www.ssoar.info What Do We Know About Gendered Violence and Ethnicity Across Europe From Surveys?

Empfohlene Zitierung / Suggested Citation: Condon, S., Lesné, M., & Schröttle, M. (2011). What Do We Know About Gendered Violence and Ethnicity Across Europe From Surveys? In R. K. Thiara, S. A. Condon, & M. Schröttle (Eds.), Violence against Women and Ethnicity: Commonalities and Differences across Europe (pp. 59-76). Opladen: Verlag Barbara Budrich. https://nbn-resolving.org/ urn:nbn:de:0168-ssoar-63382-3


What Do We Know About Gendered Violence and Ethnicity Across Europe From Surveys?
Stephanie Condon, Maud Lesné and Monika Schröttle In recent years, political and academic debate on VAW in Europe has been appropriated by an almost obsessive focus on immigrants and their descendants. The media have played a key role in relaying statements by politicians and reports of sexual harassment and rape, murder, forced marriage and honour crimes, and female genital mutilation. In the spotlight have been the victims: young women of immigrant parentage, born in Western Europe or having arrived there during childhood or adolescence, living very often in metropolitan housing estates. Most often, their assigned 'origins' (resulting from their parents' migration histories) are North or Sub-Saharan African, Turkish or South Asian. Meanwhile, in the near background are, on the one hand, the perpetrators of the violence -generally understood to be the men of 'their' groups -and, on the other hand, 'their culture'. The chapters in this Reader investigate the reasons for the sudden focus on these populations, and on the 'specific' forms of violence seen to be socially acceptable by their communities, as well as the consequences of such perceptions and representations. In many ways, the beliefs and assumptions underpinning discourseand much social practice and policy -have established themselves without recourse to representative data. Local and national governments have sought scientific validation for their policies by requesting 'hard ', statistical (i.e. 'reliable') data. Yet these requests are usually made with little knowledge about how prevalence is measured, how population statistics are compiled or how relationships between various factors, characteristics and acts are inferred. There are also assumptions regarding the boundaries of immigrant or ethnic minority categories.
A certain number of European prevalence studies have attempted to take on board these issues. Very often, the reports indicate that, whilst violence against women is a universal social problem that concerns women of all social backgrounds, ethnic groups and age groups, migrant women and black and ethnic minority women are at higher risk of experiencing violence within and outside partner and family relationships Müller and Schröttle 2004;CAHRV 2006). The studies conclude that this greater vulnerability is linked to women's less favourable socio-economic situations in relation to most of the majority population in the countries in which they live, racism, discrimination and social isolation. Since they are designed to analyse prevalence, using indicators describing the current situation of women, these surveys contain very little information on the migration trajectories or background of the women, or on cultural practices. Using examples of results from the major prevalence surveys, this chapter investigates the specificity or universality of different forms of violence, gives a critical viewpoint of statistical measurement of risks facing some groups of women and discusses important methodological considerations for future quantitative research on the subject.

Prevalence research on violence against women in Europe
Over the last decade or so, many countries around the world have started to collect data about violence against women (Gautier 2004;Hagemann-White 2001;Heise et al. 1999;Martinez and Schröttle et al. 2006a;Jaspard and Condon 2007). This was in response to the Beijing Platform call for the gathering of statistical data on the prevalence and forms of violence experienced by women throughout the world. It was clear that official crime statistics could only give a glimpse of the tip of the iceberg in relation to the scope and intensity of interpersonal violence, not just because they revealed only convictions or reporting to the police, but also because the categories used to describe acts corresponding to certain forms of violence recognised as 'crimes' (Hagemann-White 2001;Jaspard et al. 2003). These definitions of course vary from one national context to another. Whilst criminological surveys have produced detailed information on victimisation through different crimes on a national or international basis, they often cannot produce data on violence in close relationships, such as violence through partners or other family members. This is because the framework of crime victimisation is usually not an adequate framework to remember or to report cases of violence through very close persons as these are often not felt to be crimes or violent acts. One exception is the British Crime Survey with its specific self-completion module on domestic violence, sexual assault and stalking (Mirrlees-Black 1999;Walby and Allen 2004). Hence the move towards VAW prevalence surveys, whose aim was to measure the extent and types of violence experienced. Set-ting these acts or situations within their context contributed to developing international standards of what constituted violence and intolerable behaviour, to be condemned by state and legal institutions. Publications by the WHO, bringing together results from the DHS surveys conducted in many countries in the South (Heise et al. 1999;Ambrosetti et al. 2011), and international research networks on VAW like the EU Co-ordination Action on Human Rights Violations (CAHRV) have revealed the efficiency of certain indicators and reports from national surveys have contributed to building a substantial body of knowledge on the contexts and forms of violence against women.
Violence against women takes many forms. There is now a wide international consensus that it includes not only physical and sexual violence but also psychological. Another aspect of variation in the types of violence is the context in which it takes place. Earlier surveys in Europe focussed on married or cohabiting couple relationships (Römkens 1989;Gillioz, De Puy and Ducret 1997), or alternatively on the household. Violence in this context is often referred to as 'domestic violence,' although the use of this term is questioned by some who claim that it confuses violence against a partner and that towards children (as well as other violent family relationships). Studies from 1996 onwards extended the perspective to also include other contexts of violence against women in various life situations, e.g. in public space, in the work place and/or by other family members and acquaintances (Heiskanen and Piispa 1997;Jaspard et al. 2003;Lundgren and Westerstrand 2000;Müller and Schröttle 2004;Reingardiene 2003).
During this period, methodology on VAW research has improved, enabling disclosure of hitherto unreported forms of violence. For example, summarizing questions on experience of violence have been supplemented by lists of items referring to specific acts or behaviour -including a move to no longer using the term 'violence' -and combined measures of questioning (e.g. face-to-face interviews and additional written questionnaires on sensitive topics). Researchers sought to design instruments that would facilitate the reporting of violence. Furthermore, VAW research nowadays uses only specially trained (mostly female) interviewers and aims to ensure that the respondent is alone with the interviewer. Other factors, including those relating to ethical issues and to the safety of the interviewees, intervene in the design of current research instruments (Hagemann-White 2000;2001;CAHRV Reports 2006. Another trend within current prevalence research is to examine the different levels of severity of violence as well as the patterns and consequences of violence, not only throughout the life span, but also within one partner relationship (see CAHRV 2007; Schröttle and Ansorge 2009). A fragmented view of violence that differentiates only between forms of violence experi-enced by women cannot adequately reflect the reality of violence through the life course as well as within intimate relationships, as both are often marked by a combination and accumulation of various violent forms and actions. Therefore, all three forms of violence (psychological, physical and sexual) in combination as well as their severity levels and frequency have to be considered to assess whether specific patterns of violence and their consequences on social and economic participation, and health and so on, produce further victimisation. This makes it possible to distinguish, for example, severe forms of continued violence and abuse from less severe and/or isolated violent incidents and to quantify them.
Overall, prevalence data from several European surveys show that one in five to more than one in three women in several European countries have experienced at least one act of physical and/or sexual violence by a current and/or former partner. Also that one in 10 to more than one in 4 women were affected by several forms of psychological violence (Martinez and Schröttle et al. 2006b: 12-13, 23-24). A high level of overlapping of forms of violence by partners was found in all the European studies (ibid.: 29). Though domestic violence experienced by current or former partners is the most common form of violence reported in all the European studies, significant rates of violence by other (known/unknown) perpetrators were also reported: 3-11% of women reported physical violence through others than partners/expartners during their adult life and 5-8% sexual violence through other perpetrators than partners (ibid.: 33-35).
Although the methods of recording and measuring violence against women and the understanding of this complex and long invisible phenomenon have progressed, surveys carried out within the general population do not allow us to know whether the results are valid for minority groups. Most studies on prevalence of interpersonal violence in Europe have focused on violence against women, children and adolescents. There have been very few prevalence studies with a focus on violence against older people or against especially vulnerable groups such as disabled persons, homosexuals/bisexuals, and migrants. With respect to migrants and the minorities composed of immigrants and their descendants, definitions of groups involved in the samples of the studies vary from country to country, owing to differences relating to citizenship status, language spoken or ethnic backgrounds of the interviewees. Detailed results are not always published in the survey reports. Yet the projection of migrant and black/ethnic minority women to the front stage of debate on violence against women means that further effort is necessary to produce data to inform this debate. Do women migrants face specific risks in relation to violence? Do their daughters face similar risks? How can representative surveys be improved to enable us to examine such processes?

Prevalence of violence against migrant and ethnic minority women
Statistical studies of violence against women in migrant/ethnic minority communities in Europe are very rare, as are national level studies including data enabling us to situate the experiences of such groups. Accounts of violence revealed in anthropological studies or local monographs on migration are analyzed in relation to the societies in which these women -or their mothers -spent their childhood and youth and to often so-called 'traditional' values that remain intact after migration to more modern societies. These values, and their impact on women's lives, are seen as an obstacle to social integration in the 'host society'. Thus, differences in levels of fertility, in labour market participation and so on between migrant groups or ethnic minorities and the general population have been explained in terms of the maintenance or loss of such values -and their corollary, the stagnation or improvement in gender status. In the case of violence against women, few data sources have been available in order for such comparisons to be made.
During the CAHRV programme 1 , which brought together various European prevalence researchers who had conducted national surveys on violence against women, an exploratory comparison of violence against migrant and non-migrant women in Germany and France was set up (Condon and Schröttle 2006). Both national surveys had collected data on violence against migrant women and their descendants 2 . The originality of the German survey was that it included specific samples of what were defined as 'Turkish' women (that is, women born in Turkey or descendants of migrants) and women from countries of the former Soviet Union interviewed in their mother tongue language. Questionnaires were translated into French and Russian languages and interviews were conducted by mother tongue language interviewers who were of Turkish and Russian origin. Conversely, the French survey was based on a representative sample of women living in France and the analysis of violence experienced by migrant women or descendants of migrants could only be achieved by constructing sub-populations, using the questions on nationality at birth and country of birth. 3  since the survey was conducted by telephone, only women with a sufficient command of French could be interviewed. There are differences, of course, in the French and German immigration contexts. Post-colonial migrations represent a large part of immigration flows from the 1960s to the present day, particularly from North and Sub-Saharan Africa, whilst Germany's main immigration route was from Turkey, through the gastarbeiter programme. Nevertheless, present discourses in each country, constructed through the media and by politicians, bear many similarities and reveal the profound conviction -partly based on representations of the inferior status of women within these groups leading to a 'cultural' justification of violence against women -that these 'foreign' populations cannot integrate. The aim of this analysis was to inform public debate by providing some indication of the prevalence and forms of violence against certain groups of migrant or ethnic minority women at the same time as setting their experiences within the overall context of violence against women in France and Germany. Comparison of rates of reported violence between these groups and the majority group in each country, thus, goes some way to contextualising the phenomenon. Given the substantial differences in the questionnaires, the survey methods, the definition of the intimate partner relationship and also the period of reference used to measure the violence, the rates per se were not to be compared. Rather, what was analysed was the internal comparison within each data set. Higher rates and levels of violence were found for migrant women or women of immigrant parentage in both countries. A breakdown of the different forms of violence suggests that some forms present a specific risk to migrant women and their descendants. Fewer significant differences between the minority and majority populations were observed in relation to psychological or verbal violence but here results were in part contradictory as the German survey found more differences with respect to psychological violence than the French survey. However, the results from the German as well as from the French survey show significantly higher rates of male dominance and control reported by migrant women in couples or their descendants than within the majority population. Items relating to jealousy, dominance and restriction of outside contacts showed higher rates for these groups of women in both contexts. But though male dominance and control is reported significantly more often by these groups in both surveys, the surveys show that this problem is relevant for women of the majority group too and cannot be reduced simply to the experience of migrant women or their descendants.
Various items were proposed in the two surveys with the objective of exploring the extent of threats of violence. Very similar results on threats are to be found in each survey, as migrant women or women of immigrant parent-age reported significantly higher rates of threat with a weapon, and threats to kill, especially within the younger age groups. The higher rates of threats of violence against immigrant minority women in both countries may reflect the higher rates of manifest and severe violence perpetrated against women by their partners. Women of Turkish origin tend to stay longer in violent relationships than women of German origin, whose divorce-rate is generally higher. Not being able to escape such relationships has been shown to contribute to higher reported rates of violence generally .
Without casting doubt on the fact that these higher rates of violence are a problem that has to be addressed, it must be stressed that the majority of the North African or Turkish minority women in each sample did not report violence at the hands of their partners. Many factors intervene, generating higher or lower levels of reporting. Sunita Kishor (2005) has long warned against taking prevalence differentials at face value. These differentials doubtless include differentials in reporting, as some women are more likely to under-report, thus complicating the observation of actual differences. We still do not know, for example, whether migrant women living in surroundings of high acceptance of VAW tend to report violence in surveys more openly or whether, on the contrary, they report less violence because of feelings of guilt and shame or of simply not wanting to disclose intimate information to the interviewer. Social and cultural differences between interviewer and respondent may certainly intervene. It must be noted that there are high taboos regarding the reporting of experiences of violence within the majority populations of women, too, especially among more highly educated women. Whatever the national context, the shame attached to talking about violence to which one 'should not be exposed' may prevent women from reporting (Ambrosetti, Abu Amara and Condon 2011). What needs to be thoroughly investigated are the risk factors, meaning that some women are more at risk of being victims of violence than others, and also which forms of violence are specific to migrant and ethnic minority women?
Risk factors and specific vulnerability of ethnic minority women to gender-based violence Ambivalence of certain results makes the identification of 'risk factors' difficult Risk factor analysis with respect to VAW aims to find out which groups of women are more vulnerable than others, which men tend to perpetrate vio-lence against women and which factors might support the development of violence within couple relationships. There is growing interest in research and in practice in identifying risk and protective factors.
Many studies have collected relevant information, but in most cases limited resources have not permitted multi-dimensional analysis of factors that are likely to increase or reduce the risk of experiencing violence and the risk of staying in violent situations or relationships (Martinez and Schröttle et al. 2006aSchröttle et al. , 2006bSchröttle et al. , 2007. Thus, systematic research on risk and protective factors in relation to partner violence is still limited. A number of correlates emerge from these studies but it is impossible to show which factors are most important, how factors are interconnected and how the direction and interplay of various risk factors with victimisation can be identified (Schröttle and Khelaifat in this Reader). It must be stressed that if one factor is correlated with violence, this does not mean that the factor can be seen as an isolated variable. Rather, it is likely to be related to other, more relevant factors. Similarly, it cannot be deduced that the violence experienced by a majority of women is related to this factor; and in turn, nor can it be deduced that respondents giving a positive answer to the question exploring this factor have experienced violence (Schröttle and Ansorge 2009). For example, the German survey showed that although unemployment of one or both partners might be a risk factor for VAW, in most relationships in which (severe) violence had occurred, neither partner had experienced unemployment; nor had most unemployed women or men experienced or perpetrated violence within their relationship (ibid). Further research must to be carried out in order to find out under which circumstances and within which combination of factors certain risk factors lead to a higher likelihood of violence. Combining multidimensional quantitative research with qualitative research about the history of violence in women's/men's lives and couple relationships is certainly the most fruitful path towards understanding the role and interdependency of risk factors as well as of protective factors for VAW.

From the universal to the specific
The German survey offered the opportunity of exploring the reasons for higher levels of violence recorded in previous studies. The two groups of women studied (migrant women from Turkey, or having parents born there, or women born in the former Soviet republics) are affected twice as frequently by patterns of severe psychological abuse by their current partners than German women with no migration background. This suggests that, as far as psychological abuse is concerned, not ethnicity but the consequences of migration and the accompanying social tensions and strains on gender roles play a role in amplifying the risk of violence. Thus, more information about migration history -both for women and their partners (where these have migrated) -is necessary to go beyond a purely ethnicized perspective on the risk of being subjected to violence.
Before detailing the various factors that make the migration context particularly favourable to an increased experience of violence, we need to outline aspects of women's lives which, in any context, augment the risk of violence. Both the French and German surveys found that women who had suffered violence and different forms of hardship (including witnessing violence within the family) during childhood and youth were more likely to be subjected to violence during adulthood Schröttle and Müller 2004). According to the German study, the most powerful risk factor for severe violence against women in couple relationships was childhood experience of violence. Women who had experienced violence during childhood and/or adolescence had a two to threefold risk of suffering partner violence in later life; women who had been sexually abused before their 15th birthday had a fourfold risk of suffering sexual abuse in adult life. Three quarters of women affected by severe violence in current relationships had experienced physical, sexual and/or psychological assaults during childhood or adolescence and, as a consequence, are intensely affected psychologically and physically. Similar results were found in a local study 4 conducted by Maryse Jaspard and Maud Lesné in 2007 in one of the northern Paris suburbs: having suffered hardship or acts of violence, or being a witness to serious family conflict or violence, were principal factors in adult experiences of violence, whatever the geographical origins of respondents. This emphasises the key role of inter-generational transmission of violence. Although identified as a key factor in experiencing violence, this does not imply that all women who have lived through hardship or been subjected to violence during childhood will experience it later. Other factors and events often act in combination to increase the risk of experiencing violence. Migration or being part of a minoritized/ethnicized group produce specific circumstances and place women in a more vulnerable position with respect to violence.

Migration and minority status as factors of vulnerability
Migration is often perceived as a positive phenomenon, offering individuals the chance of social and economic betterment. At the same time, the process can generate a certain amount of upheaval, as people are uprooted from familiar surroundings and make their way to a destination, usually unknown or at best, imagined. Women migrants have increasingly begun to be seen as actors in the migration process, participating in family decision-making or planning their own autonomous emigration. Migration, nonetheless, continues to be a hazardous experience for many women, whether they are primary instigators of their departure or not. Encounters with immigration services in the destination, employers, housing gatekeepers, and so on can be difficult, especially when women have limited knowledge of or skills in the majority language of the destination country. Women who travel to join partners may have their immigration status settled in advance. Yet, with the passing of time and as their legal residency approaches it's end, they can find themselves in a vulnerable position -particularly if relations with their partner have become tense or violent. Those who lack resources in terms of education or capital may find themselves in great difficulty. The question of social isolation is of prime importance since many women may not have access to social networks, either compatriot networks or ones formed in the neighbourhood. Thus, they may not be aware of support facilities for women experiencing intimate partner or other types of interpersonal violence. Once again, knowledge of the majority language of the country is important for accessing information about their rights and the ways of finding their way around an administrative system about whose institutions and legal framework they remain ignorant.
The legal arrangements governing the settlement of immigrants in a state may also be a vulnerability factor for women by amplifying inequality in the marital relationship and limiting the autonomy of women. For example, in the context of family reunion in France, the conditions of obtaining and maintaining legal residence in essence limit the autonomy of women because the permit is inherently linked to the sustainability of their marital and legal status and, up until recently, have placed them 5 in a situation of dependency vis-à-vis their partners during the two years after their arrival in France. In Germany, the duration necessary for independent legal status has recently been expanded from two to three years and thus lengthens the time during which women remain dependent. Many forms of control may be exercised by the partner, including the confiscation of identity papers or any official papers necessary in obtaining a residence permit or its renewal, barring of access to financial resources and women being made to feel that they owe an allegiance to their husband for making their migration possible. Control may be exercised in other cases by a partner who has stayed in the woman's country of origin and uses the children for transnational blackmail.
As regards women born in the country of immigration or who migrated at an early age, their vulnerability is more specifically related to the social segregation and stigmatization endured by considerable proportions of minorities, as well as being victims of discrimination themselves. A recent German study found that a significant proportion of migrant women in Germany are affected by violence and health problems and have more difficulties in leaving abusive situations, also because they are undermined by social and racial discrimination within German society (see chapter of Schröttle and Khelaifat in this Reader). Despite increased awareness of discrimination and racism suffered by visible minorities, only very few surveys on violence against women, to date, have included these factors in the analysis of violence against women systematically (see chapter by Romito et al. in this Reader). Generally, in Europe, regional variations in intimate partner violence have not been a focus for attention. Media reports and declarations made by some women's associations tend to suggest, however, that in some areas of countries such as France, social deprivation and social exclusion combines with gender and ethnicity to increase the risk of interpersonal violence. 6 Accounts of sexist and sexual violence experienced by young and often adolescent women reveal a heightened vulnerability of these groups. The issue of strict social control -exemplified by the wearing of the Islamic veil by some women -has become incorporated into debates on gender, 'race' and religion (Gaspard and Khosrokhavar 1995;Amiraux 2003).
Discrimination and racism, by limiting access to goods and essential services such as improved housing or employment, serve to socially exclude minority women and discourage them to report violence, making them more vulnerable. Moreover, the trend towards the racialization of violence against women belonging to visible minorities establishes a distrust of women vis-àvis institutions. Women may face a cultural relativism questioning the abnormality and the unacceptability of violence they endure (Batsleer et al. 2002;Thiara in this Reader). Racism and discrimination can exacerbate the reluctance to confront the official institutions whose assistance is often perceived as inappropriate.

Critical points of the measurement of violence against ethnic minority women
The lack of data and the persistence of stereotypes hamper the development of strategies adapted to assist minority women in the context of policies to combat violence against women. The construction of appropriate tools for collecting data on violence against immigrant and descendants of immigrants is essential to carry out a thorough study and combat the stigma attached to certain groups and also interpretations that overlook or disregard the pervasiveness of violence against women throughout society.

Considerations for sample size and composition
If the necessary categorization of respondents depending on their migration background, their country of origin or that of their parents, as well as that of their partners, has begun to be standard in some surveys, building a sample adapted to detailed analysis is a more sensitive issue. Additionally, the choice of data collection method is linked to numerous constraints, such as residential instability or limited knowledge of the official language of the country in which the respondents live and reduce our capacity to reach a sufficient heterogeneity of women's situations. The quality of the measurement will also depend on the capacity to record the multiple forms and the dynamism of violence, depending upon the context in which they take place.
Owing to insufficient numbers, most surveys conducted in Europe have not been able to pursue analysis of results by taking account of these numerous risk factors. Studying minority groups by origin, and often what are relatively marginal situations, impose constraints of size and composition for using a representative sample of the population. A first alternative is to overrepresent randomly selected immigrants and the descendants of immigrants in general or for specific origins in the preparation of the sample. This method, already tested in the German survey, enables us to produce higher case numbers together with the ethnic minority women from the main survey and is an important basis for further investigation within target populations groups. The constraints of composition and size of the sample surveys, as well as protocols including questionnaire translation, generate extra cost and time in their implementation. Moreover, in some countries, the identification of descendants of immigrants for their over-representation may be complex. An important consideration is that the commissioners of surveys must allow sufficient time for the survey instrument to be developed and the protocol to be implemented.
Beyond these efforts in terms of sampling, we need to identify the factors that lead to the under-representation of immigrant women and their descendants in surveys. What obstacles contribute to this under-representation and how can we overcome them? In order to reach immigrants and the descendants of immigrants in the diversity of their situations, the method chosen should compensate for two difficulties: the residential instability that characterizes parts of recent migrants and also the problem in reaching young people, and people who are more isolated because of language or other problems. Indeed, considering the difficulties of settling in a new country requires extending the coverage beyond a household sample. As an illustration, the study on violence in the Seine Saint Denis local authority, mentioned earlier, attempted to widen the scope of reaching respondents using an innovative method. Face-to-face interviews were carried out with 1,600 young women aged 18-21 years contacted through universities, colleges, training centres, local job information centres, hostels for young workers, or encountered in the street or in shopping malls. Since household samples have tended to under-represent the younger populations, the aim was to find a way of minimizing selection bias and to reach women from a broad range of backgrounds. Such methods could be appropriate for contacting migrant or other mobile populations.
These two difficulties are compounded by the lack of knowledge of the majority language. However, one problem is that many migrant women, particularly those recently arrived or those suffering from social isolation or exclusion from the labour market, do not speak the language well enough to take part in the interviews conducted in the majority language. Not many surveys have been able to adopt the procedure followed by the German survey, which conducted interviews in different languages in order to increase the participation of migrant women. Such techniques may help to reduce obstacles such as the refusal to participate in surveys by women who are unlikely to respond at the time of the first contact with interviewers owing to diminished personal autonomy and thus a lack of confidence in the ability to take part .
Furthermore, as has been widely discussed in literature on empirical methods in social sciences, the social positioning and distance between interviewer and respondent is an issue to be considered. There is no real consensus on the question. Some studies have found that choosing interviewers from similar geographical origins or who also have a migration background encourages the disclosure of violence (Müller and Schröttle 2004). One argument is that it may be easier for the respondent to confidentially report about her own experience if she does not fear prejudice or stigmatisation by the interviewer. At the same time, interviewers have to be trained adequately to avoid over-identification with the respondents and insure neutral interviewing. Conversely, other studies have found that migrant or minority women may feel more comfortable about discussing intimate details (such as sexuality, virginity, female genital cutting) with women outside their community or not from the same geographical background (for example, the above-mentioned Paris suburbs study). Social differences between interviewers and respondents certainly intervene in the disclosure process, and it may be that similarity in age plays a more important role in encouraging reporting.

Questions aimed at gaining more relevant information about 'specific' vulnerability
The categorization of women according to their migration background requires systematic questioning about their country of birth and nationality at birth, as well as about other aspects of their migration trajectory (e.g. length of stay in the country of immigration). The identification of descendants of immigrants means that information on their parents' background and their own experiences in relation to this background is necessary. For a more detailed analysis that takes into account the heterogeneity of the situations of immigrant women, it is necessary to gather information on the conditions of arrival (age, marital status, residence, knowledge of the majority language) but also on their changing residency status since migration, their family circumstances and their paid work experiences. Also, despite the sensitivity of this issue, it is essential to distinguish between undocumented women and women awaiting regularization. These women living clandestinely represent a population at risk and face particular difficulties in undertaking protective action against violence.
As well as categorizing women, information must also be gathered on their partners, both present and past. Enriching the available data on former partners or marital history implies the sometimes unsettling or painful memo-rizing of past relationships. Indeed, collecting data for different partners increases all the time spent on completing the questionnaire and therefore the cost of the investigation. However, limiting the questions to the current partner relationships neglects essential contextual information on violence experienced in the past particularly as a common trend found in European surveys is for higher rates of (more severe) violence experienced at the hands of a former partner. Furthermore, such information, along with questions relating to consent to marriage, would be vital in the study of arranged or forced marriage (Hamel 2008) and how some women escape from such marriages and form new partnerships (Collet 2008). Thus marital histories must be linked up to the migration histories of both partners. Similarly, questions on aspects of family relations must be posed in order to grasp how strict social control might put women, both migrants and their descendants, at greater risk of violence.
Surveys on violence against migrant women should always include questions about discrimination through institutions and society and factors that reduce their ability to leave violent situations. They, furthermore, have to include questions on violence and social control within their families by other family members and not just partners, in order to better describe the context in which violence has been experienced and which might detain women from leaving violent situations.
Incorporating questions on the above themes into survey instruments will go some way to identifying what specific forms of violence are experienced by migrant women and their descendants, inside or outside the family home or in other interpersonal relationships. It will then be possible to locate more easily those forms of violence which are considered specific to certain national or regional groups. Immigrant women and their descendants are not a homogeneous group, neither in social terms nor in their representations of gender relations. Thus, the challenge is to track down particularizing assumptions -conscious or unconscious -while allowing the description of all types of violence in their various forms and the different risk factors. This brings up issues in relation to specific factors for different groups (questions on residency status for foreign nationals, on racial discrimination for visible minorities) and must be reconciled with a need to account for macro, contextual factors that increase risk of experiencing violence. Thus, the development of the questionnaire becomes a real challenge.

Conclusion
Despite progress in identifying the types and degrees of victimization with the development of sophisticated methods constantly refined over successive surveys conducted in Europe, research is still limited when it comes to counting and analyzing violence against women from minority groups. This lack of data is a major obstacle. It renders analysis and communication on these issues particularly sensitive. Cases reported in the media admittedly give subjective viewpoints but are no less real. Despite the existence of many localized surveys in Europe showing the universality of domestic and intimate partner violence, the risk remains that the results of such studies can be instrumentalized to reinforce racist and sexist essentialist perspectives.
Methodological precautions are essential to limit and counter the biased interpretations of the results that would only convey a stereotypical view and increase stigmatization of immigrant groups and descendants of immigrants, regarding the levels of violence they experience. Deciphering the frequency and forms of violence suffered by ethnic minority women without participating in the process of othering is a major challenge. We need to question assumptions at the same time as rendering visible the specific vulnerabilities placing migrant and ethnic minority women at risk of violence. Analysis of their social, family and economic circumstances, their education, work and migration histories must be set in context using other micro variables, such as indicators of gender relations in the household, health status, sexuality, experience of discrimination and racism, as well as using macro indicators describing the political and social context in relation to the research object (majority attitudes regarding gender roles, gender sensitive policies, level of awareness of laws protecting women against violence, anti-discrimination laws, immigration policy). The broadening of the context in which data is analysed is important to shift the focus away from women's 'cultural characteristics', enabling a more balanced perspective that brings in the characteristics of the society in which they live their lives. We need to produce the means by which to study violence against women in its various forms and to participate in a global reflection towards its eradication. This is an essential part of progressing towards gender equality.