The Evidence on Teaching

Which Instructional Practices Most Help First-Grade Students With and Without Mathematics Difficulties?

Morgan, P. L., Farkas, G., & Maczuga, S. · 2015

grade Dlongitudinalindependentunreplicated
Sample
13,883 ECLS-K students / 3,635 classrooms
Population
US first graders; math-difficulty subgroups.
Design
Correlational value-added on teacher-self-reported practices. Visible selection artifact: manipulatives/movement/music MORE used in MD-heavy classrooms (remediation targeting).
Key findings
Widely-publicized correlational result (only teacher-directed practice predicted gains for strugglers) — but grade D: teacher self-report, tiny effects, and a reverse-causation tell. The database's exhibit for symmetric standards: reform advocates' at-scale evidence rests on the same or weaker designs.
Genetic confound
Severe: practice choice responds to classroom ability composition; math ability substantially heritable; value-added only partially absorbs this.
Replication notes
No experimental replication; cite only as weak directional corroboration, never as independent support.
DOI / URL
10.3102/0162373714536608

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Teacher-directed instruction (drill/practice/worksheets), MD studentsES per SD of use+0.05-0.07 (tiny)standardizedspring grade 1nonenot-applicabledomain-skill
Student-centered / manipulatives / movement-music, MD studentsESns (student-centered helped only non-MD students)standardizedspring grade 1nonenot-applicabledomain-skill

Cited by