The alleged link between creativity and dyslexia: Identifying the specific process in which dyslexic students excel

The design has multiple problems that each independently undermine the conclusions. All participants in Study 1 attended one school in Milan chosen for its reputation welcoming dyslexic students — a self-selected population on both sides. The creativity test has only three items per subtest, each scored 1 to 3, giving a total range of 3 to 9. The group difference driving the paper is one-third of a point near the floor of this scale. Three parallel comparisons were run with no correction; only one crossed the threshold, and it would not survive correction. Study 2 uses a completely different group of ten clinical patients to rule out confounds found in Study 1's school sample — populations that may differ in severity, background, and comorbidities. The test requires students to write brief justifications for their choices (to distinguish creative from bizarre responses), but the paper never reports how these were scored — a consequential omission in a study about dyslexia, where writing ability is often impaired. The test was developed by the senior author, creating a circularity the paper does not acknowledge.

The paper runs three independent comparisons (one per creativity subtest) without adjusting the threshold. The sole positive result lands at p = .04, which does not survive even the simplest correction for three tests. No power analysis was reported; the sample can detect only large effects, and the observed difference is small. No effect sizes or confidence intervals appear anywhere. Study 2 computed roughly 30–40 correlations across three creativity subtests and multiple reading, IQ, attention, and working memory measures in ten students. Finding one correlation at .01 in a matrix that large is expected by chance alone. The specificity claim — that the advantage is limited to connecting — was never formally tested; it rests on observing that one of three comparisons crossed a threshold and two did not, which is not the same as showing the effects differ from each other. The intelligence–connecting correlation was –.53, a large negative association, but failed to reach the threshold only because ten observations cannot reliably detect any relationship. The paper treats this as "no relationship" when it means "unable to tell."

Every core claim is disconnected from its evidence. The creativity measure is a recognition task — students pick options from a list — not a production task. Choosing rare options is scored as creative regardless of reason; misunderstanding, contrarianism, or random selection would produce the same scores. The specificity claim was never tested with the comparison that would establish it. The null correlations with IQ, attention, and working memory in ten students cannot rule out those factors; the study was too small. Study 2's ten clinical patients are a different population from Study 1's school sample, so even if the correlations were trustworthy, they would not explain the school findings. The distance between "picked rarer options on three multiple-choice items at one school" and "possesses enhanced creative connecting ability" is too wide for the data to bridge.

The actual finding — a third-of-a-point difference on one subtest that does not survive correction — is transformed into "identifying the peculiar intellectual functioning associated with DD" and a basis for educational recommendations. Two smuggled assumptions do real damage: treating non-significant results on two subtests as evidence of no difference (rather than as uninformative given the small sample and coarse test), and treating null correlations in ten students as ruling out confounds (when the study lacked the power to detect even large relationships). The interpretive framework is structured so that disconfirming evidence can always be attributed to measurement limitations — if a standard creativity test finds nothing, that is because it is too coarse-grained, and the authors' test should be used instead. The opening paragraph lists Leonardo da Vinci, Einstein, and Picasso as probable dyslexics — an appeal to narrative that primes the reader before the caveat about unreliability arrives. The closing paragraph pivots to educational applications and names the first author's reading intervention program, treating a marginal result as established enough to guide practice.

The paper lacks pre-registration, open data, open code, conflict-of-interest disclosure, funding disclosure, ethics-board documentation, and sample-size justification. The creativity test was developed by coauthor Antonietti, and the study effectively promotes the instrument — the discussion argues it detects effects other tests miss — but this conflict is never disclosed. The limitations section acknowledges small sample size, comorbidity, and single-school recruitment, but omits the problems that matter most: no correction for multiple comparisons, the coauthor's stake in the test, and the inability of ten participants to support the null-result conclusions drawn from them.

The paper's evidence base depends entirely on the WCR, which has no independent validation by outside groups. At least one citation appears misattributed: Ingesson 2006, cited for claims about visual processing preferences, is actually about the stability of IQ scores in dyslexia. The opening premise — that people with dyslexia are "particularly creative" — is established through advocacy books (Davis & Braun 1994, Eide & Eide 2012), cited with "in fact" and "recognised" as though they represent established science. Neuroscience citations (Casanova et al. 2002) are stretched well beyond what the original papers support. The paper omits the critical literature challenging the "dyslexia as gift" framing. The literature review does cover the main prior studies and acknowledges some null findings, including Mourgues et al. 2014, which found the opposite pattern.

There is no industry funding or financial conflict. The conflicts are professional: the senior author developed the only creativity test used, and the first author promotes her reading intervention in the conclusion. The study uses the authors' test, finds that their test detects something other tests miss, and concludes their test is better. No independent measurement breaks this loop. No pre-registration, no independent scorer, and no blinding are reported. The scorer knew which students had dyslexia diagnoses, and in a school chosen for accommodating dyslexic students, group membership would often be apparent.