March 25, 2025
By Merel Vrancken
In Salay v. Slovakia the European Court of Human Rights held that the overrepresentation of Roma pupils in special education in Slovakia constituted discrimination. The ECtHR’s ruling in Salay v. Slovakia is very similar to that of the Grand Chamber in D.H. and Others v. the Czech Republic in 2007. Nonetheless, Salay falls short of the promise of D.H. It recognises the danger of biased testing to determine children’s enrolment in special or mainstream education, but fails to question the practice of testing itself. In addition, this blogpost argues that the divide between cases on Roma segregation in education and cases on segregated special education of children with disabilities created by the Court leads to problematic consequences.
Salay, a child of Roma ethnicity, undertook a test of school maturity in August 2004, before enrolling in primary school. In view of his results, he attended Year Zero of mainstream primary education. This year was set up for children from a disadvantaged background who did not have the requisite level of academic maturity to enrol in Year One. The next school year, Salay was enrolled in a mainstream class in Year One. He was retested in January 2006 due to the existence of learning difficulties and his transfer to a special class was recommended. The result of the test indicated a mild intellectual disability. A retesting in April 2006 confirmed this recommendation. His mother requested his transfer to a special class in June 2006.
From Year Two (2006-2007) onwards, Salay attended special classes for pupils with special educational needs. At that time all the pupils in those classes in his school were Roma. Salay was retested in June and November 2009. Again, it was held that his development was delayed and the results indicated a mild intellectual disability. He was tested again in January, June and October 2011. In September 2011, Salay’s parents requested his transfer to a mainstream class at the school, with an individual study plan. The January and October tests held that the applicant had a mild intellectual disability. The June test showed a normal development (though below average or borderline) and a possible learning disorder, but further testing was necessary to confirm this. All tests were accompanied by recommendations for Salay to remain in special education. Despite his parents’ request for a transfer to mainstream education, he completed his studies in special education.
The applicant lodged an anti-discrimination action in April 2014, indicating that an ever-growing number of Roma pupils were in special education. At that time, some 86% of all pupils in special classes were Roma. He argued that his placement in special education had been arbitrary, that his parents had not given their informed consent and that he had not been retested regularly. In addition, the physical equipment used in such classes and the curriculum followed was inferior to those of mainstream classes. His action was dismissed, as the applicant’s individual situation had been properly and repeatedly examined by the competent authorities.
Before the ECtHR, Salay complained that he had been discriminated against in the exercise of his right to education, in violation of the prohibition of discrimination under Article 14 ECHR, in conjunction with the right to education laid down in Article 2 First Additional Protocol (P1).
In its judgment, the Court accepts that the statistics on overrepresentation of Roma children in special education, together with non-culturally sensitive tests, were sufficient to ‘reveal a dominant trend of a general policy or measure exerting a disproportionately prejudicial effect on the Roma, a particularly vulnerable group’ (para 89). This created a prima facie case of discrimination.
In this respect, the Court notes that the allegations that ‘special education in Slovakia constituted an inferior standard of education and that there was persistent, widespread and systematic overrepresentation of Roma pupils in special education’, were corroborated by country monitoring reports by the European Commission against Racism and Intolerance, the European Committee of Social Rights, the Advisory Committee on the Framework Convention for the Protection of National Minorities, the CoE Commissioner for Human Rights, the UN Committee on the Rights of the Child and the UN Human Rights Committee (para 85). It moreover sums up many statistics. In 2008-2009, 86% of pupil in special classes in mainstream schools were Roma. In the period of 2008-2014, 30% of Roma pupils attended special schools and Roma pupils were 28 times more likely to be placed in special schools. In 2019, 63% of pupils in special classes were Roma, while Roma children consisted of 12.3% of all primary school pupils and the average proportion of children in special education was 3.2% (para 86).
The Court’s analysis of the objective and reasonable justification for the interference focuses on its proportionality. In contrast to the more collective proof of a prima facie case of discrimination it accepted, the Court assesses the proportionality of the interference in a very individualised manner. It focuses on the tests that laid at the basis of Salay’s enrolment in special classes. The Court affirms that as ‘the State had a positive obligation to avoid the perpetuation of past discrimination or discriminative practices disguised as allegedly neutral practices’, it must demonstrate ‘that the tests in question and the application of those tests were capable of fairly and objectively determining a person’s academic ability and intellectual capacity’ (para 97).
The ECtHR examines the tests used during the seven different moments on which the applicant was tested – although the Government had already indicated that only one test existed that took into account a pupil’s socially disadvantaged environment and that this test could not establish a mild intellectual disability, instead only serving to rule it out (para 34). The Court concludes that ‘at the very least, there is a danger that the tests used in the applicant’s case were culturally biased’ (para 107). Special safeguards needed to be applied to allow the authorities ‘to take into account the particularities and special characteristics of the Roma applicants who took those tests, in view of the high risk of discriminatory misdiagnosis and inappropriate placements’ (para 107). In this respect, there was no ‘systematic retesting with a view to monitoring any developments which might justify a pupil’s transfer back into mainstream education’, a request for transfer by the parents had not been accommodated and no other children at the school had ever transitioned back from special to mainstream education. Therefore, the placement was de facto permanent (paras 109-111).
Given all these elements, the Court concludes that Salay ‘received an education which did not offer the necessary guarantees stemming from the positive obligations of the State to undo a history of racial segregation in special education’ (para 114). The state failed to prove that it provided the safeguards needed to avoid the misdiagnosis and inappropriate placement of Roma pupils in special education with a more basic curriculum, in violation of Article 14 ECHR in conjunction with Article 2 P1.
The facts of Salay show clear parallels with those of the Grand Chamber case of D.H. and Others v. the Czech Republic. In both cases, Roma children were overrepresented in special education as a result of tests. The parallels in the Court’s analysis of both cases are also clear. As in the case of D.H., statistics are used in Salay to establish a prima facie case of indirect discrimination. These are taken from the monitoring procedures by several CoE and UN monitoring bodies, which are considered to constitute reliable evidence. The significance of this approach cannot be understated, as it considerably lowers the burden placed on an individual to prove segregation (or other types of discrimination) where such reports are available.
As in D.H., the Court critically analyses the tests used and concludes that they do not provide the necessary safeguards to ensure that Roma children are not misdiagnosed and misplaced into special education. In D.H., the Court’s consideration that ‘at the very least, there is a danger that the tests were biased and that the results were not analysed in the light of the particularities and special characteristics of the Roma children who sat them’ is of central importance to this assessment (D.H., para 201).
In Salay, the ECtHR identifies the same danger (para 207). In contrast to D.H., the Court does not conclude that the justification is insufficient because of this danger. After identifying it, the Court analyses the existence of possible additional safeguards within the placement process. It points out that systematic retesting could have been a safeguard to ensure that these children’s placement in special education was only temporary, but this safeguard was not in place in Slovakia. In stating this, it essentially holds that when the tests are not culturally sensitive, a solution would be to systemically retest the pupils.
Of course, given their inherent deficiencies, a systematic retesting of Roma pupils with culturally insensitive tests would not provide a safeguard at all. The Court implicitly proposes two solutions in this case: introducing culturally sensitive tests or systematically retesting the children. They are both problematic. The latter for the reasons just identified, the former because disproportionate testing as such is problematic from a human rights perspective. In effect, the Advisory Committee on the Framework Convention for the Protection of National Minorities has expressed its concerns about the disproportionately high testing of special needs in Roma children compared to other children in Slovakia (paras 219 and 222). The (disproportionate) testing is part of the problem, not the solution.
In Salay, the Court accepts collective proof of a prima facie case of discrimination, but takes a very individualised approach when analysing whether the interference is proportionate. This individualised approach leads to an overemphasis on the modalities of the testing procedure, instead of conducting a more encompassing analysis that could (and should) have questioned the practice of testing per se.
The Court exempts the state of the burden of proving that the tests as a whole were not discriminatory when it indicates that ‘there appears to be a consensus that diagnosing a mild intellectual disability in young children is a challenging process’ (para 98). If this is indeed the case, one may wonder why the state systematically tests these children regardless. It also raises the question of the added value of the Court’s analysis of the seven tests undertaken over the course of five years. The Court states that it was ‘called upon to ascertain whether efforts were made in good faith to implement non-discriminatory testing’ (para 100). However, discrimination can also exist in good faith: good faith is irrelevant where there is a discriminatory effect. The Court could have sufficed by concluding that the tests applied in the applicant’s case and in general were not ‘capable of fairly and objectively determining a person’s academic ability and intellectual capacity’, given the overall figures on discriminatory placement of Roma in special education.
The case of Salay v. Slovakia raises the broader question of the extent to which the segregation of children with disabilities into special education is compatible with the ECHR. If this is not the case, the consequence is that the overrepresentation of Roma pupils in special education is equally incompatible with the Convention.
In this case, the Court recognises that in special classes for children with (mild intellectual) disabilities, a more basic curriculum is followed than in mainstream education (para 114). This is problematic for the Roma children who are overrepresented in such special education. However, it does not seem to be problematic for the children with disabilities: this education is presumed to be adapted to their special educational needs. In this context, Lewis has stated that the statements made by the ECtHR in the cases of Horváth and Kiss and D.H. ‘can be read as the Court (at worst) approving of, or (at best) ignoring the fact that everyone in special schools is denied their right to education. By focusing on the misdiagnosis, these cases have had the unintended effect of legitimizing a segregated schooling system as such.’ In the case at hand the Court considers it ‘unnecessary to make a definitive ruling’ on the question of whether special education for children with special educational needs can be justified as serving a legitimate aim (para 93). The Court again ignores the question, though it reserves its discretion to decide that special education does not serve a legitimate aim.
In spite of the Court’s ambivalent stance on special education in its case law on Roma segregation in education, its case law on the right to education of children with disabilities seems to confirm Lewis’ fears. None of the principles developed within this case law indicate that segregated special education of children with disabilities is problematic under the Convention. In four (inadmissible) cases of 1989, the European Commission on Human Rights found no element of discrimination in a policy in which children with disabilities are educated in special education where necessary, ‘but rather the contrary, the special requirements of certain disabled children being catered for where necessary’ (Connolley v. UK, P.D. v. UK, P.D. and L.D. v. UK and J. and B.L. v. UK). In its decision on admissibility in Dupin v. France, the Court held that it was shown that a child with autism’s part-time educational support in special education ‘is conducive to his development’ (convient à son épanouissement, own translation). The Court was thus convinced that the segregated special education the child received was beneficial to him. This contrasts with the clear denunciation of special education under the Convention on the Rights of Persons with Disabilities (CRPD).
Under the ECHR, segregated special education for children with disabilities hence does not seem to be problematic. The Court does not seem to require any justification for the existence of educational segregation of children with disabilities. This approach contrasts with the ECtHR’s general case law on Roma segregation in education, where the Court has recognised segregation as a specific form of discrimination, requiring an objective justification for a situation of segregation (Elmazova and Others v. North Macedonia, paras 73 and 78).
By adopting two strikingly different approaches in cases of Roma segregation and disability segregation in education, the Court forces itself to adopt an either-or approach towards the ground for segregation. In this case, it means that the applicant can only have suffered discriminatory treatment in the form of segregation if he does not have a disability. This is essential to the Court’s assessment – and one of the reasons the Court legitimises (or even requires) a systematic retesting. This systematic retesting can be linked to the medical model of disability, which has been rejected under the CRPD. As a result of this approach, the applicant – and his peers – lose while winning the case. If the Court would fully accept that in addition to being of Roma ethnicity, the applicant had specific educational needs due to his socioeconomic disadvantage, he could have been granted a right to reasonable accommodation under the social or human rights model of disability. This model does not require a ‘medical impairment’ for granting specific rights and safeguards. Instead it focuses on the question whether individuals experience barriers in their participation in society. The ECtHR’s case law recognises that persons with disabilities have a right to reasonable accommodation in mainstream education (Çam v. Turkey, paras 65-67). By considering this a case of either problematic racial segregation or unproblematic segregation on the basis of disability, the applicant loses his right to reasonable accommodation along with the recognition of his socioeconomic disadvantage as an aspect of his vulnerability as a Roma child.
This blogpost criticises several aspects of the ECtHR’s reasoning in its judgment in Salay v. Slovakia. It overemphasised the importance of tests in deciding what schooling would be the best fit for children. It hereby essentially stated that where tests were inherently deficient for testing Roma children for intellectual disabilities, as they were not culturally sensitive, a systematic retesting of these pupils with those same tests would provide a safeguard against their misdiagnosis and inappropriate placement in special education. Aside from coming to this problematic conclusion, it forgot to question the practice of (disproportionately) testing Roma pupils per se. In addition, this post argues that due to the two strikingly different approaches adopted by the Court in cases of Roma and disability segregation in education, the applicant in this case has lost and won at the same time. In finding a violation in this case, the Court fragmentated the applicant’s identity (and vulnerability) by considering this a case of either problematic racial segregation or unproblematic segregation on the basis of disability. In combining both approaches, the Court could have recognised that the applicant has a right to reasonable accommodation in education along with his right not to be segregated.
Strasbourg Observers is an academic blog that discusses recent developments at the European Court of Human Rights. The editorial team of Strasbourg Observers is based at the Human Rights Centre of Ghent University and the Centre for Government and Law of Hasselt University.
Strasbourg Observers would not be possible without the valued work of countless guest bloggers. Visit the “Contact” page on our website to find out how to become one!
