The Evidence on Teaching

Interventions to Dispel Neuromyths in Educational Settings - A Review

Rousseau L · 2021

grade Dreviewindependentnot-applicablenumbers spot-checked
Sample
Narrative review; roughly 20 primary intervention studies discussed, plus two systematic reviews it leans on: Torrijos-Muelas 2021 (24 prevalence articles, 2012-2020) and Newton & Salvi 2020 (33 articles, 37 samples, 2009-2020)
Population
Educators and teacher trainees
Design
Published in Frontiers as a 'Perspective', not a systematic review - no search protocol, no inclusion criteria, no pooling, hence grade D. Reviews whether neuroscience training protects against neuromyths, whether refutation-based interventions work and endure, and - the question that matters most for practice - whether reduced belief translates into changed teaching. The author is himself a primary-study author in the field (Rousseau et al. 2018; Rousseau & Brabant-Beaulieu 2020) and cites his own work, but has no stake in any programme being evaluated.
Key findings
Prevalence of neuromyth belief among educators did not decline over the past decade (weighted mean 52.1%, range 27.3-84.5% across 24 studies; the learning-styles item alone 58-97.6%, weighted mean 89.1%, across 37 samples, with no downward trend over eleven years). Neuroscience training alone is insufficient - a single course produces no reduction, several courses produce a "modest" one. Refutation-based interventions DO reduce endorsement, and in most studies the reduction persists (1 month, 6 months, even 2 years); Ferrero 2020a/b is the exception, with a "V" pattern returning almost to baseline after 30 days. The harder problems are downstream: reduced belief has never been shown to change classroom behaviour, self-reported belief change may be social desirability, and intuitive/anecdotal evidence beats scientific evidence in teachers' own accounts of why they believe. Two findings cut against the practical case for the whole prevalence literature: award-winning teachers endorse neuromyths at the same rate as everyone else (Horvath 2018) and neuromyth endorsement does not predict pre-service teachers' grades (Krammer 2021), so no causal chain from belief to bad teaching has been demonstrated.
Genetic confound
Not applicable; outcome is educator belief.

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Neuromyth belief prevalence over the past decadeweighted mean endorsement across 24 surveys, and trend52.1% (range 27.3-84.5%), no decline 2012-2020 (Torrijos-Muelas 2021)self-report survey2012-2020nonenot-applicablenon-cognitive
Learning-styles myth prevalence among educators specificallyweighted mean endorsement across 37 samples, and trend89.1% (range 58-97.6%), no decline 2009-2020 (Newton & Salvi 2020)self-report survey2009-2020nonenot-applicablenon-cognitive
Refutation-text effect on neuromyth endorsement, immediatepre-post drop in % endorsedlarge where measured - e.g. Grospietsch & Mayer 2018, 74% to 22% on the seven targeted myths; Newton & Miah 2017, learning-styles endorsement 64% to 10%self-report surveyimmediately post-interventionbusiness-as-usualend-of-treatmentnon-cognitive
Durability of the refutation effectfollow-up status across studiesmaintained in most studies (Lithander 2021 at 1 week and 1 month; Menz 2021c at 6 months; Kowalski & Taylor 2017 at 16 weeks AND 2 years) but NOT in Ferrero 2020a/b, where endorsement rebounded almost to baseline by 30 days. Heterogeneous, not uniformly fadeoutself-report survey1 week to 2 years post-interventionbusiness-as-usualunder-1yrnon-cognitive
Belief change translating into intention to change practice (the result that cuts against the practical advice a debunked verdict implies)% still intending to use the debunked practice after a successful refutationNewton & Miah 2017: endorsement fell 64% to 10%, yet 31.6% still planned to account for learning styles in their teaching, and 89% of those agreed "even though there is no evidence base... it is my experience that their use in my teaching benefits student learning". Rousseau & Brabant-Beaulieu 2020: intention to use VAK fell only from 100% to 60% after converging scientific AND personal-anecdote counterevidenceself-report surveypost-interventionbusiness-as-usualend-of-treatmentbehaviour
Translation of reduced belief into observed classroom practicequalitative verdictnever measured - no study has behavioural classroom data on refutation-based interventions; the review calls this "a major limitation" of the whole approachunknownn/anonenot-applicablebehaviour
Does believing neuromyths actually make you a worse teacher? (finding cutting against the entire prevalence literature's premise)qualitative verdict plus two primary resultsno causal chain demonstrated. Horvath 2018 found internationally award-winning teachers endorsed neuromyths at nearly the same rate as non-award-winning teachers; Krammer 2021 found endorsement had no effect on pre-service teachers' academic grades. Rousseau: "there is still no supporting data for a causal chain from educators' beliefs in neuromyths to adoption of poor teaching practices and/or negative impact on students' academic achievement"mixedn/anonenot-applicablebehaviour
Instrument artefact in the prevalence estimates (cuts against Dekker/Macdonald headline numbers)endorsement under scenario-based vs true/false formatsTovazzi 2020 gave in-service teachers both a Dekker-style true/false survey and the multi-option Neuroscience against Neuromyths Questionnaire, in which realistic classroom scenarios offer neuromyth-based, evidence-based and distractor choices. Teachers were markedly LESS likely to pick neuromyth-derived options. Single-statement true/false surveys "might possibly overestimate the prevalence of neuromyth beliefs"mixedcross-sectionalactive-alternativenot-applicablenon-cognitive

Cited by