Review of Evidence: Arts Integration Research Through the Lens of the Every Student Succeeds Act

1,619 reports located; 135 screened by What Works Clearinghouse-certified reviewers; 44 interventions with evidence at any ESSA tier; 27 studies meeting Tier I-III design requirements contributing 122 effect sizes

US arts-integration interventions in K-12 schools, studies published since 2000

The most useful feature of this review is not its pooled estimate but its funnel. It applies the ESSA Title VIII four-tier evidence standard with WWC-certified reviewers to the whole arts-integration field, and reports how much of the field clears each bar. Note the scope: arts INTEGRATION (art used as a vehicle for teaching other subjects), not arts education for its own sake, which the companion volume covers. Graded C - a meta-analysis of mixed-quality designs, and not independent of the Crystal Bridges museum experiments, which contribute two of the 27 studies and are the sole source of the critical-thinking cell.

Only 10 of 44 catalogued arts-integration interventions met ESSA Tiers I-III, and ONE met Tier I. Thirty-four of 44 (77%) had Tier IV evidence only - meaning a rationale, with no demonstrated effect on any student outcome. Across the 27 studies that did meet design requirements, the pooled effect is g = 0.11 [0.07, 0.16], which the authors translate as four percentile points and place at the 30th percentile of interventions in the What Works Clearinghouse. By domain: academic achievement 0.11 [0.04, 0.18]; English language arts 0.11 [0.05, 0.16]; MATHEMATICS 0.06 [-0.17, 0.30], null; attitudes 0.11 [0.06, 0.17]; social-emotional learning 0.19 [0.05, 0.34]; critical thinking 0.11 [0.07, 0.15] from a single study. Within-study averages ranged from -0.90 to +0.86. A "0.3 SD" figure for this review circulates online and is wrong; the report's own gloss is four percentile points. The authors' recommendation is unusually blunt for a foundation-commissioned review: "Adopting a conservative interpretation of ESSA's evidence-based criteria might help promote interventions with a stronger likelihood of success."

Partly controlled by the ESSA tier screen, which requires a comparison group and statistical controls; but only one of 44 interventions reached the randomised-trial tier, so selection into arts-integration programmes is largely unmanaged across the corpus.

No replication of the review. Its pooled estimate is consistent in direction with the REAP verdict and with the two randomised programmes that came later (Houston, EEF), both of which are null on the achievement outcomes this review estimates at 0.11 - which suggests 0.11 is itself an upper bound inherited from the weaker designs in the pool.