Violence Against Women: Placing Evidence From a European Union–Wide Survey in a Policy Context

The Limits of Data Collection

The FRA survey on violence against women responds to calls over several years concerning the global and regional need for comparable data collection on violence against women—ranging from the United Nations and the Council of Europe, through to civil society actors such as, at the European level, the European Women’s Lobby (EWL) and Women Against Violence Europe (WAVE), as well as leading feminist researchers. The need for a multicountry survey emerged from the continued absence of comparable data on violence against women from government sources at the level of individual EU Member States, which has presented a challenge for the development of evidence-led policy and action on violence against women. As reports by the European Commission (2010), the Council of Europe (2006), and the European Institute for Gender Equality (EIGE; 2014) (an EU Agency), among others, have variously shown, at the level of police and criminal justice data, the extraction of information concerning particular offenses is not facilitated by data collection systems that are organized around different articles of law rather than specific acts, and where disaggregated data on offender/victim characteristics (including basic data on sex) are often unavailable. Given the problem of relying solely on official police and criminal justice data to estimate the extent of violence against women—due to underreporting and high attrition rates in cases of sexual assault and rape (Lovett & Kelly, 2009)—the extent and nature of violence against women in the European region has been left to estimates by international organizations, such as the Council of Europe (2006) and the World Health Organization (2013), which are based on different data sources at the country and regional level.

A number of EU Member States have made significant strides since the 1980s (Hagemann-White, 2001) in trying to document more accurately the scale and nature of violence against women through the use of criminal victimization surveys; leading examples include, among others, the Crime Survey for England and Wales (Office for National Statistics, 2015), the Finnish national survey on violence against women (Piispa, Heiskanen, Kääriäinen, & Sirén, 2006), and the Spanish national survey on violence against women (Ministry of Health, Social Services and Equality, 2015). However, it remains the case—a decade after European Union–funded research mapping the extent of violence against women surveys (Martinez & Schröttle, 2006)—that, apart from FRA’s survey, a number of Member States still do not have dedicated surveys on violence against women (for a global perspective—see United Nations, 2015). Where survey data are collected, it is often part of a broader survey on crime victimization in general—typically including both women and men in the sample—which often results in an undercount of violence against women due to a “gender-neutral” approach to data collection (as critiqued for two decades by researchers; Johnson, 1998; Smith, 1994).

At the level of the European Union, there have been efforts in recent years to improve and harmonize official data collection on crime—with Eurostat, as the statistical office of the European Union, having been given the task to work in this field alongside its more established areas for data collection such as agriculture and employment. Yet, the situation remains that we know more about the number of livestock in the European Union than we do about the scale of violence against women. The focus of Eurostat—reflecting the narrow mandate of the European Union in the area of crime, which remains largely a matter of national sovereignty (as with national security)—has been on data collection with respect to serious and organized crime that is typically cross-border in nature, and where the European Union has a clear legal mandate to act. While this has resulted in initiatives to improve European Union–wide data collection on crimes such as trafficking in women and girls for sexual exploitation, it is only recently that Eurostat’s work has been extended to consider other areas of violence against women where significant improvements in data collection are also needed—for example, with respect to domestic violence, sexual assault, and rape. This situation is reflected by the fact that there is no generic EU legislation addressing violence against women, whereas there is specific EU legislation relating to trafficking. ⁵

In recent years, Eurostat worked closely with EU Member States to gain support for a European Union–wide crime survey interviewing women and men, which was being developed in parallel to the adoption of European Union–wide legislation for victims of crime in the form of the Victims’ Rights Directive. However, this initiative to enhance European Union–wide data collection on criminal victimization to encompass common crimes such as burglary and assault, which are not typically cross-border in nature, met with some resistance at the level of the European Parliament. The European Union’s “Crime and Safety Survey” would have followed in the traditions of the European Crime Survey and the International Crime Victims Survey in providing EU Member States with alternative data on crime to be read alongside official statistics. As it stands, plans for this EU crime survey have been dropped for an indefinite period. Whereas the results of the FRA’s survey on violence against women were intended to provide detailed data on women’s experiences of violence alongside data provided by the Crime and Safety Survey, the FRA’s survey results now stand alone, in the absence of a broader crime survey, as the most comprehensive crime victimization survey—focusing on women’s experiences of violence—across the European Union’s 28 Member States.

In contrast to the current halt that has been placed on the European Union’s Crime and Safety Survey, the EU institutions successfully agreed on a package of legislative proposals for victims of crime, which sees the 2012 Victims’ Rights Directive at its heart. This has resulted in a situation whereby legislation is being enacted across the European Union to try and meet new standards for victims of crime under the Victims’ Rights Directive, whereas—in the absence of a European Union–wide Crime and Safety Survey—there exists no parallel initiative to gauge the extent and nature of crime as people experience it; the vast majority of which—as we know from existing national crime surveys—is not reported to the police and therefore does not make its way into official crime statistics. Article 28 in the Victims’ Rights Directive does refer to data collection, but simply states the following: “Member States shall, by 16 November 2017 and every three years thereafter, communicate to the Commission available data showing how victims have accessed the rights set out in this Directive.” As is the norm for EU Directives, the terms of data collection are not prescriptive and are left open to individual Member States, and, in turn, encompass all crime victims alongside the specific experiences of women as victims of crime. In contrast, the Council of Europe’s “Istanbul Convention” includes Article 11 that specifically refers to the collection of disaggregated statistical data and the need for regular population-based surveys on violence against women.

More recently, in 2016, Eurostat encouragingly indicated that it will begin to explore the possibility with selected EU Member States—on a voluntary basis—of undertaking a pilot survey on gender-based violence; however, one which would interview both women and men. FRA’s violence against women survey has served as the catalyst for the development of such a pilot initiative, which will be discussed further in the course of 2017. In parallel, the Organization for Security and Cooperation in Europe (OSCE), which is a regional security organization that encompasses work on gender-based violence, has received funding from the European Union to replicate the FRA’s violence against women survey in 10 countries that are external to the European Union but within the 57 participating States of the OSCE. Yet, alongside these developments, it is clear that comprehensive and regular data collection on violence against women—whether in the form of surveys or administrative data collection—is still someway off, and that within “Europe” (encompassing the European Union, the Council of Europe, and parts of the OSCE) there is a mismatch between law, policy, and the existence of data to inform law and policy.

In recognition of repeated calls for improved data collection, in a 2009 Resolution, the European Parliament (the directly elected parliament of the European Union) called on FRA to collect data on violence against women, which was reiterated by the Council of the European Union (made up of representatives of EU Member States’ governments). ⁶ Given FRA’s mandate to collect data and undertake survey research in key fundamental rights areas—encompassing victims of crime, access to justice, and equality on all grounds (including sex)—the Agency decided to allocate resources to develop a European Union–wide survey on violence against women. Building on the Agency’s existing work with respect to European Union–wide survey research, which includes research on criminal victimization, FRA was well placed to undertake the survey, and to bring its results to the attention of EU institutions and Member States in line with its core mandate.

A Targeted or Piecemeal Response to Violence Against Women

At the regional level of the European Union, the legal and policy response to violence against women, including the need for enhanced data collection, has been critiqued from various quarters—including by European Union–wide nongovernmental organizations such as EWL and WAVE—for the absence of a sustained and comprehensive approach (Nogaj, 2013). ⁷ For example, whereas the European Commission “Action Plan 2006-2010 for Developing an EU Strategy to Measure Crime and Criminal Justice” made “measuring violence against women” and “measuring domestic violence” objectives to support the development of a common EU framework for indicators and data collection on crime, the 2011-2015 Action Plan did not include data collection in this area—although the specific field of trafficking in human beings was addressed (which includes violence against women). Likewise, at the time of the European Commission’s midterm review of its 2010-2015 Strategy for equality between women and men, the proposed action for a European Union–wide strategy on combating violence against women was not taken up. In comparison, the focus has been on targeted action for specific groups of women and specific types of violence against women, focusing on trafficking in women and girls for sexual exploitation, forced marriage, and FGM—where, arguably, the European Union can have most impact through a focused response.

The appointment in 2010 of an EU anti-trafficking coordinator and the endorsement of an anti-FGM campaign at the highest levels of the Commission, at the end of 2013, have served to ensure that the spotlight on violence against women is focused on specific issues and on groups that are considered as being particularly “vulnerable.” This approach gels well with the European Union’s external actions to address violence against women, where particular forms of abuse—such as FGM—are addressed. However, as some critics—including EWL and WAVE—have claimed, what this focus sidesteps is the extensive violence that is experienced by women in the general population, which is clear from the results of the FRA survey on violence against women (alongside previous studies). As some have argued with respect to the need for integrated Strategies to address violence against women at the level of individual Member States (Kelly & Lovett, 2005), the “silo” approach to addressing different forms of violence against women neglects the wider causes of gender-based violence that are best addressed through a holistic response, and also risk “othering” violence as symptomatic of particular cultures (Montoya & Rolandsen, 2013). Hence, there are continued calls from some quarters for a holistic response to violence against women in the form of a dedicated Strategy, which—it has been suggested—would also serve to underline the Commission’s commitment to this area in alignment with gender equality more broadly (Lombardo & Meier, 2007; Mazey, 2002). Likewise, as referred to in the Introduction to this article, if the EU Member States agree to the European Union’s accession to the Istanbul Convention, this will have a significant influence on the future existence and content of any specific Strategy.

The legal mandate argument that has, to date, somewhat delimited the Commission’s actions in the field of violence against women is reflected in documents such as the Commission’s midterm review of its 2010-2015 Strategy for equality between women and men, where the Commission noted that while it remains committed to a strong policy response to combat all forms of violence against women, “It will keep focusing on concrete actions in areas where there is a clear legal basis for action under the Lisbon Treaty which can bring tangible results.” ⁸ Yet, recognizing that equality between women and men is one of the central pillars of Union policy, it is evident that violence against women—particularly in the form of gender-based violence that disproportionately impacts on women (such as sexual assault, rape, and sexual harassment)—is a key hindrance to women’s enjoyment of full equality. In parallel, growing recognition of and reference to fundamental rights in EU law could see violence against women positioned as a mainstream area for EU action as part of the drive toward enhanced equality as a fundamental right. Herein, the Council of Europe’s Istanbul Convention defines “violence against women” in terms that encompass equality and fundamental rights; namely (Article 13):

(a) “violence against women” is understood as a violation of human rights and a form of discrimination against women and shall mean all acts of gender-based violence that result in, or are likely to result in, physical, sexual, psychological or economic harm or suffering to women, including threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or private life.

Whereas the Council of Europe’s Istanbul Convention presents the most wide-ranging regional legal instrument to address violence against women, to date, the EU policy focus has rested with types of violence against women that disproportionately affect certain groups in the population and that are cross-border in nature. Within this framework, recent legislation has been enacted in the European Union—namely, the Regulation on mutual recognition of protection measures in civil matters ⁹ and the accompanying Directive on the European Protection Order ¹⁰ under the criminal law—that can address cross-border cases of domestic violence; the uptake of which, in practice, has to be evaluated. Alongside the above, the European Union does have legislation that addresses forms of violence against women that is not specific to certain groups or cross-border in nature—namely, the Gender Equality Directive (recast) which encompasses sexual harassment, and which is squarely within the European Union’s established legal mandate to address gender equality in the framework of employment. Yet, the relative importance of this legislation as it applies to sexual harassment and women more generally in the population—noting here the scale of sexual harassment reported by the FRA survey as impacting on average every second women in the European Union (see article by Latcheva in this special issue)—has, arguably, not been matched in recent years by a level of policy response that is comparable with responses to trafficking and FGM. Herein, the Union’s response to violence against women reflects its response to crime more generally—from the 2010 Stockholm Programme onward ¹¹ —that is, as focusing on cross-border and/or organized crime, and terrorism, where the Union is more certain of its legal mandate.

The question of the European Union’s legal mandate or “competence” to address violence against women more generally remains contested, as Montoya (2013) notes:

The characterization of the European Union as a positive contributor to domestic efforts to combat violence against women, an issue where it has yet to establish legal competency, is one for which there is mixed support in the highly critical literature on gender equality in the EU. (p. 4)

While there has been criticism of the European Union from some quarters for its “low threshold” responses to violence against women, it has to be acknowledged that significant resources have been earmarked over several years to fund initiatives at Member State level through grassroots capacity building that is aimed at addressing violence against women; for example, under the “Daphne” program (Montoya, 2008), which has as its specific objective “to contribute to the prevention of, and the fight against all forms of violence occurring in the public or the private domain, including sexual exploitation and trafficking of human beings.” ¹²

More recently, however, recognition of women as victims of crime has been enhanced in the European Union with the adoption of the 2012 Victims’ Rights Directive ¹³ —which is generic legislation for all crime victims, across the EU Member States, replacing the poorly implemented 2001 Framework Decision on the Standing of Victims in Criminal Proceedings. Importantly, while addressing all crime victims, the Directive refers at different points to victims of gender-based violence, victims of sexual violence, and victims of violence in a close relationship, alongside other vulnerable groups such as victims of hate crime. For example, Article 9 in the Directive, which covers “Support From Victim Support Services,” refers to “targeted and integrated support for victims with specific needs, such as victims of sexual violence, victims of gender-based violence and victims of violence in close relationships”; and Article 22, which addresses the need for an individual assessment of victims to identify specific protection needs, refers to “gender-based violence,” “violence in a close relationship,” and “sexual violence.” In this regard, the Commission does recognize the particular significance of particular forms of violence against women—but within a wide-ranging piece of legislation for all victims. While different standpoints exist with respect to the effectiveness of either gender-specific, gender-neutral, or generic legislation encompassing violence against women (EIGE, 2016; Weldon, 2002), the fact that EU law—in the form of the Victims’ Rights Directive—is referring explicitly to “gender-based violence” is an indication of positive change when compared with the Directive’s predecessor, the 2001 Framework Decision on the Standing of Victims in Criminal Proceedings, which made no such reference.

Having noted the above, there were encouraging signs at the end of 2014 that the new European Commission would place renewed emphasis on combating violence against women; as the Commissioner for Justice, Consumers and Gender Equality noted (in her speech to the European Parliament on the international day to eliminate violence against women), a new Commission Strategy for equality between women and men—including a chapter on gender-based violence—would be proposed and elaborated on in 2015. ¹⁴ In December 2015, the Commission published a “staff working document”—rather than a comprehensive Strategy—on strategic engagement for gender equality 2016-2019, which refers to “combating gender-based violence and protecting and supporting victims” as one of five priority areas for the period 2016-2019; a move that was critiqued by some commentators as having “watered down” the Commission’s response to gender equality, including violence against women, in the absence of a specific Strategy. ¹⁵ However, concrete action is being taken by the Commission to address violence against women, including the establishment of “action grants” to combat sexual harassment and sexual violence against women and girls, ¹⁶ and through recent legislative developments addressing violence against women—as described above—which the Commission has to monitor with respect to implementation at the level of individual Member States. Moreover, the Commission’s proposal for a Council Decision concerning the European Union’s accession to the Istanbul Convention—which outlines the legal competency framework—sends out a clear message that progress toward a more comprehensive response to address violence against women is supported. Notably, in 2016, on the eve of the International Day for the Elimination of Violence Against Women, the European Commission also announced that 2017 would be a year of focused action to fight violence against women—which again serves to underline the broadening of its response in this area. And, at the same time, the European Institute for Gender Equality (EIGE), which is a sister Agency to FRA that works with existing European datasets in the field of equality, has placed the review and analysis of EU Member State data collection on violence against women high on its agenda for the period 2015-2018.

In sum, the traditional policy framework for the European Union’s responses to violence against women has been both targeted and necessarily delimited at the same time, reflecting the European Union’s interpretation of its legal mandate, which has followed a focus on specific types of violence and on specific groups that are vulnerable to certain types of violence. The European Union has contributed in different ways to address violence against women, namely through legislation addressing various manifestations of violence against women (from trafficking through to protection orders), which, subsequently, has been adopted by Member States, and also through targeted funding to address violence against women and girls. Data from FRA’s survey on violence against women, coupled with the Council of Europe’s comprehensive Istanbul Convention, have served to move European Union—and more generally European—policy responses further forward in consideration of a more comprehensive response to violence against women.

Delimiting and Contextualizing the Scope of the Research

Reflecting repeated calls over several years for enhanced data collection on violence against women, the FRA undertook to develop a comprehensive survey instrument to collect data on women’s “lifetime” and recent experiences of violence, which would include new or newly recognized forms of violence—such as stalking and the use of information and communication technologies as new tools for inflicting abuse. To this end, examples of the most advanced international and national surveys in the field were taken as a starting point for the FRA survey, and research and survey experts on violence against women were brought on board during the development of the survey. ¹⁷

The central objectives of the FRA survey were identified as follows: to provide the first European Union–wide dataset on the extent, nature, and consequences of violence against women, as reported by women, which can be used to inform policy and action, and to highlight the manifestation of gender-based violence against women as a fundamental rights abuse in the European Union. In setting out to fulfill these objectives, the survey asked wide-ranging questions about women’s experiences of physical, sexual, and psychological violence, and employed international definitions of violence against women that have been developed under a human rights framework—including the 1993 United Nations Declaration on the Elimination of Violence Against Women, which was further developed by the 1995 Beijing Declaration and Platform for Action and, more recently at the regional level, the 2011 Council of Europe Istanbul Convention. ¹⁸

Having adopted a broad-based definition of “violence,” the resulting survey questionnaire was developed to encompass the experiences of violence among women in the general population in the European Union’s 28 Member States. Herein, it was made clear at the outset of the project that the survey set out to capture the experiences of women in the “general population,” and would not—or rather could not—capture the experiences of women in the European Union who were victims of specific forms of violence, such as trafficking in women for sexual exploitation, FGM, or forced marriage. Given that the survey was based on a random sample of women in the general population, it was not in a position to capture these “rarer” incidents of violence that specific groups in the population are prone to (such as FGM); for this, targeted studies on subgroups in the population are needed, including research of a qualitative nature.

While the survey did ask for respondent characteristics—such as citizenship status, sexuality, health and/or disability status, and age—the sample size at the country level (averaging 1,500 interviews per Member State) meant that the findings, when broken down across these grounds, were often too limited for a meaningful analysis with respect to specific respondent characteristics as they relate to different experiences of violence. Reflecting this, the Agency published key results for selected respondent groups at the EU level (FRA, 2014a, pp. 184-190), which showed that non-heterosexual women respondents and women with some form of disability or limitation in their everyday activities due to health problems were more prone to experiencing different forms of violence than women who were heterosexual or did not indicate a disability or health problem.

In addition, the survey’s sample size could not hope to capture detailed information concerning rarer incidents of violence that women may experience, particularly as the survey asked not only about women’s lifetime experiences of violence but also about their experiences in the 12 months prior to the survey interview. However, this challenge is typical to all (crime) surveys with respect to “rarer” events. Only those countries that are fortunate to have highly developed crime survey research—such as the United Kingdom (England and Wales), the United States, and Italy—are able to benefit from large enough sample sizes that can provide detailed breakdowns with respect to rarer events, and which can also look into the experiences of groups within the population to capture diversity, for example, with respect to ethnicity, migration status, and health or disability. Here, it is worth noting that around one third of the European Union’s 28 Member States have undertaken survey research on violence against women—which is often a one-off survey—using sample sizes that are equivalent to or smaller than the FRA’s survey at the national level, which necessarily limits analysis with respect to diversity and rarer events.

While the survey’s objectives were to document the extent, nature, and consequences of violence for women—through the application of a structured questionnaire format—it did not set out to provide definitive “answers” to the research findings, nor could it claim to do so. Given that the survey results presented some stark differences between the 28 countries surveyed, the FRA research posited some possible general explanatory factors for these differences (FRA, 2014a, pp. 22-26), which are being explored through further research (beginning with some of the other articles in this special issue). What the survey did achieve, as stated as one of its two main objectives, was data at the EU level—for the first time. It also produced new data at the individual Member State level, which can be examined by those working at the country level who are best placed to understand the cultural context in which violence against women takes place and is reported in a survey.

Building a Robust Dataset to Inform Policy

In embarking on the largest international survey to record women’s experiences of violence, the FRA undertook a detailed expert consultation process—over a period of 2 years—to develop the survey instrument. The Agency’s own expert staff were responsible for all stages of the survey’s design and implementation, which included, as an illustration—the development of the survey questionnaire; taking part in train-the-trainers sessions for fieldwork, including training sessions for interviewers in selected Member States; and selected on-site inspections of fieldwork progress in the Member States. The survey was pretested in 2011 and drew on national-level experience of surveys on gender-based violence, where these existed. While the technical details of the survey’s methodology, sampling, and fieldwork are available in full in a published report (FRA, 2014b), the following paragraphs serve to illustrate some of the approaches adopted and the challenges faced when undertaking a cross-national survey on this scale.

A prerequisite of the survey was to ensure that the results could be as representative as possible of the general population—of women—in each EU Member State. This was essential to ensure that the results could be used with confidence by policy makers. With the survey’s development having established at an early stage that all interviews should be conducted face-to-face in an effort to obtain the best results—which is not the methodological approach adopted with respect to all national-level surveys on violence against women—this served to frame the survey’s sampling approach. To this end, the sampling in each Member State was based on a stratified design, involving two or more stages, to ensure geographical spread, with equal probability of selection for households within clusters, and with women being selected at random from households and asked to take part in the survey. A clustered sampling design was adopted in an effort to enhance the efficiency of the data collection in the field, which meant that the survey population was broken down into small geographical areas (clusters), with clusters selected at random. While this approach is more efficient for survey research purposes where resources are finite, and is commonly adopted in survey research, the approach can be criticized for reducing the statistical reliability of samples, which, however, can be addressed with reference to reported confidence intervals—as referred to in the survey’s Main Results Report (FRA, 2014a). To further enhance quality in the sampling approach, additional measures were adopted; namely, only one woman per randomly selected household could be interviewed, and no substitutions of selected interviewees were allowed, and, where contact could not be made with a randomly selected household at the first attempt, at least three visits were required per household to try and secure an interview.

Each EU Member State has a different tradition when it comes to the generation of sampling frames from varied administrative sources, and—importantly—has a different approach and legislative framework when it comes to making information available for sampling addresses and/or individuals for the purpose of survey research. As a reflection of this reality, the survey aligned its approach for first contacting potential interviewees with established practice in different countries. This meant that in Denmark, Finland, and Sweden, randomly preselected individuals were first contacted by telephone and asked if they would like to take part in FRA’s survey with a face-to-face interview, while in Malta, Slovenia, and the United Kingdom, an advance letter was sent before a visit was made to the selected household to try and secure an interview with a randomly selected woman from within the household, and in all other Member States first contact was made in the form of a household visit based on “random route” selection criteria. In all cases, no direct reference was made at the point of first contact—regardless of the approach adopted—to the survey being about violence against women. This was done to ensure that any potential perpetrator of violence within the household did not see a letter or overhear a conversation that could have put women at risk. Instead, at the point of first contact, it was stated that the survey was about women’s well-being and safety, which was followed up with information concerning the content of the survey once a woman who was selected for interview agreed to participate.

While it could be posited that differences in the mode of first contact may result in differences in respondent selection and hence results, it is apparent that this assumption does not stand under closer scrutiny when FRA’s survey results are looked at alongside other survey research from different countries, and when the findings are explored in more detail (as outlined in other articles in this special issue). First, while countries such as Denmark, Finland, and Sweden did report some of the highest rates of violence in the FRA survey, these high rates are replicated in national surveys in these countries where very similar questions to those asked in the FRA survey, using a modified conflict tactics scale, have been used and—importantly—where other methods of first contact and questionnaire application have been applied (Andersson, Heimer, & Lucas, 2015; Balvig & Kyvsgaard, 2006; Heiskanen & Piispa, 2006; Nerøien & Schei, 2008). Second, in the FRA survey, patterns in either high or low prevalence rates do not always hold across different questions both within and between countries, and therefore the suggestion that the method of first contact consistently skews results in a certain direction can be disputed. For example, as the article by Nevala in this special issue illustrates, levels of coercive control, which can be measured on the basis of different questions asked in the FRA survey, show higher rates in countries other than Denmark, Finland, and Sweden.

Notably, when looking across established European Union–level survey research conducted for the purpose of the European statistical system (a partnership between Eurostat at EU level and national statistical institutions at Member State level), it can be observed that very different approaches are adopted with respect to how surveys are conducted at national level—including the mode of first contact—while the results of these surveys are commonly merged together and considered comparable. For example, the European Health Interview Survey—which asks respondents highly sensitive questions, as does FRA’s survey—has, in successive waves of the survey, used different sampling frames according to the country being surveyed (such as census data, population registers, and telephone lists) as well as different methodologies for questionnaire application (such as face-to-face interviews, telephone interviews, self-administered questionnaires, or a combination of these). The same differences can be observed across many EU surveys at the national level. While the FRA survey’s sampling frames were adapted according to the information available at national level (such as census data or population register), a high level of standardization was applied with respect to the method of first contact (compared with some other European Union–wide surveys), and—significantly—all interviews were face-to-face. Therefore, when viewed alongside a number of other European Union–wide surveys that are used as a source for European Union–wide data and indicators, the FRA survey took a number of additional measures to ensure that it provides a robust evidence-base for policy purposes. The practical realities of conducting survey research across several countries shows that while, in theory, adopting “one model” for survey purposes is ideal, it is currently not feasible—as the experience of those who have undertaken multicountry survey research on violence against women has shown, ranging from the WHO’s 10 country survey (Garcia-Moreno et al., 2005), through to the 11 country International Violence Against Women Survey (Johnson, Ollus, & Nevala, 2008).

Besides the FRA survey, the current diversity in approaches in the European Union to surveying violence against women at the individual country level—that has traditionally ranged from in-person and computer-assisted self-interviewing (CASI) in England and Wales through to computer-assisted telephone-interviewing (CATI) in Italy—suggests that closer attention at the policy level needs to be paid to comparison of results across surveys, within and between countries, that apply similar and different approaches for sampling and survey mode, but which, importantly, ask “matched” questions. When it comes to the development of evidence-based policy responses to address violence against women, comparing findings between surveys that ask very different questions will result in the comparison of “apples” and “oranges.” This, in itself, is not a problem providing that the differences in the questions asked are openly acknowledged and addressed, whereas, there is added-value in being able to compare and contextualize results across 28 countries based on FRA’s survey that uses the same questionnaire. At the same time, where methodologically diverse surveys show similar results—but where the content of their questionnaires is very similar—this might suggest that the mode of survey application is less significant than is generally thought to be the case, which warrants further policy consideration with respect to the use and comparison of different survey datasets, and the development of future surveys.

In sum, the survey randomly sampled women across the European Union’s 28 Member States to take part in face-to-face interviews about their experiences of violence during their lifetime (since the age of 15), in the last 12 months, and also during childhood. This resulted in 42,000 women being interviewed about their experiences of physical, sexual, and psychological violence—including the consequences of violence, reporting patterns, and their satisfaction with services received.

Capturing the Extent of Violence Against Women

The survey findings present a striking difference between what women in the general population say they have experienced as violence and what official criminal justice data are able to tell us about the extent of violence against women. As an illustration—based on responses to the survey question “How often has someone forced you into sexual intercourse by holding you down or hurting you in some way?” which was asked separately to a question on attempted rape—one in 20 women in the European Union has experienced forced sexual intercourse since the age of 15. In comparison, a review of those EU Member States for which data were available for the period 2005-2007 shows that, on average, the police recorded 11 reported rapes per 100,000 population per year (Aebi et al., 2010).

What is important to note is that the survey built up a picture of women’s experiences of violence by progressively asking them about different types of violence they may have experienced, and by describing specific acts and asking whether they had experienced each of them in turn, a version of the “conflict tactics scale.” For example, with respect to physical violence, women were asked in turn for each of the following: ‘Since you were 15 years old until now/in the past 12 months, how often has someone: Pushed or shoved you? Slapped you? Thrown a hard object at you? Grabbed you or pulled your hair? Beaten you with a fist or a hard object, or kicked you? Burned you? Tried to suffocate or strangle you? Cut or stabbed you, or shot at you? Beaten your head against something?’. Women were asked whether each of these incidents had occurred in the past 12 months, and how often. The same approach was taken with respect to incidents of sexual violence, namely, for each of the following, women were asked, ‘Since you were 15 years old until now/in the past 12 months, how often has someone: Forced you into sexual intercourse by holding you down or hurting you in some way [where needed, this was clarified as meaning forced oral sex, forced anal or vaginal penetration]? Apart from this, attempted to force you into sexual intercourse by holding you down or hurting you in some way? Apart from this, made you take part in any form of sexual activity when you did not want to or were unable to refuse? Or have you consented to sexual activity because you were afraid of what might happen if you refuse?’. The same step-by-step approach was adopted too for women’s experiences of psychological violence, stalking, and sexual harassment. For each type of violence, women were asked separately about perpetrators—namely, current partner, former partner, and other persons. The sex of perpetrators was also consistently verified, which means that the results were able to reveal the extent to which male or female perpetrators were involved. In addition, women were asked about their experiences of violence in childhood (physical, sexual, and psychological) by an adult (see article by Till-Tentschert in this special issue).

In sum, the data present a detailed picture of the nature of women’s experiences of violence, which is in line with the comprehensive coverage of the Istanbul Convention. The results of the survey show that women—in significant numbers—were able to recall their experiences of violence during the survey interviews, which was aided by the data collection approach described above.

As a result, the survey’s “headline” figures on the extent of women’s experiences of violence—when specific acts are grouped together—reveal the following:

The above paints a stark picture of women’s experiences of violence, which is compounded by the detail that is contained within these generic figures; for example, of those women who indicated that they had been forced into sexual violence by a previous partner, 32% said that this had happened once, 28% that it had happened 2 to 5 times, and 40% said it had happened 6 or more times. Although women could indicate more than one previous partner, meaning that incidents of violence could have either occurred within the same previous relationship or in different previous relationships, the fact remains that for women who do experience violence in relationships, the majority are exposed to multiple incidents of abuse. Evidence from the FRA survey clearly points to the reality that a number of women are exposed to repeat victimization—which is a particular characteristic of intimate partner violence, given that the victim and perpetrator are typically living together and there is ample opportunity for abuse. At the same time, the survey data also show that 43% of women have experienced some form of psychologically abusive and/or controlling behavior when in a relationship—which can range from belittling or humiliating behavior in private or public, through to keeping the respondent from seeing her family or friends. Of women in the survey who indicated they were currently in a relationship, 7% have experienced four or more different forms of psychological abuse by their current partner. Notably, most women who experience four or more forms of psychological violence by a current partner also indicate that their current partner has been physically and/or sexually violent against them, a finding which suggests that those who are developing policy responses to combat violence against women would be advised to look into programs that probe further about any physical and/or sexual abuse among those women who report psychological abuse. The survey data also underline the vulnerability of women who leave a violent relationship, as one in six women who had been abused by their previous partner indicated in the survey interview that they had experienced violence after the relationship ended, which, again, is evidence that policy makers can draw on for the purpose of targeted interventions—for example, by adopting programs that have successfully intervened with men, who are known to be violent, once a relationship has ended.

When looked at together, and when the individual types of violence asked about in the survey are explored in more detail (as in other articles in this special issue), the data are able to paint a clearer picture of the extent and nature of violence that women experience across the European Union, which serves to underline the vulnerability of certain women to abuse, most of which goes unreported. As hinted at above, the data can be used at various levels to inform policy, particularly in those EU Member States where there is an absence of comprehensive and/or recent survey research on violence against women. Given the importance of understanding underreporting of violence within a diverse policy context, across 28 EU Member States, in which the true extent of women’s experiences of violence is not always understood or acknowledged, the following paragraphs focus on the survey results with respect to women’s reporting and nonreporting of incidents of violence, and then briefly address how these results can be useful for policy development.

Reporting Violence

A key survey finding—which is replicated throughout the survey data irrespective of the type of violence asked about—is that the majority of women do not report their experiences of violence anywhere, including to the authorities; a finding which is corroborated extensively by other survey research (Office for National Statistics—England and Wales, 2015; Statistics Canada, 2013). For example, of those women who indicated they had experienced sexual violence by a partner, they were asked to think about the most serious incident (where women had experienced more than one) and whether they had contacted any organization or service: 61% of women indicated that they did not contact anyone—ranging from the police and medical services through to victim support. A similar pattern emerged for women who had experienced sexual violence by someone other than a partner, with 70% not contacting any organization or service. Only 15% of those experiencing sexual violence by a partner and 14% of those experiencing sexual violence by a nonpartner contacted the police. Similar levels of nonreporting were found with respect to the most serious incident of physical (nonsexual) violence experienced by women: with only 14% of partner violence and 13% of nonpartner violence reported to the police.

In an attempt to shed some light on underreporting, women were asked why they did not contact the police with respect to the most serious incident of physical and/or sexual violence they had experienced by either a partner or nonpartner. The main reasons given were that women dealt with the matter themselves or with the involvement of a friend or family member. Women also indicated that they felt shame or embarrassment as a result of the incident and that they did not want anyone to know about it, both of which were more heightened in cases of sexual violence. It is also notable that in cases of partner violence, women did not report to the police because of fear of the offender, which, again, was higher in cases of sexual violence. At the same time, 34% of women who were victims of physical violence by a partner, and 38% who were victims of physical violence by a nonpartner, indicated that one of the reasons they did not report to the police was because they did not consider the incident to be serious enough to warrant reporting, and that the idea of reporting to the police did not occur to them, a finding that is particularly worrying as it would seem to indicate that violence against women is “normalised” by significant numbers of women who are victims, and that it is still perhaps considered to be a “private” matter. Taken together with the fact that between 5% and 13% said they did not report to the police because they believed that the police would not or could not do anything, the results show that the prospect of police intervention is simply not an option for many women.

In comparison with reporting to the police, just over one third of women experiencing sexual violence by a partner and 28% experiencing sexual violence by a nonpartner had contact with either a hospital, doctor, health center or some other form of health care institution. Very few women—7% or less, depending on the service—reported to social services, a women’s shelter, or a victim support organization. This is likely to reflect a variety of factors such as the nonavailability of certain services where women live, particularly in rural areas, or the fact that seeking out particular forms of assistance, such as a women’s shelter, may be inappropriate in many circumstances.

Women who were in contact with the police or some other service after the most serious incident of physical and/or sexual violence were asked to assess how satisfied they were with the assistance they received. In sum, women are least satisfied with the police than with other services, and a consistent pattern is found of less satisfaction with assistance received by any service—ranging from the police through to victim support—when seeking help in the aftermath of sexual violence. In turn, women were also asked about their specific unmet needs as a result of the most serious incident of physical or sexual violence. The overwhelming majority of women indicated that they simply needed someone to talk to as a result of their victimization. At the same time, one in four women who were victims of sexual violence by a partner and one in five who were victims of sexual violence by a nonpartner said they were in need of protection from further victimization and harassment, which would suggest that more women are in need of protection by the police (or other services) than is currently possible given the number of women reporting to the police.

Looking at the survey results on the extent of violence against women alongside findings on reporting rates, on satisfaction with assistance received by different services, and on women’s unmet needs in the aftermath of violence, it is clear that violence against women is a pervasive problem that is grossly underreported and that women’s needs are generally not being met, a finding that is supported by other survey research (Choo, 2014; Fisher, Daigle, Cullen, & Turner, 2003). As repeat incidents of victimization are a particular problem in cases of partner violence, the scale of unmet needs is further compounded. Drawing on the above survey results, three areas for policy intervention concerning nonreporting are briefly outlined here with respect to policing, medical services, and specialist victim support services.

Low levels of reporting to the police and low levels of satisfaction with the response received from the police, relative to other services, would suggest that the police need to review and change policing culture so that incidents of violence against women are responded to in a sensitive and professional way. This, in turn, may encourage women to report to the police, which—in the long run—can assist in promoting the message that perpetrators cannot act with impunity. The necessity of ensuring an appropriate climate for victims to report to the police is underpinned in policy terms in Europe—by law—in Article 15 of the Istanbul Convention that addresses “training of professionals,” and in this context refers to the needs and rights of victims and the prevention of secondary victimization, while Article 18 in the Convention—covering “protection and support”—also refers to the avoidance of secondary victimization. In turn, different articles under Chapter two in the European Union’s Victims’ Rights Directive variously address the appropriate “provision of information and support” to victims by the competent authority, which includes the police. Given that women are least satisfied with the police in comparison with other services, as indicated in the FRA survey, it is imperative that this finding is followed through in policy terms with action to ensure that the law is applied in practice with respect to policing interventions with victims.

It is down to each EU Member State to apply the law in practice in accordance with the traditions and situation in a country. In other words, once Member States have ratified and become signatories of EU law—such as the Victims’ Rights Directive—the specifics of how it is implemented, and what policies are developed in light of legislation and broad-based European Commission “strategies,” are not determined at the level of the European Union. Useful comparisons for the development of policy can be made across countries covered in the survey to see how different police forces fare with respect to victims’ satisfaction levels having reported to the police, which, in turn, can be contextualized with respect to the implementation of police training on victims and the allocation of targeted resources to address this area. At present—as FRA’s research on the provision of victim support services has shown (FRA, 2014c)—in the absence of widespread services for victims in several EU Member States, some are establishing generic victim services in police stations. While this may be practical and resource efficient, and can serve to meet the requirements of the Victims’ Rights Directive, it may not meet the needs of some of the most vulnerable victims and those with low levels of trust in the police—such as women who are victims of violence. Therefore, there is a need to assess the effectiveness of these services, specifically for victims of violence against women, once they are established in police stations; that is, do they encourage reporting to the police, and are victims satisfied.

Whereas evidence-based assessment of the impact of (victim) policy in practice is relatively new across a number of EU Member States, already in the 1990s, the U.K.-based “one-stop-shop” initiative for the police to act as a hub for channeling information to victims—concerning the status of their case—was evaluated with mixed results (Hoyle, Cape, Morgan, & Sanders, 1998); while more recent U.K. initiatives for multiagency information sharing in high-risk cases of domestic abuse—involving different services from the police through to specialist nongovernmental organizations—has been assessed as resulting in advantages for preventing victimization if, that is, the correct approach is applied (Stanley, 2014; Steel, Blakeborough, & Nicholas, 2011). Although these models essentially provide a joined-up information sharing service for the victims and/or agencies concerned, rather than a “drop-in” service for victims where different agencies—including the police—can be directly accessed, it is apparent that assessment of new initiatives can learn from past practice where the impact of victim policy—in action—was measured (Birdsey & Snowball, 2013; Wedlock & Tapley, 2016). The importance of converting law into policy practice is clear in view of the following statement by the European Commissioner for Justice, Consumers and Gender Equality, Věra Jourová, when she announced that 2017 would be a year for heightened action at EU level to address violence against women:

I want 2017 to be a game-changing year in the fight against violence towards women. By the end of the year, I would like victims of violence to feel, among other things, that they can go to the police to get the kind of support that they deserve. ¹⁹

At the same time as striving to improve police responses to victims, it has to be acknowledged that many women—even if policing responses are improved—may not wish to report to the police, but would rather seek assistance elsewhere, particularly with medical services. In view of this finding from the FRA’s survey, it can be suggested that medical professionals could play an enhanced role in recognizing violence and abuse as experienced by patients, and could encourage women to seek further assistance or to report violence—where this could have a preventive role. Herein, Article 28 of the Istanbul Convention states:

[parties] shall take the necessary measures to ensure that confidentiality rules imposed by internal law on certain professionals do not constitute an obstacle to the possibility, under appropriate conditions, of their reporting to the competent organisation or authorities if they have reasonable grounds to believe that a serious act of violence covered by the scope of this Convention has been committed and further serious acts of violence are to be expected.

With this article in mind, countries that are signatory to the Istanbul Convention could review legislation and practitioner guidelines in consideration of how medical staff should respond—“under appropriate conditions”—with respect to “a serious act of violence.” In view of this, the survey results show that 87% of all women interviewed in the survey would find it acceptable if doctors routinely ask about violence when patients exhibit certain injuries. This would require targeted training for health care professionals that encourages women to come forward and report abuse in the knowledge that their situation will be responded to appropriately, while ensuring that women who are victims are not deterred from approaching medical services. The potential for preventing further violence—particularly in cases of partner violence—would be a key consideration for any such initiative. In support of this, the Lancet recently reported, in reference to a comprehensive research review that included an analysis of referrals for intimate partner violence, that “Consensus evidence suggests that health services need to work closely with specialist services, including the police, to enhance safety for women and children” (Garcia-Moreno et al., 2015, p. 6).

Finally, as the survey findings found that very few women contacted specialist support services, it is clear that there is a pressing need to review these services and their available resources so that they are in a position to offer targeted support for women who are victims of violence, particularly in cases of sexual assault and rape. As noted earlier, initiatives in a number of Member States to locate generic victim support services in police stations may work for victims of burglary or car crime, but may not meet the specific needs of women who have been victims of violence. However, generic services—with appropriate training—could act as hubs for referral of women to specialist services. Acknowledging the geographical scarcity of specialist services, which reflects financial and other resource challenges, may require that policy and practical initiatives are channeled more toward the provision of smartphone apps, web live-cams, and telephone hotlines for victims (as pioneered in EU Member States such as Finland, where communities are geographically widespread and ICT uptake is commonplace; FRA, 2014c). At the same time, it has to be acknowledged that these forms of support may not be appropriate for everyone, and that in-person contact should always be an option. Given that EU Member States are legally obliged to fulfill the terms of the Victims’ Rights Directive, the review of the provision of targeted services is essential. From November 2015 onward, the European Commission will assess the extent to which EU Member States are fulfilling their duties to transpose and implement the Directive into law—including Article 9(3)b in the Directive that refers to “targeted and integrated support for victims with specific needs, such as victims of sexual violence, victims of gender-based violence and victims of violence in close relationships, including trauma support and counselling.” Beyond transposition in law, the effectiveness of this Article in practice will require follow-up work to see the extent to which women—as victims—are able to utilize specific services. Also, as many women, as indicated in the survey data, may contact more than one service—ranging from a doctor through to a specialist victim support service—it is important that interinstitutional cooperation is enhanced between the public sector and various nongovernmental organizations, which is underlined by the Directive’s reference to “integrated support.”

Initiatives for harmonized and efficient exchange of data between services concerning particular individuals and cases—providing that appropriate safeguards for data exchange are in place—should serve to ensure victims’ protection from further violence. In this regard, the importance of good quality data can only be underlined as a tool to ensure the best interests of the victim, which is an area where specialist support services—whose resources are often stretched—could benefit from targeted policy interventions that focus on practical assistance in the collection of harmonized and secure data on victims (which, in turn, is aligned with EU law and policy in the field of data protection). As the European Commission has dedicated 2017 as a year for concerted action to combat violence against women, ²⁰ which includes awareness raising and the involvement of nongovernmental organizations, it is apparent that resources need to be allocated to ensure that women are aware of and can access specialist services, as and when needed, and that these services are in a position to provide support.

The scale of women’s underreporting of violence, as indicated in the survey, and particularly their underreporting to the police, corroborates the degree to which official police and criminal justice data cannot hope to capture the extent of violence, nor can these services effectively respond to women’s myriad needs. Given that doctors and health care institutions are the services that are most often contacted by women who are victims of violence, the research findings point to the need to explore other avenues that could potentially capture and effectively respond to women’s experiences of violence, while recognizing the need to bolster and provide new tools for existing specialist services—such as rape crisis intervention and domestic violence services. Multiagency partnerships—that engage different services that work to assist victims—have also been recognized as avenues for a more effective, joined-up response to violence against women, which can encompass a range of services from the police, through to the medical professions and specialist services.

The current situation of widespread underreporting serves to underline the need to enhance women’s confidence to report violence in the knowledge that their victimization is recognized as socially unacceptable, and that all reports of violence will be responded to appropriately in line with the demands of the Victims’ Rights Directive and the Istanbul Convention. In the knowledge that violence against women, particularly in the case of partner violence, is often repeated, continued underreporting of violence only serves to increase women’s risk of further violence and offenders’ impunity.