Policy Research Working Paper 10016 Intimate Partner Violence against Women Prevalence, Formal Reporting, and Risk Factors in Chile Trinidad Saavedra Manuel Contreras-Urbina Gabriela Inchauste Poverty and Equity Global Practice & Social Sustainability and Inclusion Global Practice April 2022 Policy Research Working Paper 10016 Abstract Intimate partner violence is among the most common abuse in childhood. Other factors include characteristics of forms of violence against women. In Chile, one in four partners or ex-partners associated with aggressive behav- women who have been in a partner relationship report ior in public spaces, having been a victim of intrafamily having experienced some type of partner violence in the violence in childhood, and frequent alcohol consumption. past 12 months, whether psychological, physical, sexual, The household dynamics that prevent women from partic- or economic. However, only 22 percent of female victims ipating in economic decision-making and the widespread of intimate partner violence file a formal complaint. This acceptance of inequitable gender norms also significantly study analyzes the factors that determine the likelihood increase the risk that a woman will experience intimate that a woman will be subject to violence perpetrated by partner violence. The likelihood that a woman will formally her partner or ex-partner and the factors that determine report intimate partner violence is mainly determined by the probability of reporting the abuse. Individual factors the frequency of the episodes, characteristics of the partners that increase women’s risk of experiencing intimate partner or ex-partners, economic empowerment, and whether she violence include being young, having fewer years of educa- has support networks. tion, having a disability, and having been a victim of sexual This paper is a product of the Poverty and Equity Global Practice and the Social Sustainability and Inclusion Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank. org/prwp. The authors may be contacted at tsaavedrafacusse@worldbank.org, mcontrerasurbina@worldbank.org, and ginchauste@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Intimate Partner Violence against Women: Prevalence, Formal Reporting, and Risk Factors in Chile Trinidad Saavedra, Manuel Contreras-Urbina, and Gabriela Inchauste1 JEL codes: J12, J16, K42, O15 Key words: Chile, gender-based violence, intimate partner violence, risk factors, discrete choice modeling *The authors are grateful to Jacobus De Hoop and Maria Beatriz Orlando, who thoroughly reviewed earlier drafts of the paper and provided relevant comments that helped to substantially improve it. All views expressed in this paper, and all remaining errors, are those of the authors. 1. Introduction Gender violence constitutes a serious violation of women’s human rights. UN Women defines violence against women and girls as "any act of gender-based violence that results in, or is likely to result in, physical, sexual, or mental harm or suffering to women, including threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or in private life.” Violence against women has serious short- and long-term consequences on women’s physical, sexual, and mental health and, in its most extreme form, can result in death. Gender violence not only affects the overall well-being of women and their children, but also prevents them from participating and developing their full potential in society. Violence against women is prevalent worldwide. According to the most recent data from the World Health Organization, 30 percent of women worldwide have experienced intimate partner physical and/or sexual violence or non-partner sexual violence in their lifetime (WHO, 2021). While women may be exposed to other forms of violence, this figure already represents a large percentage of women. Additionally, a report by the United Nations Office on Drugs and Crime found that nearly 60 percent of all female homicides in 2017 were committed by an intimate partner or family member (UNODC, 2018). Gender-related killings are known as "femicide" or "feminicide." Intimate partner violence against women is among the most prevalent forms of violence against women and includes psychological, physical, sexual, and economic violence, as well as controlling behaviors by an intimate partner. Over 640 million women worldwide are subject to acts of violence perpetrated by an intimate partner, which corresponds to 26 percent of all women aged 15 years and over (WHO, 2021). Additionally, in 2017, 34 percent of homicides of women were committed by their current or former partners (UNODC, 2018). This study aims to assess violence against women in Chile, focusing on intimate partner violence given its high prevalence. Specifically, the study analyzes the prevalence rates of the different forms of intimate partner violence, reporting rates, determinants of the probability that a woman will experience violence perpetrated by her partner or ex-partner, as well as determinants of the likelihood that a woman will formally report this abuse. Data from the 2020 Survey on Intrafamily Violence and Violence against Women (ENVIF-VCM) indicate that 1 in 4 women who have been in a partner relationship reports having experienced some type of intimate partner violence over the past 12 months, whether psychological, physical, sexual, or economic. Psychological violence is the most prevalent. Additionally, 26 percent of women who reported experiencing intimate partner violence in the past 12 months were subject to multiple types of violence. On the other hand, only 22 percent of women who report having been victims of intimate partner violence file a formal complaint. When disaggregated by type of violence, the data indicate 34 percent of the women victims of physical violence file a formal complaint, 21 percent the women victims of phycological violence, and 15 percent of the women victims of sexual violence. In addition, between 82 and 97 percent of the women who filed a formal complaint did it through the Chilean police force (Carabineros), highlighting this institution’s relevance for receiving and accompanying abused women. The main reasons for not reporting episodes of intimate partner violence include believing that the violence episodes were not severe, feeling ashamed to report the situation, and believing that reporting is useless. 2 The analysis of risk factors is crucial for designing policies to prevent gender-based violence. The literature identifies several factors that increase the likelihood that a woman will experience intimate partner violence. Some of these factors have been consistently identified in studies from different countries, while others are context-dependent and may vary across countries and even within the same country, for example, between urban and rural areas (Naved et al., 2005; Abramsky et al. 2011; Fulu et al., 2013; Koening et al., 2013; Jewkes et al., 2017; Ram et al., 2019). Risk factors may also vary depending on when the episode of violence occurred (over the lifetime or in the past 12 months) and the types of violence experienced (single or multiple types of violence). Among the risk factors identified in the literature are being young, having fewer years of education, having been a victim of or witness to violence in childhood, partner’s use of drugs or alcohol, low socioeconomic level of women, and acceptance of violence and inequitable gender social norms. While the literature on the determinants of violence against women is quite extensive, there are no previous quantitative studies on this subject in Chile. The risk factors associated with intimate partner violence identified in this study are consistent with the literature. The study findings also indicate that women’s economic empowerment in certain contexts leads to an increase in violence. Men who feel their masculinity threatened may try to regain their power and control over women by retaliating with violence. Finally, the household dynamics that prevent women from participating in economic decision-making and the widespread acceptance of social norms and beliefs that foster gender inequalities and violence against women are also factors that significantly increase the risk that a woman will become a victim of intimate partner violence. The likelihood that a woman will formally report intimate partner violence is mainly determined by the frequency of the episodes, characteristics of the partners or ex-partners, economic empowerment, and whether she has support networks. This study provides evidence on the phenomenon of intimate partner violence in Chile and aims to inform public policy making in this area. The study also contributes to identifying risk variables that could be included in an early warning system of high-risk violence cases to prevent reported cases of violence from ending in femicides. This analysis is organized as follows: Section 2 provides some background on violence against women in Chile. Section 3 describes the data used and presents descriptive statistics on the prevalence of intimate partner violence, reporting, and reasons for not reporting. Section 4 presents the methodology used to identify the factors that determine both the probability of experiencing intimate partner violence and the likelihood of reporting it. Section 5 presents the study findings. Finally, Section 6 discusses some relevant policy areas to guide gender-based violence prevention efforts in the country. 2. Background on violence against women in Chile In Chile, violence against women has been substantiated by the rates of intrafamily violence (private sphere), sexual violence crimes, and femicides. The Intrafamily Violence Law (Law No. 20,066, 2005) 2 defines intrafamily violence as “any abuse that affects the physical or mental integrity of an individual who is or was the spouse of the offender or who cohabitates with them; or who is related within the third degree of collateral consanguinity or affinity to the offender, their spouse or current partner.” 2 https://www.bcn.cl/leychile/navegar?idNorma=242648 3 The results of the latest ENVIF-VCM, 3 carried out by the Undersecretary for Crime Prevention, show that 2 in 5 women report having been a victim of some type of intrafamily violence in their lifetime, whether psychological or physical violence by a relative, partner or ex-partner, or sexual violence by a partner or ex-partner. 4 The intrafamily violence indicator has increased significantly over the years, from 32.6 percent in 2012 to 38.2 percent in 2017 and to 41.4 percent in 2020. This trend is mainly explained by the increase in psychological violence. 5 Additionally, 73.7 percent of reported incidents of psychological violence and 72.2 percent of those of physical violence were perpetrated by the current or ex-partner. Chilean legislation criminalized femicide for the first time in 2010, through Law No. 20,480, 6 which defines femicide as “the killing of a woman carried out by someone who is or has been her spouse or partner, or with whom she has or has had a child in common, by reason of having or having had a sentimental or sexual relationship with her without living together.” Some emblematic cases of femicide in recent years have given rise to laws and bills that seek to modify the definition of femicide, such as the "Gabriela Law," enacted in March 2020 (Law No. 21,212 7), which broadens the typification of femicide to include as the perpetrator of this crime whomever, for reasons of hatred, contempt or abuse due to gender, kills a woman, not only when it is her spouse or partner. The "Justice for Antonia" bill seeks to criminalize the incitement of suicide, also known as femicide-suicide. The National Service for Women and Gender Equity (SernamEG) reports that 43 completed femicides 8 and 151 attempted femicides 9 occurred during 2020. The COVID-19 pandemic has further exposed women's vulnerability to intimate partner violence. A recent UN Women report on the impact of Covid-19 on violence against women in Latin America and the Caribbean found that cases of violence increased considerably since the onset of the pandemic and the establishment of quarantine measures. The number of calls to helplines rose fivefold in some countries (UN Women, 2020). In Chile, the number of calls to the SernamEG hotline for guidance on violence against women nearly tripled between March and July 2020, according to information provided by the Ministry of Women and Gender Equity. 10 This situation demonstrates that for many women, the home is not a safe place. 3 The 2020 ENVIF-VCM is representative of the population of women aged 15 and over in urban areas at the national and regional levels. 4 The intrafamily violence indicator includes women aged 15 to 65 in urban areas. 5It is difficult to determine the extent to which this increase owes to rising violence rates, particularly psychological violence; and how much is due to the fact that violence is no longer normalized and, as such, women are not only more aware that they are or have been victims of violence but are more willing to report if it occurs. In recent years, the government, non- governmental organizations, and feminist movements have increased their outreach campaigns to raise awareness of gender- based violence and inequality, which may have influenced the increase in reporting. 6 https://www.bcn.cl/leychile/navegar?idNorma=1021343 7 https://www.bcn.cl/leychile/navegar?idNorma=1143040 8 At the civil society level, in 2020, the Chilean Network Against Violence Against Women registered 58 completed femicides, 3 femicide-suicides and 6 murders due to violence. These figures differ from official data because of limitations in the legal definition of femicide in Chile. Source: Femicide Registry Chilean Network Against Violence Against Women (accessed 01.11.20): https://docs.google.com/spreadsheets/d/1s_g16Ttsm0S1_9oMH2xilgDMRNaJTALjjV- wAdR5xVc/edit#gid=1952631340 9 Source: Statistical data on femicides from SernamEG (accessed on 01.11.20): https://www.SernamEG.gob.cl/?page_id=27084 10 https://www.latercera.com/politica/noticia/ministra-zalaquett-informa-aumento-del-166-en-las-atenciones-del-fono- 1455-en-5-meses-y-pide-avanzar-en-ley-de-violencia-integral-para-proteger-a-las- mujeres/LBT5HCWZNVCWFJDZXTRR5RTEXE/ 4 3. Data Data from the 2020 ENVIF-VCM was used to conduct the analysis. The ENVIF-VCM survey is carried out by the Undersecretary for Crime Prevention of the Ministry of the Interior and Public Safety and collects information on different forms of intrafamily violence such as psychological, physical, sexual, and economic violence, and on forms of violence experienced in other contexts, such as violence in public spaces, the workplace, and schools. It also gathers information on sexual crimes occurring during childhood and violence against older adults (women aged 66 and over). Additionally, the questionnaire includes questions about the sociodemographic characteristics of the women surveyed, the head of the household, and the current and/or former partner. Questions also elicit information on previous aggressive behavior of the current and/or former partner; economic dependence and the ability to dispose of one's own or the partner's financial resources; attitudes towards gender roles; and support from the extended family and participation in networks. The survey is nationally and regionally representative of the female population aged 15 and over in urban areas. The 2020 ENVIF-VCM surveyed 7,735 women; 6,775 of this sample were aged 15 to 65 while 960 were over age 66. 11 Since this study focuses on violence inflicted by a partner or ex-partner during the past 12 months, 12 the data sample was restricted to the subgroup of women aged 15 to 65 who reported that they are or have been in a relationship. This subgroup corresponds to 89 percent of the sample of women in that age group. Nearly 20 percent of the women in this subsample are between the ages of 15 and 25, 65 percent are between 25 and 55, and 15 percent are between 55 and 65. Moreover, 6 percent are foreigners, 10 percent are indigenous women, 6 percent have a disability, 30 percent have a tertiary education, and 51 percent are employed. Sixteen percent report having been a victim of childhood sexual abuse by a relative. 3.1. Prevalence One in four women (25 percent) aged 15 to 65 who have been in a relationship reported having been subject to some type of violence by their partner or ex-partner in the past 12 months, whether psychological, physical, sexual, or economic (Appendix A provides details on the episodes of violence included in each type of violence). When disaggregating the data by type of violence, psychological violence appeared to be the most prevalent. Indeed, 17 percent of respondents reported having experienced psychological violence, 11 percent economic violence, 4 percent sexual violence, and 3 percent physical violence (Figure 1). Of the total number of women aged 15 to 65 who reported having been a victim of some type of intimate partner violence in the past 12 months, 26 percent said that they had experienced more than one type of violence. In general, physical and sexual violence are accompanied by psychological violence. In fact, 17 percent reported having been victims of physical and/or sexual violence and psychological violence simultaneously, and 8 percent declared having experienced psychological and economic violence without physical or sexual abuse (Figure 2). 11 In previous versions of the survey, the study population corresponded to women aged 15 to 65. Beginning in 2020, the survey incorporated women aged 66 and over in the sample in an effort to measure violence against older women. 12 Given the way the data are collected, it is not possible to identify the perpetrator in cases of violence that occurred prior to the last 12 months. 5 Figure 1: Prevalence of intimate partner Figure 2: Distribution of women aged 15 to 65 years who violence in the past 12 months, by type of violence were victims of some type of intimate partner violence in the past 12 months, according to the type of violence experienced Psychological Psychological and economic 17% but not physical and/or sexual: Only 8,2% Only psychological: economic: Some type of 43.6% 24.5% Physical domestic Sexual Physical and/or 3% violence 4% sexual and 25% psychological Physical and economic: Physical and/or sexual 9.2% and/or sexual and and economic, psychological but not , but not psychological Economic economic Only 1.5% 11% 7.6% physical or sexual: Physical only: 1.3% 5.5% Sexual only: 4.0% Physical and sexual: 0.2% Source: Own elaboration based on data from the Source: Own elaboration based on data from the 2020 ENVIF-VCM. 2020 ENVIF-VCM. The prevalence rates of all forms of violence were analyzed according to individual characteristics of women, characteristics of partners or ex-partners, and attitudes towards inequitable gender norms. Appendix B provides details on these statistics. The data show that the prevalence of intimate partner violence is higher among women belonging to younger age groups. In fact, the prevalence of partner violence reaches almost 41 percent in the group aged 15 to 18, and 36 percent in the 19-to-25 age group. This trend remains after disaggregating the data by type of violence for all forms of violence except for economic violence, where no significant differences are observed between the different age groups. Additionally, more foreign women report economic violence than their Chilean counterparts (18 percent versus 10 percent). No significant differences are observed in the other types of violence. The prevalence of both psychological and sexual violence is higher among indigenous women than among women who do not identify as indigenous (23 percent versus 16 percent and 5 percent versus 3 percent, respectively). In the other types of violence, no significant differences are observed. Economic violence is also more prevalent among women with a disability than among those without one (16 percent versus 11 percent), as is the prevalence of physical violence (6 percent versus 3 percent). Significant differences in prevalence by education and socioeconomic levels exist in all forms of intimate partner violence. The prevalence of psychological violence among women who have completed primary education or less is 21 percent, while among women who have a tertiary education, it is 12 percent. These figures are 15 percent and 7 percent for economic violence, 5 percent and 2 percent for physical violence, and 4 percent and 3 percent for sexual violence, respectively. Similarly, the prevalence of psychological violence among women of low socioeconomic status is 22 percent, while that of women of high socioeconomic status is 8 percent. These figures, respectively, are 13 percent and 4 percent for economic violence, 4 percent and 1 percent for physical violence, and 4 percent and 3 percent for sexual violence. 6 For all types of violence, the prevalence of intimate partner violence is significantly higher among women who were victims of childhood sexual abuse 13 than those who were not. The prevalence of psychological violence among women who were victims of childhood sexual abuse is 28 percent; economic violence, 15 percent; sexual violence, 9 percent; and physical violence, 7 percent. These percentages are significantly higher than among women who were not victims of childhood sexual abuse. Of the respondents who reported having been a victim of sexual violence, 42 percent were victims of sexual abuse in childhood, and of those who reported having experienced physical violence, 33 percent were victims of sexual abuse in childhood. Thus, a strong association exists between sexual abuse in childhood and sexual abuse in adulthood by a partner. When comparing the prevalence of intimate partner violence by characteristics of the partners or ex- partners, the data show that for all types of violence, the prevalence is significantly higher among women who have/had partners who experienced intrafamily violence in childhood than among those whose partners do not/did not. Indeed, the prevalence of psychological violence is 29 percent and 13 percent for these two groups; 17 percent and 9 percent for economic violence; 6 percent and 2 percent for physical violence; and 7 percent and 2 percent for sexual violence. Similar patterns are observed in the prevalence of the different types of violence among women who have/had and those who do not/did not have partners who witnessed violence against women within their nuclear household. Additionally, the prevalence of all forms of intimate partner violence is significantly higher among women who report that their partners or ex-partners have caused harm to third parties in public spaces, with harm referring to attacking, humiliating, insulting, or belittling other people. The prevalence of psychological violence reaches 41 percent in this subgroup; 24 percent for economic violence; 16 percent for physical violence; and 13 percent for sexual violence. The prevalence of physical violence in this subgroup is especially striking given that it is eight times higher than the prevalence of physical violence among women who have partners or ex-partners who have not shown aggressive behavior in public spaces. A strong correlation is also observed between the prevalence of intimate partner violence and the frequency of alcohol consumption by the partner or ex-partner. This trend occurs in all types of violence: 39 percent of women with partners or ex-partners who consume alcohol very frequently (six alcoholic drinks per day at least once a week or almost daily) report having experienced psychological violence versus 13 percent of women with partners or ex-partners who do not drink or who drink less than once a month. The figures for the other types of violence are 23 percent versus 10 percent for economic violence, 13 percent versus 2 percent for physical violence, and 10 percent versus 2 percent for sexual violence. The data also indicate a strong association between the prevalence of certain forms of intimate partner violence and women's attitudes toward gender roles. For example, the prevalence of economic violence is 19 percent among women who agree that housework is a more appropriate task for women than for men, while it is 10 percent among women who disagree with this belief or are indifferent. Economic violence is also higher among women who agree that men should be responsible for household and family expenses than among those who disagree or are indifferent. The incidence of physical violence is 12 percent among women who agree that it is acceptable for the man to beat his partner in the case of infidelity, while it is 3 percent among women who disagree are indifferent in this regard. For sexual violence, these figures are 11 percent and 4 percent, respectively. 13 Sexual intrafamily violence in childhood includes any episode of sexual abuse occurring up to age 14. 7 3.2. Reporting Although 90 percent of the women who report having been a victim of intimate partner violence said they were aware of the Intrafamily Violence Law, 14 only 22 percent filed a formal complaint. Disaggregated by type of violence, the data reveal that 34 percent of the victims of physical violence, 21 percent of the victims of psychological violence, and 15 percent of the victims of sexual violence said they formally reported the abuse. 15 Additionally, for all types of violence, most women reported the abuse to the police force (Carabineros). Ninety-seven percent who filed complaints of sexual violence, 92 percent who filed complaints of physical violence, and 82 percent who filed complaints of psychological violence did so to the police. Figure 3 Figure 4 Distribution of victims of intimate partner Distribution of victims who formally reported intimate partner violence in the past 12 months, by type of violence violence in the past 12 months, by type of violence and and complaint institution where the complaint was filed Did not report Reported Police (Carabineros) Investigation Police (PDI) Courts Prosecutor’s Office 100 21 15 Other 80 34 Phycological 82 7 5 43 Percentage 60 85 Physical 92 8 1 40 79 66 20 Sexual 97 21 0 0 20 40 60 80 100 Phycological Physical Sexual Percentage Source: Own elaboration based on 2020 ENVIF-VCM. Source: Own elaboration based on 2020 ENVIF-VCM The percentages of informal complaints of the different types of violence, such as asking relatives, friends, or others for help when an episode of violence occurs, are relatively higher than the rates of formal complaints. Of the total number of women who report having been victims of physical violence perpetrated by their partner in the past 12 months, 50 percent state that they asked for help, either from their family, friends, neighbors, or others. This percentage is 30 percent for victims of psychological violence and 18 percent for victims of sexual violence. Additionally, of the total number of victims of intimate partner violence who report having asked for help, between 60 and 68 percent asked for help mainly from their family (Appendix C provides details). The reasons victims of intimate partner violence do not file a formal complaint vary according to the type of violence. However, some reasons are common to all kinds of violence. Nearly 40 percent of the victims of psychological violence who did not file a formal complaint said that they did not consider it severe or did not believe it necessary. Among victims of physical violence who did not file a complaint, 14 In the 2020 ENVIF-VCM, all women who report having experienced any type of violence are asked if they are familiar with the Domestic Violence Law. 15 Economic violence is not included because the survey does not collect information on complaints of this type of violence. 8 the percentage was 23 percent; and among victims of sexual violence who did not file a complaint, it was 13 percent. This result is in line with the fact that women victims of intimate partner violence are more likely to accept social norms that promote gender inequalities and violence against women and, as such, they tend to normalize violence. Additionally, 23 percent of victims of physical violence, 21 percent of victims of sexual violence and 16 percent of victims of psychological violence who did not file a complaint reported that they were ashamed to discuss the situation. Another key reason victims of intimate partner violence say they do not file a complaint is because they believe doing so is useless or because they had reported it before and nothing happened. Indeed, 22 percent of the victims of physical violence said that they did not file a complaint for this reason, while 15 percent of victims of psychological violence and 13 percent of those of sexual violence who did not file a complaint also cited this reason. This reveals victims’ mistrust and dissatisfaction with the institutions that are part of the protection system and with the laws on this issue. A significant percentage of victims of sexual violence who did not file a complaint said they did not do so because things improved or because their partner apologized and told them that they would not do it again (18 percent), or because they were afraid the partner would retaliate against them or their children or relatives (15 percent). Lastly, several victims of psychological, physical, or sexual violence report not having filed a complaint due to a reason other than those listed (Appendix D provides details). 4. Methodology A multinomial logistic regression model was used to identify the factors that determine the likelihood of experiencing intimate partner violence. Regression models enable researchers to establish the relationship between an outcome variable (dependent variable) and a series of explanatory variables or factors (independent variables). The analysis of these models makes it possible to identify not only the factors that are statistically significant in explaining the dependent variable but also the magnitude of the effect. Multinomial logistic regression is a type of regression used to model categorical dependent variables, that is, variables that take discrete values. This type of model serves to estimate the probability of occurrence of one category with respect to another, based on a series of explanatory variables. Binary logistic regression is a type of multinomial logistic regression where the dependent variable can take only two values. In the multinomial regression model, the probability that the dependent variable will take the value of a given category can be represented as follows: exp ( ′ ) � � = ( = /) = = 1, … , ∑=1 exp ( ′ ) Where represents the dependent variable or outcome variable, which can take up to J possible values; ′ = [1, 1 , 2 , … , −1 ]is the set of K explanatory variables associated with observation i and ′ = [0 , 1 , 2 , … , −1 ] is the set of regression coefficients for the j-th category of the dependent variable. 9 In terms of model estimation, the multinomial logistic regression model can be represented by J-1 independent binary logistic regression models, in which one of the categories is established as the pivot or reference category. Assuming that category J is chosen as the reference category and β_J=0 , the model can be represented through the following binary logistic regression models: Pr( = 1) = exp( ′1 ) Pr( = ) Pr( = 2) = exp( ′2 ) Pr( = ) … Pr( = − 1) = exp� ′−1 � Pr( = ) In general, it is: Pr( = ) = exp� ′ � = exp�0 + 1 1 + ⋯ + −1 −1 � = 1, … , − 1 Pr( = ) The model can be linearized by applying the natural logarithm and estimated as follows: ( = ) ln � � = 0 + 1 1 + ⋯ + −1 −1 = 1, … , − 1 ( = ) Pr( =) The ratio Pr( is known as the odds ratio or relative risk. The effect of each explanatory variable or =) factor on the relative risk is called the relative risk ratio (RRR) and corresponds to the exponential function evaluated in the coefficient associated with the explanatory variable in question. For continuous explanatory variables, the RRR represents the change in the likelihood that the dependent variable is in a particular category relative to the reference category, associated with a one-unit change in the corresponding independent variable. In the case of dichotomous variables, the RRR represents the change in the probability that the dependent variable is in a particular category with respect to the reference category, associated with exposure to the dichotomous variable or factor. Assuming that the variable is a dichotomous variable, the effect of on the relative probability is equal to: Pr( = / = 1) Pr( = / = 0) = � = exp� � = 1, … , − 1 Pr( = / = 1) Pr( = / = 0) A relative risk ratio of 1 indicates that exposure to the factor does not affect the relative probability, an RRR greater than 1 indicates a positive association and that exposure to the factor increases the relative probability of obtaining the outcome, and an RRR of less than 1 indicates a negative association and that exposure to the factor decreases the relative probability of obtaining the outcome. In this study, model's dependent variable includes three categories of violence based on the different forms of intimate partner violence to differentiate by the type and severity of the violence experienced, 10 and an additional category to account for those not having experienced any type of violence over the past 12 months, 16 for a total of four categories. The first category corresponds to women who report having experienced psychological and/or economic violence in the past 12 months but who have not experienced physical and/or sexual violence. The second category corresponds to women who report having experienced physical and/or sexual violence in the past 12 months but who do not report being subject to psychological violence. Finally, the third category corresponds to women who report having experienced physical and/or sexual violence and psychological violence in the past 12 months (see Diagram 1). It is highly likely that women who report having experienced only physical and/or sexual violence without psychological violence have also been victims of the latter but are so accustomed to the abuse that they are unaware of having been victims of emotional abuse. In any case, this group of women is considered a separate category, since the fact that they are unaware of experiencing psychological violence may be related to unobservable characteristics (for example, household dynamics that create habits toward psychological abuse) that make them different from the category of women who report both types of violence. The last category was used as the reference or comparison category. Diagram 1 Categories of the dependent variable in the multinomial regression model Psychological Physical and/or Physical and/or No type of intimate and/or economic sexual and sexual violence, but partner violence in violence, but not psychological not psychological the past 12 months physical or sexual violence Only psychological Only sexual Physical and/or sexual and psychological, but not economic Women who did not report intimate partner violence in Only physical and the past 12 months Only economic sexual but who did report intimate partner violence previously are excluded Physical and/or sexual and Only Physical and/or psychological, and sexual and economic psychological and economic, but not economic psychological Source: Own elaboration. 16 Women who report not having experienced any type of intimate partner violence over the past 12 months but who report having experienced intimate partner violence previously are excluded from the analysis. Thus, the reference category includes only women who have never experienced violence. 11 As explanatory variables, the woman’s individual characteristics are used (age, marital status, ethnicity, disability, education level, employment status, number of children from previous relationships and whether they were a victim of sexual abuse in childhood); characteristics of the household (socioeconomic level and whether the economic decisions of the household are made by the partner or other household members); characteristics of the partner or ex-partner (aggressive behavior in public spaces, victim of or witness to violence within their nuclear household and frequency of alcohol consumption); attitudes towards gender roles and beliefs that encourage violence against women (whether they agree that men should be responsible for household and family expenses, whether they agree that women should have sexual relations with their spouse/partner even if they do not want to, and whether they agree that it is acceptable for the man to beat his partner in the case of infidelity); and support from extended family or friends (whether they ask someone for help when they have a problem). Regional variables are also included as controls. The factors that determine the probability of formally reporting episodes of violence were identified through a binary logistic regression model, where the dependent variable includes the categories “reported” and “did not report”. Given that the survey provides information on psychological, physical, and sexual violence complaints, a separate model was estimated for each of these forms of intimate partner violence, conditional on having experienced the type of violence. The variables used as controls were the frequency of episodes of violence, woman’s characteristics, characteristics of the partner or ex-partner, and characteristics associated with participation in support networks. 5. Results This section presents the results of the multinomial logistic regression model, which allows identifying the factors that increase the risk of a woman experiencing different types of violence by her partner or ex-partner during the past 12 months. The binary logistic regression model results for the probability of formally reporting psychological, physical, and sexual violence, conditional on having been a victim of these types of violence, are also presented. All estimated results are expressed in terms of the relative risk ratio. 5.1. Factors that determine the probability of experiencing intimate partner violence Table 1 presents the results of the estimates of the multinomial regression model. In general, the determinants of the likelihood of experiencing only psychological and/or economic violence and those associated with the probability of being a victim of physical and/or sexual and psychological violence are similar, although the magnitude of the effect of each factor on the probabilities is generally more significant in the latter. The results indicate that age is statistically significant and negatively correlated with the probability of experiencing intimate partner violence. In other words, young women are at greater risk of experiencing intimate partner violence, even after controlling for other factors. An additional year of age reduces the risk of being a victim of psychological and/or economic violence without physical and/or sexual violence by 3 percent and the risk of experiencing physical and/or sexual and psychological violence by 5 percent. Women who identify as indigenous have 1.5 times more risk of experiencing psychological and/or economic violence without physical and/or sexual violence than those who do not identify as indigenous. Women who report having children from previous relationships have 1.2 times more risk 12 of experiencing psychological and/or economic violence without physical and/or sexual violence than those who do not. These variables are not statistically significant in the other categories of violence. Additionally, women with disabilities have 1.6 times more risk of experiencing psychological and/or economic violence without physical and/or sexual violence and two times more risk of being the victim of physical and/or sexual violence accompanied with psychological violence than women who do not have disabilities. The high propensity that women with disabilities have to suffer multiple types of violence may be associated with a greater tolerance towards violence due to the different types of discrimination these women face. Education level is negatively correlated with intimate partner violence. In other words, higher levels of education significantly reduce the risk of a woman experiencing any type of violence, particularly the risk of physical and/or sexual and psychological violence. In fact, women without a tertiary education are 3.3 times more likely to be victims of physical and/or sexual and psychological violence than those with that level of education. The findings also reveal that being employed increases the likelihood that a woman experiences intimate partner violence. Women who report being employed have 1.4 times more risk of being the victim of psychological and/or economic violence without physical and/or sexual violence and 1.7 times more risk of experiencing physical and/or sexual violence along with psychological violence than those who are not working. Although counterintuitive, this result has also been reported in other studies (Casique, 2010; Bolis, 2015; World Bank Group, 2019). In fact, some studies suggest that when women perform paid work and are more economically empowered, some men may feel their masculinity and household power status threatened and consequently seek to regain their power and control over the woman by retaliating with violence. Lastly, having been a victim of sexual abuse in childhood significantly increases the risk of experiencing all types of violence, especially the likelihood of being the victim of physical and/or sexual and psychological violence. Women who were victims of sexual abuse in childhood have a 1.9 times greater risk of experiencing psychological and/or economic violence without physical and/or sexual violence; 4.6 times more risk of physical and/or sexual violence without psychological violence; and 5.9 times more risk of physical and/or sexual violence along with psychological violence than those who were not victims of childhood sexual abuse. Regarding the characteristics of the partners or ex-partners, the results of the estimates indicate that women whose partners or ex-partners have caused harm to third parties in public spaces or whose partners or ex-partners consume alcohol very frequently have a higher risk of experiencing multiple types of intimate partner violence. Indeed, the risk of experiencing physical and/or sexual and psychological violence is 10.6 times higher among women whose partners have caused harm to third parties in public spaces than among those whose partners have not shown this type of behavior, while the risk of physical and/or sexual violence without psychological violence is 5.2 times higher and the risk of psychological and/or economic violence without physical and/or sexual violence is 2.1 times higher. The risk of experiencing physical and/or sexual and psychological violence is 6.1 times higher among women whose partners or ex-partners consume alcohol very frequently (six drinks per day at least once a week or almost daily) than among those whose partners or ex-partners do not consume alcohol or do so less than once a month. The risk of physical and/or sexual violence without psychological violence is 3.0 times higher, and the risk of psychological violence and/or economic violence without physical and/or sexual violence is 2.6 times higher. 13 Additionally, women whose partners witnessed violence against women in their nuclear household have 2.5 times more risk of experiencing physical and/or sexual and psychological violence than those whose partners did not. This factor was statistically significant for the other categories of violence. Finally, women whose partners or ex-partners were victims of abuse or insults by their parents have a 2.3 times greater risk of experiencing physical and/or sexual and psychological violence and 2.7 times greater risk of psychological and/or sexual violence and/or economic violence without physical and/or sexual violence. The estimates also indicate that women who generally accept social norms and gender roles that place women in a subordinate position with respect to men have a higher risk of experiencing intimate partner violence. Women who agree or strongly agree that men should be responsible for household and family expenses are 1.4 times more likely to be victims of psychological and/or economic violence without physical and/or economic violence than those who disagree with this belief or are indifferent. Women who report agreeing or strongly agreeing that the man is justified in beating his partner in the event of infidelity have 12.8 times more risk of experiencing physical and/or sexual violence along with psychological violence than those who disagree with this statement or who are indifferent. Additionally, women who agree or strongly agree that a woman should have sexual relations with her husband or partner even if she does not want to have 10.5 times more risk of suffering physical and/or sexual violence without psychological violence than women who do not report agreeing or strongly agreeing with this statement; 4.1 times more risk of suffering physical and/or sexual violence and psychological violence; and 3.7 times more risk of psychological and/or economic violence without physical and/or sexual violence. The economic dependence and limited bargaining power of women within the home due to social norms that legitimize asymmetric power dynamics is strongly associated with a higher risk of experiencing multiple types of intimate partner violence. In fact, women who report that their partner or other household members make economic decisions within the household have a 2.8 times greater risk of experiencing physical and/or sexual violence accompanied by psychological violence and a 1.7 times greater risk of psychological and/or economic violence without physical and/or sexual violence than women who report participating in household financial decisions. Asking for help when they have a problem, whether from immediate family members, friends, or others, significantly reduces women’s risk of experiencing multiple types of violence. Indeed, women who do not usually ask for help when facing a problem have 1.9 times more risk of physical and/or sexual and psychological violence than those who usually ask for help. This result highlights the importance of women having support networks. Table 1: Estimates of the multinomial regression model for the likelihood of having experienced some type of violence. Relative risk ratio 17 (RRR) and standard errors (in parentheses). 17The relative risk ratio (RRR) measures the association between an explanatory variable (factor) and the outcome variable. In this case, the RRR represents the change in the likelihood that a woman has experienced any of the three categories of violence in the past 12 months with respect to the probability of not having been a victim of any type of violence in the past 12 months, given exposure to a particular factor. A relative risk ratio (RRR) equal to 1 indicates that exposure to the factor does not affect the relative probability, a RRR greater than 1 indicates a positive association and that exposure to the factor increases the relative probability, and a RRR less than 1 indicates a negative association and that exposure to the factor decreases the relative probability. 14 Physical and/or Physical and/or Psychological and/or sexual violence, sexual and economic violence, but VARIABLES but not psychological not physical or sexual psychological violence Separated/Div 1.913*** 0.848 1.157 orced (0.463) (0.563) (0.474) Widow 0.979 0.904 0.593 (0.336) (0.951) (0.339) Single 1.357* 0.890 0.861 (0.220) (0.378) (0.274) Age 0.975*** 0.970* 0.948*** (0.00618) (0.0155) (0.0116) Indigenous 1.538** 1.869 1.539 (0.331) (1.342) (0.591) Number of 1.239*** 1.152 1.163 children from previous relationships (0.0811) (0.179) (0.144) Has a 1.646** 0.451 2.079** disability (0.416) (0.307) (0.722) Completed 0.735* 0.587 0.576* secondary education (0.116) (0.250) (0.175) Completed 0.454*** 0.402* 0.306*** tertiary education (0.0837) (0.206) (0.124) Working 1.384** 1.603 1.654* (0.185) (0.602) (0.432) Sexual abuse 1.946*** 4.580*** 5.896*** in childhood (0.334) (1.809) (1.773) Partner or ex- 2.169*** 5.153*** 10.59*** partner has harmed third parties in public spaces (0.634) (2.868) (3.697) Partner or ex- 2.677*** 1.722 2.257*** partner was a victim of domestic violence as a minor (0.490) (0.660) (0.633) Partner or ex- 1.246 1.056 2.452*** partner witnessed violence (0.219) (0.402) (0.680) 15 against women in their nuclear household Partner or ex- 1.293* 3.589*** 1.715* partner consumes fewer than six alcoholic drinks per day or six drinks daily at least once a month (0.171) (1.523) (0.483) Partner or ex- 2.627*** 2.974* 6.108*** partner consumes six alcoholic drinks daily at least once a week or daily (0.614) (1.943) (2.178) Does not 1.749*** 1.692 2.803*** participate in household economic decisions (0.291) (0.819) (0.786) Asks for help 0.820 0.871 0.536** when there is a problem (0.127) (0.352) (0.144) Strongly 1.424*** 1.073 1.535 agrees or (0.193) (0.382) (0.428) agrees that "men should be responsible for household and family expenses" Strongly 3.681*** 10.51*** 4.105* agrees or (1.453) (8.032) (3.450) agrees that "a woman should have sexual relations with her husband or partner, even if she does not want to" 0.496 1.010 12.79*** 16 Strongly (0.248) (1.267) (9.431) agrees or agrees that it is acceptable for "a man to beat his partner in the case of infidelity" Antofagasta 0.623* 0.871 0.486 Region (0.157) (0.658) (0.281) Atacama 0.724 0.179 0.403 Region (0.184) (0.206) (0.257) Coquimbo 0.943 2.101 1.253 Region (0.244) (1.385) (0.657) Valparaiso 0.578** 0.913 0.674 Region (0.154) (0.631) (0.384) O´Higgins 1.122 0.935 1.908 Region (0.292) (0.749) (1.025) Maule Region 1.069 0.109* 1.849 (0.291) (0.127) (0.926) Biobio Region 0.838 0.275 1.321 (0.224) (0.327) (0.726) Araucania 0.716 0.701 0.356 Region (0.193) (0.556) (0.226) Los Lagos 0.356*** 0.155* 0.348 Region (0.118) (0.148) (0.239) Aysen Region 1.273 0.773 0.815 (0.345) (0.606) (0.519) Magallanes 0.569* 0.561 0.159* Region (0.167) (0.471) (0.155) Metropolitan 0.741 0.527 0.722 Region (0.160) (0.354) (0.341) Los Ríos 1.064 0.441 1.344 Region (0.259) (0.405) (0.757) Arica and 1.081 0.802 0.733 Parinacota Region (0.283) (0.611) (0.450) Ñuble Region 0.928 1.247 1.469 (0.243) (0.921) (0.898) Constant 0.483* 0.0415*** 0.154** (0.202) (0.0450) (0.114) Observations 3,615 3,615 3,615 LR chi2() 786.9 786.9 786.9 17 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 5.2. Factors that determine the likelihood of formally reporting psychological, physical, and sexual violence perpetrated by the intimate partner The factors determining the likelihood of reporting each type of violence, conditional on having experienced each type of violence, were identified using a binary regression model. Table 2 presents the results of the estimates. Most factors associated with women’s individual characteristics and attitudes towards gender roles were not statistically significant, which indicates that, in general, there are no significant differences between women who formally report episodes of violence and those who do not, in terms of individual characteristics and level of acceptance of social norms that promote gender inequalities and violence against women. By contrast, the frequency of episodes of violence is statistically significant and is positively correlated with the probability of reporting all types of violence, although the significance of this variable in each type of violence depends on the level of frequency. In this study, victims of sexual violence were 6.2 times more likely to file a complaint if the events occurred more than once a year than if they occurred only once; victims of physical violence were 8.2 times more likely to report it if the episodes occurred more than once a month than if they occurred less frequently; and victims of psychological violence were 3.7 times more likely to file a complaint if the episodes occurred more than once a week than if they occurred less frequently . Regarding the characteristics of the partner or ex-partner, victims of physical violence were 4.5 times more likely to report abuse if their partners or ex-partners consume alcohol very frequently (6 drinks per day at least once a week or almost daily) than if they consume no alcohol or do so less than once a month. Victims of sexual violence are also 7.5 times more likely to report violence and victims of psychological violence are 1.6 times more likely to report violence if their partners or ex-partners were abused or insulted by their parents during childhood than if they were not. Additionally, the likelihood of reporting physical violence was 3.7 times higher if the woman has a paid job than if she does not earn income, and 7.5 times higher if she asks for help from third parties (relatives, friends, neighbors, among others) when she has a problem than if she does not ask for help. The likelihood of reporting sexual violence was 6.6 times higher if the woman participates in activities of groups or organizations. Finally, there is a negative correlation between the woman receiving economic support from her partner or ex-partner and the probability of reporting psychological and sexual violence. Women who are economically dependent on their partners were 0.4 times less likely to report psychological violence than those who are not. Table 2: Results of the estimates of the binary logistic regression models for the probability of reporting psychological, physical, and sexual violence, conditional on having experienced the type of intimate partner violence. Relative risk ratio (RRR) and standard errors (in parentheses). Variables Psychological violence Physical violence Sexual violence 6.190** 18 Frequency of episodes of sexual violence (more than once a year) (4.658) Frequency of episodes of physical 8.173*** violence (more than once a month) (5.523) Frequency of episodes of 3.691*** psychological violence (more than once a week) (1.070) Has a paid job 1.311 3.736** 0.893 (0.368) (2.185) (0.817) Partner or ex-partner was a victim of 1.589* 2.608 7.452** domestic violence in childhood (0.442) (2.226) (6.011) Partner or ex-partner consumes fewer 0.710 2.789 0.678 than six alcoholic drinks per day or six drinks daily at least once a month (0.224) (2.662) (0.666) Partner or ex-partner consumes six 1.214 4.462** 2.159 alcoholic drinks per day at least once a week or daily (0.423) (3.366) (2.421) Receives economic support from the 0.401** 0.363 0.0185** partner or ex-partner (0.147) (0.504) (0.0320) Asks for help when there is a problem 1.432 7.482** 4.591 (0.474) (7.392) (7.883) Participates in activities of groups or 1.223 1.205 6.609** organizations (0.322) (1.067) (5.267) Antofagasta Region 0.928 19.39 (0.512) (36.19) Atacama Region 0.519 5.835 (0.297) (11.15) Coquimbo Region 1.023 8.067 0.00653** (0.559) (13.22) (0.0153) Valparaiso Region 0.573 11.66* 0.149 (0.328) (17.37) (0.214) O´Higgins Region 0.688 3.252 0.0130* (0.394) (4.978) (0.0314) Maule Region 0.532 0.462 (0.316) (0.869) Biobío Region 1.352 1.342 0.0134** (0.708) (1.872) (0.0256) Araucania Region 1.352 0.0727* (0.747) (0.108) Los Lagos Region 0.763 10.25 0.0117* (0.651) (16.64) (0.0284) Aysen Region 0.735 1.802 0.0303* (0.464) (3.033) (0.0577) Magallanes Region 0.408 4.676 (0.365) (8.566) 19 Metropolitan Region 0.506 2.639 0.0420** (0.258) (3.587) (0.0547) Los Ríos Region 0.717 24.65* 0.0119** (0.452) (42.52) (0.0260) Arica and Parinacota Region 0.817 0.224 0.110 (0.446) (0.540) (0.192) Ñuble Region 0.880 3.390 (0.534) (5.513) Constant 0.138*** 0.00119*** 0.0524* (0.0783) (0.00254) (0.0820) Observations 763 129 129 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 6. Discussion and policy implications Eradicating violence against women requires a political agenda that addresses both the causes and the consequences of violence. The elimination of all forms of violence against women is a target of Goal 5 on Gender Equality, one of the Sustainable Development Goals adopted by United Nations member states to promote all persons' economic and social well-being. The large percentage of women who report having been victims of some type of intimate partner violence, together with the number of femicides that occurred in 2020, underscore the need to develop a policy agenda that includes prevention, protection, and reparation programs for violence against women to guarantee that women can exercise their right to live a life free from violence. To design effective programs for the prevention of violence against women, policy makers must consider the multiple underlying risk factors for violence. The risk factors associated with intimate partner violence identified in this study are consistent with the literature. Individual characteristics of the woman that increase the risk of intimate partner violence include being young, having fewer years of education, having a disability, and having experienced sexual abuse in childhood. Risk factors associated with characteristics of partners or ex-partners are aggressive behavior in public spaces, having been a victim of intrafamily violence in childhood, having witnessed violence against women within the nuclear household, and frequent alcohol consumption. Study findings also indicate that the economic empowerment of women in certain contexts may lead to an increase in violence. Men who feel their masculinity threatened may try to regain their control over women by retaliating with violence. Household dynamics that prevent women from participating in economic decision-making and the widespread acceptance of social norms and roles that foster power asymmetries between men and women are also factors that significantly increase the risk that a woman will experience intimate partner violence. Prevention programs need to consider all these factors. Education plays a key role in prevention. Given that less-educated young women are at higher risk of intimate partner violence, it is crucial to promote equal access to education and develop prevention strategies within schools, as schools represent an important space to promote initiatives that address gender inequalities from a young age. Violence prevention initiatives should raise awareness of children and young people concerning gender stereotypes and inequitable traditional gender roles, teaching them to develop egalitarian power dynamics and helping them to develop skills that foster healthy 20 intimate relationships, effective communication, and peaceful conflict resolution. The education system must teach children to build gender relations based on mutual respect and non-violence to end violence against women. Preventing violence against children and establishing adequate protection and support systems for child victims of violence are essential for reducing the future risk of becoming a victim or a perpetrator of violence. Experiences of violence in childhood, including psychological, physical, and sexual abuse, as well as witnessing violence against women within the household, are factors strongly correlated with the risk of a woman experiencing intimate partner violence and of a man becoming an aggressor. Therefore, it is vital to develop protection programs that prevent child violence and provide comprehensive support services to child survivors to break the intergenerational cycle of abuse and reduce their chances of becoming future victims or perpetrators of violence against children and women. Promoting gender equality is essential for ending violence against women. Violence against women results from structural relations of unequal power between men and women, leading to women's domination and discrimination. Social norms that foster gender inequalities and violence against women in the home are still pervasive in many Latin American countries and women who accept them are more likely to experience intimate partner violence. Although social norms can be difficult to change, it is imperative to stop normalizing violence and reject masculinity models that make women subordinate to men. To this end, governments must uphold the civil rights of women, punish perpetrators, and implement and enforce laws in this respect. Communication strategies to increase public awareness of gender inequalities, support for laws against violence against women, and women's awareness of their ability to exercise their rights are also crucial. These strategies need to reach all women, so civil society must play a fundamental role as an intermediary between the central or local government and the population. Initiatives that promote women's economic empowerment are not necessarily effective in combating violence against women if they are not accompanied by training on gender equality and changes in societal attitudes towards gender roles. This study indicates that women who do not participate in household economic decisions are at a higher risk of experiencing intimate partner violence. The findings also suggest that women who work and earn income are more likely to experience violence. Economic empowerment should provide women with the freedom to access resources and take control over their lives, so it would be expected that more empowered women would be in a better position to reject and avoid situations of violence. This study and other studies provide evidence that women's empowerment could lead to increased violence by men, particularly in countries that widely accept inequitable gender social norms. However, the results of this study also indicate that economic empowerment is a determining factor when reporting violence, which is the first step towards breaking the cycle of violence. Consequently, programs promoting economic empowerment to end violence against women should include training on gender equality and activities to encourage more equitable gender norms in society. The analysis of risk factors provides information on the variables that could be used to identify high- risk cases of violence in an early warning system. Currently, when a woman files a formal complaint, she must complete the Pauta Única form, which collects information on most of the variables this study identified as factors that increase the risk of a woman experiencing violence. This information could be used to build an indicator of the level of risk of each victim and to create an early warning system of high-risk cases to prevent future femicides. The Pauta Única would need to be filled out regardless of the woman's channel to file the complaint. This information can then be included in an integrated 21 platform that allows the different institutions involved in the protection and reparation process to access the relevant information. Ending violence against women requires a comprehensive approach and the ongoing commitment of various stakeholders. Given the multifaceted nature of the factors that influence the causes and consequences of gender-based violence, initiatives to combat it require a multisectoral approach and the sustained commitment of different key actors: the central and local governments, the community, and the civil society. Although this study focuses on intimate partner violence against women, it is essential to address the phenomenon in all its dimensions, including violence against women in other contexts and other forms of gender-based violence. 22 References Abramsky T., Watts C., Garcia-Moreno C., Devries K., Kiss L., Ellsberg M., et al. (2011). What factors are associated with recent intimate partner violence? findings from the WHO multi-country study on women's health and domestic violence. BMC Public Health, 11:109. https://doi.org/10.1186/1471- 2458-11-109. PMID: 21324186. Andersson N., Cockcroft A., Ansari N., Omer K., Chaudhry UU., Khan A., & Pearson L. (2009). Collecting reliable information about violence against women safely in household interviews: experience from a large-scale national survey in South Asia. Violence Against Women, 15(4):482-496. Bolis A. & Hughes C. (2015). Women’s economic empowerment and domestic violence: Links and lessons for practitioners working with intersectional approaches. Oxfam America. http://hdl.handle.net/10546/594563 Campbell J., Jones A., Dienemann J., Kub J., Schollenberger J., O’Campo P., Gielen A., & Wynne C. (2002) Intimate partner violence and physical health consequences. Arch Intern Med, 162(10):1157-1163. Capaldi D., Kim H., & Pears K. (2009). The association of partner violence to child maltreatment: A common conceptual framework in Preventing Partner Violence: Research and Evidence-based Intervention Strategies. Edited by: Whitaker DK, Lutzker J. American Psychological Association, Washington D.C. Contreras, M. (2013). Conflict within intimacy: a socio-demographic analysis of male involvement in physical intimate partner violence in Mexico. 10.17037/PUBS.00682337. Ellsberg M., Jansen H., Heise L., Watts C., & Garcia-Moreno C. (2008). Intimate partner violence and women’s physical and mental health in the WHO multi-country study on women’s health and domestic violence: an observational study. Lance, 371(9619):1165-1172. Fulu E., Jewkes R., Roselli T., Garcia-Moreno C. (2013). Prevalence of and factors associated with male perpetration of intimate partner violence: findings from the UN Multi-country Cross-sectional Study on Men and Violence in Asia and the Pacific. The Lancet Global Health, 1(4): e187±e207. https://doi.org/ 10.1016/S2214-109X(13)70074-3. PMID: 25104345 Garcia-Moreno C., Jansen H., Ellsberg M., Heise L., Watts C. (2006). Prevalence of intimate partner violence: findings from the WHO multi-country study on women’s health and domestic violence. Lancet, 368(9543):1260-1269. Gil-Gonzalez D., Vives-Cases C., Alvarez-Dardet C., Latour-Perez J. (2006). Alcohol and intimate partner violence: do we have enough information to act? Eur J Public Health, 16(3):278-284 Harvey A., Garcia Moreno C., Butchart A. (2007). Primary prevention of intimate partner violence and sexual violence: Background paper for WHO expert meeting May 2-3, 2007. Geneva: World Health Organization, Department of Violence and Injury Prevention and Disability. Heise L. (1998). Violence against women: an integrated, ecological framework. Violence Against Women, 4(3):262-290. Hindin M., Kishor S., Ansara D. (2008). Intimate partner violence among couples in 10 DHS countries: Predictors and health outcomes. DHS Analytical Studies No 18 Calverton, MD: Macro International. 23 Jewkes R. (2002). Intimate partner violence: causes and prevention. Lancet, 359(9315):1423-1429. Jewkes R., Levin J., Penn-Kekana L. (2002). Risk factors for domestic violence: findings from a South African cross-sectional study. Soc Sci Med, 55(9):1603±17. PMID: 12297246. Jewkes R., Fulu E., Tabassam Naved R., Chirwa E., Dunkle K., Haardörfer R., Garcia-Moreno C. (2017). UN Multi-country Study on Men and Violence Study Team. Women's and men's reports of past-year prevalence of intimate partner violence and rape and women's risk factors for intimate partner violence: A multi-country cross-sectional study in Asia and the Pacific. PLoS Med, 14(9): e1002381. doi: 10.1371/journal.pmed.1002381. PMID: 28873087; PMCID: PMC5584751. Jewkes R., Levin J., Penn-Kekana L. (2003). Gender inequalities, intimate partner violence and HIV preventive practices: findings of a South African cross-sectional study. Soc Sci Med, 56(1):125-134. Koenig M., Ahmed S., Hossain M., Khorshed Alam Mozumder A. (2003). Women’s status and domestic violence in rural Bangladesh: individual and community-level effects. Demography, 40(2):269-288. Koenig M., Stephenson R., Ahmed S., Jejeebhoy S., Campbell J. (2006). Individual and contextual determinants of domestic violence in North India. American Journal of Public Health, 96(1):1-7. Ram A., Victor C., Christy H., Hembrom S., Cherian A., Mohan V. (2019). Domestic Violence and its Determinants among 15-49-Year-Old Women in a Rural Block in South India. Indian J Community Med, 44(4):362-367. doi: 10.4103/ijcm.IJCM_84_19. PMID: 31802801; PMCID: PMC6881899. Shamu S et al. A systematic review of African studies on intimate partner violence against pregnant women: prevalence and risk factors. PLoS One, 2011, 6(3): e17591. Stith S., Smith D., Penn C., Ward D., Tritt D. (2004). Intimate partner physical abuse perpetration and victimization risk factors: A meta-analytic review. Aggression and Violent Behavior, 10(1):65-98. UNODC (2018), Global Study on Homicide. Vienna, 2018. UN Women (2020). Intensification of efforts to eliminate all forms of violence against women: Report of the Secretary-General (2020), p. 4. Vyas S., Watts C. (2009). How does economic empowerment affect women’s risk of intimate partner violence in low- and middle-income country settings? A systematic review of published evidence. J Int Dev, 21:577-602. World Bank Group (2019). Gender-Based Violence: An Analysis of the Implications for the Nigeria for Women Project (English). Washington, D.C. http://documents.worldbank.org/curated/en/731521555064925144/Gender-Based-Violence-An- Analysis-of-the-Implications-for-the-Nigeria-for-Women-Project. World Health Organization (2021). Global, regional, and national estimates for intimate partner violence against women and global and regional estimates for non-partner sexual violence against women. https://www.who.int/news/item/09-03-2021-devastatingly-pervasive-1-in-3-women-globally- experience-violence. 24 Appendix A Episodes of violence included in the different types of intimate partner violence Table A.1: Episodes of violence included in psychological violence A woman is considered to have been a victim of psychological violence if she reports having experienced at least one of the following: a) Have you been insulted? b) Have you been belittled or humiliated in front of other people? c) Have you been told that everything you do is wrong, that you are clumsy, that you are good for nothing? d) Have you been monitored or controlled? e) When they are angry with you, do they hit walls or doors or break things in the house? f) Have they destroyed your belongings? g) Have they taken your children from you, or have they threatened to take them away? h) Have they threatened to hurt or harm you or someone or something you care about? i) Have they threatened to commit suicide or kill or harm your children? j) Have they threatened to kill you? Table A.2: Episodes of violence included in physical violence A woman is considered to have been a victim of physical violence if she reports having experienced at least one of the following: a) Have you been slapped? b) Have they thrown things at you that could hurt you? c) Have you been pushed, cornered, or had your hair pulled? d) Have you been punched, kicked, or hit with something else that could hurt you, or have they bitten you? e) Have you been dragged or beaten? f) Have they tried to burn you or burned you? g) Have they tried to choke you? h) Have you been tied up or locked up? i) Have you been threatened with a gun, knife, or other weapon? j) Have you been harmed by a gun, knife, or other weapon? Table A.3: Episodes of violence included in sexual violence A woman is considered to have been a victim of sexual violence if she reports having experienced at least one of the following: a) Have they forced you to have sex when you did not want to? b) Have they physically forced you to have sex when you did not want it? c) Have they forced you to perform a sexual act that you found humiliating or degrading? d) Have they threatened to deny you money for the household if you do not have sex or engage in some kind of sexual act when they want? e) Have they forced you to stop using contraception? 25 Table A.4: Episodes of violence included in economic violence A woman is considered to have been a victim of economic violence if she reports having experienced at least one of the following: a) They have refused to give you money for household expenses at some point, even though they have money for other things b) They reproach/reproached you for living off their money c) They do not allow you to make decisions regarding household expenses, they decide what the money is spent on without considering your opinion d) You do not have money because they do not allow you to handle money e) They prohibit/prohibited you from looking for work or working f) They have taken your money or savings against your will g) They have indebted you against your will 26 Appendix B Table B1: Descriptive statistics for women aged 15-65 who have ever been in a relationship. Prevalence of the different types of violence (%, percent) and Chi-square test of independence Any type of Sample Variable intimate partner Psychological violence Economic distribution Physical violence Sexual violence violence violence % chi-square % chi-square % chi-square % chi-square % chi-square Total 100 25 17 11 3 4 Characteristics of the woman Nationality 6.65 *** 0.74 19.46 *** 0.59 0.08 Chilean 94 25 17 10 3 4 Foreign 6 31 19 18 4 4 Age group 26.93 *** 40.17 *** 1.22 13.77 *** 13.38 *** [15-18] 5 41 31 9 3 5 [19-25] 15 36 29 11 7 8 [26-35] 24 25 19 10 4 2 [36-45] 21 22 14 10 2 3 [46-55] 20 21 13 11 3 3 [56-65] 15 19 9 13 1 2 Civil status 26.59 *** 38.90 *** 0.35 6.03 *** 4.33 *** Married/Civil partnership 37 19 11 11 2 3 Separated/Divorced 11 27 19 11 4 2 Widow 2 19 9 9 2 2 Single 50 29 22 11 4 4 Indigenous 10.63 *** 17.54 *** 1.93 0.59 2.97 * Yes 10 30 23 13 4 5 No 90 24 16 11 3 3 27 Has a disability 7.03 *** 1.50 10.89 *** 7.81 *** 1.28 Yes 6 31 19 16 6 5 No 94 25 17 11 3 4 Education level 53.11 *** 27.07 *** 32.35 *** 8.73 *** 4.21 ** Complete or incomplete primary education 27 32 21 15 5 4 Complete secondary education 43 26 18 11 3 4 Complete tertiary or postgraduate education 30 17 12 7 2 3 Is working 2.50 1.19 5.07 ** 2.04 6.24 ** Yes 51 24 17 10 4 3 No 49 26 18 12 3 4 Asks for help when there is a problem 6.25 ** 4.52 ** 16.33 *** 0.93 0.11 Yes 81 24 17 10 3 4 No 19 28 19 14 4 3 Was a victim of childhood sexual abuse 76.83 *** 97.22 *** 22.70 *** 41.15 *** 115.96 *** Yes 16 36 28 15 7 9 No 84 23 15 10 3 2 Characteristics of the partner or ex-partner Has insulted or caused harm to third parties in public spaces 210.42 *** 217.01 *** 95.17 *** 272.58 *** 126.11 *** Yes 8 52 41 24 16 13 No 92 22 15 10 2 3 28 Was hit or insulted by parents in childhood 171.73 *** 159.52 *** 73.34 *** 45.37 *** 62.11 *** Yes 22 39 29 17 6 7 No 78 20 13 9 2 2 Witnessed violence against women in the nuclear household 121.79 *** 139.28 *** 72.71 *** 87.24 *** 51.17 *** Yes 22 37 28 18 8 7 No 78 21 14 9 2 3 Frequency of alcohol consumption during the past 12 months 87.71 *** 101.56 *** 39.04 *** 75.49 *** 35.04 *** Does not consume or drinks less than once per month 54 21 13 10 2 2 Has fewer than six alcoholic drinks per day or six drinks daily at least once a month 38 26 19 10 3 4 Has six alcoholic drinks per day at least once a week or daily 8 49 39 23 13 10 Characteristics of the household Socioeconomic level 47.61 *** 45.90 *** 19.91 *** 13.92 *** 2.64 * High 10 13 8 4 1 3 Medium 39 22 14 10 2 3 Low 51 30 22 13 4 4 29 Who makes the following decisions in the household? 89.61 *** 67.76 *** 54.91 *** 26.09 *** 29.96 *** The partner or other household members 17 37 26 18 6 7 The woman decides or decision is made by mutual agreement 83 23 15 10 3 3 Attitudes towards gender roles The man takes responsibility for household and family expenses 12.90 *** 5.03 ** 26.85 *** 6.42 *** 0.26 Strongly disagree/Disagree/No opinion 61 23 16 9 3 3 Agree / Strongly agree 39 28 19 14 4 4 Housework (cleaning, washing, ironing, cooking) is a task more suitable for women than for men 7.45 *** 0.07 49.56 *** 11.08 *** 3.48 * Strongly disagree/Disagree/No opinion 89 24 17 10 3 3 Agree / Strongly agree 11 29 17 19 5 5 30 A woman should have sexual relations with her husband/partner, even if she does not want to 12.30 *** 0.21 39.59 *** 4.48 ** 0.05 Strongly disagree/Disagree/No opinion 98 25 17 11 3 4 Agree / Strongly agree 2 40 15 30 7 4 It is acceptable for a man to beat his partner in the case of infidelity 1.87 ** 0.26 18.44 *** 17.09 *** 10.73 *** Strongly disagree/Disagree/No opinion 99 25 17 11 3 4 Agree / Strongly agree 1 32 20 28 13 11 Note: All prevalence estimates of violence listed in the table are statistically significant (0.05). In the chi-square test of independence, the level of significance is expressed as *** p<0.01, ** p<0.05, * p<0.1 31 Appendix C Reporting of intimate partner violence Figure C.1 Figure C.2 Distribution of victims of intimate partner violence Percentage of victims of violence who asked for help, by type in the past 12 months, by type of violence and of violence and complaint whether they asked someone for help Asked for help Did not ask for help Phycological Physical Sexual 100 68 Asked from help from family 60 68 80 50 Asked from help from 19 69 16 Percentage friends 50 60 82 Asked from help from 6 6 40 neighbors and work… 2 22 50 Asked from help from others 34 20 0 31 18 0 20 40 60 80 0 Phycological Physical Sexual Percentage Source: Own elaboration based on 2020 ENVIF-VCM. Source: Own elaboration based on 2020 ENVIF-VCM. 32 Appendix D Reasons for not formally reporting intimate partner violence. Table D.1: Reasons victims of intimate partner violence in the past 12 months did not file a formal complaint Reasons for not filing a complaint Psychological Physical Sexual I didn't know I could report this situation 10 9 14 I didn't know where to report it 7 3 3 I don't think reporting it is useful, or I reported it previously and nothing 15 22 13 happened 1 10 2 The place to report it is far from my home I was ashamed to talk about my situation 16 23 21 I was afraid that something would happen to me or my children, relatives, or 8 12 15 acquaintances Things got better or they apologized or 8 7 18 told me it wouldn't happen again Out of fear that my spouse/partner 4 7 3 would go to jail I didn't report it because of what my 4 9 3 children might think Out of fear that they would not believe 6 5 8 me, because I had no witnesses Because my husband/partner is the one who provides financial support to the 3 0 3 household Because it was not something serious and 39 23 13 I did not consider it necessary The police recommended that I not file a 2 3 1 complaint Other 33 44 27 Source: Author’s calculations based on 2020 ENVIF-VCM. Note: Respondents could select a maximum of two reasons, for which reason the percentages do not total 100. 33 Appendix E Results of the multinomial logistic regression for the likelihood of experiencing some type of violence and the binary logistic regression of the probability of reporting the abuse. Table E.1: Estimates of the likelihood of having been subject to some category of violence, coefficients Psychological and/or Physical and/or Physical and/or economic violence, sexual violence, sexual and VARIABLES but not physical but not psychological and/or sexual psychological violence Separated/Divorced 0.647*** -0.0927 0.140 (0.242) (0.599) (0.423) Widow -0.0291 0.0502 -0.649 (0.343) (0.789) (0.635) Single 0.305* -0.165 -0.138 (0.162) (0.364) (0.345) Age -0.0253*** -0.0262** -0.0575*** (0.00634) (0.0130) (0.0134) Indigenous 0.433** 0.458 0.485 (0.215) (0.659) (0.398) Number of children from previous 0.216*** 0.0816 0.177 relationships (0.0655) (0.136) (0.132) Has a disability 0.499** -0.339 0.780** (0.253) (0.622) (0.352) Completed secondary education -0.310** -0.287 -0.659** (0.158) (0.371) (0.320) Completed tertiary education -0.788*** -0.949* -1.157*** (0.184) (0.486) (0.416) Has a job 0.324** 0.532 0.483* (0.134) (0.350) (0.277) Experienced sexual abuse in 0.663*** 1.680*** 1.702*** childhood (0.172) (0.368) (0.314) Partner or ex-partner has caused harm to third parties in public 0.778*** 1.476*** 2.442*** spaces (0.292) (0.543) (0.352) Partner or ex-partner was the victim of childhood abuse 0.985*** 0.621* 0.815*** (0.183) (0.335) (0.294) Partner or ex-partner was a witness to violence against 0.223 0.0412 0.964*** women in the nuclear household (0.176) (0.335) (0.290) Partner or ex-partner consumes fewer than six alcoholic drinks per 0.257* 1.160*** 0.534* 34 day or six glasses daily at least (0.133) (0.381) (0.292) once a month Partner or ex-partner consumes six alcoholic drinks per day at 0.969*** 1.195** 1.809*** least once a week or daily (0.234) (0.545) (0.375) Does not participate in household economic decisions 0.560*** 0.545 1.059*** (0.167) (0.439) (0.292) Asks for help when there is a -0.200 -0.143 -0.669** problem (0.155) (0.353) (0.280) Strongly agree or agree that "men 0.353*** 0.172 0.417 should be responsible for (0.136) (0.317) (0.296) household and family expenses" Strongly agree or agree that "a 1.297*** 1.935** 1.512* woman should have sexual (0.394) (0.780) (0.876) relations with her husband or partner, even if she does not want to" Strongly agree or agree that it is -0.706 1.784 2.271*** acceptable for "the man to beat his (0.495) (1.095) (0.875) partner in the case of infidelity" Antofagasta Region -0.472* -0.186 -0.679 (0.252) (0.734) (0.582) Atacama Region -0.321 -1.787 -0.866 (0.255) (1.151) (0.644) Coquimbo Region -0.0538 0.658 0.287 (0.258) (0.638) (0.528) Valparaiso Region -0.547** 0.00359 -0.463 (0.267) (0.656) (0.586) O´Higgins Region 0.118 -0.0856 0.695 (0.260) (0.775) (0.547) Maule Region 0.0637 -0.669 0.490 (0.273) (0.842) (0.525) Biobío region -0.182 -0.384 0.129 (0.267) (0.826) (0.562) Araucania Region -0.336 -0.325 -1.098* (0.269) (0.737) (0.658) Los Lagos Region -1.038*** -1.211 -1.244* (0.333) (0.827) (0.734) Aysen Region 0.245 -0.371 -0.130 (0.271) (0.775) (0.631) Magallanes Region -0.563* -0.645 -1.773* (0.293) (0.851) (0.961) Metropolitan Region -0.298 -0.575 -0.313 35 (0.216) (0.636) (0.480) Los Ríos Region 0.0663 -0.814 0.324 (0.243) (0.899) (0.576) Arica and Parinacota Region 0.0607 0.442 -1.081 (0.263) (0.663) (0.691) Ñuble Region -0.0786 0.454 0.250 (0.262) (0.658) (0.659) Constant -0.723* -3.404*** -1.735** (0.419) (0.976) (0.773) Observations 3,615 3,615 3,615 LR chi2() 813.6 813.6 813.6 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table E.2: Likelihood of filing a complaint, conditional on having experienced the type of violence, coefficients VARIABLES Psychological violence Physical violence Sexual violence Frequency of episodes of sexual 1.087*** violence (more than once a year) (0.371) Frequency episodes of physical 1.207*** violence (more than once a month) (0.356) Frequency episodes of psychological 0.737*** violence (more than once a week) (0.166) Does paid work 0.137 0.768** -0.0782 (0.156) (0.318) (0.442) Partner or ex-partner was a childhood 0.279* 0.640 1.086*** victim of domestic violence (0.154) (0.423) (0.417) Partner or ex-partner consumes fewer -0.178 0.508 -0.238 than six alcoholic drinks per day or six drinks daily at least once a month (0.173) (0.469) (0.444) Partner or ex-partner consumes six 0.134 0.848** 0.417 alcoholic drinks per day at least once a week or daily (0.199) (0.423) (0.544) Receives economic support from the -0.550*** -0.613 -2.151** partner or ex-partner (0.188) (0.610) (0.868) Asks for help when there is a problem 0.173 1.096** 0.843 (0.180) (0.463) (0.654) Participates in activities of groups or 0.109 -0.0202 1.068*** organizations (0.148) (0.416) (0.357) Antofagasta Region -0.0575 1.593* 36 (0.300) (0.946) Atacama Region -0.456 0.239 (0.317) (1.065) Coquimbo Region 0.0482 1.108 -2.601*** (0.299) (0.837) (0.997) Valparaiso Region -0.316 1.357* -1.028 (0.315) (0.741) (0.717) O´Higgins Region -0.224 0.532 -2.207** (0.315) (0.754) (1.003) Maule Region -0.400 -0.392 (0.314) (0.903) Biobío Region 0.169 0.0254 -2.329*** (0.291) (0.686) (0.865) Araucania Region 0.190 -1.450* (0.308) (0.754) Los Lagos Region -0.136 1.274 -2.143* (0.459) (0.894) (1.178) Aysen Region -0.202 0.404 -1.857** (0.345) (0.847) (0.927) Magallanes Region -0.390 0.760 (0.459) (0.956) Metropolitan Region -0.377 0.551 -1.764*** (0.275) (0.678) (0.650) Los Ríos Region -0.134 1.688** -2.150** (0.338) (0.828) (1.049) Arica and Parinacota Region -0.175 -0.714 -1.172 (0.301) (1.120) (0.871) Ñuble Region -0.0645 0.548 (0.334) (0.843) Constant -1.156*** -3.696*** -1.730** (0.302) (0.929) (0.741) Observations 787 133 131 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 37