At the same time, there are legitimate concerns as to the interests of protection of personal data. 'Examples of misuse of data for the purpose of persecution are well documented and partially explain the reluctance of some countries to compile ethnic data There is a need to' strike a balance between these protection needs and the requirements to measure implementation of minority policies. Recommendation 28 tries to take some of these concerns into account as it Provides that data collection should be conducted in an ethnically sensitive manner and on a voluntary basis, consistent with the right of minorities to self-identification, and with full respect for the privacy and anonymity of the individuals concerned and in accordance with international standards of personal data protection. The Applied Research Association recalls that everyone must be left free to declare, or not declare, their ethnicity. The decision to rely on self-declaration of ethnic or national origin is dictated by the will to respect free consent. The subjective belonging is indeed an important element in the determination of ethnic origin. Self-estimation can however result in figures that do not reflect the reality in the sense that perception may be the source of discrimination rather than identity. Monitoring implementation on the basis of ethnic data collected on the basis of self-identification may not include those who are being discriminated on grounds such as ethnic origin. There is also a methodological deficiency when it comes to statistics based on self-identification. Identification to a varying degree depends on the context in which data is Collected, for instance by police or at social encounters:Most importantly, data gathered in such a manner can rarely be replicated. The potential for stigmatization and the interpretation of ethnic data may also give rise to concerns. A disproportionately high number of certain groups or minorities in detention centres for example will reveal police and court practices, such as severer sentencing, and not be representative for the criminal propensity of this particular group. Finally, the Applied Research Association also notes that the lack of readily available data does not mean that empirical patterns and trends on minority cannot be identified. Other methods, such as algorithms and survey techniques may be used to extract information on Minority inclusion. aradpd hopes that these thoughts will inform the drafting of the final recommendations of this Forum. Thank you for .your attention. 2