Factors influencing work participation of adults with developmental dyslexia: a systematic review

Qualitative metasummary is a reasonable fit for cataloging what the literature reports about dyslexia and work. The search strategy is clearly described, inclusion criteria are explicit, two reviewers screened independently with a third for disputes, and quality thresholds were set for both qualitative and quantitative studies. Table 3 displays every extracted factor, its source study, and its direction code, which is unusually transparent. What lowers confidence: the core analytic step — extracting 318 factors and coding each as positive or negative — has no reported inter-rater reliability. The reported Kappa scores (0.82 for qualitative, 0.66 for quantitative) measure agreement on study quality, not on factor extraction. The paper acknowledges classification was "occasionally difficult" and a third reviewer was "consulted several times," but provides no agreement statistics for the step that generates every number in the results. Participants across the included studies were recruited from dyslexia clinics, advocacy organizations, and universities — well-adapted adults who never sought services are absent, inflating negative findings.

The review performs no statistical tests. Its sole quantitative metric, the "manifest frequency effect size," is a proportion: the number of quality-passing studies mentioning a factor in a given category divided by 21 (the total qualifying studies). Despite the name, this is not an effect size — it carries no information about how large, how common, or how consequential any factor is. A category mentioned once in passing in 15 studies scores higher than a category with a large, well-measured association in 5 studies. No confidence intervals are provided. Several quantitative studies in the review reported regression coefficients and group comparisons, but the review discards all of that, reducing each study to a binary: was a factor mentioned, and was it positive or negative? The personal factors category scored 100%, but this reflects the category's breadth (it includes age, gender, coping strategies, self-esteem, and stress), not any meaningful finding. The quality threshold for quantitative studies was set by scaling proportionally from the qualitative threshold, a method the authors invented without validation.

The paper's three core claims are all weaker than presented. First, the claim that dyslexia affects "nearly all domains" of work functioning is close to inevitable given the method: extracting 318 factors from 33 studies and sorting them into a comprehensive taxonomy will populate most categories regardless of the topic. No comparison against another condition was run to test whether this breadth is specific to dyslexia. Second, "mostly negative" depends entirely on the ratio of minus-to-plus codings, but the coding step has no measure of reliability, and the paper flags cases where the coding was debatable — "occupation choice" was coded negative because dyslexia constrained the choice, though the authors weren't sure participants experienced this negatively. Third, the claim that dyslexia's impact "increases over the life course" requires longitudinal evidence, but only two of 33 studies were longitudinal, and neither tracked whether dyslexia spreads across ICF domains over time. The paper infers this from the difference between childhood literature (where the main issues are reading and spelling) and its adult findings (where factors span all domains), but that comparison cannot distinguish developmental progression from selection effects or cohort differences.

The life-course claim — "within each domain the impact of DD increases over the life course" — is stated in the conclusion as a finding. It is not. The review collected cross-sectional snapshots of adults. The claim is a narrative bridge between background literature on children and the review's adult data, which cannot distinguish genuine developmental progression from the fact that working life introduces demands (social relationships, career management, workplace politics) that childhood doesn't. The "mostly negative" conclusion rests on counting minus-sign versus plus-sign codes, but the direction coding has no reported reliability, and the paper concedes that some codings were uncertain. The "lack of positive attitudes" claim mistakes a skew in factor counts for an absence — the review itself reports positive factors (perseverance, creativity, self-esteem), and the apparent teacher exception may simply reflect that teachers were the only occupation studied in multiple dedicated papers.

The search strategy, inclusion criteria, and quality assessment instruments are fully documented — enough to attempt replication. The limitations section is unusually candid, flagging population heterogeneity, direction-coding ambiguity, the range of meanings within factors, and the ICF's inability to capture causal mechanisms. However, the factor extraction data, coding rules, and disagreement records are not publicly available, and no inter-rater reliability is reported for the step that produces every number in the paper. There is no pre-registration, no open data, and no funding disclosure.

The 33 included studies are appropriately characterized, and Table 3 allows readers to trace each claim to its source. The problems are in the framing. Background claims about dyslexic strengths — visual thinking, pattern recognition, creativity — cite practitioner books (Reid and Kirk 2001, West 1997) rather than empirical studies, and present contested findings as settled. The discussion cites Logan (2009) to support the claim that "many employees with DD start their own businesses and become entrepreneurs." We reviewed Logan 2009 and rated it nonsense: its headline finding that 35% of entrepreneurs are dyslexic came from an unvalidated screening checklist with a 7% response rate, and its coping-strategy comparisons showed no difference between groups. The entrepreneurship claim is peripheral here, but its inclusion as established evidence shows uncritical citation of a source whose problems — a tiny response rate, null results — were visible in the original paper.

No competing interests are declared and no funding source is mentioned. One co-author (Heerkens) developed both ICF framework extensions the paper uses and served as the third reviewer on classification disputes, creating a mild self-validation loop. This is routine in academic research and does not visibly shape the findings.