The Evidence on Teaching

The Aftermath of Accelerating Algebra: Evidence from District Policy Initiatives (Charlotte-Mecklenburg)

Clotfelter, C. T., Ladd, H. F., & Vigdor, J. L. · 2015

grade Bnatural-experimentindependentreplicatednumbers spot-checked
Sample
N=35,339 (policy cohorts); Guilford County verification
Population
NC middle/high schoolers; 8th-grade algebra for median students jumped 51%->85%, then reverted.
Design
IV/diff-in-diff across cohorts x prior-ability deciles; state EOC outcomes; falsification tests pass.
Key findings
Best-identified evidence on universal acceleration: pushing unprepared students into earlier algebra HARMS them at every ability level tested, worst for the weakest. The district reversed its own policy. The correlational case for early algebra was pure selection.
Genetic confound
The design's core point: high-ability students select into early algebra; remove selection and the 'benefit' becomes harm.
Replication notes
Verified in a second district; extended to 10 NC districts; converges with California (Domina) and Chicago (Allensworth).
DOI / URL
10.3368/jhr.50.1.159

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Moderate performers accelerated into 8th-grade algebraSD-0.32 on Algebra I EOC; Geometry on-time passage -10ppstandardizedend of course to +4yrbusiness-as-usualover-2yrdomain-skill
Low performers accelerated into 9th-grade algebraSD-0.5; Geometry pass 50%->40%standardizedend of coursebusiness-as-usualend-of-treatmentdomain-skill
OLS (selection) vs IV (causal)contrastOLS shows early-takers do BETTER (+0.18); causal IV flips the sign — textbook ability-selection artifactstandardizedn/aunclearnot-applicabledomain-skill

Cited by