The Effects of Team-Based Learning on Social Studies Knowledge Acquisition in High School
Wanzek J, Vaughn S, Kent SC, Swanson EA, Roberts G, Haynes M, Fall A-M, Stillman-Spisak SJ, Solis M · 2014
grade Brctdeveloper-ledmixed
Sample
26 eleventh-grade social studies classes randomly assigned in a blocked design
Population
Eleventh-graders in US high school social studies (US History) classes
Design
Component-dismantling trial: it strips PACT down to just its team-based-learning parts (TBL comprehension checks and TBL knowledge application) across three 15-day units, to ask how much of the package's knowledge effect comes from the discussion/retrieval structure rather than the content scaffolding. Randomised blocked design at class level. The outcome is the same developer-built content acquisition measure used throughout the PACT programme. Read at abstract depth, cross-checked against the developers' summary in Roberts et al. (2023). The companion eighth-grade TBL trial (Wanzek, Kent et al. 2015, 24 classes) found NO significant effect, so the component result does not generalise down an age band.
Key findings
Team-based learning alone produced a small but significant gain in US history content knowledge, Hedges g = 0.19 - roughly half the effect of the full PACT package at the same grade level (0.36), which is a rare and useful piece of dismantling evidence: the discussion/retrieval machinery carries part of the effect but not most of it. The moderator result is the one to carry forward and it runs against the equity case usually made for these programmes: students with HIGH or MODERATE pretest scores benefited, while students with LOW pretest scores did not benefit at all. Knowledge-building instruction here compounded existing advantage rather than closing gaps - the same direction as the Cabell et al. (2025) kindergarten interaction.
Genetic confound
Low. Class-level randomisation in a blocked design; ability-relevant alleles balanced in expectation.
Replication notes
Mixed. The parallel eighth-grade TBL trial by the same team (Wanzek, Kent et al. 2015, 24 classes) found no significant effect on content acquisition, so the component effect replicated at grade 11 but failed at grade 8. No independent group has evaluated it.
DOI / URL
10.1080/19345747.2013.836765
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| US history content acquisition (team-based learning components only) | Hedges g | 0.19 (statistically significant) | researcher-designed | after three 15-day units | business-as-usual | end-of-treatment | domain-skill |
| Content acquisition, by prior attainment | subgroup response | high and moderate pretest groups benefited; LOW pretest group did not benefit | researcher-designed | after three 15-day units | business-as-usual | end-of-treatment | domain-skill |
| Same components at eighth grade (companion trial, 24 classes) | significance test | no significant effect on content acquisition | researcher-designed | after three 15-day units | business-as-usual | end-of-treatment | domain-skill |
Cited by
- Should history be built on content knowledge or on generic historical-thinking skills?moderate supportconf: mediumgc: low