A child can need help in a classroom without possessing the certificate that makes help an enforceable entitlement. That gap is the subject of Rahkyung Kim and SeonYeong Yu’s comparative article on special-education eligibility in South Korea. Its central argument is not that Korean law simply recognises too few kinds of disability. It is that the administrative route to recognition may favour conditions that can be documented consistently across education, welfare, employment quotas and university admissions.
The paper gives this proposal a name: the administrative certification model. It is a useful addition to familiar medical, social and human-rights accounts of disability because it asks a different question. Not only what disability is, but who is authorised to certify it, for which purposes, and under what institutional pressures?
The puzzle is not only the percentage
In 2025, South Korea identified about 2.2 per cent of school students as eligible for special education. The published figures cited by Kim and Yu are much higher in several comparison systems: roughly 15 per cent in the United States, 7.5 per cent in Germany, 19.5 per cent in England and 27 per cent receiving educational adjustments in Australia. The authors repeatedly warn that these indicators are not directly equivalent. Countries count different age groups, services and categories. Their point is therefore not to construct a league table, but to explain why Korea remains an unusually low-identification case even after its statutory categories have expanded.
The composition of the identified population sharpens the question. Intellectual disability, autism and developmental delay account for more than four-fifths of Korean students formally eligible for special education. Learning disability accounts for less than one per cent and emotional or behavioural disorder for about 1.4 per cent. Kim and Yu read this concentration as consistent with a system in which diagnostically clear categories travel more easily through administrative decision-making than functional or context-dependent educational need.
This is where the paper moves beyond the claim that a narrow definition produces a narrow service. Korean law already requires two judgements: a student must fall within a recognised category and be assessed as needing special education. Yet the practical infrastructure of diagnosis overlaps with a wider disability-registration system. Registered status can affect welfare benefits, whether an employer satisfies a disability-employment quota and access to supernumerary university places. A classification made credible across all these settings may need to be more standardised and medically defensible than a classroom judgement about what would allow a particular student to participate.
A mechanism, not a conspiracy
The article is strongest when it refuses an intentionalist explanation. The authors do not claim that employment quotas or admissions routes were designed to restrict educational support. Their proposal concerns institutional layering. Distinct policies acquire a shared dependence on disability-related status; officials then face pressures for consistency, entitlement integrity and defensible thresholds. The cumulative effect can be restrictive even when each policy has a compensatory purpose.
The international comparison is used to clarify this mechanism. The United States separates school eligibility under IDEA from Social Security disability benefits and employment protections. Germany separates pedagogical identification from the card used for tax and employment-quota purposes. Australia records educational adjustments independently of welfare and income-support eligibility. Finland’s support system has historically relied more heavily on pedagogical need than on diagnosis. Korea and Japan are presented as more tightly coupled cases, both with low formal identification rates and a stronger concentration in particular diagnostic categories.
The paper also shows what non-certification does. A student outside the special-education category lacks an entitlement to an individual education plan, specialist teaching and mandated accommodation. Difficulty may instead be recorded as underachievement or maladjustment. Kim and Yu call this second-order educational exclusion: exclusion from the system designed to prevent exclusion. It affects the institutional record as well as the child. Needs that never enter official statistics cannot easily become evidence for the next allocation of staff and services.
What the evidence can—and cannot—carry
This is a conceptual and comparative policy analysis, not a causal evaluation. Its strengths lie in bringing law, official statistics and disability theory into one explanatory proposal; in separating formal systems from their practical interdependence; and in treating uncategorised need as an institutional outcome rather than a personal failure. The discussion of medically fragile and twice-exceptional students also prevents the argument from being reduced to one disputed percentage.
But the evidence does not observe certification decisions as they happen. There are no interviews with clinicians, committee members, families or students whose referrals were declined or redirected. The proposed pressure towards medically verifiable thresholds is inferred from the architecture and the cross-national pattern. The authors acknowledge this clearly, describing their model as an evidence-informed interpretation requiring direct investigation.
The coupled–decoupled contrast may also be too clean. Educational support in nominally decoupled systems still depends on categories, school resources, parental advocacy and uneven professional judgement. Conversely, some Korean students may receive meaningful general-education support without appearing in special-education counts. Stigma, family preferences, teacher preparation, school incentives and national reporting conventions could also help explain the pattern. These alternatives do not invalidate the model, but they mean that administrative coupling has not yet been shown to be the decisive cause.
There is a further normative tension. Separating educational identification from other disability entitlements may widen access to support, but it could also create multiple classifications that people must navigate, repeat and contest. Decoupling is not automatically humane. Its value would depend on whether it reduces gatekeeping without multiplying incompatible assessments or weakening rights elsewhere.
An autopoietic ecology reading: when need becomes a status
My autopoietic ecology (AE) interpretation begins with the transformation of an educational difficulty into an administratively actionable object. A teacher may recognise that a pupil is struggling. A family may document the effects. A clinician may supply a diagnosis. A committee may confer eligibility. These are not interchangeable descriptions. The AE Concept Guide uses semantic transduction for the question of what changes as an expression moves between settings. Here, need becomes influential only if it survives translation into a status that can alter a later decision.
The paper’s double exclusion is especially important for AE. If an unmet need does not enter the official record, the system not only withholds present support; it removes part of the evidence through which future provision might be revised. The activity of classification helps reproduce the informational conditions of the next classification. Low recorded eligibility can then appear to justify a limited service, while the limited service has fewer routes for recording what it excludes.
This is an EdgeLab extension, not a finding demonstrated by Kim and Yu. A conventional street-level-bureaucracy or disability-policy analysis may explain the same process without AE. AE adds value only if following recognition, translation and later allocation reveals how one decision changes the conditions of the next. If decision-level evidence shows that broader entitlements play no role in educational thresholds, the proposed link—and the AE reading built around it—would need substantial revision.
What would test the argument?
The most informative evidence would begin before certification. Researchers could follow referrals through diagnosis, committee review, service assignment and appeal, recording who is rejected, redirected or informally supported. Comparable cases could identify whether similar classroom needs receive different decisions when a diagnosis carries consequences outside education. Interviews would reveal whether officials actually consider consistency with welfare, employment or admissions systems, rather than researchers inferring that pressure from institutional design.
Longitudinal comparisons would be stronger still. If a jurisdiction separated educational support from non-educational certification, did identification broaden, did access become earlier, and were students better able to participate? Did families face fewer barriers, or merely a new sequence of assessments? Measures should include waiting time, appeal, continuity of support and the labour required from families and teachers—not only the number of certificates issued.
For practice, the paper suggests a cautious principle: provide a route to support on evidence of educational need while questions of high-stakes status are still being resolved. Record unmet and redirected need as carefully as successful eligibility. Otherwise, the apparent precision of certification can conceal a simple institutional failure: a child’s difficulty is recognised by everyone who encounters it but never becomes the kind of fact that changes what happens next.
Reference: Kim, R., and Yu, S. (2026). Exclusion Through Eligibility: How Administrative Disability Certification Limits Access to Special Education in South Korea. Education Sciences, 16(9), 1565. Published 20 September 2026. https://doi.org/10.3390/educsci16091565.
EdgeLab is an independently organised research network. It is not formally recognised as a research group, centre or unit by the Faculty of Education or the University of Cambridge. Institutional affiliations refer to individual contributors and do not imply institutional endorsement.
Leave a Reply