The effects of coding on children's planning and inhibition skills
Arfé, B., Vardanega, T., & Ronconi, L. · 2020
grade Crctdeveloper-involvedunreplicated
Sample
179 first graders across 5 schools
Population
Italian grade 1 children.
Design
A cluster-randomised trial in which the comparison group received standard STEM instruction — an active-ish comparison, which is a real strength and rarer than it should be in this literature. The dose was 480 minutes (8 hours) of Code.org, roughly a third of one percent of a school year. Standardized executive-function outcomes. Read at secondary depth through the authors' own later meta-analysis, so the effect sizes below are tabulated rather than extracted from the paper.
Key findings
Inhibition d = -0.71, planning d = 1.27, and problem solving d = 1.31 explicitly labelled near transfer. A planning effect of d = 1.27 from eight hours of Code.org against an active STEM comparison is larger than essentially any effect in this archive. Either this is the most cost-effective cognitive intervention ever documented, or it is an artefact; the two independent, larger, validated-instrument trials in this topic point to the latter.
Genetic confound
Low. Cluster randomisation.
Replication notes
The same laboratory as Arfé et al. (2019) and as the Montuori et al. meta-analysis that pools it. No independent replication.
DOI / URL
10.1016/j.compedu.2020.103807
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Planning | Cohen d | 1.27 | standardized | post-test | active-alternative | end-of-treatment | far-transfer |
| Response inhibition (negative d means fewer errors) | Cohen d | -0.71 | standardized | post-test | active-alternative | end-of-treatment | far-transfer |
| Problem solving (labelled near transfer by the authors) | Cohen d | 1.31 | researcher-designed | post-test | active-alternative | end-of-treatment | near-transfer |
Cited by
- Does learning to code improve general thinking?no effectconf: mediumgc: low