A Bayesian Network Meta-Analysis to Synthesize the Influence of Contexts of Scaffolding Use on Cognitive Outcomes in STEM Education.
Level 1 - systematic review of randomized trials
Systematic review and Bayesian network meta-analysis (level assigned by design analogy for educational research).
PubMed 29200508 · doi:10.3102/0034654317723009
What was done
The authors conducted a Bayesian network meta-analysis of 56 studies to synthesize within-subject (pre-post) cognitive outcomes resulting from computer-based scaffolding in STEM education. The posterior distribution was estimated using 20,000 Markov Chain Monte Carlo samples to examine effects across student populations, STEM disciplines, instructional models, and educational levels.
What was found
Scaffolding demonstrated consistently strong effects across student populations, STEM disciplines, and assessment levels, as well as across most problem-centered instructional models (exceptions: inquiry-based learning and modeling visualization) and educational levels (exception: secondary education). The abstract reports no overall numerical effect sizes, but notes a particularly large effect size for students with learning disabilities (ḡ = 3.13).
Why it matters
This review synthesizes previously overlooked pre-post within-subject literature on computer scaffolding, showing robust benefits across most STEM contexts while identifying specific exceptions and highly responsive populations such as students with learning disabilities.
Limits
The meta-analysis is restricted to within-subject (pre-post) designs rather than randomized between-subject comparisons, which are vulnerable to maturation and testing effects. The abstract does not report the total participant sample size, confidence intervals, or numerical effect sizes for the main comparisons.
Cited by
- supports There is good scientific evidence showing that a linear, scaffolded approach is most effective for educational learning.