They note that small, incremental reductions may not be cost-effective — framing the policy debate around magnitude and targeting.

2013
International

Sweden — Long-Run Wages & Cognition

Fredriksson, Öckert & Oosterbeek follow Swedish students from primary school into their 40s. A 7-student class reduction yields +0.23 SD cognitive gains at age 13 and meaningful adult wage gains. Internal rate of return: ~18%. One of the most compelling long-run cases for class-size investment in international research.

2017
International

Germany — Class Size and Grade Repetition

Argaw & Puhani analyze 258,098 students in Hesse, Germany. Using the cap-at-25 rule as an instrument, they find each additional grade-1 student increases repetition probability by 0.4 percentage points. Even in a system with already-small classes, further reductions produce measurable benefits.

2019
Socioemotional

Austria & Netherlands — Bullying Dynamics

Garandeau et al. study 2,911 students across 136 classrooms and find bullying can appear higher in smaller classes — partly a measurement artifact of peer-nomination scales. An important caveat: bullying prevention requires more than size reduction alone; social norms and bystander interventions matter too.

2020
Quasi-Experimental

California — Composition Effects Revisited

Gilraine, Macartney & McMillan evaluate California’s massive K–3 reduction using 914,514 observations. Beyond direct math gains of ~0.11 SD, school composition changes amplified the benefits — making the policy more cost-effective than direct effects alone would suggest.

2020
Socioemotional

West Bank — Class Size, Bullying & Violence

Hallaq uses regression discontinuity across ~6,000 students in grades 5–9. A 10-student increase corresponds to 12.9 pp higher bullying-victim probability. Mental-health difficulties also rise with class size, demonstrating psychological harm from overcrowded classrooms.

2021
Socioemotional

Quebec — Nonlinear Effects on Social-Emotional Skills

Connolly & Haeck study 108,151 kindergarteners using Quebec’s cap-at-20 rule. Reducing class size improves emotional maturity (+0.12 SD), social competence (+0.07 SD), and communication (+0.08 SD). Effects are nonlinear and strongest in disadvantaged areas — key evidence that targeting matters.

2025
International

Japan — Class Size and Peer Effects

Tanaka & Wang use student-teacher pair fixed effects across 145,264 student-year observations in Japanese elementary schools. A 10-student reduction yields +0.087 SD in math. Importantly, the study shows that peer composition interacts with class-size effects — who is in the classroom matters alongside how many.

2025
Socioemotional

Denmark — Class Size and Student Distress

Beuchert & Nandrup establish a causal link between larger K–2 classes and increased physical and emotional distress, with effects concentrated among low-income students. The latest evidence that class size policy has consequences for student welfare far beyond test scores.

The evidence continues to grow

Across 12 countries, 4 continents, and 46 years of research, the conclusion is consistent: smaller class sizes — especially in early grades and for disadvantaged students — produce meaningful, lasting benefits in both academic achievement and socioemotional well-being.