Abstract
This research set out to determine prevalence, recurrence, types, and risk factors for intimate partner violence (IPV) among women of childrearing age across Israel, attending to diversity in these factors by ethnicity and immigration status. The first nationwide study of its kind, this research was based on a stratified proportional cluster sample of 1,401 Arab, and Jewish immigrant and nonimmigrant women (aged 16-48 years) who visited 63 maternal and child health (MCH) clinics between October 2014 and October 2015. Female research staff interviewed women face-to-face in a private room at the MCH clinics using a structured questionnaire in the women’s main language (Arabic or Hebrew). We measured IPV using a 10-item questionnaire used for screening at some MCH clinics. Response rate was 74%. In the multivariate analysis, we used generalized estimating equations (GEEs) to adjust for the MCH clinic cluster effect. We found marked differences in the prevalence of IPV among Arab, and Jewish immigrant and nonimmigrant women (67%, 30%, and 27%, respectively). Types (physical, verbal and social) and recurrence of IPV were significantly higher among Arab women compared with the other two groups. In the GEE analysis, compared with IPV among Jewish nonimmigrants, IPV among Arab women persisted after considering socioeconomic, sociodemographic, and reproductive factors (odds ratio = 3.83; 95% confidence interval = [2.55, 5.72]). Low family income was the main risk factor for IPV for all women. Among Arab women, younger age, high religiosity, and living in urban settings were associated with higher IPV. These results suggest that diversity (ethnicity and immigration status) should be considered when developing tailored policies and interventions to protect women from IPV.
Introduction
The World Health Organization (WHO) estimates that one in three women will face some form of violence during her lifetime (Krug, Dahlberg, Mercy, Zwi, & Lozano, 2002). This violence is far more likely to be perpetrated by a romantic or intimate partner than by a stranger (Breiding et al., 2014). About half of female homicides are perpetrated by intimate partners (Breiding et al., 2014; Krug et al., 2002). More prevalent among women than men (Breiding et al., 2014), intimate partner violence (IPV) has been defined as the systematic abuse of persons by their current or previous partner that may be physical, social, emotional, or economic (Jewkes, 2002). A WHO study that included 22 countries estimated prevalence of IPV among women as between 10% and 52% (Garcia-Moreno, Jansen, Ellsberg, Heise, & Watts, 2006).
While IPV is a global and cross-cultural problem that occurs in socioeconomically and racially diverse households, multiple hindrances have prevented researchers from establishing the prevalence of IPV among ethnic minority women, including immigrants. These hindrances include a lack of population-based studies, variations in study design, different measures for IPV and its types (physical, sexual, or emotional; Taillieu & Brownridge, 2010); disclosure issues (Al-Modallal, 2017), and the complexity of culturally based attitudes toward IPV (Haj-Yahia, 1998). While some studies show that IPV disproportionally affects minority women (Benson, Wooldredge, Thistlethwaite, & Fox, 2004; McFarlane et al., 2005; Sampson & Wilson, 1995; Stockman, Hayashi, & Campbell, 2015), other research of minority women actually reports lower IPV (Daoud, Uriqua, et al., 2012). Establishing prevalence of IPV among immigrant women is similarly fraught. Some find low IPV prevalence in this group compared with nonimmigrants (Daoud, O’Campo, Urquia, & Heaman, 2012; Daoud, Uriqua, et al., 2012; Hicks, 2006). But these lower rates have been attributed by some to problems in detection due to disclosure issues related to cultural attributes and patriarchy (Ahmad, Riaz, Barata, & Stewart, 2004; Daoud, Uriqua, et al., 2012). One possibility, then, is that immigrant women may be less likely to disclose IPV due to fear and because they must weigh the implications of disclosure for their lives (Ahmad et al., 2004).
Ethnicity and immigration status also exist in tension with class, which may impact prevalence of IPV. Some suggest that ethnicity might be confounded by socioeconomic position (SEP), as minority women are more likely to be situated at a lower SEP, which exposes them to higher risk of abuse (Benson et al., 2004). In Canada, for example, SEP was a major contributor in explaining inequalities in any abuse and IPV between Aboriginal and non-Aboriginal women (Daoud, Smylie, Urquia, Allan, & O’Campo, 2013). Indeed, when adjusting for economic strain among mid-income African Americans and non-Hispanic Whites in the United States, disparities in IPV disappeared or were dramatically diminished (Schollenberger et al., 2003).
Israel, a country highly stratified by ethnicity, culture, and SEP, as well as a receiving nation for waves of Jewish immigrants from diverse cultural backgrounds, presents a pressing case for the need to establish prevalence of IPV as a path to adequate protective policies for women. Israel’s Jewish and Arab populations are not uniform, and each might present different exposures to and experiences of IPV that are not yet understood. In this study, we sought to address this diversity and fill a gap in the literature by determining prevalence, recurrence, types, and risk factors of IPV among Arab and immigrant and nonimmigrant Jewish women of reproductive age across Israel, and to identify risk factors associated with the gaps in IPV.
Background
For a variety of reasons, nationwide data on IPV in Israel are limited (Muhlbauer, 2006). To date, only reports of extreme cases, such as homicide and police reports are available, while other important forms of IPV go undocumented. The most recent annual report of the Israeli Parliament on violence against women (Mizrahi Simon, 2015) showed that 72 women were murdered between January 2013 and October 2015. Most were killed by their intimate partner or by family members. One third were Arab women, and another fifth were Jewish immigrants from Ethiopia and the former Soviet Union. In the same period, women filed 19,100 complaints with police regarding familial violence; Jewish women made 70% of these complaints.
Until now, the only national survey on domestic violence in Israel was conducted in 2000-2001 (Eisikovits, Winstok, & Fishman, 2004), but that study did not look at IPV and risk factors within ethnic and immigrant groups of women. The survey found that 6% of women had experienced physical violence in the previous year, while 56% had experienced emotional (verbal) violence (Eisikovits et al., 2004). Domestic violence was higher among less educated women in younger families, Muslims, religious Jewish women, unemployed men and women, those married fewer years, those facing economic hardship, and those holding attitudes that justify violence (Eisikovits et al., 2004). As nearly two decades have passed since that study was compiled, an overview of IPV for the entire population and a study of risk factors within each of Israel’s minority groups is in order.
The Jewish majority in Israel currently comprises 75% of the Israeli-born population (Central Bureau of Statistics, 2013). They are mostly descendants of immigrants who arrived shortly after the establishment of the state, many as Holocaust survivors and refugees from Europe and from Arab countries (Shuval & Anson, 2000). The State invested many resources in providing these immigrants with employment, housing, and health care (Shuval & Anson, 2000). As well, since establishment, for ideological reasons, Israel had received new Jewish immigrants from many countries (Shuval & Anson, 2000). Today, the largest proportion of recent immigrant Jews in Israel (who make up 20% of the total population) come from the Former Soviet Union (FSR) and Africa (mainly Ethiopia). As with the earlier settlers, state policies are supportive of these diverse immigrants, who receive stipends, housing, health care, education, and employment services (Shuval & Anson, 2000).
Arabs make up 21% of Israel’s population (Central Bureau of Statistics, 2013). For historical reasons, 85% of Arabs live apart from their Jewish counterparts, in separate villages and towns. The rest live in mixed towns (Al-Haj, 1993). Since the establishment of the state of Israel, Arab society has undergone dramatic transitions from agricultural rural village life to semi-urban living (Al-Haj, 1993). Some argue that such rapid urbanizations might help to explain increased social and community violence in this group. In addition, political violence (Clark et al., 2010) continues to emerge from the long-standing Palestinian–Israeli conflict, which elevates violence among Arabs in Israel and could also affect IPV, not only among Arabs, but in all groups within Israeli society, though this remains an underexamined area of research.
One unique aspect of the Arab citizenry in Israel is that it is a largely native-born demographic. Unlike elsewhere, where ethnic minority status is confounded by immigrant status, for Arabs in Israel, minority status arose through the establishment of the state in 1948 (Yiftachel, 1991). Thus, IPV among Arabs can be examined in relation with socioeconomic status without being confounded by immigration (Ahmad et al., 2004; Martinez, 2000; Yick, 2000). This makes Arabs in Israel a unique case for studying IPV, as IPV can be examined among them separately from immigration. The political status of Arabs and discriminatory policies of the governments have affected their SEP compared with Jewish counterparts (Carmi & Rosenfeld, 1992): Arabs have lower education (Okun & Friedlander, 2005), lower income (Svirsky, Konner-Attias, & Ophir, 2014), and are more likely to be unemployed (Institution for Social Security, 2015).
We assume that different historical experiences as well as different cultural, political, and social standing among ethnic and immigrant groups in Israel will expose women to different levels of IPV, and that risk factors for IPV will also vary by ethnic group.
Method
This study was approved by the Ethics Review Board at Ben-Gurion University of the Negev (BGU) and the Public Health Division at the Israeli Ministry of Health (MCH). Study tools and data collection were prepared and carried out according to WHO guidelines for research on domestic violence against women (WHO, 2001).
Study Design and Sampling
We conducted a cross-sectional study of women of childbearing age selected through a stratified proportional cluster sampling procedure. Women were recruited and interviewed individually during a visit to maternal and child health (MCH) clinics in five large districts of Israel’s MCH (Beer Sheva, Ashkelon, Center, North, and Haifa). Interviews were conducted over a 1-year period—2015). Participating clinics were selected from each health district proportional to the number of births and ethnic composition (Arabs vs. Jews) of women in the district to reach the sample size of 1,401 women (age 16-48). In the end, 63 MCH clinics were included in the study; 21 located in Arab localities, 33 in Jewish localities, and nine in mixed Arab–Jewish localities.
Recruitment and data collection
Prior to data collection, in an attempt to increase participation by our target population of women visiting MCH clinics, we took steps to increase collaboration on the part of clinic nurses. After receiving approval from the Ministry of Health and the ethics committee, the researchers met with medical directors or head nurses of participating health districts and shared with them the study aims, data collection procedures, and questionnaire. The nursing directors then informed nurses at the MCH clinics about the study. Furthermore, 1 month before data collection began, the research team used flyers and posters to advertise the study at the clinics. To put women at their ease, only female research assistants gave out the flyers (in Hebrew and Arabic), which were titled “Family Relations, Violence and Health.” To reach out to women who cannot read, research assistants also provided oral descriptions of the study, emphasizing the right to refuse or agree to participate, and assuring women that their choice would not affect any medical treatment they were eligible for.
Due to the diversity of our study population, the principal investigators also prepared data collection and study tools in both languages. All women who attended the clinic for either prenatal care or well-baby follow-up were approached and asked to participate in the study. Interviewers introduced the study, explained that it was being conducted by researchers from BGU, and noted that information obtained in the study would be used for study purposes only and would not be shared with clinic staff. Women who agreed to participate were invited to a private room at the MCH clinic, where they were asked to sign an informed consent form.
Trained female interviewers conducted the interviews face-to-face in participants’ preferred language (Arabic or Hebrew). The interviewers were recruited from student’s association at the universities, local nongovernmental organizations (NGOs), and women’s organizations. All have at least high school education and many of them had previous experience with interviewing. The principal investigators did the initial training for these interviewers, followed by more training by research coordinators. Training took place in both languages in each of the health districts at least 1 week prior to the interviews. Training lasted 6 hr and included specific information and guidance on IPV, highlighting the sensitivity of interviewing on violence against women, as well as methods for increasing women’s responses while assuring safety and confidentiality. Interviewers were provided with a special kit that included a questionnaire and written instructions on how to approach and recruit women, and how to stick to the study questionnaire to minimize study bias.
At the end of the interview, all study participants received written information about support services for women who are victims of violence, including IPV. Women who disclosed abuse or IPV were encouraged by the interviewer to inform the MCH clinic nurse for professional support. In total, we interviewed 1,401 women. The response rate was 73% among Jewish women and 76% among Arab women.
Measures
The study questionnaire was prepared in Hebrew, translated into Arabic, back translated, and finally piloted in both languages before use to assure similarity and clarity of terminology.
Ethnicity and immigration status
This was a composite variable created from two variables: Ethnicity, which was self-determined (categories: 1. Arab and 2. Jewish), and Immigration Status, determined by a question on country of birth (answers: 1. Israel, 2. Other). Women not born in Israel were assigned as immigrants. For the present analysis, we constructed a three-category variable: 1. Arab, 2. Jewish immigrant, and 3. nonimmigrant Jew.
Any IPV
Any IPV was assessed using a 10-item questionnaire about acts perpetrated by a participant’s intimate partner (categories: 1. never, 2. seldom, 3. often, 4. always). Questions were based on those used in the Preventive Services Task Force Family screening tool on IPV (US Preventive Services Task Force, 2004). A similar questionnaire is used in some MCH clinics to screen for IPV against women during pregnancy and after birth. A woman was defined as a victim of any IPV if she had a positive answer to any of the following 10 items on abuse perpetrated by her intimate partner, creating a scale from 0 to 10:
Are you fearful of drastic changes in your partner’s mood?
Does your partner blame you or your environment for his problems?
Does your partner try to isolate you from your family and friends?
Do you require your partner’s approval for all daily financial expenses?
Is your partner jealous in an extreme manner, to the point that he behaves obsessively, for example, following you, calling you frequently, or needing to know where you are at all times?
Your partner has hit you, kicked you, pushed you, or thrown things at you?
Your partner has threatened to intensify the violence against you if you tell anyone?
Your partner has threatened that he will commit suicide or hurt himself if you leave him?
You partner has forced you to have sex with him against your will?
You live with a constant sense of danger?
Types of IPV
Types were determined based on factor analysis for the above 10 items and categorized as follows: Physical or sexual violence: Items 6, 9, 10; Emotional or verbal violence: Items 1, 2, 7, 8; and Social or economic violence: Items 3, 4, 5.
Recurrence of IPV
Recurrence was determined by women’s selection of response categories (1. seldom, 2. Often, and 3. always) for the above 10 questions on IPV. As the number of answers in the last two categories was small, we summed the answers for these two categories together. We calculated the number of times a woman had answered as 1. seldom, or 2. often and always, for the total 10 acts of IPV (any IPV) and for different types of IPV.
Independent Variables (Predictors)
These were as follows:
Age (categories: 16-24, 25-34, and 35-48).
Marital status (categories: 1. married, and 2. not married, including single, divorced, separated, not-cohabitating, or other).
Women’s status at the time of the interview, which was a composite variable based on answers to two questions: 1. Are you currently pregnant (yes or no), and 2. Do you have children? If a woman had children, she was asked how many (categories: 1. pregnant with no children, 2. pregnant with children, 3. not pregnant with 1-2 children, and 4. not pregnant with 3 and more children).
Women’s education (categories: 1. high school or less, 2. postsecondary education, and 3. university education [bachelor’s, master’s, or doctorate]).
Partner’s education (categories: 1. high school or less, 2. postsecondary education, and 3. university education [bachelor’s, master’s, or doctorate]).
Partner’s employment: current work outside the household (yes/no).
Woman’s current employment: current work outside the household (yes/no).
Family source of income (categories: 1. work only, 2. social allowances only, or 3. other source, which included any combination of work and social allowances, and work and other resources, like grants, family support, land, etc.).
Religiosity level was self-identified variable (categories: 1. religious or very religious, 2. traditional, 3. not religious). It intended to capture women’s level of religiosity across religious and cultural groups (Haj-Yahia, 2002; Steinmetz & Haj-Yahia, 2006). “Traditional” women in Arab or Jewish society adhere to cultural norms through dress, marriages customs, and (sometimes) religious belonging, while “religious” or “very religious” women additionally observe religious commandments and obligations, such as fasting, keeping Shabbat (for Jewish women), and daily prayer (for Arab women).
Type of locality: Answers were categorized into urban or rural, based on participants’ place of residence.
Statistical Analysis
We conducted the analysis using SPSS version 23. We calculated prevalence (%), recurrence, and types of IPV for the total sample and compared them by women’s ethnic group (Arab and Jewish) and immigration status. We then conducted univariate analysis for the associations between socioeconomic and demographic variables and any IPV, along with types of IPV, for the total sample. In the multivariate analysis, we adjusted for the MCH clinic cluster effect by conducting GEE analysis. The GEE analysis was conducted first for the total sample, while adjusting for ethnicity and immigration status and variables that were associated (p < .05) with any IPV in the univariate analysis (age, women’s status at the interview, women’s education, family source of income, level of religiosity, type of locality). We then repeated the GEE analysis for each of the groups in the study (nonimmigrant Jews, immigrant Jews, and Arabs). In this model, we adjusted for potential risk factors (age, women’s status at the interview, women’s education, family source of income, religiosity). The multivariate analysis (GEE) for the Jewish women did not include type of locality, as few Jewish women in our sample lived in rural areas.
Before conducting the multivariate GEE analysis, we examined correlations between the study’s independent variables. We found a correlation of R > .4 between women’s education and husband’s education, women’s education and women’s current employment, and husband’s employment and family source of income. Therefore, we excluded the variables of husband’s education, husband’s employment, and women’s employment from the GEE analysis. Also, husband’s employment did not differ significantly across the three study groups. Marital status was also excluded from the GEE analysis, as it was not associated with the main variable of any IPV in the univariate analysis. Furthermore, we examined interactions between ethnicity, IPV, and each of the independent variables. While we found some significant interactions, none was significant in the GEE analysis. Finally, to assess risk factors for the three groups in the study (Arab, Jewish immigrant, and nonimmigrant Jew), we conducted separate GEE analyses for each group, while adjusting for the same independent variables that were used in the GEE for the total sample.
Results
Table 1 presents the characteristics of the study sample. Arab women were significantly younger, more likely to be married, and more often pregnant at the time of the interview than women in the other two groups of immigrant and nonimmigrant Jewish women. The distribution of age and marital status was almost the same among immigrant and nonimmigrant Jewish women. Most participants had children and were not pregnant during the interview.
Characteristics of the Study Sample by Ethnic Group and Immigration Status.
Arab women had lower socioeconomic status compared with the other two groups of Jewish women. Two thirds of Arab women and three quarters of their husbands had high school education or less, while this was the case for just one quarter of Jewish women and one third of their partners. Only one quarter of Arab women reported that they worked outside their household at the time of the interview, compared with three quarters of nonimmigrant Jewish women and two thirds of Jewish immigrant women. Compared with Jewish immigrant and nonimmigrant women, a higher percentage of Arab women reported that their family income was from social allowances (4%, 2%, and 12%, respectively).
Arab women reported higher religiosity than the other two groups. Two thirds of Arab women self-identified as traditional, compared with one third of the other two groups. One quarter of Arab women and nonimmigrant Jewish women self-identified as religious. Almost half of immigrant Jewish women self-identified as not religious (secular), compared with few Arab women (about 14%). About half of Arab participants lived in rural areas, compared with less than 5% of nonimmigrant Jewish and 2% of immigrant Jewish women. Notably, partner employment did not differ significantly between the three groups of women (Table 1).
The prevalence of any IPV in the total sample was about 40%. About 4.6% of participants reported physical violence, 28.6% reported emotional violence, and 26.1% reported social violence. Figure 1 presents prevalence and types of IPV by ethnicity and immigration status. We found marked differences between Arab, immigrant, and nonimmigrant Jewish women regarding the prevalence of any IPV (67%, 30%, and 27%, respectively). The prevalence of different types of IPV was significantly higher among Arab compared with Jewish immigrant and nonimmigrant Jewish women; physical IPV was 10.6%, 2.9%, and 1.9%, respectively; emotional or verbal violence was 49.7%, 20.6%, and 18.5%; and social or economic violence was 49.3%, 18.1%, and 14.6%.
Prevalence (%) of any IPV and types of IPV by ethnicity and immigration status.
The supplementary appendix presents the distributions of women’s responses for specific acts of IPV. Arab women reported significantly more acts of IPV than the other two groups. Differences were not significant between the study groups in just two instances: Acts 8 and 9 (partner threatening to commit suicide or hurt himself if she left, or having been forced to have sex against her will).
Figures 2 to 5 present results on recurrence of any IPV and types of IPV. Arab women reported higher frequency of all 10 IPV acts and each type of IPV, while nonimmigrant Jewish women reported the lowest frequency of IPV acts.
Recurrence (%) of IPV among nonimmigrant Jewish women (n = 722).
Recurrence (%) of IPV among Jewish immigrant women (n = 243).
Recurrence (%) of IPV among Arab women (n = 436).
Recurrence (%) of IPV among all study participants (N = 1,401).
Table 2 presents univariate associations between the study variables and any IPV and IPV types. Most of the associations were significant. Higher rates of any IPV as well as all types of IPV were significantly associated with younger age, being pregnant at the time of the interview, women’s lower education, partner’s lower education, women not working, having an unemployed partner, households with income from social allowances, traditional level of religiosity, and rural place of residence compared with urban. Exceptions included marital status, which was not significantly associated with any of the IPV variables, as well as age, partner employment, and level of religiosity, none of which was significantly associated with physical IPV.
Associations Between Study Variables and Any IPV and Types of IPV (Physical, Verbal, and Social).
Note. IPV = intimate partner violence.
We did find significant associations between the different types of IPV: of the women who reported physical IPV, 92% reported verbal IPV, and 73.4% reported social or economic IPV. Among women who reported verbal IPV, 57.1% also reported social IPV.
Table 3 presents results of the multivariate GEE analyses of any IPV for the total sample of women and for each of the study groups. For the total sample of women, compared with nonimmigrant Jewish women, Arab women reported having experienced any IPV nearly 4.5 times more often (odds ratio [OR] = 4.38, and 95% confidence interval [CI] = [2.91,6.60]), while no significant difference was found for any IPV between immigrant and nonimmigrant Jewish women. Women who reported that their income was from other sources (not social allowances or work only) and religious women reported about 1.5 times higher any IPV compared with women in the other categories. The variables of age, women’s education, women’s status at the interview, and type of locality were not significant in the final model, and therefore it did not contribute to explaining any IPV.
Multivariate Analysis (GEE) for Any IPV Among the Total Sample of Women and for Each Study Group.
Note. GEE = generalized estimating equations; IPV = intimate partner violence; OR = odds ratio; CI = confidence intervals.
p < .05. **p < .01. ***p ≤ .001.
To explore risk factors for IPV in the study ethnic and immigrant groups, we conducted separate GEE analysis for each group. We found that low income was associated with higher IPV in all groups. However, the significance of the category of income (which makes a difference in IPV prevalence) was different in each of the three groups. Among nonimmigrant Jewish women, compared with income from work, income from other sources (not social allowances or work only) was associated with higher IPV (OR = 1.64, 95% CI = [1.10, 2.45]). Among Jewish immigrant women, income from social allowances was significantly associated with higher IPV (OR = 4.03, 95% CI = [1.20, 13.56]) compared with income from work. Among Arab women, both income from other sources and from social allowances were associated with almost 2 times higher IPV compared with income from work (OR = 2.06, 95% CI = [1.07, 3.98], and OR = 2.08, 95% CI = [1.00, 4.31]). Also among Arab women, older age groups of 25 to 34 and 35 to 48 years were protected from IPV compared with the younger age group of 16 to 24 years. Religious and traditional women had 2.5 times and 3.5 times higher IPV, respectively. Finally, Arab women living in urban localities were almost twice as likely to report IPV (OR = 1.76, 95% CI = [1.07, 2.89]).
Discussion
This is the first study in Israel that compares prevalence, types, and recurrence of IPV among three groups of women at the childbearing age in Israel (Arab, immigrant, and nonimmigrant Jew) and explores risk factors for IPV within each of these groups. We found marked differences between Arab women and the two groups of Jewish women regarding prevalence, recurrence, and types of IPV. Arab women reported twice the rate of any IPV compared with immigrant and nonimmigrant Jewish women (67%, 30%, and 27%, respectively), while IPV was around 40% for the total study population. These results are in line with research findings from other countries showing higher prevalence of IPV among minority women. For example, higher prevalence has been found in the United States among non-Hispanic Black and Native Americans (Breiding et al., 2014), Black women (Field & Caetano, 2004), and Native American women (Bohn, 2002; Lipsky, Caetano, & Roy-Byrne, 2009); in Canada among Aboriginal women (Daoud et al., 2013; Muhajarine & D’Arcy, 1999); and in New Zealand among Maori women (Fanslow, Silva, Robinson, & Whitehead, 2008).
It is not clear why native ethnic minority women experience higher IPV than majority-group women across settings. In the current study, the multivariate analysis for all women showed that only source of income (not from work) and higher religiosity were significant in the final model. However, even after adjusting for these factors and others in the multivariate analysis, Arab women still had about 4.5 times higher IPV. Some studies in Canada and the United States have found that SEP makes an important contribution to explaining ethnic disparities in IPV (Daoud et al., 2013; Schollenberger et al., 2003). However, in our study, as well as in the Canadian study (Daoud et al., 2013), SEP did not fully explain the ethnic disparity in IPV. Clearly then, IPV among native minorities is multidimensional (Daoud et al., 2013; West, 2004) and can be confounded by the community context (Benson et al., 2004). The ethnic gap in IPV in Israel cannot be understood in isolation from the social class and political position of the Arab minority, or from political violence resulting from the long-standing Palestinian–Israeli conflict (Clark et al., 2010). We know that social and structural factors in the living environment or at the neighborhood level affect IPV (O’Campo et al., 1995). Add to this the impact on IPV and domestic violence of state- and local-level policies to detect and treat community violence and we can see that many dimensions must be considered to fully understand IPV in Israel across ethnic groups.
Religiosity might be tied to the patriarchal sociocultural structure of Arab society (Haj-Yahia, 2000), which might elevate IPV in this community by keeping women low in the social hierarchy. Religiosity was also one of the factors that was significantly associated with higher IPV among Arab women. Our GEE analysis for the separate study groups showed that among Arab women, four main factors predict IPV: younger age, low income, self-definition as traditional or religious, and living in an urban locality. One hint as to why this is lies in research on age and IPV. Younger age was an important risk factor for IPV in previous studies in the United States (Saltzman, Johnson, Gilbert, & Goodwin, 2003) and Canada (Daoud, Uriqua, et al., 2012). The average age of Arab women in Israel at marriage is relatively lower compared with the other two groups of women (Central Bureau of Statistics, 2010). Therefore, we think age might make a significant difference in the association with IPV. Younger women have less life experience that can help protect them from IPV. Therefore, the younger age of Arab women in our sample might make them more vulnerable to IPV compared with the other subgroups.
Our results regarding the lower-than-expected prevalence of IPV among Jewish immigrant women are consistent with a previous study in Israel (Eisikovits et al., 2004), and with other studies showing that IPV is not more prevalent among immigrant women, for example in Canada (Ahmad et al., 2004; Daoud, Uriqua, et al., 2012; Hyman, Forte, Du Mont, Romans, & Cohen, 2006) and in the United States among Chinese women (Hicks, 2006). However, we cannot be certain if these results represent low disclosure, as some previous studies have suggested (Ahmad et al., 2004; Daoud, Uriqua, et al., 2012), or whether they are affected by cultural perceptions of IPV (Eisikovits et al., 2004), as immigration is a high-stress period for families and can increase vulnerability to IPV (Raj, Liu, McCleary-Sills, & Silverman, 2005).
Our finding that women’s education, being pregnant, and number of children at the time of the interview made no significant contribution in the multivariate analysis differs from previous research that shows that these are important risk factors for IPV (Daoud, Uriqua, et al., 2012; Saltzman et al., 2003). However, these factors were associated with IPV in our univariate associations with directions, as per our hypothesis. Marital status has often been found to be a risk factor for IPV (Daoud, Uriqua, et al., 2012; Saltzman et al., 2003). In our study, the proportion of unmarried women was relatively low, and unmarried status was reported only by Jewish women, most of whom are not single mothers, but cohabitate with their partners without being officially married. Women’s education is also an important factor associated with IPV in previous research, but here did not reveal any significant effect. This means that women in different educational groups in Israel are exposed to similar experiences of IPV, and that in the Israeli context, pregnancy and number of children do not affect the likelihood of IPV.
While in Israel national data on IPV relate mainly to extreme cases of women’s homicide and police files, our study shed light on the importance of studying other forms of IPV. A major finding in our study was that women experienced higher emotional (or verbal) and social IPV (28.6% and 26.1%, respectively) than physical IPV (4.6 %). This result clearly shows a need for more study of the variety and copresence of IPV types. Many studies to date have focused primarily on physical or sexual violence when studying IPV (Lown & Vega, 2001; Martin, Mackie, Kupper, Bacchus, & Maracco, 2001; Rachana, Suraiya, Hirsham, Abdulaziz, & Haj, 2002; Saltzman et al., 2003; Yick, 2000). Among our participants, meanwhile, women who disclosed one type of IPV were likely to report recurrence of the problem and to report other types of IPV. For example, most women who reported physical IPV also reported verbal, and social or economic IPV. These results are consistent with some previous research (Mason & O’Rinn, 2014) and might indicate true prevalence.
However, the gap in types of IPV might also reflect women’s perceptions of some acts of violence queried by our survey; that is, women in our study might be less likely to perceive some emotional or verbal and social acts as a form of partner violence (Al-Modallal, 2017; Montalvo-Liendo, 2009), and were therefore more anxious to disclose these compared with physical violence. Women’s attitudes toward and perceptions of IPV are culturally and socially embedded, formulated by socialization processes determined by gender relations and power, especially among women living in a climate of patriarchal ideology (Haj-Yahia, 1998). Previous research in Israel showed that women’s attitudes toward violence were associated with higher prevalence of IPV (Eisikovits et al., 2004; Haj-Yahia, 1999, 2000). It is also possible that women in our study were more likely to disclose emotional, verbal, social, or economic IPV than physical IPV, as the latter could bring legal consequences they wished to avoid.
Strengths and Limitations
Our study’s greatest strength was our sample, which was large, nationwide, and fairly representative of ethnic and immigrant groups of women in Israel. We also built in steps to ensure participants’ privacy; women were interviewed individually in a safe setting. To the best of our knowledge, this is the largest, most extensive recent study on violence against women and IPV in Israel.
The study, however, has some limitations. This research was conducted in MCH clinics, and as such represents women who use MCH services. This should not have introduced a significant selection bias, however, as most women of childbearing age in Israel use these services for well-baby follow-up and immunizations, which are free of charge and geographically accessible for the large majority of women. Another selection bias might have been introduced if a significant portion of women could not be interviewed in either Hebrew or Arabic, but only a few women could not be interviewed due to a language barrier.
It is also possible that women in our study were reluctant to disclose IPV because they feared the consequences of IPV disclosure. Research has shown that fear is the most common factor that prevents women from disclosing IPV, especially women from minority groups (Montalvo-Liendo, 2009). In one Jordanian study, the desire to maintain the family unit and patience toward the abusive partner in the hopes that he would change were the main answers by participants when asked why women would not disclose IPV (Al-Modallal, 2017). Our interviews took place at MCH clinics located close to women’s homes, in their own neighborhoods. Some women may have felt that disclosure of IPV would betray their partners to authorities.
As well, in Israel, a health professional cannot make a report to police based solely on a woman’s disclosure of IPV. Such a report only follows a hospital emergency room visit after a physical attack based on IPV. This means other types of IPV do not meet the standard for police reports and go unreported. For this reason, we think our study sample leads to an underestimation of IPV.
This underestimation could also stem from other factors. While our sample is representative of all women in Israel of childbearing age, as most women in Israel visit the MCH clinics for immunizations and follow-up for their children, the poorest sector of Israeli society is still not included, as this demographic does not visit MCH clinics on a regular basis (Daoud, O’Campo, Anderson, Agbaria, & Shoham-Vardi, 2012; Daoud & Shoham-Vardi, 2015). This means their possible experiences of IPV were left out of the study. As well, we found high prevalence (40%) of IPV in the study total sample, but this finding is still lower than that found in a previous study conducted almost two decades ago on family violence in Israel (Eisikovits et al., 2004). Yet although both studies include a representative sample of women, we cannot conclude that IPV has decreased, as the studies are difficult to compare due to different study settings, measures, and data collection methods. For example, while the previous study used a telephone survey, ours was based on face-to-face interviews.
Conclusion
We believe this work fills a gap in the literature by providing some explanations regarding the inconsistency of the prevalence of IPV among diverse ethnic and immigrant groups of women in Israel. The risk factors we have identified for the different study groups can help inform tailored policies and interventions for both prevention and protection of women from IPV. Future research should delve yet further into the contextual factors—patriarchy, family structures, and the social-political and built environment—that help shape the experiences of IPV of these diverse women.
Footnotes
Acknowledgements
We would like to thank Professor Itamr Grotto, director of the Public Health Division at the Israeli Ministry of Health (MCH); Mira Chimovitch, head nurse at the MCH; the Medical and Nursing directors of Ministry of Health Districts; and nurses at the MCH clinics for the collaboration with the study. Thanks to research assistants Alex Polsky, Areen Agbariah, Nitza Shwartz, and Samira Elfayomi-Zeadna. Many thanks to all the women who agreed to participate in the study.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Israel Science Foundation (ISF) Grant 881/13.