Abstract
The purpose of this study was to examine the perceptions of Palestinian adults toward different dimensions of wife abuse. A cross-sectional survey, using a combination of self-administered questionnaires and interviews, was conducted among a systematic random sample of 624 adult Palestinian men and women from the West Bank and Gaza Strip (18 years or older). Study results indicated a strong tendency to justify wife beating in different situations, such as when the wife is perceived as having an affair with another man or as physically attacking her husband. Participants considered the following acts of husband’s violence against wife as most severe: using a weapon (86%), having sex with the wife against her will (67%), and hitting her with his fist (57%). The majority of participants thought that wife beating should be considered a crime (82.3%). Traditional marital role expectations was the main significant predictor for all of the study criterion variables. Gender, place of residence, age, and marital status were significant predictors of some of the criterion variables.
Keywords
Introduction
Cultural views toward violence against women have become a major concern throughout the world, yet information is still lacking about several dimensions of domestic violence in most Arab societies in spite of significant research accomplishments in the past 10 years in some Arab societies. Beliefs about wife abuse across Arab communities may vary; however, they seem to share common views such as violence against women being acceptable, by some people, in specific situations and violence being considered as a family rather than a society problem (Dhaher, Mikolajczyk, Maxwell, & Krämer, 2010; Djabari, 1998; Haj-Yahia, 1998a, 1998b, 2003, 2005; Zurayk, Sholkamy, Younis, & Khattab, 1997). As in most traditional and patriarchal societies, cultural codes in Arab societies strictly regulate female behavior and sexuality and view men as responsible for imposing such control to protect the family and the society at large (Bates & Rassam, 2001).
Researchers have reported prevalence rates of physical abuse among Arab women between 30% and 53% (El-Zanaty, Hussein, Shawky, Way, & Kishor, 1996; Haj-Yahia, 1999, 2000a, 2000e). In the first and second Palestinian National Surveys (Haj-Yahia, 1999, 2000a, 2000b), 91%, 53%, and 39% of the women reported experiencing psychological abuse, physical violence, and sexual abuse at least once in the past year, respectively. In Jordan, 44.7% of the women reported spousal physical violence during lifetime, 17.4% reported experiencing such violence in the past year, and 14.2% of the women reported experiencing such violence during pregnancy (Khawaja & Barazi, 2005). Likewise, in Egypt, one out of three women reported that they had experienced marital physical violence at least once in their lifetime. Of those women, 45% and 17% had experienced physical violence at least once and at least 3 times in the past year, respectively (El-Zanaty et al., 1996).
The purpose of this study was to examine the perceptions of Palestinian men and women from the West Bank and Gaza Strip toward physical violence against the wife as well as their attitudes toward the severity, harm, and criminality of wife abuse. It is hoped that this study will contribute to increasing our understanding of cultural views toward violence against women in Arab societies and pave the way for more comprehensive studies on different dimensions and issues related to domestic violence in these societies.
Literature Review
Beliefs About Wife Abuse in Arab Societies
A few studies have examined the beliefs of Arab men and women about wife abuse. Available research has shown that nonnegligible portions of the adult population in Arab societies tend to justify wife beating in certain circumstances and to blame battered women for violence against them. Moreover, studies have found that adults in Arab societies show a general tendency to believe that violent husbands are not responsible for their behavior against their wives (Amowitz, Kim, Reis, & Iacopino, 2004; Btoush & Haj-Yahia, 2008; Dhaher et al., 2010; El-Zanaty et al., 1996; Haj-Yahia, 1998a, 1998b, 2003, 2005). For example, whereas about half of the Palestinian men and women indicated that there is no excuse for a man to beat his wife, about 70% indicated that it would be acceptable if the wife is sexually unfaithful. Furthermore, an average of 56% of the men and 41% of the women in Palestinian society believed that wife beating is acceptable if a wife challenges her husband’s manhood or insults him in front of his friends (Haj-Yahia, 1998a, 1998b). Studies have revealed that even though a large majority of women and men (69% and 62%, respectively) agreed that husbands should be held responsible for their violent behavior, they still believed that husbands are not completely responsible for the violence because the wives and the husbands’ daily life circumstances play a major role in causing violence (50% of the women and 62% of the men; Haj-Yahia, 1998a, 1998b). In fact, it was found that 43% of women and 49% of men in Palestinian society believe that wives are responsible for the violence against them.
Studies have shown that justification of violence and blaming the wife for violence against her were significantly associated with lower level of education, higher level of religiosity, holding more traditional views about women, as well as nonegalitarian expectations of marriage and gender roles (Haj-Yahia, 1997, 1998a, 1998b, 2002a, 2003). Justification of wife beating was also found to be significantly higher among residents of rural areas and unemployed women (El-Zanaty et al., 1996). Nonetheless, the tendency to hold the husband responsible for violent behavior is significantly associated with high levels of education and low levels of religiosity as well as with positive, liberal, and egalitarian views about marriage and gender roles. Justifying wife beating and views against holding the husband responsible for violence against the wife have also been significantly associated with lack of positive communication between a couple and witnessing violence in the family of origin (Haj-Yahia, 1997). Studies have shown that among Palestinians and Jordanians, there is very low support for the punishment of violent husbands and for the use of social services and/or police intervention (Btoush & Haj-Yahia, 2008; Haj-Yahia, 1998c, 2002a, 2003). This could be related to the patriarchy oriented cultural view that wife abuse is a personal and family problem rather than a social problem, which should be dealt with inside the family without using formalized interventions from “outsiders” (Haj-Yahia, 2000d, 2000e). In addition, the low tendency to seek assistance from social and/or law enforcement authorities in combating violence against women might also be attributed to mistrust of these services and mistrust of local authorities in general (Haj-Yahia, 2002b).
Finally, according to the sociopolitical structure and cultural context in Arab societies, family violence, specifically against women and children, is usually not considered a crime (Ahmed, 2004; Araji & Carlson, 2001; Haj-Yahia, 2000e, 2002b). In a survey of opinions on the severity of crime scenarios in Israel (Herzog, 2004), Arab participants ranked marital assault as the 13th most serious crime among a list of 14 scenarios, including shop robbery, apartment burglary, false testimony, inflated bill, etc. The only crime scenario considered less serious than marital assault was tax evasion. These studies reflect a societal view that clearly does not consider wife beating a crime or at least not a serious crime.
Perspectives of Wife Abuse in Non-Western Societies
Studies on wife abuse in several non-Western patriarchal societies have revealed perspectives that are similar to those prevailing in Arab societies. Surveys of several countries in Asia (Rani & Bonu, 2009) have revealed that rates of acceptance of wife beating among women were 32% in Armenia, 30% in Kazakhstan, 49% in Turkey, 29% in Nepal, 56% in India, and 35% in Cambodia. By comparison, the rates of acceptance of wife beating among men were 42% in Armenia, 27% in Kazakhstan, 56% in Turkey, 34% in Nepal, and 37% in Bangladesh. Moreover, acceptance of wife beating was found to be associated with income, level of education, age, and employment status. In Armenia, Nepal, and Turkey, men were significantly more likely to justify wife beating. In a different study, Turkish women reported the following reasons for wife beating: failure to fulfill domestic duties, failure to fulfill husband’s sexual will, the wife wanting to work, and the husband’s own social problems (Kocacik, Kutlar, & Erselcan, 2007).
In Singapore (Choi & Edleson, 1996), 6% of the men and women participating in a survey agreed that “sometimes it is all right for a husband to use physical force on his wife,” and 7% of them did not consider wife assault to be a crime. Accepting wife beating was significantly associated with being unmarried, living in middle- and low-status housing, and among participants who tended to rate various actions by a husband as reflecting less use of force (lower scores on the perception of force index). They considered the following acts to be wife assault: using a weapon (97%), hitting with a fist (94%), banging her head against the wall (89%), kicking her (89%), having sex with the wife against her will (74%), slapping her (70%), or pushing/shoving her (67%). In the same study, participants were more accepting of wife beating if the wife abuses their children (33%) or if she is having an affair (25%). Accepting wife beating in these two situations were mostly associated with younger age, being unmarried, and with lower scores on the perception of force index (Choi & Edleson, 1996).
Finally, in a study of Turkish and Brazilian men and women (Glick, Sakalli-Ugurlu, Ferreira, & de Souza, 2002), men were significantly more likely to justify wife beating. The findings of that study revealed that legitimizing and accepting violence was significantly associated with views that reflect hostile sexism. The approach of Arab societies toward wife abuse shares several similarities with most traditional and patriarchal societies, such as some of those mentioned above, especially in accepting and justifying wife abuse.
Conceptual Model
The present study was based on an integrative conceptual framework that combines predictive factors derived from the patriarchal ideology perspective (e.g., DeKeseredy & Dragiewicz, 2007; Dobash & Dobash, 1979, 1992; Moghadam, 1992) and social learning theory (e.g., Bandura, 1977; Carter, Stacey, & Shupe, 1988; Kalmuss, 1984; O’Leary, 1988), along with sociodemographic characteristics. Researchers have used the patriarchal ideology perspective to explain beliefs about wife beating and actual wife abuse and battering (e.g., Dobash & Dobash, 1992; Haj-Yahia, 2003, 2005). Patriarchy includes hierarchical social structures and organized social relations. Patriarchal ideology reinforces the continuity and acceptance of patriarchal structures, which support gender inequality and the male dominance perspective. Children in patriarchal societies are raised according to gender-role stereotypes. According to this perspective, violence against wives is caused by the subordinate position of women in relation to men (Dobash & Dobash, 1979). In a patriarchal society, social relationships routinely reinforce the dominant status of men and sanction violence against women to maintain the status of men. This ideology is embedded throughout the society, including its religious, political, and economic systems (Moghadam, 1992).
The idea of intergenerational transmission of violence originated with Albert Bandura’s social learning theory (1977). According to that theory, exposure to child abuse or interparental violence during childhood increases the likelihood and acceptance of aggression during adulthood, although it should be highlighted that not all people who had been exposed to family violence as children grow-up to be abusive and violent during adulthood (Haj-Yahia & Uysal, 2008; O’Leary, 1988). According to this theory, individuals learn and modify new behaviors through process of modeling. Reinforcement serves to strengthen new behaviors vicariously by observing the consequences of others’ behaviors as well as by direct experience. According to this theory, adults tend to be more violent and/or accepting of violence if, as children, they were exposed to violence within the family (either against them or between their parents).
On the basis of these theories, our integrative theoretical framework in this study includes egalitarian versus patriarchal expectations of marriage, as a variable that represents patriarchal ideology, and witnessing interparental violence and experiencing parental violence as variables that represent social learning theory. This study examines the extent to which views about wife abuse are influenced by patriarchal ideology and/or exposure to family violence.
The Purpose and Questions of the Study
The purpose of this study was to examine the perceptions of Palestinian men and women from the West Bank and Gaza Strip toward physical violence against the wife, its severity and harm, and criminality of wife abuse. Predictor variables were derived from the patriarchal ideology perspective (e.g., Dobash & Dobash, 1979, 1992; Moghadam, 1992) and social learning theory (e.g., Bandura, 1977; Carter et al., 1988; Kalmuss, 1984; O’Leary, 1988), along with sociodemographic characteristics. The study examined the participants’ perceptions about justification of physical violence against women, severity of violence, degree of harm of body injuries, and criminality of physical violence, specific acts of violence, and specific body injuries. Accordingly, the study also examined possible relationships between the aforementioned perceptions, as criterion variables, and marital role expectations, knowing a woman who has been physically abused by her husband, witnessing father-to-mother violence, and experiencing parental violence, as the main predictor variables of the study. In essence, it was hypothesized that patriarchal marital role expectations, witnessing father-to-mother violence, experiencing family violence during childhood and adolescence, and knowing a woman who has been physically abused by her husband would be associated with increased justification of wife abuse, a lower tendency to perceive specific acts of wife abuse as severe, and a lower tendency to consider specific body injuries as causes of severe harm to the abused wife. In addition, it was hypothesized that patriarchal marital role expectations, witnessing father-to-mother violence, experiencing family violence during childhood and adolescence, and knowing a woman who had been physically abused by her husband would be associated with a lower tendency to perceive wife abuse as well as specific acts of physical violence against wives as a crime and a lower tendency to perceive specific body injuries that result from wife abuse as a criminal problem. The relationship between the participants’ sociodemographic variables (i.e., age, gender, marital status, place of residence, religion, income, and level of education) and each one of the above-mentioned criterion variables were also examined.
Method
Participants and Procedures
A systematic random sample of 624 adult Palestinian men and women from the West Bank and Gaza Strip (18 years or older) were drawn for the study. The participants were recruited from a sample of three cities, four villages, and two refugee camps in the West Bank and one city, one village, and one refugee camp in the Gaza Strip. Based on consultation with community figures in the localities that were chosen to take part in this study, cities, villages, and refugee camps were divided into three sections: central, intermediate (i.e., between center and periphery), and periphery. In every city, three streets were chosen (one from each section). In every village and refugee camp, one street was chosen. Female research assistants were recruited from universities in these regions, and they were trained on several matters, including sampling, interviewing skills, cultural sensitivity in talking and interviewing about sensitive matters such as domestic violence, and dealing delicately with some aspects related to the study, such as people who refuse to participate in the study and husbands who expressed interest in being present while the wife is filling out the questionnaire or interviewed, and so on. With regard to sampling, research assistants were instructed to select one home at random on both sides of every street for the sample and then were instructed to enter every 10th home on the list they chose and ask to speak with an adult aged 18 or older. In their conversations with the adult household members, the research assistants introduced themselves, gave a general description of the research aims, and asked the person whether he or she would like to participate. Afterwards, they asked potential participants whether they knew how to read and write. To ensure the generalizability of the sample, research assistants were instructed to interview about 7% to 15% of the illiterate participants. Furthermore, due to the sensitivity of the problem, it was assumed that self-administered questionnaires would provide a more reliable indication of the participants’ attitudes toward wife abuse. Therefore, only 9% of the participants were interviewed by research assistants at mutually convenient times. The interview was conducted privately in one of the rooms of the house, which was chosen by the participant. Hence, husbands who were present in the home during the interview were asked not to be in the same room where the interview was conducted. The final response rate was 91% (N = 624).
Participants who indicated that they can read and write were given the study questionnaire with an envelope and were assured that the information would remain completely confidential. Research assistants informed the participants that they would return within a week to collect the completed questionnaire. In the event that the questionnaire was not completed, the research assistant returned to the home 3 days later. The research assistants were instructed not to pressure participants who failed to complete the questionnaire after the second round of collection. Participants were instructed to complete the questionnaire independently and privately at their convenience. However, due to the small size of the apartments or houses of some participants, about 24% (n = 104) of the married participants stated that their spouse was with them in the same room when they completed the questionnaire, and 88% of them thought that the spouse’s presence did not influence their answers. All participants who revealed that the presence of their spouse influenced their answers (n = 12) indicated that the influence was only in explaining some unclear words, items, or questions. In response to the research assistant’s question, “Was your spouse involved in choosing the answers?,” none of the participants indicated that their spouse was involved in choosing the answers or in anything else. It should be noted that this question was presented to the participants when they were alone (i.e., the spouse was not present). The first author conducted weekly meetings with all research assistants for the purpose of debriefing as well as for supervision in dealing with any personal, methodological, cultural, or sociopolitical issues or concerns that they encountered during the process of data collection. None of the research assistants indicated in those meetings that she felt worried about her personal safety during any phase of data collection.
The study was approved a priori by the ethics committee at the academic institution of the first author, and was also approved retrospectively by the IRB (institutional review board) of the second author’s academic institution. The study was conducted during the last trimester of 2001.
Instruments and Measurement
The study questionnaire was prepared in three stages. In the first stage, a mixed Arabic and English version of the questionnaire was prepared. The Arabic part included description of the purpose of the study in the cover page of the questionnaire, and at the introduction of each one of the parts of the questionnaire as well as an explanation of their confidentiality and voluntary participation. It also included questions about several sociodemographic characteristics (e.g., participant’s age, gender, religion, marital status, occupation, level of education, average monthly income in New Israeli Shekel, and type of locality, that is, village, city, or refugee camp, etc.), two questions about exposure to family violence, and several open-ended questions. Furthermore, the Arabic version included a question in which the participants indicated whether their spouse was present when they completed the questionnaire and whether the presence of their spouse affected their answers. The English part of the questionnaire included several measures that were taken from Choi and Edleson (1995, 1996) as well as a short version of the Marital Role Expectations Inventory (Dunn, 1960; Dunn & DeBonis, 1979), as will be explained later. To ensure valid and robust procedures in translating and adapting the English part of the questionnaire, it was translated from English into Arabic by four Palestinian professionals fluent in English, and who specialize in the social, behavioral, and mental health sciences. Each of them was asked to translate the English part of the questionnaire into Arabic, taking into consideration the cultural context of the Palestinian society, and then they worked together with the first author to prepare a unified Arabic version of all four translations of that part of the questionnaire. To maintain comparability of the questionnaire in both languages, the unified Arabic version of the questionnaire was back-translated into English by a professional translator. Afterwards, the final English version of the questionnaire was submitted to the ethics committee at the institute of the first author, together with the planned procedures for choosing the sample and collecting the data, in order to obtain ethical clearance for conducting the study. Herein we refer only to those parts of the questionnaire that are relevant to this article (in addition to the sociodemographic questions that were indicated earlier).
Marital role expectations
This was one of the predictor variables of the study. It was measured using a short form (nine items) of the Marital Role Expectations Inventory, developed by Dunn in 1960 and revised by Dunn and DeBonis in 1979 (Dunn, 1960; Dunn & DeBonis, 1979). This instrument measures marital role expectations and behavior orientations (companionship-egalitarian vs. traditional-patriarchal) in eight areas of marriage. Examples of statements in this instrument are as follows: “When there is a disagreement of opinion between a husband and wife, the outcome should be the husband’s view and decision,” and “The husband has the primary responsibility to provide for the family’s financial needs.” The Spearman–Brown reliability coefficient for this measure has been reported at .975 on a split-half correlation analysis (Dunn & DeBonis, 1979). The Cronbach’s alpha coefficient of the Arabic version has been reported at .86 (Haj-Yahia, 2003), and the internal reliability of the short form used in this study was .96. Responses to all nine items were based on a 4-point Likert-type scale ranging from 1 (strongly agree) to 4 (strongly disagree). Higher scores on this instrument indicated more egalitarian and liberal expectations of marriage, whereas lower scores indicated more traditional and patriarchal expectations. A score was generated by summing up the responses for all 9 items and then dividing the result by 9.
Exposure to family violence
Participants were asked about their exposure to family violence using two questions, which were used as two additional predictor variables in this study. The first question asked whether the participants personally experienced family violence by their parents before they were 18 years old (i.e., their childhood and adolescence). The second question asked whether they witnessed their father hitting their mother during that period of time. Both questions were developed for this study. Participants’ responses were obtained using the following scale: 1 (never), 2 (a few times), and 3 (many times)
Knowing a woman who was physically abused by her husband
Participants were asked whether they knew a woman who had been physically abused by her husband, which is another predictor variable in this study. Participants’ responses were obtained using a dichotomous scale, 1 (yes), 2 (no).
Perceived justification of physical violence against women
Perceived justification of husband-to-wife physical violence, a criterion variable in this study, was measured using eight scenarios, which are based on the work of Choi and Edleson (1995, 1996). Instrument items were scored on a 4-point Likert-type scale with response options ranging from 1 (strongly agree) to 4 (strongly disagree), and as shown in Figure 1. The Cronbach’s alpha coefficient for the Arabic version of this instrument used in this study was .94.
Perceived justification for physical violence (ranked)
Perceived severity of violence against women
This is also a criterion variable in this study, and for this item, the participants were asked about their perceptions of the severity of eight acts of husband-to-wife physical violence, based on the work of Choi and Edleson (1995, 1996). Items were scored on a 3-point scale ranging from 1 (mild) to 3 (severe), as shown in Figure 2. The Cronbach’s alpha coefficient for the Arabic version of this instrument used in this study was .99.
Perceived severity of specific acts of violence (ranked from mild to severe)
Perceived degree of harm of body injuries
This item was used as a criterion variable, and in this regard, the participants were asked to estimate the degree of harm for eight body injuries related to husband-to-wife physical violence, based on the work of Choi and Edleson (1995, 1996). Items were scored on a scale ranging from 1 (minimal harm) to 3 (severe harm), and as shown in Figure 3. The Cronbach’s alpha coefficient for the Arabic version of this instrument as used in this study is .994.
Perceived degree of harm for specific body injuries (ranked)
Perceived criminality of physical violence
Another criterion variable, this item was assessed by asking participants whether they personally considered the husband’s physical violence against the wife a crime in general, with three response options: yes, no, or not sure. Participants were asked the same question more specifically, using yes/no responses for the eight acts of physical violence (Figures 2 and 4) and for the eight body injuries related to physical violence mentioned above (Figures 3 and 5). In this study, the Cronbach’s alpha coefficients were .98 for the eight acts of physical violence and .99 for the eight body injuries related to physical violence. A score for each one of the criterion variables was produced by summing up the responses on all eight items that measure that variable and then divided by the number of items.
Perceived criminality of specific acts of violence (ranked)
Perceived criminality of specific body injuries (ranked)
Data Analysis
In this study, we used the SPSS software to conduct all of the needed analyses. Initially, we conducted descriptive analysis to determine ranked proportions, using frequencies and Freidman’s test (see Figures 1, 2, 3, 4, and 5) for each of the above-mentioned perceptions of wife abuse, as the criterion variables of the study. Friedman’s test was conducted to examining whether there were significant differences in ranking items or acts that were used in measuring the above-mentioned perceptions. In addition, we used regression and multiple regression analyses to examine associations among predictor and criterion variables in multivariate models. For that purpose, hierarchical regression and multiple regression analyses were conducted for each of the perceptions of wife abuse. To successfully carry out these analyses, first the sociodemographic characteristics were entered into each one of the six regression formulas; after that, the main predictor variables were entered into each formula (see Tables 1 and 2; Kachigan, 1986). However, religion was dropped from the regression analysis because a very small portion of the study participants were Christian (6.7%) and overwhelming majority were Muslims (93.3%). Tests of multicollinearity were conducted and were within acceptable parameters (Agresti & Finlay, 2009). The variance inflation factor (VIF) values ranged between 1.043 and 1.535.
Regression and Multiple regression Analyses on Perceived Justification of Violence, Perceived Severity of Specific Acts of Violence, and Perceived Degree of Harm for Specific Body Injuries
Note: N = 624. GEN = gender; AGE = participants’ age; MAS = marital status; EDU = level of education; INC = income; RES = residence; WRK = work status; FHM = witnessing father-to-mother violence; EFV = experiencing parental violence; KNW = knowing a physically abused woman; MRE = marital role expectations.
Statistically significant at p < .05
Regression and Multiple Regression Analyses on Perceived Criminality of Physical Violence, Perceived Criminality of Specific Acts of Violence, and Perceived Criminality of Specific Body Injuries
Note: N = 624. GEN = gender; AGE = participants’ age; MAS = marital status; EDU = level of education; INC = income; RES = residence; WRK = work status; FHM = witnessing father-to-mother violence; EFV = experiencing parental violence; KNW = knowing a physically abused woman; MRE = marital role expectations.
Statistically significant at p < .05
Results
This study was conducted among a sample of 624 Palestinian men and women from the West Bank and Gaza Strip, of which 53.8% were men and 46.2% were women, between the ages of 18 and 76 years (M = 33, SD = 10.97). The majority of participants were married or ever married (70%) and Muslims (93.3%). The residence distribution was 49% urban, 36% rural, and 15% refugee camps. Half of the participants had community college or university education, 33% had high school education, and 17% had less than high school education. Only 49% of participants were employed at the time of the study (67% of the men and 27% of the women were employed). The average monthly income of the participants was about NIS 2,406 (SD = 1,400), which is equivalent to about US$680, at the year of conducting the study.
About half of the participants (51.4%) knew a woman who had been physically abused and battered by her husband. In addition, 42% of the participants reported experiencing parental physical violence a few times, whereas 3% reported experiencing parental physical violence many times during childhood and adolescence. Furthermore, 28% and 3% reported witnessing their father hitting their mother a few times and many times, respectively, during that period of time.
Perceived Justification of Physical Violence Against Women
As shown in Figure 1, husband’s physical violence toward his wife is considered justified when she is perceived having an affair with another man (65%), when she physically attacks her husband (60%), if she always reminds the husband of his mistakes or weaknesses (36%), when she is perceived as a nagger or a complainer (38%), and if she refuses to have sex with the husband (31%). Figure 1 shows the ranking and mean ranks for the participants’ responses to this questions using Freidman’s test (χ2 = 987.311, df = 7, Asymp. Sig.= .001).
The results of regression and multiple regression analyses in Table 1 show that 47.5% of the variance in justifying the use of husbands’ use of physical violence against wives can be attributed to our conceptual model. Justifying the use of husbands’ use of physical violence against wives was significantly associated with marital status (β = .067, p = .032), knowing a physically abused woman (β = –.067, p = .028), and marital role expectations (β = .677, p = .001). In other words, those who are ever married, do not know a physically abused woman, and hold more nonegalitarian and patriarchal expectations of marriage had significantly higher tendency to justify the use of husbands’ use of physical violence against wives. Gender, age, educational level, income, residence, work status, and experiencing and witnessing of family violence did not contribute significantly toward explaining this perception.
Perceived Severity of Violence Against Women
As shown in Figure 2, study participants ranked the following acts of husband-to-wife physical violence as the most severe: using a weapon (86%), having sex with the wife against her wishes (67%), hitting her with his fist (57%), shoving against the wall (46%), and kicking (45%). An alarming finding here is that a nonnegligible portion of participants considered these violent behaviors as mild acts of violence (11%, 14%, 15%, 27%, and 18%, respectively). Figure 2 shows the ranking and mean ranks for the participants’ responses to this questions using Freidman’s test (χ2 = 1026.476, df = 7, Asymp. Sig. = .001).
The results of regression analyses in Table 1 show that 11.8% of the variance in participants’ perceptions of severity of violence against wives can be attributed to our conceptual model, where in particular this criterion variable was significantly associated with participants’ age (β = .128, p = .001), marital status (β = .119, p = .004), and marital role expectations (β = .236, p = .001). In other words, older and ever married participants as well as those with more nonegalitarian and patriarchal expectations of marriage showed a significantly higher tendency to consider the above-mentioned acts of husband-to-wife physical violence as less severe. Gender, education level, income, residence, work status, experiencing and witnessing family violence, and knowing a physically abused woman did not contribute significantly to explaining this perception (Table 1).
Perceived Degree of Harm of Body Injuries
As shown in Figure 3, the overwhelming majority of participants considered the following body injuries as causes of severe harm to the wife: skull fracture (97%), skin cut from a knife (94%), broken nose (91%), broken ribs (87%), broken arm (85%), and black eye (74%). Figure 3 shows the ranking and mean ranks for the participants’ responses to this questions using Freidman’s test (χ2 = 1226.025, df = 7, Asymp. Sig. = .001).
The results of regression analyses in Table 1 show that 16.0% of the variance in participants’ perceptions of the degree of harm of body injury caused to women as a result of violence against them can be attributed to our conceptual model, where this criterion variable was significantly explained only by participants’ marital role expectations (β = .377, p = .001). In other words, participants with more nonegalitarian and patriarchal expectations of marriage showed a significantly lower tendency to consider the above-mentioned body injuries as causes of severe harm to the wife. Other predictors in the conceptual framework of the study did not contribute significantly to explaining this perception (Table 1).
Perceived Criminality of Physical Violence, Specific Acts of Violence, and Specific Body Injuries
When participants were asked whether or not they personally consider physical violence against the wife a crime, 82.3% of them said yes, 7.5% said no, and 10.2% were not sure. As shown in Figure 4, the majority of participants indicated that the following acts of husband-to-wife physical violence should be considered a crime: husband’s use of a weapon against his wife (89%), hitting her with his fist (75%), having sex with the wife against her will (75%), kicking her (72%), and shoving her against the wall (70%). Figure 4 shows the ranking and mean ranks for the participants’ responses to this question using Freidman’s test (χ2 = 693.986, df = 7, Asymp. Sig. = .001).
As shown in Figure 5, the overwhelming majority of participants thought that the following body injuries from physical violence should be considered a crime: skull fracture (94%), skin cut from a knife (92%), broken ribs (92%), broken nose (90%), broken arm (90%), black eye (83%), and bruises (79%). Figure 5 shows the ranking and mean ranks for the participants’ responses to this questions using Freidman’s test (χ2 = 522.955, df = 7, Asymp. Sig. = .001).
Table 2 presents the results of regression and multiple regression analyses for the variables in this section. In the first model, the results of the regression analyses show that 3.3% of the variance in the participants’ perceived criminality of physical violence can be attributed to our conceptual model, where this criterion variable can be significantly explained by marital status and participants’ marital role expectations (β = .083, p = .050, and β = .082, p = .046, respectively). In other words, participants who were ever married and had more nonegalitarian and patriarchal expectations of marriage showed a significantly lower tendency to consider physical violence against wife a crime. Other study variables did not contribute significantly toward explaining this perception. The results also revealed that 18.3% of the variance in the participants’ perceived criminality of specific acts of physical violence can be significantly explained by their age (β = .089, p = .021), marital status (β = .174, p = .001), place of residence (β = .128, p = .001), and marital role expectations (β = .315, p = .001). In other words, participants who are older, ever married, residents of rural areas, and hold more nonegalitarian and patriarchal expectations of marriage had a significantly lower tendency to consider specific acts of physical violence against the wife (listed in Figure 4) as a crime. Other study variables did not contribute significantly toward explaining this perception. Finally, the results revealed that 12.2% of the variance in the participants’ perception of criminality of specific body injuries can be significantly explained by place of residence and marital role expectations (β = .079, p = .044, and β = .311, p = .001, respectively). In other words, participants who live in rural areas and those who hold more nonegalitarian and patriarchal expectations of marriage had a significantly lower tendency to consider specific body injuries (listed in Figure 5) as a crime. Other predictors in the study conceptual framework did not contribute significantly toward explaining this perception.
Discussion
This study examined the perceptions of Palestinian men and women from the West Bank and Gaza Strip about justification of physical violence against the wife as well as perceptions of the severity of specific acts of violence, the degree of harm caused by specific body injuries from physical violence, and the criminality of physical violence in general and as related to specific acts of physical violence and the criminality of specific body injuries from physical violence. The study findings reveal a significant tendency to justify physical abuse against the wife, especially in situations that are considered to conflict with cultural beliefs about the role of women in marital relationships (e.g., having an affair with another man, physically attacking the husband, being a “nagger” or a “complainer,” reminding the husband of his mistakes or weaknesses, or refusing to have sex with the husband).
These findings are consistent with studies conducted among Arab men and women in Palestine, Jordan, and Israel (Haj-Yahia, 1998a, 1998b, 2002a, 2003; Khawaja, Linos, & El-Roueiheb, 2008). The participants’ perceived justification for physical violence against the wife is associated with the patriarchal orientation of Arab culture, especially related to family structure, marital relationships, and rigid gender roles. Despite recent changes in Arab societies, “this culture still perpetuates the husband’s superiority and dominance, whereas the wife remains (and is expected to be) inferior and submissive” (Haj-Yahia, 2003, p. 203). Therefore, women are expected not to contest these clearly defined gender roles and to accommodate the marital relationship through excessive selflessness and sacrifice, which makes them emotionally and personally vulnerable.
In this study, there was a significant association between higher tendency to justify physical violence and marital status, knowing a physically abused woman, and holding more nonegalitarian and patriarchal expectations of marriage and the role of men and women in a marital relationship but not between other variables in the current study’s conceptual framework. The significance of views about and expectations of marriage in predicting attitudes toward wife abuse is consistent with other studies (Haj-Yahia, 1998a, 1998b). However, the results of this study revealed a nonsignificant association between participants’ sociodemographic characteristics and their tendency to justify violence against women. This is in contrast to the results of previous research that reported significant associations between participants’ age, gender, area of residence, level of education, employment, and religiosity and their willingness to justify wife beating (Haj-Yahia, 1998a, 1998b, 2002a, 2003, 2005).
Alarmingly in this study, a small but significant portion of participants considered using a weapon, having sex with the wife against her will, punching, shoving, and kicking as mild acts of violence. Similarly, a small but significant portion of the participants in the study thought that these acts should not be considered a crime. The serious damage that these beliefs can cause to women has been highlighted in studies that have documented the significant consequences of wife abuse and beating not only in the form of body injuries but also in terms of depression, anxiety, stress, and low-self esteem (Campbell, 2002; Haj-Yahia, 2000a, 2000c) and in terms of negative implications for their marital relations (Haj-Yahia, 2002c). These beliefs can also cause serious damage to children, who may witness their father hit their mother and/or experience abuse and violence from parents and/or siblings (Maker & deRoon-Cassini, 2007).
Not surprisingly, the participants’ nonegalitarian marital role expectations was the most significant predictor of the criterion variables examined in the study (e.g., the participants’ perceptions about justification of physical violence, perceptions of the severity of specific acts of violence, perceptions of the degree of harm caused by specific body injuries, as well as overall perceptions of the criminality of physical violence and of specific acts of violence and body injuries). Accordingly, these results are consistent with the patriarchal ideology perspective (Dobash & Dobash, 1979, 1992). Similar to other studies (Haj-Yahia, 2003, 2005; Haj-Yahia & Uysal, 2008), this study reaffirms the strong connection between a more accepting view about violence against women and nonegalitarian and patriarchal expectations of marriage and traditional attitudes toward the role of men and women in marital relationships. Hence, there is a need for more comprehensive studies on the correlation between other variables that derive from patriarchal ideology (e.g., attitudes toward women, sex-role stereotypes, chauvinist attitudes, attitudes toward the status of women in the private and public spheres, etc.) and perceptions about wife abuse.
However, the results of the study did not establish the role of social learning theory (Bandura, 1977; Carter et al., 1988; Kalmuss, 1984; O’Leary, 1988) in explaining views about violence against women, contrary to what other studies have documented (Haj-Yahia & Uysal, 2008). Our results reaffirm Haj-Yahia’s (1997) conclusion that although social learning theory may contribute to views about violence against women, it may not be as strong as patriarchal ideology in explaining beliefs about wife beating. A possible explanation for the lack of support for social learning theory in explaining views about wife abuse is that we used only one question asking about experiencing family violence and one about witnessing the father hitting the mother. Another explanation for the lack of support for social learning theory is recall bias, where participants have forgotten or repressed some of the violence they experienced as children.
We believe the study results might have been different if we had used a more detailed questionnaire about exposure to family violence, as will be discussed later. Hence, we strongly recommend that future studies include more detailed questions about experiencing and witnessing different patterns and types of violence by father, mother, and siblings. Furthermore, we recommend conducting studies that examine why some people do not learn violent behaviors and perceptions that justify wife abuse despite their exposure to family violence. In accordance with the social learning theory, variables such as self-efficacy (Bandura, 1977) and some resiliency factors as well as personal, familial, and social protective factors may play significant roles in inhibiting that learning.
Regarding the strengths and limitations of the study, a systematic random sample of participants was used to ensure accurate representation of Palestinian men and women in terms of religious affiliation, place of residence, level of education, age, and so on. However, because the sample included only a small number of residents of refugee camps, we were unable to compare residents from all types of localities. Therefore, we recommend that similar studies in the future plan targeted recruitment to ensure an adequate number of participants in each type of locality are screened.
The present study dealt with topics that are sensitive in many Arab societies, including Palestinian society. Therefore, there is a risk that the participants’ responses were affected by a certain tendency toward social desirability, which detracts from the validity of the results; however, we do not know how strong that tendency was. We made efforts to minimize the effect of social desirability by using self-administered questionnaires and maintaining anonymity by not asking for names or any other identifiers. Nonetheless, we cannot ascertain the extent to which the use of self-administered questionnaires enabled us to completely overcome this tendency. Moreover, 24% of the participants (n = 104) indicated that their spouse was present while filling out the questionnaire, and 12% of them (n = 12) indicated that the presence of their spouse only helped them to understand some unclear words, items, or questions. All of them indicated that the presence of their spouse did not affect their answers to any of the questions or items. Yet, the very fact that some spouses were present while filling out the questionnaire could have the potential for implicitly influencing some of the participants’ answers. This source of potential confounding needs to be better controlled in future studies. In addition, to ensure the generalizability of the results, we included illiterate participants (9%) who were interviewed by research assistants. By limiting the number of illiterate participants to 9%, we attempted to achieve a balance between ensuring generalizability on the one hand and minimal effect of social desirability related to the use of interviewing rather than self-administered questionnaires on the other. This limitation can be minimized in future studies by including a social desirability measure in the study instrument.
Furthermore, a major limitation of the study is related to the measurement of experiencing parental violence and witnessing father-to-mother violence, where each one of these two major predictor variables in this study was measured by only one question with just three possible response options. Undoubtedly, this approach to measurement cannot ensure comprehensive probing of participants’ exposure to (i.e., witnessing and experiencing of violence) different patterns, types, severities, and frequencies of abuse, assault, aggression, and violence in the family. It should also be noted that because people have different perceptions and definitions of abuse, which is the word used in each of the two questions mentioned above, it is possible that different participants reported about exposure to different types and patterns of emotional, verbal, physical, and sexual abuse, and thus, we could never capture which type of abuse they reported.
In addition, participants were asked about their exposure to abuse in their families of origin before the age of 18, when this age group was considered as one life stage, without measuring the extent of exposure to abuse during different periods of childhood and adolescence. Furthermore, participants were not asked about the context in which they were exposed to abuse, nor were they asked how were their interpretations of exposure to abuse during childhood and adolescence compared with their current interpretations. Hence, the relevance of the historical interpretations of exposure to family abuse and violence to the participants’ perceptions about wife abuse in the past and present was not examined in this study. Consequently, the participants may have forgotten or repressed some of the abuse and violence that they witnessed and/or experienced during childhood and/or adolescence, and their reports may have been partial. This may have had an impact on the strength of the relationship between their exposure to different types and patterns of abuse and their perceptions of wife abuse. Future research on the topic can take these limitations into account. This can be accomplished by utilizing comprehensive measures of different frequencies, types, acts, and patterns of abuse, assault, violence, and aggression in future research. Moreover, it can be achieved by inquiring about the contexts in which the participants were exposed to these types and patterns of violence and by asking them about their past and present interpretations of the different types and patterns of violence to which they were exposed.
Finally, to measuring marital role expectations this study relied on a scale developed in a Western, postindustrial society. Although this instrument was culturally adapted and validated in Arab society, further research is needed to establish the reliability of this instrument as well as other measures that were used in this study to measure perceptions of justifying physical violence against women, the severity of violence against women, the degree of harm caused by body injuries, and the criminality of physical violence.
We also recommend that future studies examine the extent to which attitudes toward wife abuse are learned through other sources such as the media and community exposure. Another source, the influence of religion, should be assessed in terms of its degree of influence by measuring religiosity level instead of religious affiliation in itself. Moreover, general worldviews, family values, sex-role stereotypes, and conflict resolution approaches may be of value in understanding views toward family violence and wife abuse. This brings us to the need to understand cultural views about wife beating and abuse in a more holistic and integrative way by incorporating other theories such as family systems theory, family stress theory, personality and intrapersonal theory, and social exchange theory.
The present study as well as other studies on violence against women in Arab societies can contribute to enhancing understanding of this phenomenon in a unique culture characterized by widely varying and, sometimes conflicting, political, sociocultural, economic, religious, and educational contexts. The results of this study have important implications for prevention and intervention strategies in Arab societies, including community campaigns to counteract the social acceptability and approval of violence against women, increasing awareness of the emotional and physical consequences of violence against women, beliefs and attitudes that legitimize male dominance, and traditional attitudes toward women and marriage.
Footnotes
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The authors received the following financial support for the research, authorship, and/or publication of this article: The study was partially funded by the Ford Foundation, Cairo Branch, and conducted through the Bisan Center for Research and Development, Ramallah, Palestinian Authority.