5. Challenges to Collecting Ethnic Data on the Roma Minority in Europe

This was addressed in Chap. 1 (Sect. 1.​3), when introducing the two main themes of the book.

Chapter 3 analysed the key elements, rules and principles of ethnic data collection. Chapter 4 focused on the benefits, risks, data sources and methods of this human rights tool.

Council of Europe (CoE), Roma and Statistics (22–23 May 2000), paras. 9 and 19. Messing (2014), pp. 811, 821 and 822. ERRC (2013b), pp. 7 and 55. FRA (2012a), p. 175. Guy et al. (2010), pp. 5 and 10 and 11. McDonald and Negrin (2010), p. 10. Hollo (2006), p. 4. Ringold et al. (2005), p. 187. Petrova (2004), p. 6. Landman and Carvalho (2010), p. 116. Limitations to data comparability will be considered in Sect. 5.6.2. See also Chap. 4 (Sect. 4.​1.​4) on implementing, monitoring and evaluating policies as one of the benefits of ethnic data collection.

Council Recommendation on effective Roma integration measures in the Member States (9 December 2013), art. 3.4. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 12. CoE, Roma and Statistics (22–23 May 2000), paras. 26 and 27. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 11. ERRC (2013b), p. 37. Guy et al. (2010), pp. 10 and 11. McDonald and Negrin (2010), p. 10. Ringold et al. (2005), pp. 175, 187 and 188. Krizsán (2001), p. 179. See Chap. 4 (Sect. 4.​2.​1) for a discussion on the exposure of bad government policies and insufficient measures as a fear or risk surrounding ethnic data collection.

Guy et al. (2010), pp. 5, 10 and 11. The identification of good practices was mentioned as one of the benefits of ethnic data collection in Chap. 4 (Sect. 4.​1.​2).

CoE, Roma and Statistics (22–23 May 2000), paras. 14, 26 and 27. Hollo (2006), p. 11. Krizsán (2001), pp. 178 and 179. See Chap. 4 on the uncovering of indirect discrimination (Sect. 4.​1.​2) and on the use of ethnic data to support discrimination claims in legal proceedings (Sect. 4.​1.​5) as benefits of ethnic data collection.

For instance, to confront governments about the lack of progress made. Abdikeeva (2014), p. 25. Gray (2009), p. 55. FRA (2008), p. 111. Ringold et al. (2005), pp. xxiv, xxv, 166 and 172. Project on Ethnic Relations (PER) (2000), p. 10. Section 5.8.1 will zoom in on awareness-raising as a key principle of ethnic data collection on Roma. Awareness-raising was identified as the third benefit of ethnic data collection in Chap. 4 (Sect. 4.​1.​3).

In addition to the international and European bodies mentioned in this section, it also concerns non-governmental organisations (NGOs), including the European Network Against Racism, Open Society Foundations (OSF) and the European Roma Rights Centre (ERRC). The role of NGOs will be discussed further in Sect. 5.2.2 on the added value of research data on Roma. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 15.7. Farkas (2017), p. 32. European Union Agency for Fundamental Rights (FRA) (2011), p. 22. UNICEF (2011), pp. 39–41. OSF (2010), pp. 29, 30 and 40. Halász (2008), p. 26. Open Society Institute (OSI) (2006), p. 2. Ringold et al. (2005), pp. 187 and 188. European Commission (2004), p. 47.

For instance, the 1999 meeting on the Roma and crime data was co-organised by the PER. There was also a 2-day round table on the issue of Roma and statistics in 2000. CoE, Roma and Statistics (22–23 May 2000), para. 18. PER (2000), pp. 34 and 35.

CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 46. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 13. ERRC (2013b), p. 37.

See, for example: CESCR Committee, Concluding observations on Greece (27 October 2015), paras. 9, 10, 28, 34 and 40. CESCR Committee, Concluding observations on Ireland (5 June 2002), para. 33.

Report of the Independent Expert on Minority Issues on the Implementation of General Assembly Resolution 60/251 of 15 March 2006 entitled “Human Rights Council” (2 February 2007), para. 91.

ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 12. ECRI, Fifth Report on Hungary (19 March 2015), paras. 106–108.

ACFC, Fourth Opinion on Slovakia (3 December 2014), para. 26. ACFC, Second Opinion on the Czech Republic (24 February 2005), paras. 10, 34, 37, 51 and 191. ACFC, Third Opinion on Hungary (18 March 2010), paras. 56, 60–62, 127, 134 and 135.

McDonald and Negrin (2010), pp. 35 and 36. Hollo (2006), p. 39. Barton (2004), p. 29. The importance of ethnic data to implement, monitor and evaluate policies was previously highlighted in Chap. 4 (Sect. 4.​1.​4).

Brüggemann and Bloem (2013), p. 535.

The 12 participating States were Albania, Bosnia and Herzegovina, Bulgaria, Croatia, the Czech Republic, Hungary, Macedonia, Montenegro, Romania, Serbia, Slovakia and Spain. Moldova, Norway, Slovenia, and the United States of America had observer status. FRA and United Nations Development Programme (UNDP) (2012), p. 8. In Chap. 4, the acquiring of information on social and economic integration levels (Sect. 4.​1.​1) and the uncovering of discrimination and the identification of good practices (Sect. 4.​1.​2) were cited as two of the five main benefits of ethnic data collection.

For instance, data on school enrolment, attendance rates and completion rates facilitate the monitoring of primary school completion rates among Roma. Brüggemann and Bloem (2013), pp. 521 and 535. Waldron (2011), p. 47. McDonald and Negrin (2010), pp. 19 and 20. Gray (2009), p. 61. See Chap. 4 (Sect. 4.​1.​4) for a discussion on the implementation, monitoring and evaluation of policies as a benefit of ethnic data collection.

This includes both new and already existing data. The Decade Indicator Working Group proposed a unified methodology to track and report on progress made in the inclusion of Roma. The Experts Group on Data and Measurement identified gaps in existing data and suggests improvements. See, for instance: Decade of Roma Inclusion Steering Committee (2004), pp. 5–7.

A 2008 a results-based monitoring workshop supported by UNDP helped the Macedonian government to reformulate the monitoring of their action plan. The Serbian Presidency to the Decade organised a similar workshop in 2009 and the Slovakian Presidency in 2010. The World Bank, the UNDP, the CoE and the Organization for Security and Co-operation in Europe (OSCE) Office for Democratic Institutions and Human Rights supported the Decade, as did so non-governmental organisations such as the ERRC and OSF. The European Union Platform for Roma Inclusion coordinated. McDonald and Negrin (2010), p. 22. OSF (2010), p. 49.

Commission Communication, An EU Framework for National Roma Integration Strategies up to 2020 (5 April 2011). ERRC (2013b), p. 42. FRA and UNDP (2012), p. 9.

In addition to experts from Bulgaria, the Czech Republic, France, Finland, Hungary, Italy, Romania, Slovakia, Spain and the United Kingdom (UK), the ad hoc working group includes experts from the European Commission, Eurofund and UNDP. Commission Communication, Steps forward in implementing National Roma Integration Strategies (26 June 2013), p. 3. FRA (2013a), p. 191.

The LERI programme (2013-ongoing) is carried out in 22 localities in 11 Member States (Bulgaria, the Czech Republic, Finland, France, Greece, Hungary, Italy, Romania, Slovakia, Spain and the United Kingdom). FRA (2014), p. 171.

Section 5.8.2 points to active participation of Roma in ethnic data collection as a prerequisite to render such practices effective and successful. See also Chap. 1 (Sect. 1.​2.​4) on the key role of cultural identity in Roma inclusion, Chap. 2 (Sect. 2.​3.​2) on the notion social inclusion, Chap. 9 (Sect. 9.​2.​4) on the need for consultation and participation of local communities in positive action for Roma, and Chap. 10 (Sect. 10.​1) on the importance of a bottom-up approach to positive action.

See Sect. 5.1.1 on international and European calls for data on Roma.

Brüggemann and Bloem (2013), pp. 520 and 535. Ringold et al. (2005), p. 187. OSF (2010), pp. 32 and 69. Gray (2009), pp. 55 and 60 and 61. Davidović and Rodrigues (2010), pp. 169 and 170.

This is, for instance, the case in Serbia, Slovakia and Romania. ERRC (2013b), pp. 50, 67, 75 and 76.

For example, official population figures on Roma are available in Bulgaria and the Netherlands, but unavailable in Albania, Belgium, Finland, France, Germany, Italy, Lithuania, Spain and Sweden. FRA (2013b), p. 7. OSF (2010), p. 28. Ringold et al. (2005), pp. xviii and 28. The limitations of official statistics on Roma will be considered in Sect. 5.2.1. For a general consideration of official statistics as a source of ethnic data, see Chap. 4 (Sect. 4.​3.​1).

For instance, the Greek government has data on Roma in primary education while there is no official data collection and local Bulgarian administrations have informal data on the geographic location of Roma communities. CoE, Roma and Statistics (22–23 May 2000), para. 21. O’Reilly (2014). OSF (2010), pp. 20 and 21. PER (2000), pp. 5, 10 and 11.

Ede and Enschede unofficially collected data on Roma without their knowledge and consent. De Groene Amsterdammer (2010). Vroon (2010).

The file contained data on approximately 5000 Roma. Mansel (2013). Ghosh (2013).

Examples include Albania, Romania and Slovakia. Because of the divergent estimates, the CoE uses minimum, maximum and average estimates when dealing with data on Roma population across Europe. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 12. Messing (2014), p. 822. Chopin et al. (2014), p. 49. Ivanov (2012), p. 80. OSF (2010), pp. 7, 10, 11, 20, 21, 64, 65 and 78. Rughiniș (2010), p. 342. Gray (2009), p. 61. Covrig (2004), p. 90.

This is even the case for countries where research on Roma is done regularly. For example, a 2011 regional Roma survey from UNDP, World Bank and the European Commission indicates the employment rate of Roma in Hungary stands at 23%, compared to 36% in a 2011 survey from FRA on the situation of Roma in 11 EU Member States (FRA and UNDP 2012) and 20% in national surveys. Messing (2014), pp. 822 and 823. Clark (1998). Druker (1997), pp. 22 and 23. Research data will be considered as an important data source on Roma in Sect. 5.2.2. Chapter 4 (Sect. 4.​3.​2) includes a general overview of the advantages and disadvantages of research data.

NGO research on Roma is irregular or lacking in several Member States, including in Croatia. McDonald and Negrin (2010), pp. 31 and 33 and 34. OSF (2010), pp. 20, 21, 27, 32, 40, 82 and 85. FRA (2009c), p. 272. Farkas (2007), p. 36.

For instance, the divergence between official and estimated data on the Roma population in the Decade of Roma Inclusion countries ranges from 45% to 99%. CoE, Roma and Statistics (22–23 May 2000), para. 15. Corsi et al. (2010), pp. 104 and 105. OSF (2010), pp. 10, 11, 23 and 53. Ringold et al. (2005), pp. xiii, xiv and 28. Babusik (2004), pp. 17 and 18. Milcher and Ivanov (2004), pp. 9 and 10. Clark (1998). Druker (1997), pp. 22 and 23. The four complementary data sources of ethnic data collection were analysed in Chap. 4 (Sect. 4.​3).

See, for example: Decade Watch (2007), p. 19. Gray (2009), p. 61.

McDonald and Negrin (2010), pp. 9, 11, 25, 28–31 and 33. OSF (2010), pp. 7 and 55.

McDonald and Negrin (2010), p. 21.

This is, for example, the case for the Czech Republic, Macedonia, Montenegro, Romania, Serbia and Spain. OSF (2010), pp. 33, 47, 54, 55, 61, 67 and 81. McDonald and Negrin (2010), p. 28.

For instance, UNDP provided “guidelines and set clear indicators for monitoring the effects and impact of the Decade action plans and planning policies”. McDonald and Negrin (2010), pp. 33 and 36.

McDonald and Negrin (2010), p. 21.

I McDonald and Negrin (2010), p. 21. Gray (2009), p. 61.

McDonald and Negrin (2010), p. 28. As was explained in Chap. 4 (Sect. 4.​3.​1.​3), the absence or unreliability of official statistics is not an excuse for States to escape responsibility and they should look for alternative or additional means to assess discrimination and monitor progress.

The added value of research data and international and European monitoring and litigation will be considered in Sects. 5.2.2 and 5.2.4 respectively. See also Chap. 4 (Sect. 4.​3) for a general overview of the four main data sources of ethnic data collection and their complementarity (Sect. 4.​3.​1.​3).

Chopin et al. (2013), pp. 63 and 64. Lamberts et al. (2014), pp. 22 and 24.

Exceptions include Austria (studies as baseline), Bulgaria (two nationally representative surveys), Croatia (special steering group), Estonia (informal working group; survey mapping situation Roma), France (set of indicators), Finland (steering group), Hungary (set of indicators; regular reporting; research; data collection) and the Netherlands (baseline qualitative study). Other factors include lack of budget setting and/or the absence of the assignment of responsible authorities/contacts. Council Recommendation on effective Roma integration measures in the Member States (9 December 2013), para. 3.4. Commission Communication, National Roma Integration Strategies: a first step in the implementation of the EU Framework (21 May 2012), pp. 12 and 13. Commission Communication, Steps forward in implementing National Roma Integration Strategies (26 June 2013), pp. 8 and 15. FRA (2014), p. 172. Chopin et al. (2014), p. 43.

Commission Communication, Report on the evaluation of the EU Framework for National Roma Integration Strategies up to 2020 (4 December 2018), p. 11.

Chopin et al. (2014), p. 64. European Structural Funds will be discussed in Chap. 9 (Sect. 9.​2.​5) when analysing how funding limitations affect positive action for Roma.

Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), arts. 1 and 3. Chapter 1 (Sect. 1.​2) includes reflections on the situation of Roma in Europe.

The data can be manipulated by different actors, depending on the political context in which they are used. CoE, Roma and Statistics (22–23 May 2000), para. 9. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 27. European Roma Information Office (ERIO) (2009), p. 31. Milcher and Ivanov (2004), p. 8. Clark (1998).

The Irish government did not adopt a Traveller health action plan following the Traveller Health Study. Abdikeeva (2014), p. 27. This problem also arises with regard to other types of equality data. For instance, Farkas (2017, pp. 5 and 6) refers to Denmark, where the thorough collection of data on gender does not automatically mean that this information is used to promote equality for women. The notion equality data collection was defined in Chap. 1 (Sect. 1.​3.​1) when introducing ethnic data collection as one of the main topics of this book.

FRA (2013a), p. 190. FRA (2011), p. 15. Kate (2010), p. 11. Barton (2004), p. 29.

CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 12. ECRI, Fifth Report on Hungary (19 March 2015), paras. 25, 53–56 and 106–108. Brüggemann and Bloem (2013), p. 520. ERRC (2013b), p. 55. UNICEF (2011), pp. 39–41. Halász (2008), p. 26. Hollo (2006), p. 39. OSI (2006), p. 2. Ringold et al. (2005), pp. 187–189. European Commission (2004), p. 47. Acquiring of information on social and economic integration levels and uncovering of discrimination were cited as two of the main benefits of ethnic data collection in Chap. 4 (Sects. 4.​1.​1 and 4.​1.​2).

Hollo (2006), p. 28. Clark (1998).

For instance, the rise in Roma self-identification in the 2011 Montenegrin Census resulted into higher spending of public funds to improve the situation of Roma. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 12. Abdikeeva (2014), p. 28. ERRC (2013a). Jovanovic and Haliti (2012). Kate (2010), p. 11. Hollo (2006), p. 29. Clark (1998).

Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 5. McDonald and Negrin (2010), p. 12. Petrova (2004), pp. 5 and 6. Ramsay (2006), p. 4.

CoE, Roma and Statistics (22–23 May 2000), para. 12.

See Chap. 4 (Sect. 4.​2) on the five main risks of ethnic data collection.

Chapter 1 (Sect. 1.​2) reflected on the situation of Roma in Europe.

Abdikeeva (2014), pp. 25 and 26. Corsi et al. (2010), pp. 10, 103 and 141. Gray (2009), pp. 59 and 60. Ramsay (2006), p. 4. The lack of uniform status of Roma across Europe was discussed in Chap. 1 (Sect. 1.​1.​3).

Clark (1998). The heterogeneity and diversity of Roma groups and communities was addressed in Chap. 1 (Sects. 1.​1 and 1.​2).

Council Recommendation on effective Roma integration measures in the Member States (9 December 2013), recital 8. ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 15(d) and (e). Farkas (2014), pp. 5 and 15. Jovanovic and Haliti (2012). Rorke (2011). Corsi et al. (2010), pp. 101–103. FRA (2008), p. 86. Kurtic (2006). Ringold et al. (2005), pp. xv, xvi and 84. Covrig (2004), p. 92.

Council Recommendation on effective Roma integration measures in the Member States (9 December 2013), recital 19. Messing (2014), pp. 821–823. Kate (2010), p. 11. Ringold et al. (2005), p. 25. Farkas (2004), p. 22. Petrova (2004), p. 6. Krizsán (2001), p. 192. See Sect. 5.3 on challenges to the construction of ethnic categories for Roma, Sect. 5.4 on the appropriateness different of ethnical identification approaches for Roma, Sect. 5.5 on methodological challenges to conducting surveys on Roma populations, and Sect. 5.6 on four factors impacting the analysis of data sets on Roma populations.

McDonald and Negrin (2010), p. 10. Farkas (2004), p. 22. Petrova (2004), p. 6. Milcher and Ivanov (2004), p. 8. Krizsán (2001), p. 192.

Section 5.8.3 will further consider the need for genuine political will. See also Chap. 6 (Sect. 6.​3.​3) on the need for political will and funding to implement effective positive action schemes and Chap. 9 (Sect. 9.​2.​2) on how the lack of political will limits the use of positive action for Roma. The other two elements are awareness-raising and active participation. See Chap. 12 (Sect. 12.​2.​3).

The survey was conducted via e-mail, phone and face-to-face. Roma from most European countries were reached for the survey, while special consideration was given to a proportional number of answers from Western and Eastern-European countries. The results are based on 102 received answers from Roma represented in (inter)governmental organisations, NGOs, and other sectors such as freelancers. The survey contained three questions: (1) Do you support data collection on ethnic grounds? (2) Do you think ethnic data would be a good instrument to fight discrimination against Roma? (3) Would you agree to disclose your ethnicity for statistics and surveys in your country in relation to education, employment, health care, housing and social services? ERIO (2009), pp. 17–31.

The survey shows that Roma at grassroot level and those working outside the NGO sector tend to be more reluctant and negative towards ethnic data collection than Roma working at national or European level and inside the NGO sector. The Roma living in Western Europe also appear to be more negative towards ethnic data collection practices compared to Roma living in Central and Eastern European countries. The behaviour of the majority population towards data collection on Roma is also very mixed, with larger NGOs generally being more open to the idea than smaller NGOs, statistical and data protection institutes, political leaders. CoE, Roma and Statistics (22–23 May 2000), para. 3. ERIO (2009), pp. 17, 18, 27–29 and 31. PER (2000), p. 4.

ERIO (2009), pp. 4, 5, 17, 18 and 21–28. See also: CoE, Roma and Statistics (22–23 May 2000), paras. 4 and 5. Chopin et al. (2014), p. 62. Krizsán (2001), p. 158. PER (2000), pp. 4, 9, 12 and 13. The over-representation of Roma in crime data will be addressed in Sect. 5.2.3.

ERIO (2009), pp. 23, 30 and 31. See, similarly: UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 41. Ringold et al. (2005), pp. xxv, xxvi and 116–118. Abdikeeva (2014), p. 28. Jovanovic and Haliti (2012). Waldron (2011). ERIO (2009), pp. 23, 24 and 26. Simon (2007), p. 57. Milcher and Ivanov (2004), p. 8. Krizsán (2001), p. 157. PER (2000), pp. 4, 5, 11, 17 and 29. See Chap. 4 (Sects. 4.​2.​3 and 4.​2.​4) for a discussion on the misuse of the data and discriminatory ethnic profiling by public bodies as possible risks of ethnic data collection.

ERIO (2009), p. 22.

During the 2000 Roma and Statistics Roundtable it was reported that Roma are more reluctant towards the collection of administrative data by the police than towards data collection in population Censuses. This will be discussed further in Sect. 5.5.4 on the appropriateness of the different ethnical identification approaches for Roma and in Sect. 5.8.2 on the need for active participation of all relevant stakeholders in data collection practices. CoE, Roma and Statistics (22–23 May 2000), paras. 12 and 16. ERIO (2009), pp. 21, 23, 25 and 26.

ERIO (2009), p. 22.

National differences were registered. Roma respondents were interviewed in seven Member States (Bulgaria, Czech Republic, Greece, Hungary, Poland, Romania and Slovakia) in 2008. FRA (2009c), p. 158. The second survey (EU-MIDIS II) was conducted in 2015–2016 and asked more general questions regarding Roma’s everyday life.

Farkas (2017), p. 32. Krizsán (2001), p. 157. PER (2000), pp. 4 and 17. See also Chap. 4 (Sect. 4.​2.​3) on the misuse of sensitive data to discriminate against ethnic groups as a risk of ethnic data collection.

CoE, Roma and Statistics (22–23 May 2000), para. 4. Jovanovic and Haliti (2012). Makkonen (2010), p. 226. Negrin (2003). Seltzer and Anderson (2001), p. 503. PER (2000), p. 4.

Roma were deported to, among others, Auschwitz, Dachau, Mauthausen and Ravensbrück. Chopin et al. (2014), p. 59. Waldron (2011). Clark (1998). Suggested further reading: Kenrick and Puxon (1995), pp. 124–151.

Lauder (2014). Waldron (2011). Uzunova (2010), p. 299. Clark (1998).

There are a limited numbers of exceptions. For instance, a system of compensation for Roma Holocaust survivors is in place in the Czech Republic, Germany and Hungary. In several other countries, including Croatia, (2014), Hungary (2005), Poland (2011) and Ukraine (2004), national parliaments have officially acknowledged the Roma genocide and declared 2 August as national Roma Holocaust Memorial Day. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 30. Resolution of the European Parliament on the occasion of International Roma Day—anti-Gypsyism in Europe and EU recognition of the memorial day of the Roma genocide during World War II (15 April 2015). Prague Daily Monitor (2016). Velinger (2016). Uzunova (2010), p. 299. Krizsán (2001), p. 170. While the Roma Holocaust was not included in the 2009 Holocaust commemoration ceremony on the International Day in the Memory of the Holocaust, some reference was made to WWII. ECRI, General Policy Recommendation No. 3: Combating Racism and Intolerance against Roma/Gypsies (6 March 1998), preamble (“(p)aying homage to the memory of all the victims of policies of racist persecution and extermination during the Second World War and remembering that a considerable number of Roma/Gypsies perished as a result of such policies”).

The data collection did not lead to the adoption of positive action but to the forced eviction of Roma from informal settlements in three regions in (Lombardia, Lazio and Campania) where a decree of May 2008 declared a state of emergency. The Italian government implemented guidelines to soften the initial measures, thereby avoiding an infringement procedure by the European Commission. The highest Italian administrative court struck down the measures in 2011, though not on the basis of its discriminatory character. Hermanin (2013), pp. 97 and 98. European Network of Legal Experts in the Non-Discrimination Field (2009), p. 53. ERRC et al. (2009).

For instance, data uncovering a high unemployment rate among Roma could be used to support the stereotype that Roma are lazy. CoE, Roma and Statistics (22–23 May 2000), paras. 4 and 21. Abdikeeva (2014), p. 28. ERIO (2009), pp. 22–24. Gray (2009), p. 58. Covrig (2004), p. 93.

Waldron (2011). Covrig (2004), p. 93. PER (2000), p11.

CoE, Roma and Statistics (22–23 May 2000), paras. 4 and 21. Mansel (2013). Ghosh (2013). ERIO (2009), pp. 21 and 22. Covrig (2004), p. 93.

Section 5.6.4 will consider dissemination challenges. Anti-Gypsyism was discussed in Chap. 1 (Sect. 1.​2.​1) on the particular vulnerability of Roma in Europe today.

CoE, Roma and Statistics (22–23 May 2000), para. 11. Jovanovic and Haliti (2012). Waldron (2011). ERIO (2009), pp. 21 and 22. Covrig (2004), p. 93. Farkas (2004), p. 22. PER (2000), pp. 10 and 30.

Simoni (2011), p. 13. Issues of terminology and the changing denomination of Roma were considered in Chap. 1 (Sects. 1.​1.​1 and 1.​1.​2). See also Sect. 5.3.1 on the complexities surrounding the ethnical classification of Roma for data collection purposes.

The law also foresaw harsh punishments in case the rules were not followed. France, Law on the exercise of itinerant professions and the regulation of the movement of nomads (Loi du sur l’exercice des professions ambulantes et la réglementation de la circulation des nomades) (16 July 1912). Simoni (2011), p. 13.

ECRI urged France to review its provisions requiring Travellers with an itinerant lifestyle to present a travel permit (carnet or livret de circulation) to the police of places where they temporarily settle. This practice is considered to be disproportionate and discriminatory. The administrative regime was abolished in November 2016. ECRI, Fourth Report on France (29 April 2010), paras. 98 and 103. Cosse (2016).

Fraser (2010).

The European Commission started an infringement procedure against France for violating EU rules.

Abdikeeva (2014), p. 28.

Waldron (2011). Covrig (2004), p. 94.

Section 5.8.1 will focus on the importance of awareness-raising in relation to ethnic data collection. Awareness-raising was also considered in Chap. 3 (Sect. 3.​7.​5) as the fifth operational and organisation principle to render personal data processing success and in Chap. 4 (Sect. 4.​1.​3) as the third benefit of ethnic data collection.

The promotion of ethnic data collection for equality and anti-discrimination purposes by international and European actors was addressed in Chap. 3 (Sect. 3.​1.​2). Official statistics and ethnical classification were discussed in Chap. 4 (Sects. 4.​3.​1 and 4.​4.​2 respectively).

This is the second Common Basic Principle on Roma Inclusion discussed at the first meeting of the integrated European platform for Roma inclusion in April 2009. Council Recommendation on effective Roma integration measures in the Member States (9 December 2013), art. 1.1. Common Basic Principles on Roma Inclusion, as annexed to the Council Conclusions on Inclusion of the Roma (8 June 2009), principles 1 and 4. Part II of the book will focus on positive action.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 14. European Roma and Travellers Forum (2015), p. 6. Corsi et al. (2010), pp. 10 and 122. Surdu and Surdu (2006), pp. 73 and 74. Intersectional discrimination faced by Roma was introduced in Chap. 1 (Sect. 1.​2.​3).

For instance, Roma women mainly take care of the family and their access to day-care is limited. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 16.4. Lamberts et al. (2014), pp. 16 and 17. Corsi et al. (2010), pp. 10, 122 and 105. Ringold et al. (2005), p. 65.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), paras. 12 and 18. Report of the European Parliament on Gender Aspects of the European Framework of National Roma Inclusion Strategies (10 December 2013). European Roma and Travellers Forum (2015), p. 6. Tremlett (2014), pp. 832 and 833. FRA and UNDP (2012), p. 17. O’Higgins (2012), pp. 6, 16, 17, 22, 26, 27 and 30. Corsi et al. (2010), pp. 10, 105, 106, 114 and 139. Farkas (2007), 13. European Parliament (2006). Ringold et al. (2005), p. 169. The particular situation of Roma women was briefly introduced in Chap. 1 (Sect. 1.​2.​3.​2) when discussing intersectionality.

Results from the EU-MIDIS I survey indicate that, with the exception of Poland, Roma generally suffer comparable from discrimination as the majority population. European Roma and Travellers Forum (2015), p. 6. Guy et al. (2010), p. 10. FRA (2009c), pp. 258 and 260. Hollo (2006), p. 10.

Census data in the Czech Republic and in Hungary are disaggregated by ethnicity and by gender. Regarding research, one notable exception is a number of national reports from the Network of Experts in Gender Equality, Social Inclusion, Health and Long-term Care on Roma women. Where available, data on Roma women is often not comparable. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 40. Corsi et al. (2010), pp. 103, 104 and 139. OSF (2010), p. 43.

ERRC (2013b), p. 55. FRA and UNDP (2012), p. 10. Corsi et al. (2010), pp. 10 and 139.

ERRC (2013a). ERRC (2013b), pp. 6 and 39. Corsi et al. (2010), pp. 6 and 103. Guy et al. (2010), pp. 10 and 11. Hollo (2006), p. 10. Ramsay (2006), p. 5.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 1.

ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 15(c). European Roma and Travellers Forum (2015), p. 5. Guy et al. (2010), p. 10. For an overview of the main benefits of ethnic data collection, see Chap. 4 (Sect. 4.​1).

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 14.

ECRI, General Policy Recommendation No. 10: Combating racism and racial discrimination in and through school education (15 December 2006) para. 5. CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), paras. 6 and 46. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), paras. 17, 23, 24 and 45. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), arts. 12 and 16.1.

Decade of Roma Inclusion 2005–2015, Terms of Reference—Decade Declaration (2 February 2005), p. 3. See also: Corsi et al. (2010), pp. 13 and 14. McDonald and Negrin (2010), p. 19. The notion equality mainstreaming will be discussed in Chap. 6 (Sect. 6.​1.​3.​4).

The EU Framework for NRIS mentions gender three times and women four times in relation to disparate statistics on Roma employment and health. Most Member States disregard the gender dimension in their NRIS. Exceptions include Croatia, Finland, Greece, Hungary, Italy, Portugal, Romania, Slovakia, Spain and Sweden, which all integrated—to varying degrees—a gender dimensions in their NRIS. Report of the European Parliament on Gender Aspects of the European Framework of National Roma Inclusion Strategies (10 December 2013), H and p. 28. Commission Communication, Report on the evaluation of the EU Framework for National Roma Integration Strategies up to 2020 (4 December 2018), pp. 3, 10 and 11. Commission Communication, An EU Framework for National Roma Integration Strategies up to 2020 (5 April 2011), p. 8. European Roma and Travellers Forum (2015), pp. 7–9 and 11–22.

European Roma and Travellers Forum (2015), pp. 10, 13, 14, 20, 21 and 24. Government of the Republic of Croatia, National Roma Inclusion Strategy from 2013 to 2020 (2012), pp. 8, 37, 44, 56–58, 76, 91, 102, 110, 111, 118–123 and 142. Office of the Plenipotentiary of the Slovak Government for Roma Communities, Strategy for the Integration of Roma up to 2020 (2012), pp. 8 and 12. Swedish Government, A coordinated long-term strategy for Roma inclusion 2012–2032 (2012), pp. 49 and 64. Hungarian Ministry of Public Administration and Justice, National Social Inclusion Strategy—Extreme Poverty, Child poverty, the Roma 2011–2020 (2011), pp. 26–29 and 128.

At average and per yearly assessment, the European Commission mentions the words gender, women and girls only three times. European Roma and Travellers Forum (2015), p. 7.

Report of the European Parliament on Gender Aspects of the European Framework of National Roma Inclusion Strategies (10 December 2013), pp. 12 and 15.

CoE, Roma and Statistics (22–23 May 2000), para. 14. PER (2000), p. 10.

The four complementary data sources of ethnic data collection were analysed in Chap. 4 (Sect. 4.​3).

For a brief summary of the particular vulnerability of Roma in current and former times, see Chap. 1 (Sects. 1.​2.​1 and 1.​2.​2).

See Sect. 5.2.4 on the significance of international and European monitoring and litigation. See also Chap. 4 (Sect. 4.​3.​2) on research data as one of the four complementary data sources of ethnic data collection.

Official statistics as a data source of ethnic data collection were covered in Chap. 4 (Sect. 4.​3.​1).

This is the case in several Member States, including Denmark, Finland, the Netherlands and Sweden. Makkonen (2006), p. 108.

This is, for instance, the case in Austria (6000 people ticked the box Romani language in the 2001 Census), Bosnia (8964 people identified as Roma in the 1991 Census), Bulgaria (312,000 Roma self-identified in the 1992 Census), Croatia (463 people identified as Roma in the 2001 Census), the Czech Republic (the 2011 Census counted 12,444 Roma), Greece (7429 Roma in the 1951 Census), Hungary (205,720 Roma in the 2001 Census), Ireland (29,573 people self-identified as Irish Traveller in the 2011 Census), Latvia (8205 self-identified as Roma in the 2000 Census), Montenegro (the 2011 Census counted 6251 Roma), Poland (2001 Census counted 12,900 Roma), Romania (619,007 Roma in 2011), Serbia (the 2002 Census counted 147,604 Roma), Scotland (category Gypsy/Traveller included since the 2011 Census), Slovakia (the 2008 Census counted 104,034 Roma), Slovenia (3246 Roma in 2002) and the UK (58,000 Roma self-identified in the 2011 Census). Abdikeeva (2014), pp. 14 and 25. Chopin et al. (2014), p. 75. ERRC (2013b), p. 14. Jovanovic and Haliti (2012). United Kingdom Office for National Statistics (2012), pp. 1–15. Pavee Point Traveller and Roma Centre (2011), pp. 10, 25, 28 and 35. Corsi et al. (2010), pp. 104 and 105. OSF (2010), pp. 31, 36, 43 and 78. Jacobs et al. (2009), p. 82.

For instance, this is the case for the Bosnian Census that counted 8864 Roma in 1991, while research from 2007 estimates that the Roma population in Bosnia stands around 76,000. OSF (2010), pp. 17 and 53.

Abdikeeva (2014), p. 25. FRA (2013b), p. 6. McDonald and Negrin (2010), p. 28. OSF (2010), p. 31. Gray (2009), p. 60. UNDP (2005), p. 7. Ringold et al. (2005), pp. 60 and 61. Covrig (2004), p. 92. Krizsán (2001), p. 192. Section 5.1.3 briefly discussed Roma reluctance towards ethnic data collection practices. Problems with the self-identification approach in the context of the Roma minority will be covered in Sect. 5.4.1. Under- and over-reporting issues in relation to official statistics were addressed in Chap. 4 (Sect. 4.​3.​1).

Abdikeeva (2014), p. 25. Corsi et al. (2010), pp. 104 and 105. OSF (2010), pp. 57 and 72. UNDP (2005), p. 7. Covrig (2004), p. 91. Krizsán (2001), pp. 191 and 192. PER (2000), p. 27.

Jovanovic and Haliti (2012). Corsi et al. (2010), pp. 104 and 105. OSF (2010), p. 43.

Unofficial estimates in Montenegro range up to 20,000 Roma. The Slovak Government for Roma communities estimates that the Roma population ranges from 320,000 to 380,000 people. Jovanovic and Haliti (2012). OSF (2010), pp. 59 and 78.

Ringold et al. (2005), pp. xiii, xiv, 28, 95 and 156–158. Babusik (2004), pp. 17 and 18. Csepeli and Simon (2004), p. 137.

These estimates vary greatly with the 619,007 Roma counted in the 2011 Romanian Census. Covrig (2004), p. 91. ERRC (2013b), p. 14. OSF (2010), pp. 64 and 65.

Csepeli and Simon (2004), p. 137.

Id.

These estimates vary greatly from the 12,444 Roma counted in the 2001 Czech Census. OSF (2010), p. 36.

Surdu (2019), p. 12.

Id.

Messing argues that “Roma living in a segregated, marginalized settlement may be less reluctant to express ethnic belonging because they are considered as Roma by the rest of the society anyway”. Messing (2014), p. 818.

This is the case in Bulgaria, Hungary, Poland and Romania. FRA (2013b), pp. 6 and 7.

UNDP (2005), p. 7. Milcher and Ivanov (2004), pp. 8 and 9.

UK Office for National Statistics (2012), p. 10.

Abdikeeva (2014), p. 30.

For instance, the 2001 Czech Census indicated that the unemployment rate of Roma stood at 57.3%, compared to 7.9% following a survey of the workforce by the Czech Statistical Office in 2005. CoE, Roma and Statistics (22–23 May 2000), para. 49. OSF (2010), pp. 36 and 37. PER (2000), p. 27. See also Sect. 5.4.2 on how proxies for ethnicity produce insufficient data on Roma. Indicators will be considered in Sect. 5.6.2.

OSF (2010), p. 28.

See Chap. 4 (Sect. 4.​3.​1) on official statistics as a data sources of ethnic data collection. See also: PER (2000), p. 24.

75% (Roma) compared to 25% (national average). Abdikeeva (2014), p. 12.

Ringelheim (2013), pp. 53 and 54.

Clark (1998).

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 39. FRA (2007), pp. 45 and 46. Ramsay (2006), p. 4.

FRA (2014), p. 172.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 39. Messing (2014), pp. 818 and 820. Bond et al. (2010), p. 9. Milcher and Ivanov (2004), p. 8. This will be discussed further in Sect. 5.5.1 on sampling frames and the risk of under- or over-representing Roma.

This is due to their reliance on proxies for ethnicity. Alidadi (2017), pp. 23 and 24. Farkas (2017), pp. 16–18 and 47. See Chap. 4 (Sect. 4.​3.​2.​2) on the disadvantages of research data. Proxies for Roma ethnicity will be considered in Sect. 5.4.2.

This goes for some Croatian ministries. OSF (2010), p. 28. Banton (2001), pp. 65 and 66.

Farkas (2017), pp. 26 and 27. Chopin et al. (2014), pp. 46, 54, 55, 74 and 75.

The Former Yugoslav Republic of Macedonia was officially renamed the Republic of North Macedonia in February 2019. McDonald and Negrin (2010), p. 30. OSF (2010), p. 49.

For example, the Ministry of Health collected data on the socio-economic, education, and health of approximately 102,000 Roma in 50 municipalities in Serbia. OSF (2010), pp. 67–69.

Part II of the book focuses on positive action. More specifically, Chap. 10 (Sect. 10.​1) underlines that a sectorial bottom-up approach to positive action leads to tailored and proportionate measures.

The context-dependency of attitudes towards ethnic data collection among Roma was highlighted in Sect. 5.1.3 on the multiple causes for the widespread lack of (reliable) data on Roma. See also Chap. 1 (Sects. 1.​1 and 1.​2) for more on the diversity among Roma communities.

The Commissioner is an independent, ombudsman-like institution. Krizsán (2001), pp. 168, 170 and 171.

The author identifies active participation as a key element of ethnic data collection on Roma. See Sect. 5.8.2.

Babusik (2004), p. 15.

Id.

See Sect. 5.8.2 on the active participation of Roma in ethnic data collection as a key principle to ethnic data collection on this ethnic minority.

The limitations of official statistics were addressed in Sect. 5.2.1. Furthermore, Chap. 4 (Sect. 4.​3.​1) also considered the advantages and disadvantages of official statistics as one of the data sources of ethnic data collection.

See, for example: ACFC, Second Opinion on the Czech Republic (24 February 2005), paras. 10 and 37. ACFC, Second Opinion on Slovenia (26 May 2005), para. 46. ACFC, Second Opinion on Hungary (9 December 2004), paras. 31, 35 and 53. Outline for reports to be submitted pursuant to Article 25 paragraph 1 of the Framework Convention for the Protection of National Minorities (30 September 1998), para. 4.5. CoE, Roma and Statistics (22–23 May 2000), para. 49. McDonald and Negrin (2010), pp. 29 and 30. FRA (2007), p. 78. Ramsay (2006), p. 5. Ringold et al. (2005), p. 44. The need to combine different data sources was previously stressed in Chap. 4 (Sect. 4.​3.​1.​3).

PER (2000), p. 27. As explained in Sect. 5.1.2, international and European calls for data on Roma largely remain unanswered. See also Chap. 4 (Sect. 4.​3.​2.​1) for an analysis of the advantages of research data as a source of ethnic data collection.

This is not always the case. For instance, an Irish self-report survey on discrimination could not be adequate analysed because of the low number of respondents self-identifying as Travellers. Additionally, other elements such as language barriers, cultural barriers and health problems might also limit the participation of Roma in research and surveys. Bond et al. (2010), p. 9. Russell et al. (2010), p. 29.

CoE, Roma and Statistics (22–23 May 2000), para. 49. The risk of overlooking certain groups in data collection efforts was previously mentioned in Sect. 5.1.3 when considering the different reasons for the widespread lack of (reliable) data on Roma in Europe.

FRA and UNDP (2012), p. 8. Ivanov (2012), p. 91. Ringold et al. (2005), pp. xxv and 19. Milcher and Ivanov (2004), p. 10. The issue of accountability was addressed in Chap. 4 (Sect. 4.​2.​1) on governments’ fear of exposure of bad policies and insufficient measures taken.

PER (2000), p. 27.

Section 5.1.3 looked into the multiple causes of the widespread lack of (reliable) ethnic data on Roma. ERIO (2009), p. 21.

Bulgaria is an example. FRA (2009c), p. 272. The disadvantages of research data were considered in Chap. 4 (Sect. 4.​3.​2.​2).

The methodological challenges of collecting ethnic data on Roma will be addressed in Sects. 5.3–5.6. Ivanov (2012), p. 91. Rughiniș (2010), p. 339. Ringold et al. (2005), p. 3. Ladányi and Szelényi (2001), pp. 79–89.

FRA (2011), p. 19. McDonald and Negrin (2010), p. 31. OSF (2010), p. 82. FRA (2008), p. 85. FRA (2007), pp. 45 and 46. Ringold et al. (2005), pp. xiv, 2 and 3. See, for instance: Feischmidt et al. (2010), pp. 1–85. Fundación Secretariado Gitano (2009), pp. 1–177.

Unofficial and independent research and surveys is sometimes supported by the government or conducted in partnership with the government, as is done in Spain. In Slovenia, data generated through a 2008–2009 study on health care services among Roma women and children were used to create adequate programmes. European Roma and Travellers Forum (2015), p. 21. Rughiniș (2010), p. 338. OSF (2010), pp. 33, 39 and 40. FRA (2009a), p. 4. FRA (2009c), p. 19. Krizsán (2001), pp. 192 and 193.

See, for instance, the survey conducted by ERIO in 2008, which was discussed in Sect. 5.1.3. ERIO (2009), p. 3.

Such a Romanian survey from 2010 “showed that only 54% of the respondents agreed with the idea of having a Roma as co-worker”, which is considerably lower than their perception of other minorities. FRA (2011), p. 17.

This has been successfully done in Hungary (employment and healthcare), Portugal (housing), Romania (healthcare) and Spain (healthcare). FRA (2013a), pp. 196 and 197. Lamberts et al. (2014), p. 15. Russell et al. (2010), p. 21. FRA (2008), p. 85. FRA (2007), pp. 9, 85 and 86.

PER (2000), p. 29. Large gaps in complaints data on Roma will be addressed in Sect. 5.2.3.

FRA (2014), pp. 179 and 180. FRA (2013a), p. 197. ERRC (2013a). ERRC (2013b), pp. 6, 55 and 59. FRA (2011), p. 19. Amnesty International (2010). OSF (2010). ERRC (2007). Ringold et al. (2005), p. 189. Rona and Lee (2001).

FRA and UNDP (2012), p. 9. FRA (2008), p. 111. World Bank (2008). Feischmidt et al. (2010), pp. 1–85. OSF (2010), p. 37. FRA (2007), p. 89. Ringold et al. (2005). UNDP (2005), p. 10. Milcher and Ivanov (2004), p. 10. UNDP (2003). UNDP (2002).

FRANET was mentioned when presenting research data as the second source of ethnic data collection in Chap. 4 (Sect. 4.​3.​2). FRA (2008), p. 111. FRA (2007), p. 89.

Topics include violence against women, discrimination in employment. FRA (2014), p. 172. FRA (2012a), p. 31. FRA (2009a), p. 4. FRA (2009b), p. 2. FRA (2009c), p. 271.

The acquisition of information on ethnic groups, the uncovering of discrimination and awareness-raising among the majority population were cited as benefits of ethnic data collection in Chap. 4 (Sects. 4.​1.​1–4.​1.​3). FRA (2009a), p. 4. FRA (2009b), p. 2. FRA (2009c), p. 271.

FRA (2009c).

Within the framework of the World Bank/UNDP/European Commission regional survey, 22,203 Roma and non-Roma were interviewed in Albania, Bulgaria, Bosnia and Herzegovina, Croatia, the Czech Republic, the Former Yugoslav Republic of Macedonia, Hungary, Montenegro, Moldova, Romania, Serbia and Slovakia. The FRA pilot survey was conducted in Bulgaria, the Czech Republic, France, Greece, Hungary, Italy, Poland, Portugal, Romania, Slovakia and Spain. Data on these countries was previously collected through the EU-MIDIS survey in 2008. FRA and UNDP (2012), p. 10.

Sections 5.5.1–5.5.3 will focus on challenges to sampling Roma. FRA (2013b), pp. 6 and 15. FRA (2012a), p. 168.

14,925 Roma and 7278 non-Roma were interviewed. For instance, big differences were recorded in employment in France, Italy and Portugal. Higher pre-school and kindergarten attendance rates are reported for Roma pupils in Hungary and Spain compared to Greece and Slovakia. FRA (2012a), p. 168. FRA and UNDP (2012), pp. 10, 13, 16 and 26. The present-day situation of Roma in Europe was briefly introduced in Chap. 1 (Sect. 1.​2.​1).

O’Higgins (2012), pp. 49 and 50.

Commission Communication, Report on the evaluation of the EU Framework for National Roma Integration Strategies up to 2020 (4 December 2018), p. 4.

Goodwin (2004), p. 1439. Discrimination testing was discussed in Chap. 4 (Sect. 4.​3.​2).

FRA (2013a), p. 193.

Sabatauskaité and Urbonaité (2013), pp. 23, 27, 28, 31, 41, 42 and 54. FRA (2012a), p. 171. FRA (2008), p. 50. FRA (2007), pp. 9, 57, 58 and 60.

Krizsán (2001), p. 180.

FRA (2012a), p. 171. FRA (2008), p. 50. FRA (2007), pp. 57 and 58 and 60. Krizsán (2001), p. 180.

Examples of this practice are found in the Czech Republic and Hungary. FRA (2007), pp. 9, 150 and 151.

Examples include the Czech Republic, Hungary and Lithuania. Sabatauskaité and Urbonaité (2013), pp. 41 and 42. FRA (2012a), p. 171. Farkas (2011), p. 47. FRA (2008), p. 50. FRA (2007), pp. 57 and 60. Goodwin (2004), p. 1439. Krizsán (2001), p. 180.

Goodwin (2004), pp. 1439 and 1440.

Id.

See Sect. 5.8.1 on awareness-raising among Roma and non-Roma communities.

Complaints data were defined in Chap. 4 (Sect. 4.​3.​4).

The other categories are Central American and the Caribbean, East European and Russia, Spanish, Maghreb, Sub-Saharan, Andean South American, South American, and Others. Farkas (2017), p. 24.

The Hungarian Equal Treatment Authority independently monitors the implementation of minority rights and investigates complaints. In 2009, 54 of the complaints brought by Roma resulted in a decision on the merits. In Romania, the National Council for Combating Discrimination yearly records five to ten cases of discrimination of Roma by local authorities, public figures, newspapers or companies. In Slovakia, out of 1571 cases received by the National Center for Human Rights in 2009, 15 were brought by Roma. From 2004 to 2010, the Swedish Equality Ombudsman received 230 complaints from Roma. FRA (2014), p. 177. FRA (2012a), p. 173. OSF (2010), pp. 12, 45, 66 and 80. Ringold et al. (2005), pp. 131 and 132.

The National Council for Combating Discrimination is a governmental agency in charge of the National Antidiscrimination Plan. OSF (2010), p. 66.

OSF (2010), p. 12.

Farkas (2017), p. 24.

Id.

In Poland, data on hate crime is collected by the office of the Prosecutor General since 2012. For the Czech Republic and Sweden, data on anti-Roma crime is collected by Office for Democratic Institutions and Human Rights (ODIHR). FRA (2014), p. 158. Chopin et al. (2014), p. 75. ODIHR (2013), pp. 55–60. FRA (2012b), p. 9.

McDonald and Negrin (2010), p. 32.

In 2010, Montenegro lacked anti-discrimination legislation and Bosnia, Macedonia and Serbia still had to establish an equality body, while equality bodies in Croatia and Spain had yet to publish their first report. McDonald and Negrin (2010).

In the Czech Republic, the Public Defender of Rights also does not record the ethnicity of complainants. McDonald and Negrin (2010), p. 32. OSF (2010), p. 37.

The Bulgarian Commission for Protection of Discrimination reports on the numbers of complaints on ethnic grounds in its annual report, but not on cases brought solely by Roma. OSF (2010), p. 25.

This issue was previously highlighted in Chap. 4 (Sect. 4.​3.​2.​2) on the disadvantages of complaints data as a data source for equality and anti-discrimination purposes.

FRA (2009b), pp. 5 and 12. FRA (2009c), pp. 169 and 170. Davidović and Rodrigues (2010), p. 153. Farkas (2007), p. 13.

In 2008, national numbers published by the Dutch police show the registration of four complaints on Roma. From 2006 to 2008, the notification points registered one to four complaints of Roma discrimination per year. In 2005, the Public Prosecutor’s Office reported that discrimination of Roma concerned 1% of its caseload. FRA (2009b), p. 9. Davidović and Rodrigues (2010), pp. 155 and 171.

FRA (2014), p. 172.

The EU-MIDIS surveys, which were conducted in 2008 and 2015–2016 respectively, indicate that reporting rates among Roma remain low and have not improved. FRA (2016), p. 40. FRA (2009b), p. 5. FRA (2009c), pp. 14 and 17.

Reporting rates vary according to incident and across Member States. 34% of Roma in the Czech Republic report incidents of discrimination, compared to medium-level rates in Hungary (34%), Poland (33%) and Slovakia (28%) and lower rates in Bulgaria (8%) and Greece (10%). Reporting rates of crimes against Roma also varies considerably across Member States, with higher numbers in Slovakia (31%), Poland (28%), Romania (25%) and the Czech Republic (24%) and considerably lower numbers in Greece (11%), Bulgaria (12%) and Hungary (15%). Discrimination in education is most likely to be reported upon. FRA (2009b), pp. 8 and 15. FRA (2009c), pp. 14, 155 and 163.

The results are based on the question “(l)ast time you felt discriminated against because of your Roma background when [domain], did you report or make a complaint about the incident?”, which was asked to each of the Roma respondents who stated that the last incident of discrimination based on skin colour or ethnic origin took place in the past 12 months. As was the case in EU-MIDIS I, national differences were also identified in EU-MIDIS II. FRA (2016), p. 40.

FRA (2016), p. 40. Davidović and Rodrigues (2010), pp. 153, 155 and 156. FRA (2009b), pp. 3, 6, 8, 9 and 12. FRA (2009c), pp. 13, 17, 158, 159, 163, 165 and 273. Farkas (2007), p. 13. Rodrigues and Matelski (2004), pp. 32–36.

FRA (2013a), p. 268. FRA (2012b).

ODIHR (2014).

Only the non-profit foundation Triana in Utrecht responded by establishing a hotline and involving volunteers, but failed to secure regional funding. Davidović and Rodrigues (2010), p. 156.

See Sect. 5.8.2 on the importance of raising awareness among Roma and non-Roma communities on ethnic data collection for equality and anti-discrimination purposes.

Simon (2007), p. 15.

CoE, Roma and Statistics (22–23 May 2000), paras. 11 and 21–23. Simon (2007), p. 15. Age-old stereotypes of Roma were addressed in Chap. 1 (Sect. 1.​2.​1) on the particular vulnerability of Roma in Europe today.

Simon (2007), p. 15.

Simon (2007), p. 15. As will be discussed in Sect. 5.6.4, such over-representation must also be correctly reported on when disseminating data sets in order to avoid that the data promote anti-Gypsyism. Discriminatory ethnic profiling was cited as a risk of ethnic data collection in Chap. 4 (Sect. 4.​2.​4). It was also explained in Chap. 4 (Sect. 4.​1.​5) how ethnic data can help prove indirect discrimination claims in the context of ethnic profiling.

Simon (2007), p. 15.

These include reports from States, statistical evidence, testimonies and information on trends and patterns from domestic stakeholders, NGOs and civil society as well as comparative data from international organisations. The other sources include Sect. 5.2.1 on the limitations of official statistics, Sect. 5.2.2 on the added value of surveys, research and discrimination testing, and Sect. 5.2.3 on the unreliability of complaints data and the over-representation of Roma in crime data. The four complementary data sources of ethnic data collection were first introduced and analysed in Chap. 4 (Sect. 4.​3). The role of international and European monitoring bodies in ethnic data collection was previously highlighted in Chap. 3 (Sect. 3.​1.​2) on international and European calls for ethnic data and in Chap. 4 on the benefits of ethnic data collection (Sect. 4.​1.​4) and when considering research as a data source (Sect. 4.​3.​2).

CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 11. Farkas (2014), p. 40.

Goodwin (2004), pp. 1440 and 1441.

See, for instance: CAT Committee, Concluding Observations on Hungary (6 February 2007), para. 19 (discrimination by law enforcement officials). CERD Committee, Concluding Observations on Italy (4 April 2012), para. 15 (housing discrimination, targeted, forced evictions and spatial segregation). CERD Committee, Concluding Observations on Austria (31 August 2012), para. 17 (high dropout rates and over-representation in special schools). CERD Committee, Concluding Observations on Finland (23 October 2012), para. 15 (high dropout rates and over-representation in special schools; discrimination in access to employment). CESCR Committee, Concluding Observations on Slovakia (8 June 2012), paras. 9, 13 and 26 (persistent discrimination in employment and education). CESCR Committee, Concluding Observations on Greece (7 June 2004), paras. 11, 15, 28, 32, 43–46 and 50 (discrimination in housing, health and education and police violence). HR Committee, Concluding observations on Lithuania (31 August 2012), para. 7. (Roma poverty, discrimination and lower education and employment levels).

See, for instance: ACFC, Third Opinion on Romania (21 March 2012), para. 27 (segregation in education). ECRI, Fourth Report on Finland (21 March 2013), paras. 65, 107–115 and 175 (insults and acts in schools and employment and relationship with the police). ECRI, Fourth Report on Latvia (9 December 2011), paras. 104–109 (over-representation in special needs schools). ECRI, Fourth Report on Portugal (21 March 2013), paras. 88–99 and 122–127 (housing and anti-Gypsyism by police). ECRI, Fourth Report on Sweden (19 June 2012), paras. 97 and 110–116 (access to employment). Report by the Commissioner for Human Rights following his visit to Portugal from 7 to 9 May 2012 (10 July 2012), paras. 37, 43, 44 and 47–63 (discrimination in housing and education and hate speech by the police).

See, for instance: CAT Committee, Concluding Observations on the Czech Republic (3 June 2004), para. 5(a) en (k) (absence of adequate investigations into these incidents). Report by the Commissioner for Human Rights following his visit to Greece from 28 January to 1 February 2013 (16 April 2013), paras. 103, 106, 109 and 133.

Goodwin (2004), pp. 1440 and 1441.

Farkas (2014), p. 41. Makkonen (2006), p. 30. This was previously mentioned in Chap. 2 (Sect. 2.​1.​2) on the difference between direct and indirect discrimination and the link with statistics and in Chap. 4 (Sect. 4.​1.​5) on supporting indirect discrimination claims in legal proceedings as one of the benefits of ethnic data collection.

Farkas (2014), p. 40. Makkonen (2006), p. 30. Official statistics were considered in Sect. 5.2.1, surveys, research and discrimination testing in Sect. 5.2.2 and complaints and crime data in Sect. 5.2.3.

EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), para. 134. Ignăţoiu-Sora (2011), p. 69. ERIO (2009), p. 32.

In 1999, Romani NGOs sent questionnaires to the head teachers of the eight special schools and 69 primary schools in the town of Ostrava. EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), paras. 134 and 191. Ignăţoiu-Sora (2011), pp. 69 and 70. ERIO (2009), p. 32. Goodwin (2004), p. 1437.

EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), paras. 103, 134, 190 and 191. Ignăţoiu-Sora (2011), p. 69. ERIO (2009), p. 32. Ringelheim (2006/7), pp. 60 and 61.

Ignăţoiu-Sora (2011), p. 70.

EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), para. 188. Farkas (2014), p. 40. Ignăţoiu-Sora (2011), pp. 67 and 71. ERIO (2009), pp. 32 and 33. The notion indirect discrimination and the role of statistics in proving indirect discrimination claims was described in Chap. 2 (Sect. 2.​1.​2). See also Chap. 4 (Sect. 4.​1.​5) on the support of indirect discrimination claims in legal proceedings as one of the benefits of ethnic data collection.

The EctHR ruled by 13–4 votes that the Czech Republic violated art. 14 European Convention for the Protection of Human Rights and Fundamental Freedoms (4 November 1950) (ECHR) read in conjunction with art. 2 Protocol 1 to the ECHR (20 March 1952). EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), para. 103. ERIO (2009), pp. 32 and 33.

Some measures were taken, such as the employment of teaching assistants and preparatory classes. It was also acknowledged that better information on Roma in schooling collected through statistical surveys could be helpful to implement better and more effective measures. The role of teaching assistants will be addressed in Chap. 11 (Sect. 11.​7.​1) on inter-cultural mediation to enhance Roma inclusion. Romea (2015). FRA (2014), p. 175. Ignăţoiu-Sora (2011), pp. 73 and 74. ERRC (2013c), pp. 8, 9, 11 and 12.

Anagnostou (2010) underlines the importance of European human rights case law, the implementation of which can—under certain conditions—result in legal and policy change at the national level. Relevant factors in this regard are repeated litigation and legal mobilisation by interested actors that are in a position to exert pressure and link the EctHR’s judgments to policy problems as well as support from political and other influential actors. Suggested further reading on the factors that impact on the domestic implementation of human rights judgments in Europe, see: Schneider (2015), pp. 155–196 and 232–241. Anagnostou and Mungiu-Pippidi (2014), pp. 205–227.

OSF (2016).

Id.

Id. at p. 80.

Ivanov (2017).

The EctHR confirmed later that a claim of indirect discrimination can be formulated without relying on statistical data but that such data help to establish a prima facie case of discrimination. See: EctHR, Oršuš and Others v. Croatia, Judgment (16 March 2010, GC), paras. 152 and 153. This was previously discussed in Chap. 2 (Sect. 2.​1.​2) on the use of statistical data to prove indirect discrimination claims.

This argument was introduced in Chap. 2 (Sect. 2.​1.​2) when discussing the link between indirect discrimination and ethnic statistics. ERIO (2009), pp. 32 and 33. Hollo (2006), p. 24.

EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), paras. 190 and 191. ERIO (2009), pp. 32 and 33. As explained in Chap. 4 (Sect. 4.​3.​1.​3), the absence of official statistics is not an excuse for States to avoid responsibility.

European Committee of Social Rights (ECSR), ERRC v. Greece, Decision (8 December 2004), paras. 27, 28 and 40. Later confirmed in: ECSR, ERRC v. Italy, Decision (7 December 2005), paras. 22 and 23 (2005).

ECSR, ERRC v. Greece, Decision (8 December 2004), paras. 27, 28, 40 and 41. Ringelheim (2006/7), pp. 60 and 61.

See Chap. 4 (Sect. 4.​3.​1.​3) on research data as a source of ethnic data collection.

FRA (2008), p. 16. EctHR, D.H. and Others v. the Czech Republic, Judgment (13 November 2007), paras. 83 and 88–90. The use of gender statistics in the case law of the CJEU was mentioned in Chap. 4 (Sect. 4.​1.​5) when discussing the support of indirect discrimination claims in legal proceedings as the fifth benefit of ethnic data collection.

While not about statistics, the CJEU delivered a Roma-specific ruling in the case CHEZ Razpredelenie Bulgaria AD v. Komisia za zashtita ot diskriminatsia on 16 July 2015. The case concerned the systematic placement of electricity meters out of reach of households in a predominantly Roma neighbourhood in Bulgaria. While the applicant was not a Roma herself, the CJEU ruled that she was a victim of discrimination based on racial or ethnic origin as she was wronged by the electricity provider’s discriminatory practice. CJEU, CHEZ Razpredelenie Bulgaria AD v. Komisia za zashtita ot diskriminatsia, Judgment (16 July 2013). For an analysis of this case, see: Nyitray and Cox (2015). Another prior case on Roma discrimination was declared inadmissible in 2013 because the referring body, the Bulgarian Equality Body, was not considered to be a court or tribunal within the meaning of the Treaty. See: CJEU, Valeri Hariev Belov v. CHEZ Elektro Balgaria AD and others, Judgment (31 January 2013).

ERRC (2004a). Goodwin (2004), p. 1440.

Krizsán (2001), p. 179.

Id. at p. 178.

Ignăţoiu-Sora (2011), p. 70. Makkonen (2006), p. 30.

If definitions and methodologies are outdated or inconsistent, the data resulting from it will be outdated. Difference in definitions and methodology also hinders comparisons between the data. Messing (2014), pp. 811, 812, 821, 822, 824 and 825. Rughiniș (2010), pp. 337 and 339. Ringold et al. (2005), p. 25. Challenges to data comparability will be considered in Sect. 5.6.2. Chapter 4 (Sect. 4.​4) analysed issues relating to ethnical categorisation on a general level.

Section 5.3.1 looks at how the complex and fluid nature of Roma identities influences ethnical classification. Section 5.3.2 puts forward open-ended categories, multiple affiliations and re-classifications as possible solutions to facilitate the ethnical classification process. The discussion focuses on the construction of ethnic categories, because most Member States collect data on Roma as an ethnic group. This is, however, not the case in Cyprus, Greece and Luxembourg, with the former two including Roma in the religious category Turkish Muslim. Farkas (2017), p. 38.

It concerns Sect. 5.4 on ethnical identification approaches and Sect. 5.5 on other methodological challenges to collection ethnic data on the Roma minority.

Ethnic classification as a complex, context-dependent and evolving process was discussed in Chap. 4 (Sect. 4.​4.​2). See also Chap. 3 (Sect. 3.​2.​2) where it was explained that ethnicity is a multidimensional process that may change over time under the influence of various circumstances.

Messing (2014), pp. 824 and 825. Ivanov (2012), p. 81.

Messing (2014), p. 816. Ivanov (2012), p. 79. Csepeli and Simon (2004), p. 131. Chapter 1 (Sect. 1.​2) reflected on the situation of Roma in Europe, including the discrimination this minority faces. Assimilation and other policies targeting Roma were also considered in Sect. 5.1.3 when considering the multiple reasons for the widespread lack of (reliable) data on Roma. The notion assimilation was discussed in Chap. 2 (Sect. 2.​3.​3).

Messing (2014), pp. 816 and 824 and 825.

Farkas (2017), p. 40. The UK Census categories were cited in Chap. 4 on ethnic classification as a complex, context-dependent and evolving process (Sect. 4.​4.​2) and on the wide variety in terminology and categories used (Sect. 4.​4.​3).

Surdu (2019), p. 1.

Messing (2014), pp. 812 and 813. See Chap. 1 (Sect. 1.​1) on the ambiguous conceptualisation of Roma.

Messing (2014), pp. 812, 813, 832 and 833. Tremlett (2014), p. 832. FRA and UNDP (2012), pp. 10, 29 and 30. Ivanov (2012), pp. 80 and 94. Rughiniș (2010), p. 346. FRA (2009b), p. 14. Sigona (2005), pp. 744 and 745. See Chap. 1 for reflections on the manifold definitions used that reflect the diversity among Roma (Sect. 1.​1.​1) and on the ambiguous and changing denomination of this ethnic minority (Sect. 1.​1.​2).

Tremlett (2014), p. 832. CoE (2012), pp. 6–10.

Messing (2014), p. 813.

Hungarian Roma self-identify as Hungarian and/or Musician Roma and speak exclusively Hungarian; Vlach Roma speak Romani/Lovari self-identify as Hungarian and Rom or Vlah Gypsies; Boyash (Bea’s) people, who speak Boyash (based on archaic Romanian dialect) and identify as Boyash Gypsies. Messing (2014), p. 813.

Matras (2005), pp. 53–55. Csepeli and Simon (2004), p. 129. Variations in denominations used were previously highlighted in Chap. 1 (Sects. 1.​1.​1 and 1.​1.​2).

In Hungary, the term Roma is used politically. Minority self-governments, government documents and the Minorities Act use the term Gypsy (cigány). The Romungro or Magyar Cigányok (Hungarian Gypsies) or the Beás often do not call themselves Roma. Rughiniș (2010), pp. 344, 365 and 366. Babusik (2004), p. 16. Covrig (2004), p. 95. PER (2000), p. 26.

The term Gypsy is sometimes used to exclusively define Roma or to refer more broadly to commercial nomads. Brüggemann and Bloem (2013), p. 520. Matras (2005), pp. 53–55. Csepeli and Simon (2004), p. 129.

Hermanin (2013), p. 97. Hepworth (2012), p. 434. European Network of Legal Experts in the Non-Discrimination Field (2009), p. 53.

Sigona (2005), pp. 744 and 745.

Sigona (2005), pp. 744 and 745. Csepeli and Simon (2004), pp. 134 and 135.

CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 3. ECRI, General Policy Recommendation No. 3: Combating Racism and Intolerance against Roma/Gypsies (6 March 1998). The importance of active participation of Roma in ethnic data collection practices will be discussed in Sect. 5.8.2.

This helps to builds trust. Farkas (2017), p. 45. Chopin et al. (2014), p. 47. This was explained in Chap. 4 (Sect. 4.​4.​3) when considering the wide variety in terminology, categories and answer formats used to collect ethnic data. Active participation as a key principle of ethnic data collection on Roma will be highlighted in Sect. 5.8.2.

Abdikeeva (2014), pp. 28 and 29.

See Chap. 4 (Sect. 4.​4.​3) for a general consideration of the wide variety in terminology, categories and answer formats used in ethnic data collection practices.

Chopin et al. (2014), p. 47. Corsi et al. (2010), p. 141. Haug (2001), p. 305. The lack of uniform status of Roma communities across Europe was discussed in Chap. 1 (Sect. 1.​1.​3). This risk was previously highlighted in Chap. 4 (Sect. 4.​4.​3) on the wide variety in terminology, categories and answer formats used.

Surdu (2019), p. 12. The importance of consulting with Roma communities on the terms to include was highlighted in Sect. 5.3.1.

For instance, respondents of the Hungarian census can specify a group or language that is not included in the list. UK Office for National Statistics (2012), p. 4. Rughiniș (2010), p. 346. Krizsán (2001), p. 191.

Travellers were formally recognised as an ethnic minority in Ireland in March 2017. The Irish Census does not include a Roma category. Abdikeeva (2014), pp. 25 and 26.

The Constitutional Court ruled that, because existing statistics legislation could not provide sufficient guarantees regarding the secrecy of the data, citizens could not be obliged to fill in detailed census questionnaires. German Constitutional Court, BverfGE 65, 1ff (1983). Negrin (2003). OSI (2002), pp. 158 and 159.

Messing (2014), pp. 811, 814, 824 and 825.

Messing (2014), p. 814. Simonovits and Kézdi (2014), p. 4.

It is important that reports and analyses of the data reflect such multiple identifications. Simonovits and Kézdi (2014), p. 9. Rughiniș (2010), pp. 346 and 352. Krizsán (2001), p. 191.

This is called co-equal dual identification in Hungary. Corsi et al. (2010), p. 102. Rughiniș (2010), p. 352.

Messing (2014), pp. 824 and 825. Ivanov (2012), p. 79. PER (2000), p. 26. See Sect. 5.6.2 for more on limitations to data comparability.

Ivanov (2012), p. 82. Csepeli and Simon (2004), pp. 129 and 148. See Sect. 5.4.1 on the limitations of the self-identification approach and Sect. 5.5.4 on how the type of interview, the interviewer and the questions asked affect the quality of the data collected. For a general consideration of the context-dependency and variability of ethnicity, see Chap. 3 (Sect. 3.​2.​2).

Historically T(s)igan was used as an umbrella term, but Roma was promoted and widely used after 1990 by Roma NGOs and the CoE because the former notion acquired pejorative connotations. In Hungary, however, the term Roma is highly controversial. Messing (2014), pp. 812 and 813. Rughiniș (2010), p. 344. Csepeli and Simon (2004), p. 148. Suggested further reading: Woodcock (2007), pp. 493–515.

Ivanov (2012), p. 82. As explained in Chap. 1 (Sect. 1.​1.​2), notions may obtain a negative connotation over time, requiring them to be replaced with another term, including for ethnical categorisation purposes.

See, for instance, the evolution of the categories used in the UK Census, as discussed in Chap. 4 (Sect. 4.​4.​2) on ethnic classification as a complex, context-dependent and evolving process. Farkas (2017), p. 39.

The importance of close co-operation between experts will be considered in Sect. 5.7.3.

See Sect. 5.3 on challenges to the construction of ethnic categories for Roma.

Brüggemann and Bloem (2013), p. 520. Rughiniș (2010), pp. 339, 340 and 342.

Id.

Challenges to ethnical classification of Roma were addressed in Sect. 5.3.

Messing (2014), pp. 815 and 816.

Messing (2014), pp. 815 and 816. Chapter 4 (Sect. 4.​5) analysed the main four ethnical identification approaches.

This was explained in Chap. 4 (Sect. 4.​5.​2.​1) on ethnical identification as the preferred approach.

CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 3. Ringold et al. (2005), pp. 25 and 26.

Messing (2014), pp. 817, 818, 824 and 825. FRA (2013a), pp. 196 and 197. Research data was presented as the second source of ethnic data collection in Chap. 4 (Sect. 3.​2).

Babusik (2004, p. 14) points out that this “fundamentally changed the system of data gathering, and consequently any research concerning Roma”. The Central Statistical Office performed its first data gathering directly focusing on Roma in 2005–2006.

Covrig (2004), p. 95.

Abdikeeva (2014), p. 28. Messing (2014), pp. 814 and 816. FRA (2013b), p. 6. Ringelheim (2013), p. 54. Makkonen (2010), pp. 236 and 237. OSF (2010), p. 21. Rughiniș (2010), pp. 339 and 366. See Sect. 5.1.3 on the multiple reasons for the widespread lack of (reliable) ethnic data on Roma and Sects. 5.2.1–5.2.3 on the multifarious data sources on Roma.

Roma often self-identify as Slovakian or Hungarian in Slovakia and as Romanian or Hungarian in Romania to avoid negative connotations associated with Roma identity. Rughiniș (2010), p. 350. Covrig (2004), pp. 94 and 95. Clark (1998).

Abdikeeva (2014), p. 28. FRA and UNDP (2012), p. 29. Ivanov (2012), pp. 80 and 81. Krizsán (2001), p. 192.

This is based both on former and current negative experiences as well as with a lack of confidence to self-identify. This is, for instance, the case for Travellers in Ireland. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 7. Abdikeeva (2014), pp. 25 and 28. Brüggemann and Bloem (2013), p. 520. FRA and UNDP (2012), p. 29. ERIO (2009), pp. 22, 26, 30 and 31. Hollo (2006), p. 29. Covrig (2004), p. 99. Clark (1998). Waldron (2011). OSF (2010), p. 76.

For instance, many of the approximately 300,000 Roma from Eastern European countries in the UK self-identify by nationality rather than ethnicity, resulting in uncertain statistics. Farkas (2014), p. 22. OSF (2010), p. 76. The different meanings terms may have in different contexts were considered in Chap. 4 (Sect. 4.​4.​3).

Beis (2012).

FRA and UNDP (2012), p. 29. Challenges to ethnical categorisation in the Roma context were addressed in Sect. 5.3.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 7. Krizsán (2001), p. 192.

“Being rid of the Roma identity might also means being rid of the bad resonance it has.” For example many Roma in Hungary are not proud of being Roma because of the bad resonance it has and therefore they sometimes ignore their roots in order to assimilate into the majority population. Rughiniș (2010), p. 350. Babusik (2004), p. 18. Covrig (2004), p. 95. Csepeli and Simon (2004), p. 135. Anti-Gypsyism was defined and described in Chap. 1 (Sect. 1.​2.​1).

Csepeli and Simon (2004), p. 135.

The importance of open-ended categories, multiple affiliations and re-classifications was stressed in Sect. 5.3.2.

Messing (2014), pp. 814, 817, 818, 824 and 825. Chopin et al. (2014), p. 47. McDonald and Negrin (2010), p. 36. Rughiniș (2010), p. 352. Csepeli and Simon (2004), pp. 135 and 136. For more on the importance of awareness-raising and active participation when collecting ethnic data on Roma, see Sect. 5.8.

Makkonen (2010), pp. 236 and 237.

Ringold et al. (2005), pp. xviii and 28.

Russell et al. (2010), p. 29.

Rughiniș (2010), p. 351. FRA (2009c), p. 158.

Messing (2014), p. 819. Rughiniș (2010), p. 354.

This is for example true for self-identification by various Roma communities in the Romanian Census. Simonovits and Kézdi (2014), p. 13 and 14. Rughiniș (2010), p. 351. See Sect. 5.2.1 for a discussion on the limitations of official statistics as the first data source on Roma.

Simonovits and Kézdi (2014), pp. 13, 14 and 26. Rughiniș (2010), p. 351. The challenges to ethnical classification were addressed in Sect. 5.3.

The latter refers to the question whether they might benefit or suffer from self-identifying. Covrig (2004), pp. 92, 96, 97 and 99. Csepeli and Simon (2004), p. 148. The impact of the choice of interviewers and questions will be considered in Sect. 5.5.4 and the active participation of Roma as a key principle of ethnic data collection will be addressed in Sect. 5.8.2.

This can be explained by cultural and social-psychological factors. Research in Bulgaria and Hungary demonstrates that more respondents identify their parents as parents than the number of respondents who self-identify as Roma, which could suggest possible weakening of ethnic identity over time. Ringold et al. (2005), p. 28. Csepeli and Simon (2004), pp. 143 and 144.

Research in Hungary and Romania indicate that self-identification might be easier in a Roma environment where the Roma identity is reinforced on a daily basis. Simonovits and Kézdi (2014), pp. 19–21. Csepeli and Simon (2004), pp. 143 and 144.

Changes in socio-economic status appear to be a powerful predictor for changes in reported ethnicity across surveys. Research conducted in Bulgaria, Hungary and Romania shows that the stigma of poverty often relates to those willing to self-identify. Those who want to do improve their socio-economic status are more likely not to self-identify to overcome negative associations. Messing (2014), p. 818. Simonovits and Kézdi (2014), pp. 2–5. Covrig (2004), p. 95. Csepeli and Simon (2004), pp. 148 and 149. This will be discussed further on in Sect. 5.4.5 on the context-dependency of the appropriateness of different ethnical identification approaches and in Sect. 5.5.4 when considering how the type of interviews, the interviewers and the questions asked affect the quality of the data.

Simonovits and Kézdi (2014), pp. 13, 14 and 26. Rughiniș (2010), p. 351.

Ringelheim (2013), pp. 53 and 54. Babusik (2004), p. 14. Covrig (2004), p. 95. The benefits and limitations of the self-identification approach in the Roma context were reviewed in Sect. 5.4.1. Self-identification was considered on a more general level in Chap. 4 (Sect. 4.​5.​2).

McDonald and Negrin (2010), p. 36.

Participation in Roma minority education is used a proxy for Roma ethnicity in Hungary and Romania, Some Censuses in Central and Eastern Europe do not only inquire about respondents’ national or ethnic minority, but also about mother tongue as a proxy for ethnicity. CoE, Roma and Statistics (22–23 May 2000), para. 18. Chopin et al. (2014), pp. 13, 49, 54, 55, 58 and 76. FRA (2013b), p. 6. McDonald and Negrin (2010), p. 36. Farkas (2007), p. 36. Krizsán (2001), p. 192.

Krizsán (2001), p. 192.

See Chap. 4 (Sect. 4.​5.​4) expanded on the use of proxies.

Several countries expanded their ethnic questions to enhance their system and to achieve better results. For example, up to 1941 Hungary collected ethnic data on the basis of language. In 2001, an ethnicity question inquiring about nationality was introduced and changed against afterwards to enhance the system. The Czechoslovak and Romanian Censuses of 1970 and 1980 assessed the Roma ethnic affiliation in several ways, such as language, housing, way of life and anthropological features. In the 1990s, surveys in Bulgaria, Hungary and Romania identified Roma on the basis of an electoral list in combination with objective criteria such as the traditional way of life. Corsi et al. (2010), p. 103. Simon (2007), p. 54. PER (2000), pp. 14 and 16.

Section 5.3.1 focused on the complex and fluid nature of Roma identities. See also Chap. 1 (Sect. 1.​1.​2) on the ambiguous conceptualisation of Roma.

Messing (2014), p. 816.

Brüggemann and Bloem (2013, pp. 535 and 536) explain that experience with data disaggregation for Romani-speaking students in Slovakia in the framework of PISA 2009 shows that only 1.2% of 4555 participating students say they speak Romani as the first language at home. Underlying reasons for this low percentage are multiple, such as low school attendance rate among Roma or placement in special schools not included in the research. Moreover, respondents may not speak Romani at home and if they do, be reluctant to say so out of fear for discrimination or because they do not consider it to be a formal language. Ringold et al. (2005), pp. 25 and 26. Krizsán (2001), p. 190.

Corsi et al. (2010), pp. 104 and 105.

Brüggemann and Bloem (2013), pp. 525–530. Sections 5.5.1–5.5.3 look at challenges to sampling Roma. See also Sect. 5.6 on bias and limited comparability and representativeness.

Chopin et al. (2014), pp. 49 and 62. PER (2000), p. 14.

The police terminated the practice after investigation by the Hungarian Minority Commission. Krizsán (2001), p. 189. Data misuse was previously addressed in Sect. 5.1.3 on the multiple reasons for the widespread lack of reliable data on Roma. See also Chap. 4 (Sect. 4.​2.​4) where discriminatory ethnic profiling by public bodies was identified as a risk of ethnic data collection.

See Chap. 4 (Sect. 4.​5.​4) on proxies for ethnicity.

Farkas (2017), p. 6.

Messing (2014), pp. 815, 818, 819, 824 and 825. Farkas (2004), p. 20. Alternative or supplementary approaches to ethnical identification for equality and anti-discrimination purposes were considered in Chap. 4 (Sect. 4.​5.​3).

Chopin et al. (2014), pp. 12 and 13. Krizsán (2001), p. 175.

Babusik (2004), p. 14.

For example, the educational development of Roma children in Hungary was tracked on the basis of statistics gathered via the perception of teachers. Babusik (2004), pp. 14, 17 and 18. Chopin et al. (2014), pp. 54 and 63. Rughiniș (2010), p. 356. Ringold et al. (2005), pp. 25 and 26. Csepeli and Simon (2004), p. 145. Krizsán (2001), pp. 170, 171, 192, 193 and 197.

Messing (2014), p. 815. Csepeli and Simon (2004), pp. 137, 139 and 148. Covrig (2004), p. 95. Krizsán (2001), p. 197.

Rughiniș (2010), pp. 356 and 366.

Id. at pp. 356 and 357.

Id.

Csepeli and Simon (2004, pp. 148 and 49) add that it is difficult to compare interviewers’ image of Roma with the image held by the general population of this minority, because the socio-demographic background of former differs from the latter: generally speaking, interviewers are younger, more educated and mainly female. Additionally, age, education and religiousness have an impact on anti-Gypsy prejudice levels.

While overlap was acceptably high in Bulgaria, perceptions of Roma identity strikingly dissimilar in Hungary and Slovakia. Messing (2014), pp. 816, 817 and 820. Rughiniș (2010), pp. 355 and 356.

CoE, Roma and Statistics (22–23 May 2000), para. 17. Messing (2014), pp. 816 and 817. Rughiniș (2010), pp. 354 and 367. Ladányi and Szelényi (2001), pp. 81, 82 and 86.

Messing (2014), pp. 812 and 813. Rughiniș (2010), pp. 354 and 367. Simon (2007), p. 41. Ladányi and Szelényi (2001), 81 and 88.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 41. Rughiniș (2010), pp. 366 and 367. Ringold et al. (2005), p. 28.

Poverty is interpreted as a sign of ethnicity. Messing (2014), p. 815.

Milcher and Ivanov (2004, p. 10) explain that this is because the worst-off are recognisably Romani and most unlikely to be integrated in majority communities. Messing (2014), pp. 816, 817, 822 and 823. Simonovits and Kézdi (2014), p. 8. Rughiniș (2010), pp. 359 and 360. Simon (2007), pp. 54 and 55. UNDP (2005), p. 10. Babusik (2004), p. 16. Farkas (2004), p. 21.

FRA (2013b), p. 8. Rughiniș (2010), pp. 356 and 361. Babusik (2004), p. 15. Krizsán (2001), p. 196.

Babusik (2004), pp. 17 and 18. Krizsán (2001), pp. 170, 171 and 197.

In case of disagreement, the ethnicity of a child was determined to be majority population. OSF is of the opinion that this methodology can only be replicated in Bulgaria. The Roma minority self-government is elected by local Roma. It represents the interests of Roma as a national minority in various fields such as education and culture. Chopin et al. (2014), pp. 31–33, 62 and 63. Krizsán (2001), p. 197.

Ringelheim (2006/7), pp. 70 and 71. Farkas (2004), p. 19.

OSF (2010), p. 41. This example involves self-identification as well as identification by members of the group. The combination of different ethnical identification approaches will be addressed in Sect. 5.4.5 as the fifth approach to ethnic data collection.

The scholarship covers full tuition and a living stipend. Eligibility is limited based on citizenship. Central European University explains the rules on its website: https://​www.​ceu.​edu/​admissions/​funding-fees/​ref (Accessed 12 May 2019).

The letter must briefly describe the organisation and expand on the connection with the applicant (how long the organisation has known the applicant and in what capacity) and his/her involvement in Roma related activities. In addition to the letter of recommendation, Roma students must also submit an essay in which they address, among others, how their personal and professional goals will contribute to the collective advancement of the Roma community, whether they consider themselves to bea role model for younger Roma, what defines them as Roma and what makes them feel a part of the Roma community. For more information, see the webpage of the scholarship on the website of the Central European University: https://​www.​ceu.​edu/​admissions/​funding-fees/​ref (Accessed 12 May 2019).

Rughiniș (2010), p. 356. Section 5.4.3 addressed external identification through visual observation. Alternative or supplementary approaches to ethnical identification for anti-discrimination purposes were analysed in Chap. 4 (Sect. 4.​5.​3).

See Sects. 5.4.1–5.4.4.

Messing (2014), pp. 814, 815 and 820. Ringelheim (2013), pp. 53 and 54. Ivanov (2012), pp. 80 and 81. Rughiniș (2010), pp. 355 and 356. Csepeli and Simon (2004), pp. 136–138. Farkas (2004), p. 21. Milcher and Ivanov (2004), p. 13. Krizsán (2001), p. 176.

Babusik (2004), pp. 15 and 16. Csepeli and Simon (2004), pp. 135 and 36. For the advantages and disadvantages of proxies for Roma ethnicity, visual observation and identification by other Roma, see Sects. 5.4.2–5.4.4.

Hungary, Act No. 4 (IV) on the Criminal Code, 1978, art. 174/B (Any person who assaults another person for being part, whether in fact or under presumption, of a national, ethnic, racial or religious group, or of a certain societal group, or compels him by applying coercion or duress to do, not to do, or to endure something, is guilty of a felony punishable by imprisonment for up to 5 years). Farkas (2004), p. 21.

Farkas (2004), p. 22.

The survey was carried out during the school year 2009–2010. Farkas (2017), p. 34. D.H. and Others was analysed previously in Sect. 5.2.4 on the significance of international and European monitoring and litigation.

Data were collected based on observation by a third party, which was informed by indirect criteria provided by class teachers. Farkas (2017), p. 34.

Messing (2014), p. 816. FRA (2013b), p. 12. Ivanov (2012), pp. 92 and 93. O’Higgins (2012), p. 9. Rughiniș (2010), pp. 352 and 353. UNDP (2005), p. 7. Milcher and Ivanov (2004), pp. 8 and 9. FRA and UNDP (2012), pp. 10, 29 and 30. The self-identification approach was discussed in Sect. 5.4.1.

Messing (2014), p. 816. Ivanov (2012), p. 92.

Id.

Ivanov (2012), p. 94. UNDP (2005), p. 7.

There are also variations between countries: 5% of selected respondents in Romania rejected Romani identity, compared to 14% in Bulgaria. Messing (2014), p. 816. Rughiniș (2010), pp. 352 and 353.

Milcher and Ivanov (2004), pp. 9 and 10.

Their responses do, however, implicitly suggest they are Roma but that they prefer not to reveal their ethnic background. Milcher and Ivanov (2004), pp. 9 and 10.

Farkas (2017), p. 25.

Id.

Simon (2007), pp. 47, 54 and 55. Krizsán (2001), pp. 190 and 191.

Simon (2007), pp. 54 and 55. Krizsán (2001), pp. 190 and 191.

Simon (2007), pp. 54 and 55.

Waldron (2011).

The list contains 16 predefined categories; Roma is not a separate category but can be included under the category any other. UK, Police and Criminal Evidence Act (PACE) Code A (1984), art. 4.3(a) and 4.5. UK, PACE Code A Revised code of practice for the exercise by: police officers of statutory powers of stop and search (2013), 18, 22A and Annex B. Discriminatory ethnic profiling by public bodies was cited as one of the risks of ethnic data collection in Chap. 4 (Sect. 4.​2.​4).

Both the answer of the respondent and the answer of the officer must be recorded. Respondents must be explicitly presented with the option not to state their ethnicity and indicate the reason.

FRA (2010), p. 53. It was explained in Sect. 5.4.3 that visual observation can be an appropriate approach when gathering data on those perceived to be Roma by others, as is the case when researching ethnic profiling by police officers. The uncovering of discriminatory ethnic profiling was also cited as a benefit of ethnic data collection in Chap. 4 (Sect. 4.​2.​4).

FRA and UNDP (2012), p. 30. Ivanov (2012), pp. 91 and 92.

This was previously highlighted in Sect. 5.4.1 on the factors that influence whether or not Roma self-identify as being Roma. The importance of the context in ethnic data collection will be discussed further in Sect. 5.5.4 on the impact of the type of interview, interviewers and questions asked on data quality, and in Sect. 5.8.2 on active participation as a key principle of ethnic data collection on Roma.

See Sect. 5.3 on challenges to the construction of ethnic categories for Roma and Sect. 5.4 on the appropriateness of different ethnical identification approaches for Roma.

Rughiniș (2010), p. 358. Babusik (2004), pp. 15, 17 and 18.

Milcher and Ivanov (2004), p. 9. The multifarious data sources on Roma were analysed in Sect. 5.2.

Messing (2014), p. 818. FRA (2013b), p. 6. The limits of official statistics were cited in Sect. 5.2.1.

The use of electoral lists for sampling may create a bias because it is possible that the number of Roma that do not register their residence is possibly proportionally higher compared to other ethnic groups. FRA and UNDP (2012), p. 29. Rughiniș (2010), pp. 353 and 354. Milcher and Ivanov (2004), pp. 8 and 13.

Respondents for custom-built Roma samples designed to be representative of the Romani population are found by a targeted search of selected settlements. Rughiniș (2010), pp. 340 and 361.

Rughiniș (2010), p. 366.

Id. at p. 361.

Messing (2014), p. 818. Rughiniș (2010), p. 359.

Rughiniș (2010), p. 359.

UNDP (2005), p. 7. Milcher and Ivanov (2004), pp. 8, 9 and 13. See Sect. 5.2.1 on official statistics.

In the 2011 UNDP/World Bank/EC survey, “the methodological decision was taken to assume (…) that the propensity to under-report was identical for each region within an individual country”. FRA (2013b), pp. 9 and 10. Ivanov (2012), pp. 92 and 93. UNDP (2005), pp. 7–9.

It is argued to be representative for roughly about 85% of Roma in each country, thereby providing a good basis for quantitative socio-economic indicators, such as quality of life, life expectancy, income and access to services of Roma. Messing (2014), p. 820. UNDP (2005), p. 9. Limited representativeness of surveys on Roma will be addressed in Sect. 5.6.3. For more on data dissemination, see Sect. 5.6.4.

Messing (2014), p. 820. FRA (2013b), pp. 9 and 10. Rughiniș (2010), p. 361. UNDP (2005), p. 9. See Sect. 5.5.5 for a discussion on the impact of time and financial considerations.

Messing (2014), p. 825.

FRA (2013b), p. 8. FRA and UNDP (2012), p. 29. Rughiniș (2010), pp. 358 and 367. Babusik (2004), p. 18. Milcher and Ivanov (2004), p. 13. The different data sources on Roma were analysed in Sect. 5.2.

See the discussion on the multifarious data sources on Roma in Sects. 5.2.2–5.2.4.

For example, within the framework of the FRA Roma Pilot Survey, associations and NGOs helped to identify a sufficient number of camps for interviews in France. In Italy, the overwhelming majority of local Roma/Sinti associations helped to collect geo-demographic information to identify households in different municipalities. Messing (2014), p. 825. FRA (2013b), pp. 7 and 8. Ivanov (2012), pp. 92 and 93. Rughiniș (2010), pp. 358, 360, 361 and 367. FRA (2008), p. 51. UNDP (2005), pp. 7 and 8. Babusik (2004), pp. 7–18. Milcher and Ivanov (2004), pp. 10–13. This was previously highlighted in Sect. 5.2 on the multifarious data sources on Roma.

When samples overestimate the size of Roma populations, an insufficient numbers of interviews can be conducted. Also, settlements may disappear or be dismantled before the interviewing process starts, as was the case in France during the FRA Roma Pilot Survey. FRA (2013b), pp. 11 and 15.

This was solved by decreasing the Roma sample in Poland and in Italy from 1000 interviews to 600 interviews with Roma and increasing the sample size to 1100 interviews in other countries participating in the survey (except for France). FRA (2013b), p. 10.

FRA (2013b), pp. 10 and 11.

These include considerable learning disadvantages, disadvantaged socio-economic background, lesser pre-school experience, higher representation in vocational tracks and higher repetition rates. Brüggemann and Bloem (2013), pp. 525, 529, 530, 535 and 536. The different factors that impact on the analysis of data sets on Roma will be reviewed in Sect. 5.6.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 39. Financial restrictions will be considered in Sect. 5.5.5.

Farkas (2017), p. 36. For more on privacy and data protection concerns, see Sect. 5.7.

For example, the World Bank over-sampled Roma when researching poverty among Roma in Bulgaria, Hungary, and Romania for its study Breaking the Poverty Cycle. Brüggemann and Bloem (2013), p. 537. Ringold et al. (2005), pp. 27 and 28.

Rughiniș (2010), p. 342. Sample boosters were previously mentioned in Chap. 4 (Sect. 4.​3.​2.​2) when discussing the disadvantages of research data as a source of ethnic data collection.

Messing (2014), p. 819.

This may be due various factors, such as reluctance among Roma respondents to be interviewed by an unknown interviewer and/or reluctance among interviewers to enter Roma dense residential areas. Messing (2014), p. 819.

FRA (2013b), p. 12.

Messing (2014), p. 825.

Id. at p. 818.

FRA (2013b), p. 9.

The EU-MIDIS survey adopted four different sample designs but used only one approach per country. The FRA Roma Pilot survey adopted a multi-stage probability sampling design that reflected differences at the Member State level to get “a reasonably good coverage of the target population on the national level given the time and resources available”. FRA (2013b), p. 9. FRA (2009c), pp. 23 and 24.

FRA (2013b), p. 12. Rughiniș (2010), p. 358.

For the FRA Roma Pilot Survey, adult meant 16 or older. Respondents were chosen randomly using the last birthday method, or—if not known or celebrated—selection rotated between eldest/youngest and male/female household members on the basis of the last identification number digit of the questionnaire. ERRC (2013b), p. 60. FRA (2013b), pp. 9 and 14.

For instance, the ERRC’s Hidden Health Crisis report in Romania (2013b, p. 60) asked “all persons aged 15 and above who are at home at the time of the interview (…) to answer questions pertaining to their persons” and “in each household certain data for each member of the household, including all children, was requested from an adult”. See also: FRA (2013b), p. 9.

See Sect. 5.8.3 on the importance of active participation of Roma in ethnic data collection practices.

FRA (2009b), p. 13. Milcher and Ivanov (2004), p. 9.

ERRC (2013b), p. 59. FRA (2013b), pp. 9 and 12. FRA and UNDP (2012), p. 30. FRA (2009b), pp. 4 and 13. FRA (2009c), p. 24. Milcher and Ivanov (2004), pp. 11 and 12.

This works where dwellings have street names, intersections and number and where households live in one dwelling, though different households may live in the same dwelling. FRA (2013b), p. 13. FRA (2009c), p. 24.

FRA (2013b), pp. 7 and 8. FRA and UNDP (2012), p. 30.

In order to facilitate quality control, interviewers had to document their sampling activity on maps and complete matching route administration sheets. FRA (2009c), p. 24.

This is done because proximity of residence means that non-Roma share certain characteristics of the local environment with Roma. ERRC (2013b), p. 59. FRA (2013b), pp. 9 and 14. FRA and UNDP (2012), p. 29.

Milcher and Ivanov (2004), p. 12.

FRA (2013b), pp. 12 and 13.

Id.

ERRC (2013b), p. 59. FRA (2013b), pp. 12 and 13.

FRA (2013b), pp. 12 and 13.

FRA (2013b), p. 9. FRA and UNDP (2012), p. 30. FRA (2009c), p. 24.

FRA (2013b), p. 13. FRA (2009c), p. 24.

FRA (2009c), p. 24.

FRA (2013b), p. 15. FRA and UNDP (2012), p. 10. Milcher and Ivanov (2004), pp. 8 and 11.

FRA (2009c), pp. 23 and 24.

FRA (2009c), p. 28.

FRA (2013b), pp. 18 and 19.

For the FRA Roma Pilot survey (2013b, pp. 18 and 19), interviewers were selected on the basis of “experience with simple random route sampling; experience in communicating with vulnerable minority groups; being comfortable with/not prejudiced against Roma; overall interviewing experience; communication skills; responsibility and professional ethics; previous experience through EU-MIDIS if possible”. See also: FRA (2009c), p. 29.

FRA (2013b), p. 15.

For more on the need for active participation of Roma in ethnic data collection, see Sect. 5.8.2.

FRA (2013b), p. 16.

The importance of active participation as a key principle governing ethnic data collection will be addressed in Sect. 5.8.2, and the limited comparability of data on Roma in Sect. 5.6.2.

See Chap. 2 (Sect. 2.​3.​2) on social inclusion: a multi-dimensional concept with varying meanings.

These include, among others, lack of trust, awareness and participation of Roma, the too restricted interpretation of data protection legislation and the absence of positive action measures for Roma.

For instance, those collecting the data may assume that the way of life of majority populations is the common standard and that integration equals adaptation to this way of life by Roma. The notion integration was analysed in Chap. 2 (Sect. 2.​3.​3).

Active participation of Roma as a key principle of ethnic data collection makes up the focus of Sect. 5.8.2. Suggested further reading on the importance of a bottom-up approach to investigating social inclusion: Wolff and De-Shalit (2013).

This was done with EU-MIDIS and the FRA Roma Pilot Survey. FRA (2013b), p. 16. FRA (2009c), p. 21.

UNDP’s Faces of Poverty Faces of Hope survey included a field control on 10–15% (depending on the country) of the sample. The FRA Roma Pilot survey included random quality-control checks on interviews. FRA (2013b), p. 15. UNDP (2005), p. 10.

FRA (2009c), p. 29.

UNDP (2005), p. 10.

FRA (2009c), p. 30. UNDP (2005), p. 10.

Messing (2014), p. 820. Goodwin (2004), p. 1439. This was previously addressed in Sect. 5.5 on methodological challenges to conducting surveys on Roma populations.

Goodwin (2004), p. 1439.

Messing (2014), p. 820.

FRA (2009c), p. 271.

Messing (2014), p. 820.

For instance, the 11 Member States covered by the FRA Roma pilot survey (Bulgaria, the Czech Republic, France, Greece, Hungary, Italy, Poland, Portugal, Romania, Slovakia and Spain) (2013b), p. 6) were chosen on the basis of the large size of their Roma populations and the assumption that Roma live in concentrated geographical areas, which was considered necessary for probability sampling “at reasonable cost and within a limited time frame”. Messing (2014), pp. 820 and 821.

Surdu (2019), p. 12.

PER (2000), p. 10. This point was previously made in Chap. 4 (Sect. 4.​3) when introducing the different data sources of ethnic data collection.

For example, the All-Ireland Traveller Health Study from 2014 was not used to create a Traveller health action plan in Ireland. CoE, Roma and Statistics (22–23 May 2000), para. 9. Abdikeeva (2014), pp. 27 and 30. Chopin et al. (2014), p. 62. PER (2000), p. 10. The multiple reasons for the widespread lack of (reliable) ethnic data on Roma were considered in Sect. 5.1.3.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 27. ERIO (2009), pp. 21 and 22. The lack of (reliable) data on Roma in Europe and the multiple causes thereof were discussed in Sect. 5.1.

Messing (2014), p. 824. PER (2000), pp. 10 and 11.

Ringold et al. (2005), p. 57.

Id. at pp. xix and 57.

Ringold et al. (2005), p. 26. The particular vulnerability of Roma in former and current times made up the focus of Chap. 1 (Sections 1.​2.​1 and 1.​2.​2).

Ringold et al. (2005), p. 26.

PER (2000), p. 16.

Ringold et al. (2005), p. 26. The over-representation of Roma in crime data was addressed in Sect. 5.2.3. It will be considered in Sect. 5.6.4 that dissemination of data on Roma may not contribute to or promote anti-Gypsyism.

See Chap. 4 (Sect. 4.​3) for an overview of the four main data sources of ethnic data collection.

Chalenges to the construction of ethnic categories for Roma were analysed in Sect. 5.3. See also Sect. 5.4.1 on the context-dependency of Roma’s willigness to self-identify as being Roma.

FRA (2013a), p. 190. PER (2000), p. 27. The added value of surveys, research and discrimination testing was mentioned in Sect. 5.2.2.

Messing (2014), pp. 811, 821–823 and 825. FRA and UNDP (2012), p. 17. FRA (2011), p. 17. Ringold et al. (2005), pp. 25 and 26. See Sect. 5.3 on the challenges to the construction of ethnic categories for Roma, Sect. 5.4 on the appropriateness of different ethnical identification approaches for Roma, and Sect. 5.5 on the methodological challenges to interviewing Roma populations.

Messing (2014), pp. 812 and 824.

Messing (2014), p. 812. Indicators were mentioned throughout Sect. 5.1 on large data gaps on Roma and in Sect. 5.2.1 on the limitations of official statistics as a data source on Roma.

Messing (2014), pp. 821, 822 and 825.

Id.

Messing (2014), pp. 821–825. O’Higgins (2012), p. 18.

Guy et al. (2010), pp. 8 and 9.

Using the same questionnaires to interview Roma in different countries can improve data comparability. Within the framework of EU-MIDIS, comparability between data on minorities and majority populations was improved by using survey structures and questions from other international surveys such as the Eurobarometer. See: FRA (2013b), p. 17. FRA (2009c), pp. 28, 29, 31 and 255. Milcher and Ivanov (2004), pp. 8, 10 and 11. UNDP (2003), pp. 9 and 86.

For instance, the application of an identical sampling design methodology for surveys in the framework of UNDP’s ‘Avoiding the Dependency Trap’ report produced comparable cross-country results. McDonald and Negrin (2010), p. 37. Milcher and Ivanov (2004), p. 8. Ringold et al. (2005), p. 189.

It concerned Bulgaria, Czech Republic, Hungary, Romania and Slovakia. The use of the same sample methodology made it possible to develop a common dataset on core indicators and to ensure comparability and consistency of results. O’Higgins (2012), p. 8. FRA (2013b), pp. 9 and 15.

FRA (2013b), p. 9. FRA (2012a), p. 168. FRA and UNDP (2012), p. 31. Sampling and interview training were mentioned in Sect. 5.5.

See, for example, UNDP (2005): a regional survey of Roma in the Central European and Western Balkan countries were used as baseline data for participating countries in the Decade of Roma Inclusion. See also: FRA and UNDP (2012), pp. 30 and 31.

The EU-MIDIS surveys use the same standard questionnaire to conduct interviews in all Member States and across all groups selected. Given that Roma were a target group in both EU-MIDIS surveys, trend analysis and results’ comparison is possible between EU-MIDIS I and EU-MIDIS II. FRA (2017), p. 14. FRA (2016). Furthermore, some general level comparisons of results on employment status of Roma from EU-MIDIS I are possible to those of the Eurobarometer and the European Social Survey from 2008. FRA (2012b). FRA (2009a), pp. 14 and 17. FRA (2009b), p. 13. FRA (2009c), pp. 20, 27, 30, 242, 269 and 271.

The context-specificity of methodological choices when collecting ethnic data on Roma was previously highlighted in Sects. 5.3–5.5.

This was previously stressed in Sect. 5.3.1 when considering the challenges to the construction of ethnic categories for Roma. See also Chap. 1 (Sect. 1.​1.​2) on the ambiguous conceptualisation of the notion Roma.

FRA (2009b), p. 14.

Milcher and Ivanov (2004), p. 10. The author previously addressed the question who ‘the Roma’ are in Chap. 1 (Sect. 1.​1).

Ringold et al. (2005), p. 92.

On the other hand, for resource allocation purposes that are usually based on headcount, consensus should be found at the national level on the number of Roma. UNDP (2005), p. 9.

Brüggemann and Bloem (2013), p. 536. FRA (2013b), p. 28. Methodological challenges to interviewing Roma populations were reviewed in Sect. 5.5.

ERIO (2009), p. 27. Section 5.5.2 addressed how small sampling sizes affect quality of data on Roma.

Messing (2014), p. 825. OSF (2010), p. 41. UNDP (2005), p. 9.

FRA (2009b), p. 13. FRA (2009c), pp. 20, 30 and 154.

FRA (2013b), p. 9.

The same goes for the non-Roma respondents in the survey. FRA (2013b), p. 9. FRA and UNDP (2012), pp. 10 and 29. UNDP (2005), p. 10. FRA (2016, p. 8) uses this reasoning explicitly to describe the representativeness of the EU-MIDIS II survey results on Roma, by stating that “the data are representative for Roma living in the nine EU Member States in geographic or administrative units with density of Roma populations higher than 10%, who self-identify as “Roma” or as members of one of the other groups covered by this umbrella terms. Therefore, the findings presented in this report reflect the living conditions, enjoyment of fundamental rights and experiences with discrimination of up to 80% of Roma living in the nine EU Member States surveyed”.

For example, while the FRA survey on the situation of Roma in 11 EU Member States (FRA and UNDP 2012, p. 16) indicates clear differences between the rates of Roma and non-Roma in paid employment, further investigation is needed to explain the low employment rates of Roma compared to non-Roma. This was briefly mentioned in Sect. 5.2. FRA (2009c), p. 271. Ringold et al. (2005), p. 41.

See, for example: FRA (2010), p. 31.

Id. at p. 28.

FRA (2010), p. 28. FRA (2009c), p. 20.

FRA (2009c), p. 161.

Id.

FRA (2009c), p. 270.

Id.

This is called ethnogenesis. Makkonen (2010), p. 236. Yanow (2003), pp. 12 and 13.

Anti-Gypsyism was defined and briefly addressed in Chap. 1 (Sect. 1.​2.​1).

The importance of awareness-raising will be discussed in Sect. 5.8.1. The role of the media was briefly addressed in Chap. 1 (Sect. 1.​2.​1) the particular vulnerability of Roma in Europe today.

Ivanov (2013).

The violation of privacy and data protection rules was cited as one of the risks of ethnic data collection in Chap. 4 (Sect. 4.​2.​5). Problems due to misinterpretation or too restrictive interpretation of the applicable European data protection rules were also highlighted in Chap. 3 when introducing the core data protection rules (Sect. 3.​3) and when discussing the special rules that apply to sensitive data (Sect. 3.​5).

See Chap. 3 (Sects. 3.​3–3.​7) for a discussion of international and European data protection rules and principles.

See: Convention for the Protection of Individuals with regard to Automatic Processing of Personal Data (28 January 1981), including changes introduced by the Protocol amending Convention 108 (18 May 2018) (Convention 108+), art. 2(b). Regulation 2016/679 of the European Parliament and of the Council on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (27 April 2016) (GDPR), art. 4.2. The definitions included in these articles were included in full in Chap. 3 (Sect. 3.​3).

For instance, the Hungarian Census includes a voluntary question on ethnicity since 2001. ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 12. Krizsán (2001), pp. 190 and 192.

Hermanin (2013), pp. 105 and 106.

For example, Hungarian Labour Law allows employers to collect data on their employees to adopt an equal opportunities plan for a fixed time period, provided they respect arts. 2.2 and 3.2 of Act No. 63 (LXIII) on the Protection of Personal Data and the Publicity of Data Public Interest (1992). Makkonen (2006), p. 100. Farkas (2004), p. 22.

Farkas (2004, p. 19) refers to art. 19(1)b of Hungarian Act No. 125 (CXXV) on Equal Treatment and the Promotion of Equal Opportunities (2003), which requires complainants to establish his racial or ethnic origin or national or ethnic minority characteristic in order for the burden of proof in cases of ethnic or racial discrimination to be reversed. Moreover, when the case concerns indirect discrimination, the plaintiff must also establish the comparator, which becomes very hard or even impossible due to the data protection provisions. Goodwin (2004), p. 1439. See Chap. 3 (Sect. 3.​1.​1) on the argument made at UN level that the right to be free from discrimination includes the right to access information that could help to prove discrimination.

Vivaldi (2014). Data subjects’ rights in relation to the processing of their personal data were discussed in Chap. 3 (Sect. 3.​4.​3).

Lawful processing was considered in Chap. 3 on the general rules that apply to all sorts of personal data (Sect. 3.​4.​1) and on the special rules that apply to sensitive data (Sect. 3.​5).

For an analysis of the special data protection rules included in Convention 108+ and the GDPR, see Chap. 3 (Sect. 3.​5).

The general and sensitive data protection rules included in these instruments were analysed in Chap. 3 (Sects. 3.​3–3.​5).

Alidadi (2017), p. 20. Farkas (2017), p. 36. This will be discussed further in Sect. 5.7.2 on the need for strong data security.

See Chap. 3 (Sect. 3.​5) on special data protection rules applying to sensitive data categories.

Farkas (2017), p. 45.

Lack of political will was previously mentioned in Chap. 3 (Sects. 3.​3 and 3.​5) and Chap. 4 (Sect. 4.​2.​5). It will be highlighted as a key element in the framework of ethnic data collection in Sect. 5.8.3.

ERIO (2009), pp. 30 and 31. Makkonen (2006), p. 109. This was discussed in Sect. 5.1 on the large quantitative and qualitative data gaps on Roma communities in Europe.

ERIO (2009), pp. 30 and 31. The reluctance among Roma to self-identify was addressed in Sect. 5.4.1.

Simon (2007), p. 57. PER (2000), p. 16. Section 5.1.3 underlined the context-dependency of attitudes among Roma towards ethnic data collection. This will be discussed further in Sect. 5.8.1 on the importance of awareness-raising when collecting data on Roma.

PER (2000), p. 25. Purpose specification was cited as the fourth general data protection rule in Chap. 3 (Sect. 3.​4.​4).

Hermanin (2013), p. 109. PER (2000), pp. 11 and 16.

Hermanin (2013), p. 109. Krizsán (2001), pp. 170 and 171. Integrity and confidentiality through appropriate security measures was mentioned as the eighth general data protection rule in Chap. 3 (Sect. 3.​4.​8). See also the five operational and organisational principles for personal data processing in Chap. 3 (Sect. 3.​7).

Hermanin (2013), p. 109. Vroon (2010). Davidović and Rodrigues (2010), p. 170. Krizsán (2001), pp. 170 and 171. These cases were first introduced in Sect. 5.1.3.

This was the case in Ede and Enschede. Damman (2010). De Groene Amsterdammer (2010). Vroon (2010).

De Groene Amsterdammer (2010). Vroon (2010).

Id.

The victims each received 5000 Swedish Krona (approximately 512 euro) in damages for privacy invasion. Eleven people, eight adults and three children, sued for further damages and were awarded each 30,000 Swedish Krona (approximately 3074 euro) by the Stockholm district court in June 2016. Albert (2017). The Local (2016).

ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 12. ERIO (2009), p. 21. Anonymisation was discussed in Chap. 3 when analysing the storage limitation rule (Sect. 3.​4.​7) and the data security rule (Sect. 3.​4.​8).

FRA and UNDP (2012), p. 30. FRA (2009a), p. 20. FRA (2009b), p. 14.

Krizsán (2001).

Pavee Point Traveller and Roma Centre (2013).

This is the case in the Hungarian Census since 2001. OSF (2010), p. 40. Krizsán (2001), p. 190. PER (2000), p. 25.

This was discussed in Chap. 3 (Sect. 3.​4.​7) on anonymisation as the seventh general data protection rule. Within the framework of Census, such removal can be done upon completion of reliability checks. In surveys and research, computer programmes could eliminate and eventually also destroy names and addresses contained in the original survey forms. CoE, Roma and Statistics (22–23 May 2000), para. 43. EURoma (2009), p. 5.

For instance, sensitive data collected by Hungarian employers to adopt an equal opportunities plan may not be kept beyond the last day of the validity period of the plan. OSF (2010), p. 41. Ringelheim (2006/7), pp. 70 and 71. Makkonen (2006), p. 100.

Oppenheimer (2008), p. 750.

Krizsán (2001), pp. 173 and 174.

See Chap. 3 on the storage limitation rule (Sect. 3.​4.​7) and on professional secrecy and confidentiality of data processing (Sect. 3.​7.​1).

Hermanin (2013), p. 109. PER (2000), pp. 11 and 16.

PER (2000), p. 20. Data accuracy and the erasure and rectification of inaccurate data were mentioned in Chap. 3 (Sects. 3.​4.​3.​2 and 3.​4.​6).

See Chap. 3 for an overview of data subjects’ rights, including their right to rectification (Sect. 3.​4.​3.​2), and for a review of the protection of personal data through the right to private life (Sect. 3.​6).

ERIO (2009), p. 25. Covrig (2004), pp. 99 and 100.

PER (2000), p. 30. The operational and organisational principles of sensitive data processing, including professional secrecy and confidentiality, political independency, ethical norms and standards, trained staff and management involvement, and active and meaningful involvement of all relevant stakeholders, were discussed in Chap. 3 (Sect. 3.​7).

See Chap. 3 (Sect. 3.​3) on the introduction of the two sets of core data protection rules and Chap. 4 (Sect. 4.​2.​5) on fear of violating privacy and data protection rules as a commonly used excuse not to collect ethnic data for equality and anti-discirmination purposes.

This was previously briefly mentioned in Sect. 5.1.1 on international and European calls for data on Roma. It will be also be stressed in Sect. 5.8.3 on the need for political will to make ethnic data collection for Roma work. See also Chap. 6 (Sect. 6.​3.​1) on the need for an analytical, legal and regulatory framework on positive action, and Chap. 9 (Sect. 9.​2.​2) on the need for political will to use positive action to advance the situation of Roma in Europe.

Chopin et al. (2014), p. 63. ERRC (2013b), p. 42. Waldron (2011). Simon (2007), p. 55. Hollo (2006), p. 29.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 45.

Such joint supervision is put in place in the UK in relation to the collection of data on ethnicity of pupils for national anti-discrimination policies. Waldron (2011).

McDonald and Negrin (2010), p. 36. As explained in Sect. 5.3.2, close co-operation with social scientists is useful to determine the feasibility and operationalibility of different methodological options. The importance of active and meaningful involvement of different stakeholders was stressed in Chap. 3 (Sect. 3.​7.​5).

Pavee Point partnered with other anti-racism NGOs to actively advocate for the inclusion of a Traveller question in the Census. The Irish Census does not include a Roma category. Abdikeeva (2014), p. 25. Healy (2013).

See Sect. 5.8.2 on active participation of all relevant stakeholders in ethnic data collection.

ERIO (2009), p. 21.

See Sect. 5.1 on large data gaps on Roma in Europe, Sect. 5.2.2 on the added value of research data, Sect. 5.2.3 on the shortcomings of complaints and crime data, Sect. 5.4.1 on Roma reluctance towards self-identifying as Roma, Sect. 5.4.5 on combining different ethnical identification approaches for more accurate results, Sect. 5.5.4 on the impact of type of interview, interviewers and questions asked on data quality, and Sect. 5.7.3 on the need for close co-operation between experts.

The situation of Roma in Europe was introduced in Chap. 1 (Sect. 1.​2).

Chapter 3 (Sects. 3.​3–3.​7) includes the analysis of the core data protection rules and principles.

This was addressed in Sect. 5.1.3 on the multiple reasons for the widespread lack of (reliable) data on Roma in Europe, Sect. 5.2.1 on the limitations of official statistics as a data source, Sect. 5.2.3 on the unreliability of complaints data due to under-reporting, and Sect. 5.4.1 on Roma reluctance towards self-identification.

See Sect. 5.1.3 on the reasons for the widespread lack of (reliable) data on Roma, Sect. 5.2.3 on the unreliability of complaints data, and Sect. 5.4.1 on the reluctance of many Roma to self-identify as Roma. The EU-MIDIS survey demonstrates that awareness levels vary among Roma and across Member States. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 15.2. Abdikeeva (2014), p. 28. FRA and UNDP (2012), pp. 12 and 29. Ivanov (2012), pp. 80 and 81. ERIO (2009), pp. 23, 30 and 31. FRA (2009b), pp. 3, 6–9 and 12. FRA (2009c), pp. 13, 17, 158, 159, 163, 165, 244 and 273. Covrig (2004), p. 94. Petrova (2004), p. 5. Krizsán (2001), pp. 157 and 192. PER (2000), pp. 4, 5 and 29.

ERIO (2009), p. 31.

See, for example: ACFC, Second Opinion on Slovakia (26 May 2005), para. 27. ACFC, Second Opinion on Hungary (9 December 2004), para. 32. Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 15.2. CoE, Roma and Statistics (22–23 May 2000), para. 37. Ringelheim (2013), p. 54. ERIO (2009), p. 3. FRA (2009c), p. 13. Chapter 3 (Sects. 3.​3–3.​7) analysed the general and special data protection rules and principles at CoE and EU level.

Abdikeeva (2014), p. 25. ERIO (2009), p. 30 and 31. Hollo (2006), p. 29. Covrig (2004), pp. 99 and 100. Banton (2001), p. 65 and 66. Krizsán (2001), p. 187. PER (2000), pp. 21, 22 and 29.

ACFC, Second Opinion on the Czech Republic (24 February 2005), para. 37. ACFC, Second Opinion on Slovakia (26 May 2005), para. 27. ACFC, Second Opinion on Hungary (9 December 2004), paras. 31 and 32. CoE, Roma and Statistics (22–23 May 2000), para. 50. Abdikeeva (2014), p. 29. Chopin et al. (2014), p. 75. Krizsán (2001), p. 158. Chapter 11 will focus on inter-cultural mediation to enhance Roma inclusion.

Ringelheim (2006/2007), pp. 69 and 70. Covrig (2004), p. 100.

Ringold et al. (2005), pp. 127 and 128.

ERIO (2009), p. 31.

Abdikeeva (2014), p. 25.

See Sect. 5.8.2 on the importance of active participation of Roma in data collection practices.

For instance, the number of Roma doubled in the 2011 Montenegrin Census following a grassroots campaign. In Romania, the number of self-identifications rose from 535,140 in 2002 to 621,573 in 2011. A Serbian grassroots campaign led to a 40% increase in the official number of Roma in the 2011 Census. In Montenegro, the official number rose from 2501 to 6251 Roma in the 2011 Census. CoE, Roma and Statistics (22–23 May 2000), paras. 46–49. ERRC (2013b), p. 14. Jovanovic and Haliti (2012). Rorke (2011). OSF (2010), pp. 64 and 65. PER (2000), p. 21.

The campaign received support from the European Commission, the Dutch Embassy and the OSF Roma Initiatives Office. Abdikeeva (2014), pp. 28 and 29.

In total, 83 Roma enumerators were involved in the project. Abdikeeva (2014), pp. 28 and 29. The added value of Roma enumerators will be addressed in Sect. 5.8.2 on the participation of Roma as a key principle to render ethnic data collection truly effective. Chapter 11 will zoom in on the importance of inter-cultural mediation and on the role of Roma mediators.

The self-identification rate rose with 56% in targeted communities compared to 16% at the national level. Geographical limitations of the project were due to budgetary constraints. Abdikeeva (2014), pp. 28 and 29.

Chopin et al. (2014), p. 75.

For instance, Roma self-identification increased by 16% in the 2011 Romanian Census, but still approximately 1.4 million people do not self-identify. Additionally, Roma without identity documents and those without legal residence are not included in Census results. CoE, Roma and Statistics (22–23 May 2000), paras. 48–49. Abdikeeva (2014), p. 30. Jovanovic and Haliti (2012). PER (2000), p. 27.

Abdikeeva (2014), p. 30.

Awareness-raising among majority populations and empowerment of minority populations were cited as benefits of ethnic data collection in Chap. 4 (Sect. 4.​1.​3). The impact of stereotypes and anti-Gypsyism was also addressed in Chap. 1 (Sect. 1.​2.​1) on the particular vulnerability of Roma in Europe.

In Montenegro, the amount of public funds attributed to the Decade of Roma Inclusion was proportional to the official number of Roma. In Serbia, quotas in employment for public administration and the police are determined on the basis of Census data. Positive action will be discussed in Part II of the book. Jovanovic and Haliti (2012).

Makkonen (2010), p. 227. See the introduction to Chap. 3 (Sect. 3.​7) on the operational and organisational principles for sensitive data processing, and to general introduction to Chap. 4.

As explained in Chap. 4 (Sect. 4.​4.​3), revealing one’s minority or ethnic affiliation should be done voluntarily.

Ivanov (2012), pp. 90 and 91. Corsi et al. (2010), p. 122. Ringold et al. (2005), pp. xv, xvi, 12, 13, 116, 117 and 176. The particular vulnerability of Roma was briefly summarised in Chap. 1 (Sect. 1.​2.​1).

ERRC (2004b), p. 39.

CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 9.

ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 15(b). ECRI, Fourth Report on France (29 April 2010), para. 105. ERIO (2009), p. 31. This will be discussed further in Chap. 11 on inter-cultural mediation to enhance Roma inclusion.

FRA (2007), pp. 150 and 151. Goodwin (2004), p. 1440. The added value of discrimination testing as a data source on Roma was addressed in Sect. 5.2.2. See also Chap. 4 (Sect. 4.​3.​2) where discrimination testing was identified as one of the sources of ethnic data collection.

Such a roundtable discussion took place in Hungary in 2000 involving representatives of the Roma minority, expert sociologists and data protection experts. Healy (2013). As stressed in Sect. 5.6.4, data dissemination may not promote anti-Gypsyism.

This has been done successfully in Spain. Ringold et al. (2005), pp. 172 and 173. Inter-cultural mediation to enhance Roma inclusion makes up the focus of Chap. 11.

ECRI, General Policy Recommendation No. 3: Combating Racism and Intolerance against Roma/Gypsies (6 March 1998).

Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 15.8.

CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 14. ODIHR (2014). Problems with data collection on complaints of discrimination and hate crimes against Roma were addressed in Sect. 5.2.3.

Other groups include migrants, people with intellectual disabilities and LGBT. FRA (2014), p. 135.

For instance, reports should avoid blaming Roma communities as a whole when reporting on incidents involving individuals of such communities. ECRI, General Policy Recommendation No. 3: Combating Racism and Intolerance against Roma/Gypsies (6 March 1998). The role of the media was addressed in Chap. 1 (Sect. 1.​2.​1).

ECRI, General Policy Recommendation No. 3: Combating Racism and Intolerance against Roma/Gypsies (6 March 1998).

Past wrongs and lack of acknowledgment were cited in Sect. 5.1.3. See also Chap. 1 (Sect. 1.​2.​1) on the particular vulnerability of Roma in Europe.

Swedish Ministry of Employment (2014).

This is not only the case with Roma, but more broadly with racial and ethnic minority communities in Europe. Decade of Roma Inclusion 2005–2015, Terms of Reference—Decade Declaration (2 February 2005), pp. 3 and 5. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), para. 26. Farkas (2017), pp. 42 and 43. FRA (2014), p. 171. Messing (2014), pp. 814, 817, 818, 824 and 825. Chopin et al. (2014), p. 47. Ringelheim (2013), p. 54. FRA and UNDP (2012), p. 9. McDonald and Negrin (2010), pp. 19, 20, 33, 34 and 36. Rughiniș (2010), p. 352. Ringelheim (2006/7), p. 69. Hollo (2006), p. 4. Ringold et al. (2005), p. 187. Covrig (2004), p. 100. Csepeli and Simon (2004), pp. 135 and 136. Milcher and Ivanov (2004), pp. 12 and 13. Organization for Security and Co-operation in Europe (2000), p. 61. The need for active participation of Roma was stressed repeatedly throughout this chapter (see Sects. 5.1, 5.3, 5.4.1, 5.4.5, 5.5.4 and 5.7.3). Active and meaningful involvement of target groups was cited as the fifth organisational principle of ethnic data collection in Chap. 3 (Sect. 3.​7.​5). Suggested further reading on working with Roma, including on challenges to meaningful participation and on participatory action research: FRA (2018).

Ringold et al. (2005), p. 188.

CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 43. CoE, Roma and Statistics (22–23 May 2000), para. 53. Ringold et al. (2005), p. 187. PER (2000), p. 23. Suggested further reading on communication: FRA (2018), pp. 29–36.

In practice, only 1.5% gave multiple answers per question. Simon (2007), p. 55. Krizsán (2001), p. 196. Challenges to ethnical categorisation on Roma were analysed in Sect. 5.3.

The tenth Common Basic Principle on Roma Inclusion stresses the importance of full participation in public life and active participation of the Roma in the design, implementation and evaluation of policy initiatives to enhance the effectiveness of policies. Milcher and Ivanov (2004, p. 12) promote the mainstreaming of consistent Roma participation in general policies and in Roma policies. This will be discussed further in Chap. 6 (Sect. 6.​3.​2) on the importance of active participation in positive action. Common Basic Principles on Roma Inclusion, as annexed to the Council Conclusions on Inclusion of the Roma (8 June 2009), principle 10. See also: Decade of Roma Inclusion 2005–2015, Terms of Reference—Decade Declaration (2 February 2005), p. 5. ACFC, Second Opinion on Finland (2 March 2006), para. 18 and 144–146 and 168. Report of the Independent Expert on Minority Issues on the Implementation of General Assembly Resolution 60/251 of 15 March 2006 entitled “Human Rights Council” (2 February 2007), para. 89. CERD Committee, General Recommendation No. 27: Discrimination against Roma (16 August 2000), para. 43. CERD Committee, Thematic Discussion on the Question of Discrimination against Roma by the Committee on the Elimination of Racial Discrimination (16 August 2000), paras. 8 and 34. CoE, Roma and Statistics (22–23 May 2000), paras. 13, 20, 29 and 44. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 43. Jovanovic and Haliti (2012). McDonald and Negrin (2010), p. 19. Guy et al. (2010), pp. 4 and 11. Davidović and Rodrigues (2010), pp. 172 and 173. Hollo (2006), pp. 5 and 35. Ringold et al. (2005), p. 188. Krizsán (2001), p. 195. PER (2000), pp. 11, 25 and 34.

See Sect. 5.8.3 on genuine political will a key principle of ethnic data collection on Roma.

Resolution 1740 of the Parliamentary Assembly on the situation of Roma in Europe and relevant activities of the Council of Europe (22 June 2010), art. 9. Farkas (2017), p. 41. FRA (2014), p. 171. LERI was briefly mentioned in Sect. 5.1.1 on international and European calls for data on Roma.

See Sect. 5.3 on challenges to the construction of ethnic categories for Roma, Sect. 5.4.1 on Roma reluctance towards self-identification, and Sect. 5.5.4 on how the type of interview, interviewers and questions asked affect data quality.

CoE, Roma and Statistics (22–23 May 2000), paras. 29 and 43. ERIO (2009), p. 21. Ringold et al. (2005), p. 188. Milcher and Ivanov (2004), pp. 9 and 13.

See, for example: Ringold et al. (2005), p. 188.

Ringold et al. (2005), p. 188. Farkas (2017, p. 42) also highlights the importance of building confidence in data collection.

This was addressed in Sect. 5.1.3 when discussing the multiple reasons for the widespread lack of (reliable) data on Roma, Sect. 5.4.1 on Roma’s reluctance towards self-identification, and Sect. 5.5.4 on the importance to carefully consider the type of interviews, interviewers and questions asked when collecting data on Roma. ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 2. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 43. Ringold et al. (2005), p. 188. Milcher and Ivanov (2004), pp. 12 and 13.

In the Czech Republic, Roma helped to collect data on socially excluded Roma. The results generated in a society-wide dialogue on the issues at hand and in recommendations to ameliorate the situation on the basis of European Social Fund funding allocation. Funding of measures promoting Roma inclusion will be addressed in Chap. 9 (Sect. 9.​2.​5). ECRI, General Policy Recommendation No. 13: Combating Anti-Gypsyism and Discrimination against Roma (24 June 2011), para. 2. Messing (2014), pp. 814, 817 and 818. Guy et al. (2010), p. 35. Hollo (2006), p. 29. Covrig (2004), p. 100.

It concerns the Ministries of Health and Education. OSF (2010), pp. 68 and 69. Inter-cultural mediation to enhance Roma inclusion will be discussed in Chap. 11.

Kurtic (2006).

Abdikeeva (2014), pp. 26 and 27. Bond et al. (2010), pp. 9 and 10.

Abdikeeva (2014), p. 26. Healy (2013). Roma reluctance to self-identify was previously mentioned in Sect. 5.1.3 as one of the reasons for the lack of (reliable) ethnic data on Roma in Europe, Sect. 5.4.1 on the limitations of the self-identification approach in the Roma context, and Sect. 5.5.4 on the impact of the type of interview, interviewers and questions asked on the quality of the data collected..

McDonald and Negrin (2010), p. 36. Rughiniș (2010), p. 354.

Report of the Independent Expert on Minority Issues on the Implementation of General Assembly Resolution 60/251 of 15 March 2006 entitled “Human Rights Council” (2 February 2007), para. 77. McDonald and Negrin (2010), 36. Rughiniș (2010), p. 354. Hollo (2006), p. 29.

For the study, Roma respondents were located in settlements selected with the help from local Romani organisations. It thus combines self-identification with external identification by members of the group. Babusik (2004), p. 18. The appropriateness of the different ethnical identification approaches to collect data on Roma was reviewed in Sect. 5.4.

CoE, Roma and Statistics (22–23 May 2000), para. 43.

Report of the Independent Expert on Minority Issues on the Implementation of General Assembly Resolution 60/251 of 15 March 2006 entitled “Human Rights Council” (2 February 2007), para. 77. CoE, Roma and Statistics (22–23 May 2000), para. 54. Abdikeeva (2014), p. 29. Messing (2014), p. 814. FRA (2013b), p. 19. Guy et al. (2010), pp. 5, 8 and 9. McDonald and Negrin (2010), p. 36. UNDP (2005), p. 10. Hollo (2006), p. 29. Babusik (2004), p. 18. PER (2000), p. 29.

Both Member States engaged Roma enumerators in the 2011 Census. In Romania, Roma leaders accompanied censors in some Roma communities. Active participation of Roma and awareness-raising activities led to an official increase of 16—56% in municipalities where active campaigning took place—in self-identifications among Roma in Romania. Farkas (2017), p. 43. Abdikeeva (2014), p. 29. Chopin et al. (2014), p. 75.

FRA (2013b), p. 19. Health mediators were also trained as enumerators for the 2011 Romanian Census. Farkas (2017), p. 43. Inter-cultural mediation will be considered in Chap. 11.

UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 43. Milcher and Ivanov (2004), pp. 11 and 12.

UNDP (2005), p. 10. Milcher and Ivanov (2004), p. 12.

Milcher and Ivanov (2004), p. 12.

Involvement of civil society in the design, development, implementation and evaluation of Roma inclusion policy initiatives is the ninth Common Basic Principle on Roma Inclusion. They can mobilise expertise and disseminate the necessary knowledge to develop public debate and accountability. Common Basic Principles on Roma Inclusion, as annexed to the Council Conclusions on Inclusion of the Roma (8 June 2009), principle 9. Farkas (2017), p. 43. Milcher and Ivanov (2004), p. 12.

Farkas (2017), p. 43.

As explained in Sect. 5.1.4, insufficient regard for the gender dimension risks overlooking intersectionality. See also Chap. 1 (Sect. 1.​2.​3.​2) for a brief summary of the situation of Roma women in Europe. The importance of not only addressing ethnicity, but also gender, was also stressed in Chap. 3 (Sect. 3.​1.​2).

Hollo (2006), p. 35.

Kurtic (2006).

Roma women are also best interviewed without the presence of a male family member. UNDP, Gender Aspects and Minority Data: An Illustrative Case of Roma Women in Southeast Europe (12 September 2006), para. 43.

Involvement of regional and local authorities in the design, development, implementation and evaluation of Roma inclusion policy initiatives is the eight Common Basic Principle of Roma Inclusion. Common Basic Principles on Roma Inclusion, as annexed to the Council Conclusions on Inclusion of the Roma (8 June 2009), principle 8.

For instance, the Minibus project in the Czech Republic shares work methods and a database containing prospects employers and potential clients to address long term employment of mostly Roma. Guy et al. (2010), pp. 8 and 9. FRA (2007), p. 75.

Commission Communication, Steps forward in implementing National Roma Integration Strategies (26 June 2013), p. 14.

Report of the Independent Expert on Minority Issues on the Implementation of General Assembly Resolution 60/251 of 15 March 2006 entitled “Human Rights Council” (2 February 2007), para. 91.

Id.

Chopin et al. (2014), p. 33. This was briefly addressed in Sects. 5.2.2 and 5.2.3 when reviewing the multifarious data sources on Roma. See also Chap. 4 on how ethnic data can help to uncover discrimination and identify good practices (Sect. 4.​1.​2), on research data (Sect. 4.​3.​2), and on complaints data (Sect. 4.​3.​4).

See Sect. 5.7.3 on the need for close co-operation between experts. The link between ethnic data and positive action has been highlighted repeatedly. It will also be stressed in Part II on positive action. See: Chap. 6 (Sect. 6.​3.​4) on the notion positive action, Chap. 9 (Sect. 9.​2.​2) on positive action for Roma, and Chap. 11 (Sect. 11.​5) on inter-cultural mediation in the Roma context.

The multiple reasons for the large data gaps on Roma in Europe were summarised in Sect. 5.1.3. See also Chap. 3 (Sect. 3.​1.​2) on strong calls for ethnic data collection for equality and anti-discrimination purposes.

See Sect. 5.1.2 on the absence of reliable data on Roma in Europe and Sect. 5.1.3 on the multiple reasons for the widespread lack of such data.

PER (2000), p. 30.

Hollo (2006), pp. 4 and 8.

This was previously highlighted in Chap. 1 (Sect. 1.​2.​2) when considering the position of Roma in European society.

PER (2000), p. 12.

See Chap. 6 (Sect. 6.​3.​3) on the lack of political will to implement positive action. Specifically in relation to the Roma minority, see Chap. 9 (Sect. 9.​2.​2) and Chap. 11 (Sect. 11.​5). See also the final conclusions on the need for genuine political will in Chap. 12 (Sect. 12.​2.​3).