The Evidence on Teaching

PISA 2015 Results (Volume II): Science Teaching Practices and Scientific Literacy

OECD · 2016

grade Dlongitudinalindependentfailed
Sample
~519,000 students, 72 systems
Population
15-year-olds across 72 systems; standardized PISA science literacy.
Design
Cross-national CORRELATIONAL analysis; self-reported practice frequency; severe confounding (more hands-on inquiry in weaker/less-disciplined classes; reverse causation).
Key findings
The headline 'inquiry is negatively related to achievement' — across 72 systems more inquiry correlated with LOWER science scores, teacher-directed with higher. But purely correlational and heavily confounded; it collapses under Jerrim 2019's within-country controls.
Genetic confound
Severe: purely correlational/self-report; ability sorting and reverse causation. Never causal.
Replication notes
The raw negative attenuates to null with better controls (Jerrim 2019). Use as confounded correlational evidence only.
DOI / URL
10.1787/9789264267510-en

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Frequent inquiry-based instruction -> science literacyassociationNEGATIVE in nearly all systems (e.g. New Zealand ~ -14 points)standardizedcross-sectionalnonenot-applicabledomain-skill
Teacher-directed instruction -> science literacyassociationPOSITIVE (strongest of the practice indices)standardizedcross-sectionalnonenot-applicabledomain-skill

Cited by