The relative effects and equity of inquiry-based and commonplace science teaching on students' knowledge, reasoning, and argumentation

58 students aged 14-16, individually randomised to two arms

14-16 year olds in a controlled, out-of-classroom ("laboratory-based") instructional setting in the US.

The cleanest head-to-head inquiry-versus-not contrast in school-age science, and also the smallest and least generalisable. Fifty-eight students were individually randomised; BOTH arms were taught toward the SAME learning goals by the SAME teacher, one from materials organised around the BSCS 5E Instructional Model (Engage, Explore, Explain, Elaborate, Evaluate — guided inquiry with a scripted sequence, not open discovery) and one from materials organised around commonplace teaching strategies as defined by national teacher survey data. Holding teacher, goals and time constant is exactly the design the field usually lacks, and it makes this a test of INSTRUCTIONAL SEQUENCE rather than of teacher quality or curriculum coverage. Two hard limits: (a) n=58 with one teacher in a non-classroom setting means the estimate is imprecise and the ecological validity is low; (b) all four authors are at BSCS, which owns and promotes the 5E model, so this is developer-led evaluation of the developers' own instructional model. Outcomes were built for the study across three constructs — knowledge, scientific reasoning and scientific argumentation — at immediate post-test and four weeks later. We were unable to obtain the paper itself and therefore did not extract the effect sizes; the direction and significance below come from the published abstract.

Students taught with the 5E inquiry-based materials reached significantly higher levels of achievement than students taught with commonplace strategies, and the advantage held across all three outcome constructs (knowledge, reasoning, argumentation) and at both time points (immediately and four weeks later). A separate finding worth carrying: commonplace instruction produced a DETECTABLE ACHIEVEMENT GAP BY RACE while the inquiry-based materials did not — an equity claim from n=58, so treat it as hypothesis-generating only. This study is the strongest existing warrant for "well-structured guided inquiry beats ordinary science teaching when everything else is held constant", and it is a 58-student developer-led trial. That gap between how load-bearing it is and how thin it is, is itself a finding about this literature.

Low. Individual random assignment with a common teacher; the threats are sample size, developer allegiance, and a contrived setting, not heredity.

No independent replication of this specific contrast located. It is pooled into Furtak 2012's inquiry meta-analysis. The BSCS 5E model itself has no large independently evaluated cluster RCT in this archive.