Loading the library…
instruction, resources & composition
The operative claimSubstantial class-size reduction in the early grades (e.g., to 13-17 students, as in Tennessee's Project STAR) produces small-to-moderate short-term gains in student achievement, with larger benefits for low-income and minority students; long-term and at-scale effects are smaller, inconsistent, and contested.
Class-size reduction shows a real but modest short-term effect that is genuinely contested in magnitude and scalability: a respected RCT (STAR) on one side, a cautious systematic review and pointed econometric critiques on the other.
Anchored by one large randomized experiment (Project STAR, N~11,600, grades K-3) showing roughly a 4-percentile first-year gain, with the biggest gains for Black and free-lunch students; long-term follow-ups link small-class assignment to higher college-test-taking and a narrowed Black-white gap, though not reliably to adult earnings. Observational and systematic-review evidence finds at most a small reading effect and no reliable mathematics effect, and statewide scale-ups largely failed to reproduce STAR. The effect is concentrated at large reductions (to 13-17) and in the early grades.
Krueger, A. B (1999). Experimental estimates of education production functions. Quarterly Journal of Economics (NBER WP 6051), 114(2), 497-532.
Experimental · N = 11,600 · ages ~5-9 (K-3)
In the randomized STAR experiment, assignment to a small class (13-17) raised standardized-test performance ~4 percentile points in the first year (then ~1 pt/yr); effects were larger for minority and free-lunch students; teacher aides had little effect; gains were not a Hawthorne effect.
Bearing on this claim: Project STAR RCT; ~4-percentile first-year gain, larger for minority/low-income students.
doi.org/10.1162/003355399556052Krueger, A. B., & Whitmore, D. M (2001). The effect of attending a small class in the early grades on college-test taking and middle school test results: Evidence from Project STAR. The Economic Journal, 116(4), 1491-1527.
Experimental · N = 11,600 · grades K-8 + college-entrance exams
Attending a small class in early grades was associated with higher later test performance and a greater likelihood of taking a college-entrance exam, narrowing the Black-white gap in college-test-taking by ~54%; estimated internal rate of return ~5.5%.
Bearing on this claim: Long-term STAR follow-up: more college-test-taking, narrowed Black-white gap, ~5.5% IRR.
doi.org/10.1111/1468-0297.00586Chetty, R., Friedman, J. N., Hilger, N., Saez, E., Schanzenbach, D. W., & Yagan, D (2011). How does your kindergarten classroom affect your earnings? Evidence from Project STAR. Quarterly Journal of Economics, 126(4), 1593-1660.
Experimental · N = 11,571 · followed to age ~27
Small-class students were significantly more likely to attend college and showed improvements on other outcomes; class size had no statistically significant effect on age-27 earnings (imprecisely estimated). Test-score gains faded in later grades while non-cognitive gains persisted.
Bearing on this claim: Adult-outcome follow-up: more college attendance but no significant earnings effect; test gains fade, non-cognitive persist.
doi.org/10.3386/w16381Counter-authority (2). Studies an adversary may cite — surfaced deliberately so the grade is honest about its limits.
Filges, T., Sonne-Schmidt, C. S., & Nielsen, B. C. V (2018). Small class sizes for improving student achievement in primary and secondary schools: A systematic review. Campbell Systematic Reviews, 14(1), 1-107.
Systematic review · grades K-12
Across the broader literature the evidence suggests at best a small effect of class-size reduction on reading achievement and a negative but non-significant effect on mathematics; the authors caution that small classes are costly and may even be counterproductive for some students. STAR-based estimates were more positive but did not change the overall finding.
Bearing on this claim: Campbell systematic review: at best a small reading effect, no reliable math effect; cautions on cost.
doi.org/10.4073/csr.2018.103 supporting · 2 counter verified sources
STAR's implementation had attrition and some non-compliance (re-assignment), which critics argue bias estimates upward (Hanushek 1999); effects fade after return to regular classes (Chetty 2011); the math effect is unreliable and may be null (Filges 2018); and scale-ups did not reproduce STAR, so external validity is weak. The rating is of a modest causal short-term effect, not of cost-effective or scalable policy.
Last reviewed June 26, 2026
Hanushek, E. A (1999). Some findings from an independent investigation of the Tennessee STAR experiment and from other investigations of class size effects. Educational Evaluation and Policy Analysis, 21(2), 143-163.
Review
Argues non-experimental evidence shows no consistent achievement gain from class-size reduction, that STAR's design/implementation issues bias results upward, and that any effect is limited to large reductions in K/grade-1. (Critical/dissenting view.)
Bearing on this claim: Critique: non-experimental evidence shows no consistent gain; STAR estimates argued to be biased upward.
doi.org/10.3102/01623737021002143