Rape and sexual violence occur in all societies, and cut across all social classes. Prevalence estimates of rape victimisation range between 6 and 59% of women having experienced sexual abuse from their husbands or boyfriends in their lifetime. Two population-based studies from South Africa have found that 28% and 37% of men, respectively, have perpetrated rape. Estimates of rape perpetration from high-income countries seem to be lower than those from low- and middle-income countries; however, current data make it impossible to confirm this. Women and girls are much more likely to be the victims and men the perpetrators and, in most instances, the perpetrator is known to the victim. Children are particularly vulnerable to sexual abuse, with girls being at greater risk, especially while at school and at home. High rates of child sexual abuse are emerging from the research, with an increasing understanding of the effect of child sexual abuse on later perpetration and victimisation, highlighting the importance of primary prevention for sexual violence to address childhood exposures to violence. Much of our knowledge about sexual violence has historically been based on research undertaken in high-income countries. This, however, is changing with the emergence of good-quality studies from other settings, particularly in Africa, alongside an increasing number of multi-country studies looking at interpersonal and sexual violence. Most countries lack population data on perpetration of sexual violence, across all categories, including children, and a major gap exists in research on sexual violence among sub-groups and populations. Much of the existing research has limitations that affect cross-study comparability, owing to differences in definitions, research tools, methods and sampling used. Improved research is essential. Research priorities for understanding the magnitude of sexual violence prevalence include assessment of the prevalence and patterns of sexual violence victimisation and perpetration in a range of settings, across a range of acts of sexual violence, in men and women, in adults and children, using methodologies based on best practice in gender-based violence research and standard measures of different forms of sexual violence; research on the social context of sexual violence perpetration and victimisation by both men and women; and methodological research to measure sexual violence for particular population sub-groups or violence types, such as child perpetrators or young child victims, or sexual harassment at work and school.
Introduction
Sexual violence is a profound human rights violation and public health concern. It cuts across class and race, and occurs in peace and conflict settings. Perpetrators are most commonly men known to the victims, and often an intimate partner or, in the case of child sexual abuse, a trusted family or community member. Perpetrators of sexual violence may also be women and children.
Sexual violence has been defined in the World Report on Violence and Health as
‘any sexual act, attempts to obtain a sexual act, or acts to traffic for sexual purposes, directed against a person using coercion, harassment or advances made by any person regardless of their relationship to the victim, in any setting, including but not limited to home and work.
Sexually violent acts take place in many different circumstances and settings. These include the following: rape in marriage or dating relationships; rape of non-romantic acquaintances; sexual abuse by those in positions of trust, such as clergy, medical practitioners or teachers; rape by strangers; multiple perpetrator rape; sexual contact involving trickery, deception, blackmail or of persons who are incapacitated or are too drugged, drunk or intoxicated to consent; rape during armed conflict; sexual harassment, including demanding sex in return for work, school grades or favours; unwanted sexual touching; rape of men in prisons; unwanted exposure to pornography; sexual abuse of mentally or physically disabled people; sexual abuse of boys and girls; and violent acts against sexual integrity, including female genital mutilation, inspections for virginity, forced anal examination and forced trafficking of people for the purpose of sexual exploitation. Thus, the diversity of sexual violence encompasses a range of different victim perpetrator relationships, a range of different sexual acts, a range of forms of coercion and contexts of vulnerability and it occurs in a range of settings.
What these examples have in common is that they are all sexual acts or acts directed against women and men’s sexuality, and they all occur in circumstances where consent is not given or not given freely. Laws across countries differ in the aspects of sexual violence that they regard as offences. Great variation exists in the scope of legal definitions of ‘rape’, and even though marital rape, for example, is a criminal offence in many countries and a crime under international law, 127 countries fail to criminalise marital rape.
In this chapter, we review research on the prevalence of sexual violence against women. It does not address all types of sexual violence; for example, it does not discuss sexual harassment or female genital mutilation. Rather, the focus of the paper is on available prevalence estimates on rape and sexual abuse, from peacetime or post-conflict settings. We also discuss some of the challenges experienced when researching sexual violence, provide information on available data sources, and outline what we know about the scope of the problem of rape and sexual violence in terms of victimisation and perpetration of rape.
Researching sexual violence: challenges and data sources
‘The measurement of rape or sexual assault represents one of the most serious challenges in the field of victimization research.
Measuring the extent of sexual violence presents challenges. Sexual violence is an unusual violation in that, irrespective of setting, the victim is often roundly blamed for its occurrence. Furthermore, rape is commonly regarded as defiling. As a result, the experience of sexual violence is seen as stigmatising and shameful, which makes it difficult for victims to share their stories. Social consequences of sexual violence may be extreme; for example, a family may feel that the victim has brought dishonour to them and victims may be forced to marry the perpetrator, ostracised, or even killed. A further challenge that is rooted in rape stigma, as well as women’s internalisation of strong, culturally rooted ideas about male sexual entitlement, is that they may, after the sexually violent act, reclassify acts of sexual violence as ‘not rape’, and thus avoid perceiving they have to take action against the perpetrator and enable them to pass as ‘not raped’. In South Africa, for example, it is estimated that one in 25 women who have been raped have ever reported it to the police, compared with the USA, where higher reporting is found, but still only an estimated 40% of rapes are reported.
Research shows that higher levels of reporting of sexual violence are found when questions are framed around behaviourally specific acts; for example, ‘being forced into sex against your will’. Notions of ‘rape’, ‘violation’ or ‘abuse’ are highly subjective, and questions that use these words always result in low levels of reporting. Considerable debate exists around the use of popular constructions such as ‘having sex’ compared with formal constructions such as ‘sexual intercourse’, with clarification of vaginal and anal penetration. Researchers in the USA have developed and widely used the Sexual Experiences Survey (SES). The SES uses precise specification of acts. It seems likely that the acceptability of this may show cultural diversity, as has been seen in South Africa. When the SES was tested in South Africa, men indicated that they found the anatomically precise questions offensive. They preferred to be asked about the vernacular version of ‘having sex’, and indicated that they all interpreted that as an act of vaginal intercourse.
Questions also need to ask about a range of victim and perpetration contexts, such as when a woman is too drunk to consent, and also multiple perpetrator (gang) rape. Furthermore, they need to be framed around perpetrator type for proper ascertainment. In particular, it is important to ask about sexual violence from intimate partners specifically, as this form is often under-reported unless it is specially asked. The number of questions needed to measure sexual violence exposure will differ, depending on the operational definition; however, all validated scales measuring rape, attempted rape, sexual coercion or sexual harassment use multiple questions.
Ascertainment is also sensitive to the contexts in which research is undertaken, particularly when face to face interviews are used, and in the training and support of fieldworkers. Much higher levels of ascertainment are found in surveys that use small numbers of highly trained and supported fieldworkers. For perpetration surveys, the use of computerised technology for questionnaire completion has also been extremely valuable. These differences influence our ability to compare prevalence estimates across settings. The figures provided in the literature need to be interpreted with these limitations in mind.
The World Health Organization (WHO) Multi-Country Study (MCS) has assisted the field greatly by developing and formalising operational definitions for intimate partner violence. It also provides measures of sexual violence by a non-partner and child sexual abuse, although these are limited in scope and thus cannot be seen as gold-standard measures. The operational definitions and questionnaires provide us with the tools to compare rates of sexual violence across time and settings. The operational definitions used in this study for the different categories of sexual violence researched in the WHO MCS are presented in Table 1 .
| Category | Operational definition |
|---|---|
| Sexual violence by an intimate partner | Physically forced to have sexual intercourse when she did not want to. |
| Had sexual intercourse when she did not want to because she was afraid. | |
| Forced to do something sexual that she found degrading or humiliating. | |
| Sexual violence from the age of 15 years a (by non-partner) | Forced to have sex or to perform a sexual act she did not want to. |
| Childhood sexual abuse (before age 15 years) | Touched sexually or made to do something sexual that she did not want to |
| Forced first sex | First sexual intercourse, forced or rape. |
a , It is suggested that, in future studies, the 15-year cut-off should be changed to 18 years to reflect legal definition of adulthood, which in most parts of the world is 18 (with some exceptions).
The WHO MCS focused on victimisation, and did not include men. It is increasingly understood, that for us to gain insight into the global scope of the problem, the causes of sexual violence and how best to strengthen prevention programmes, it is important to also include men in our research. A methodology for measuring prevalence of perpetration was developed for research in South Africa, and has been used in the multi-country IMAGES study and further refined for the Change Project run by the United Nations Development Project’s Partners for Prevention. It is currently being used in research with over 15,000 men in seven countries in the Asia and Pacific Region ( Table 2 ).
| Category | Operational definition |
|---|---|
| Rape of a partner (current or previous wife or girlfriend) | Forced to have sex with you when she did not want to. |
| Sex when he knew she didn’t want it but believed she should agree because she was a wife or partner. | |
| Forced her to watch pornography when she didn’t want to. | |
| Forced her to do something sexual that she did not want to do | |
| Rape of a non-partner (a woman who was not your wife or girlfriend at the time) | Forced to have sex. |
| Sex with a woman or girl when she was too drunk or drugged to say whether she wanted it or not. | |
| Multiple perpetrator rape | Sex with a woman when she didn’t consent to sex or was forced by ‘you and other men’. |
| Sex with a woman when she was too drunk or drugged to stop it by ‘you and other men’. |
Country-specific national surveys are also a source of sexual violence prevalence data, as are country-specific studies. For example, the USA has created the National Intimate Partner and Sexual Violence Survey (NISVS). The NISVS was started in 2010 and involves annual interviews with adults in English and Spanish. The Demographic and Health Surveys (DHS) and Reproductive Health Surveys also have country-specific data available on violence against women. The Reproductive Health Survey and DHS data have limitations. For example, DHS data have been shown to under-report violence against women consistently, compared with findings from the WHO MCS and country-based studies specifically looking at violence against women. Efforts are being made to strengthen the Violence Against Women module in the DHS, but these do not overcome the problems of surveys that have large numbers of fieldworkers who are not particularly well trained and interested in gender, and do not have adequate support for vicarious trauma.
Researching sexual violence: challenges and data sources
‘The measurement of rape or sexual assault represents one of the most serious challenges in the field of victimization research.
Measuring the extent of sexual violence presents challenges. Sexual violence is an unusual violation in that, irrespective of setting, the victim is often roundly blamed for its occurrence. Furthermore, rape is commonly regarded as defiling. As a result, the experience of sexual violence is seen as stigmatising and shameful, which makes it difficult for victims to share their stories. Social consequences of sexual violence may be extreme; for example, a family may feel that the victim has brought dishonour to them and victims may be forced to marry the perpetrator, ostracised, or even killed. A further challenge that is rooted in rape stigma, as well as women’s internalisation of strong, culturally rooted ideas about male sexual entitlement, is that they may, after the sexually violent act, reclassify acts of sexual violence as ‘not rape’, and thus avoid perceiving they have to take action against the perpetrator and enable them to pass as ‘not raped’. In South Africa, for example, it is estimated that one in 25 women who have been raped have ever reported it to the police, compared with the USA, where higher reporting is found, but still only an estimated 40% of rapes are reported.
Research shows that higher levels of reporting of sexual violence are found when questions are framed around behaviourally specific acts; for example, ‘being forced into sex against your will’. Notions of ‘rape’, ‘violation’ or ‘abuse’ are highly subjective, and questions that use these words always result in low levels of reporting. Considerable debate exists around the use of popular constructions such as ‘having sex’ compared with formal constructions such as ‘sexual intercourse’, with clarification of vaginal and anal penetration. Researchers in the USA have developed and widely used the Sexual Experiences Survey (SES). The SES uses precise specification of acts. It seems likely that the acceptability of this may show cultural diversity, as has been seen in South Africa. When the SES was tested in South Africa, men indicated that they found the anatomically precise questions offensive. They preferred to be asked about the vernacular version of ‘having sex’, and indicated that they all interpreted that as an act of vaginal intercourse.
Questions also need to ask about a range of victim and perpetration contexts, such as when a woman is too drunk to consent, and also multiple perpetrator (gang) rape. Furthermore, they need to be framed around perpetrator type for proper ascertainment. In particular, it is important to ask about sexual violence from intimate partners specifically, as this form is often under-reported unless it is specially asked. The number of questions needed to measure sexual violence exposure will differ, depending on the operational definition; however, all validated scales measuring rape, attempted rape, sexual coercion or sexual harassment use multiple questions.
Ascertainment is also sensitive to the contexts in which research is undertaken, particularly when face to face interviews are used, and in the training and support of fieldworkers. Much higher levels of ascertainment are found in surveys that use small numbers of highly trained and supported fieldworkers. For perpetration surveys, the use of computerised technology for questionnaire completion has also been extremely valuable. These differences influence our ability to compare prevalence estimates across settings. The figures provided in the literature need to be interpreted with these limitations in mind.
The World Health Organization (WHO) Multi-Country Study (MCS) has assisted the field greatly by developing and formalising operational definitions for intimate partner violence. It also provides measures of sexual violence by a non-partner and child sexual abuse, although these are limited in scope and thus cannot be seen as gold-standard measures. The operational definitions and questionnaires provide us with the tools to compare rates of sexual violence across time and settings. The operational definitions used in this study for the different categories of sexual violence researched in the WHO MCS are presented in Table 1 .
| Category | Operational definition |
|---|---|
| Sexual violence by an intimate partner | Physically forced to have sexual intercourse when she did not want to. |
| Had sexual intercourse when she did not want to because she was afraid. | |
| Forced to do something sexual that she found degrading or humiliating. | |
| Sexual violence from the age of 15 years a (by non-partner) | Forced to have sex or to perform a sexual act she did not want to. |
| Childhood sexual abuse (before age 15 years) | Touched sexually or made to do something sexual that she did not want to |
| Forced first sex | First sexual intercourse, forced or rape. |
a , It is suggested that, in future studies, the 15-year cut-off should be changed to 18 years to reflect legal definition of adulthood, which in most parts of the world is 18 (with some exceptions).
The WHO MCS focused on victimisation, and did not include men. It is increasingly understood, that for us to gain insight into the global scope of the problem, the causes of sexual violence and how best to strengthen prevention programmes, it is important to also include men in our research. A methodology for measuring prevalence of perpetration was developed for research in South Africa, and has been used in the multi-country IMAGES study and further refined for the Change Project run by the United Nations Development Project’s Partners for Prevention. It is currently being used in research with over 15,000 men in seven countries in the Asia and Pacific Region ( Table 2 ).
| Category | Operational definition |
|---|---|
| Rape of a partner (current or previous wife or girlfriend) | Forced to have sex with you when she did not want to. |
| Sex when he knew she didn’t want it but believed she should agree because she was a wife or partner. | |
| Forced her to watch pornography when she didn’t want to. | |
| Forced her to do something sexual that she did not want to do | |
| Rape of a non-partner (a woman who was not your wife or girlfriend at the time) | Forced to have sex. |
| Sex with a woman or girl when she was too drunk or drugged to say whether she wanted it or not. | |
| Multiple perpetrator rape | Sex with a woman when she didn’t consent to sex or was forced by ‘you and other men’. |
| Sex with a woman when she was too drunk or drugged to stop it by ‘you and other men’. |
Country-specific national surveys are also a source of sexual violence prevalence data, as are country-specific studies. For example, the USA has created the National Intimate Partner and Sexual Violence Survey (NISVS). The NISVS was started in 2010 and involves annual interviews with adults in English and Spanish. The Demographic and Health Surveys (DHS) and Reproductive Health Surveys also have country-specific data available on violence against women. The Reproductive Health Survey and DHS data have limitations. For example, DHS data have been shown to under-report violence against women consistently, compared with findings from the WHO MCS and country-based studies specifically looking at violence against women. Efforts are being made to strengthen the Violence Against Women module in the DHS, but these do not overcome the problems of surveys that have large numbers of fieldworkers who are not particularly well trained and interested in gender, and do not have adequate support for vicarious trauma.
Sexual violence prevalence: what do we know?
Sexual violence is rooted in gender inequality and discrimination. It occurs at an alarmingly high rate in many settings. The WHO MCS found that between 6 and 59% of women have experienced sexual violence from their husband or a boyfriend in their lifetime ( Table 3 ). Prevalence estimates of sexual violence victimisation from other population-based studies from 39 countries estimate that between 0.3% and 39% of women report intimate partner or non-intimate partner sexual violence at some point in their lives. A review of violence against women in Latin American and Caribbean countries, using data from Reproductive and Demographic Health Surveys from six countries, c
c Paraguay, Nicaragua, Jamaica, Guatemala, El Salvador and Ecuador.
found that between 5.8% and 13.4% of women reported forced intercourse in their lifetime, with up to one-third of first experiences occurring before the age 15 years. South African studies estimate prevalence rates for sexual violence across all categories, ranging from between 12 and 28% of women ever reporting rape.
