Inquiry-based science instruction—what is it and does it matter? Results from a research synthesis years 1984 to 2002

138 studies (101 addressing conceptual understanding; 42 with multiple treatment arms, of which 34 rated moderate/high rigor)

K-12 science, predominantly United States, 1984-2002; outcome of interest was science content/conceptual learning.

This is the single most-cited "inquiry works" reference in science education and it is NOT an effect-size meta-analysis. The authors state plainly that "traditional effect sizes were not calculated"; the analysis is vote-counting plus logistic models predicting whether a study reported a positive result. The corpus is mostly NOT experimental: chi-square tables show 30 experimental, 35 quasi-experimental and 73 non-experimental studies. Rigor was weak — 71% of studies gave no or poor information on sampling strategy, 80% of quantitative studies mishandled or failed to report missing data, and 30% of studies were rated low rigor overall. Measure quality is the decisive weakness for this archive: 56% of studies did not say whether their instrument was new or existing, and of those that did, ONLY 11% (16 of 138 studies) used an established instrument; 62% never reported reliability for their sample; only 26% demonstrated any validity, and that exclusively content validity. "Inquiry" is operationalised as "inquiry saturation" — a composite of how much of the investigation cycle (question/design/data/conclusion/communication) was present and how much student active thinking and responsibility was emphasised — so guided inquiry, open discovery, hands-on labs and computer simulations are all pooled into one continuous dose variable.

51% of the 138 studies showed a positive impact of some level of inquiry instruction on content learning. But in the 101 studies of conceptual understanding, there was NO statistically significant association between amount of inquiry saturation and increased conceptual learning; only after model refinement did two specific components — emphasis on student ACTIVE THINKING (b=2.351, p=0.040, ~11x odds) and emphasis on DRAWING CONCLUSIONS FROM DATA (marginal, p=0.095) — predict positive outcomes. Among the 34 moderate/high-rigor studies that directly compared a higher-inquiry to a lower-inquiry arm, 19 (56%) favoured higher inquiry, 8 (23%) found no difference, 4 (12%) were mixed and 1 (2%) favoured LOWER inquiry. The authors' own summary is the honest one: "The evidence of effects of inquiry-based instruction from this synthesis is not overwhelmingly positive."

Low relevance: contrasts are within-study instructional comparisons, not between-family. The dominant confound is measure quality (89% non-established instruments) and publication/vote-count bias, not heredity.

Minner explicitly cross-references Schroeder 2007 as directionally concordant, and the finding that guidance/active thinking rather than inquiry dose is the operative variable matches Furtak 2012 (teacher-led > student-led) and Lazonder & Harmsen 2016.