Introduction

Violence against women is a health, mental health, social, and criminal problem. Because health practitioners are often on the front lines for identifying and assessing battered women, it is often assumed that they can be the first ones to provide assistance to these women. In essence, when health practitioners encounter women who are victims of intimate partner violence (IPV), they are expected to perform several tasks in cooperation with other health and mental health practitioners. These tasks include conducting routine screenings and inquiring about IPV, as well as about other types of abuse and neglect in the family (Gerbert et al. 2002; Haj-Yahia 2010, 2013; Hamberger and Patel 2004).

In addition, health practitioners are expected to perform the following tasks: documenting cases of IPV; maintaining the records; and referring women who are victims of IPV to other physical and mental health practitioners. They are also expected to refer these women to shelters, to officials in the criminal justice system, and to community resources such as women’s advocacy and domestic violence organizations. Health practitioners are also expected to be involved in formulating a follow-up plan and future visits that may include coordination with other agents in the community (Gerbert et al. 2002; Haj-Yahia 2010, 2013; Hamberger and Patel 2004).

Studies conducted in recent decades, mainly in Western countries, have revealed that health practitioners are becoming increasingly aware of various types and patterns of family violence, including IPV. As such, they have been intensifying their involvement in providing assistance to victims of IPV, including battered women. However, the research evidence indicates that, in some cases, health practitioners are still reluctant to perform these tasks and functions, and that they sometimes refrain from intervening in cases of IPV (Bair-Merritt et al. 2006; Colarossi et al. 2010; Haj-Yahia 2010, 2013; Hamberger and Patel 2004; John et al. 2011; Minsky-Kelly et al. 2005).

For example, a study conducted among health practitioners in outpatient settings revealed that only 5 % of the pediatricians and 8 % of the family practitioners examined in the study had routinely screened a patient for different patterns of family violence, mainly IPV (Borowsky and Ireland 2002). In another study, the physicians reported that they had screened fewer new patients for violence against women (19 %) than for other risk behaviors such as tobacco use (98 %), alcohol abuse (90 %), or HIV/STD risks (47 %). Furthermore, only 13 % of the physicians reported that they asked regular or returning patients about IPV, compared with 82, 61, and 27 % who reported that they asked regular or returning patients about tobacco use, alcohol abuse, and HIV/STD risks, respectively (Gerbert et al. 2002).

There is a growing body of research on health practitioners’ attitudes toward the problem of violence against women in general, and toward their involvement in screening, identifying, assessing, and intervening in cases of IPV in particular. Findings of these studies have brought to light problems regarding the failure of health care providers to treat the problem of domestic violence seriously. For instance, one study conducted in the US revealed that even when treating injured patients, 45.2 % of the clinicians in the sample had seldom or never assessed domestic violence (Sugg et al. 1999). In a study conducted among surgery residents in Canada, only about 10 % of the participants reported that their clinical setting has procedures in place for assessing and responding to IPV (Sprague et al. 2013). In addition, 18 % of them believed that victims “must get something from the abusive relationship”, and 11 % believed that “victims chose to be victims” (Sprague et al. 2013).

There are several obstacles that may prevent health practitioners from performing the above-mentioned tasks in their work with women and families who might be experiencing violence. These obstacles include: traditional attitudes and misconceptions about abuse and violence; fear of invading individual privacy; discomfort about talking with battered women; fear of offending the women; a sense of powerlessness; exclusive emphasis on physical health; limited formal education in the field of IPV; lack of professional knowledge and skills for dealing with cases of IPV; shortage of supportive facilities and services; and the feeling that identification of abuse and violence against women is not part of their role (Bair-Merritt et al. 2006; Berkowitz 2005; Colarossi et al. 2010; Djikanovic et al. 2010; Haj-Yahia 2013; Hamberger and Patel 2004; Han 2008; John et al. 2011).

Personal factors, such as discomfort with the topic and negative beliefs about battered women; powerlessness (e.g., frustration and perceived lack of competency to intervene); loss of control (e.g., feeling that control of the outcome and circumstances are in the hands of the patients, feeling that their intervention is useless); and time constraints, may also prevent physicians from becoming involved in fulfilling the above-mentioned tasks (Hamberger and Patel 2004; Sugg and Inui 1992). Other studies have revealed that some physicians who intervene in cases of violence against women pay a financial price because they become marginalized by their colleagues and by the organizations that employ them (Cohen et al. 1997).

Despite the disturbing findings about problems with assessment, detection, and attitudes about IPV among health care providers, there remains a dearth of research on this problem. In particular, we know little about this problem among health practitioners from non-Western societies, including Arab societies. Like their Western counterparts, scholars have charged that Arab health practitioners’ lack of involvement in detecting battered women is alarming (Douki et al. 2003). Based on their clinical experience, Douki et al. (2003) concluded that health practitioners in the Arab world tend to deny, interpret as delusional, and ignore women’s reports of violence, in addition to incliniations to label women experiencing abuse “masochistic” and “self-defeating,” and minimize both the experience and consequences of domestic violence. Indeed, one study found that 44, 29, and 10 % of the Palestinian physicians agreed with the following statements, respectively: “A very small percentage of Palestinian wives are abused by their husbands,” “wives are abused because of the abnormal way they treat their husbands,” and “most abused wives feel relieved after their husbands batter them” (Haj-Yahia 2010, p. 422). In a survey conducted among Sudanese physicians, 70 % of the participants indicated that they would not intervene with battered women beyond providing essential treatment (Ahmed et al. 2003).

Furthermore, studies on this topic rarely incorporate theoretical perspectives aimed at explaining the patterns of beliefs and responses regarding IPV among health care providers. In light of the serious shortage of empirical, theory-based research on the attitudes, beliefs, and perceptions of health practitioners toward the problem of violence against women in the Arab world, this study uses two theoretical frameworks to examine the beliefs of Palestinian physicians about wife beating.

Theoretical Framework

A combination of two theories was adopted in this study to examine and explain Palestinian physicians’ beliefs about wife beating: exposure to family violence (as one major component of social learning theory), and patriarchal ideology. Social learning theory argues that violent behavior as well as beliefs and attitudes that condone such behavior are learned through various socialization processes that occur mainly in childhood and adolescence, and are transmitted from one generation to the next (Jasinski 2001; Mihalic and Elliot 1997). Accordingly, when people are exposed to violent behavior as witnesses and/or through personal experience, that behavior will be justified, socially accepted, internalized, and emulated if it is not denounced by society at large, and by significant figures in particular (Jasinski 2001). Based on this theory, the present study examined exposure to family violence, with emphasis on witnessing interparental violence and on experiencing violence from parents as predictors of beliefs about wife beating.

The second theoretical component, patriarchal ideology, is reflected in values, beliefs, and norms that justify male dominance in all social areas and domains (Yllö and Straus 1990). Theories about patriarchal ideology explain how social and political arrangements privilege males as a group, and enable many forms of domination, including IPV. As is the case in many other societies, patriarchal ideology is deeply rooted in Arab societies, where the general notion of father rule and brother rule (i.e., dominance of men over women) is prevalent (Haj-Yahia 1998a, b, 2002a, b). Among other arrangements around gender, patriarchy denotes specific traditions that, because they are based on patrilocal residence and patrilineal descent, entrench the domination of men over women across generations. In the public sphere, power is shared by male patriarchs. In the private, family sphere, the senior man wields power over everyone else, including younger men, and exercises modes of subordination and control over women, which transcend cultural and religious boundaries (Haj-Yahia 2002a, b).

Thus, within the traditional arrangements, the husband is culturally accepted as the central authority in the family, whom the wife and children must ultimately respond to and obey. Accordingly, societies exhibit lenient attitudes toward men’s intimate violence against women throughout history in order to maintain the male’s advantage in conjugal power relations (Dobash and Dobash 1979; Haj-Yahia 2002a). Hence, participants’ sex-role stereotypes, attitudes toward women, attitudes toward women’s social involvement, and marital role expectations were examined as predictors of beliefs about wife beating.

Coleman and Stith (1997) conducted a study among nursing students, and found that the more the students maintained egalitarian sex-role attitudes as a possible measure of patriarchal ideology, the more sympathetic they were toward victims of domestic violence. Although they used a model that examined the combined impact of patriarchal perspectives and social learning on students’ attitudes toward violence against women, only patriarchal perspectives were found to have a significant effect. Hence, as measures of social learning theory, witnessing and personally experiencing family violence did not significantly correlate with the nursing students’ attitudes toward victims of domestic violence (Coleman and Stith 1997).

The findings of studies conducted among medical students from Sri Lanka (Haj-Yahia and de Zoysa 2007) and Turkey (Haj-Yahia and Uysal 2008), as well as among nursing students from Turkey (Haj-Yahia and Uysal 2011), are also noteworthy. These studies have revealed that nonegalitarian expectations of marriage and traditional attitudes toward women, as well as exposure to family violence during childhood and adolescence, correlated significantly with their negative beliefs about wife beating (e.g., the tendency to justify wife-beating, and the belief that battered women deserve to be beaten).

In addition, a study conducted among Palestinian physicians revealed that the more they witnessed interparental violence and experienced violence from parents, and the more they maintained a patriarchal ideology (i.e., sex-role stereotypes, negative attitudes toward women’s social involvement, traditional attitudes toward women, and non-egalitarian expectations of marriage), the greater their tendency to approve of violence against women at various levels of severity (Haj-Yahia 2010). As indicated, except for Haj-Yahia’s studies on Palestinian physicians (2010, 2013) and the above-mentioned study on Sudanese physicians (Ahmed et al. 2003), as well as the clinical insights on Arab health practitioners (Douki et al. 2003), there is a serious lack of research on the approach of health practitioners in the Arab world toward violence against women. Therefore, the present study on beliefs about wife beating among Palestinian physicians should contribute to a much-needed subject of scholarship.

Research Questions

Based on data from a larger study conducted among Palestinian physicians (for more details, see Haj-Yahia 2010, 2013), this paper focuses on two research questions. (1) To what extent do Palestinian physicians tend to justify wife beating, blame battered wives for beating, believe that abused women benefit from battering, and believe in helping (or not helping) battered women? and (2) To what extent can those beliefs be explained by the physicians’ patriarchal ideology and their exposure to violence in their families of origin, over and above the variance in those beliefs that can be attributed to physicians’ sociodemographic characteristics?

Method

Sample

The study was conducted among a convenience sample of 396 Palestinian physicians from the West Bank and East Jerusalem, who work in four major hospitals in those areas. The mean age of the participants was 39.50 years (range 27–62, SD = 8.40); and 73.6 % of the participants were male; 19 % were single, 78 % were married, 1.5 % were either divorced or separated, and 1.5 % were widowed at the time the study was conducted. Regarding the participants’ religion, 93.4 % of the participants were Muslim, and the remaining 6.6 % were Christian. In addition, 59.6 % of the participants were living in urban areas, 32.4 % were living in rural areas, and the remaining 8 % were living in refugee camps at the time the study was conducted. Thus, the sample was heterogeneous with regard to the variables age, gender, religion, and place of residence. Nonetheless, it should be noted that owing to the absence of data on the characteristics of the overall population of Palestinian physicians, there is no basis for comparison of the overall population with the sample of participants in this study. Similarly, there is no available information on the characteristics of all Palestinian hospitals that could facilitate a comparision of those hospitals with the four hospitals from which the present research sample was drawn. In addition, the physicians who participated in this study were hospital-based rather than office-based. However, there is no information on the characteristics of both groups of physicians (i.e., hospital-based and office-based) that can be used as a basis for comparison of those groups.

Questionnaire

A self-administered questionnaire in the Arabic language was used in this study, which consisted of several sections. As indicated, because this paper presents partial results of a larger project (see Haj-Yahia 2010, 2013), we will only describe the relevant sections and scales of the questionnaire.

Beliefs about Wife Beating

The questionnaire consisted of 37 items based on a 5-point Likert type scale ranging from 1 (strongly agree) to 5 (strongly disagree). The items were derived from a revised version of Saunders et al. (1987) Inventory of Beliefs About Wife Beating (IBWB), and were utilized to measure the following four beliefs: (1) justifying wife beating; (2) blaming women for their beating; (3) the belief that women benefit from beating; and (4) helping battered women (for sample items, see Table 1). In examining the scale’s dimensionality, reliability, and validity, Saunders et al. (1987) found that the factors that were derived had strong construct validity. The Arabic version of the IBWB, as utilized in previous studies, proved to have very good internal reliabilities (e.g., Haj-Yahia 1998a, b, 2003, 2005). The Cronbach’s alpha internal reliability coefficients for all four subscales of the IBWB in this study ranged from .91 (justifying wife beating) to .78 (the belief that women benefit from beating).

Table 1 Beliefs about wife beating (sample items): percentages, mean, and standard deviation (N = 396)
Full size table

Sex-role Stereotypes (SRS)

An Arabic version of Larsen and Long’s (1988) scale for measuring sex-role stereotypes was used in this study. The scale consisted of 20 statements, where responses were based on a 6-point Likert type scale ranging from 1 (strongly agree) to 6 (strongly disagree). Larsen and Long (1988) maintained that the SRS has excellent part-whole correlations and corrected split-half reliability coefficients, as well as promising concurrent and construct validity. The original English version of this scale was translated from English into Arabic by two Palestinian professionals who specialize in the social and behavioral sciences and have an excellent command of English. Each of them was asked to translate the scale into Arabic independently, taking into consideration the cultural context of Palestinian society. Afterwards, they worked together with the principal investigator to prepare a unified Arabic version of both translations of the scale. To maintain comparability of the scale in both languages, the unified Arabic version was back-translated into English by a professional translator. Furthermore, varimax-rotated factor analysis for the Arabic version of this scale revealed one major factor that accounted for 84.6 % of the variance, with factor loadings ranging from .44 to .79 for all 20 items. Cronbach’s alpha internal reliability coefficient for the Arabic version of the questionnaire used in this study was .88.

Attitudes Toward Women’s Social Involvement (WSI)

The scale was developed specifically for this study, in order to measure participants’ attitudes toward women’s social involvement and integration. Participants were asked to indicate the extent of their agreement with each item on a 6-point Likert type scale ranging from 1 (strongly disagree) to 6 (strongly agree). Sample items of the measure: “Women have the right to be involved in establishing non-governmental organization to serve their communities,” “Women can contribute to the success of their neighbourhood committees no less than men.” (The scale can be obtained upon request from the first author.) In the first stage of developing the measure, the principal investigator compiled the list of 15 items based on interviews conducted with six Palestinian professionals who have extensive experience with research in the social and behavioral sciences in general and in women’s studies in particular. All of the professionals were requested to provide operational sentences that could be used to measure attitudes towards women’s social involvement. In the next stage of developing this version of the measure, it was pilot-tested twice among another group of seven Palestinian professionals who have extensive research experience in the above-mentioned fields. First, it was tested with each one of the professionals individually, and then it was tested with all seven of them as a focus group. Based on the results of this procedure, an advanced and refined version was developed. Afterwards, its face validity and content valdity were examined by another group of four Palestinian professionals. Factor analysis based on the results of this study revealed that nine out of 15 items belonged to one major factor, with factor loadings ranging from .41 to .84; and varimax-rotated factor analysis revealed that this major factor accounted for 86.2 % of the variance. The Cronbach’s alpha measure of internal consistency for all nine items was .89.

Attitudes Toward Women (ATW)

This study used a short version of 14 statements in Spence and Helmreich’s (1978) Attitudes Toward Women scale (ATW). The scale measured traditional-patriarchal versus liberal-egalitarian attitudes toward women. Participants were asked to indicate the extent to which they agree or disagree with each item on a 4-point Likert type scale ranging from 1 (strongly agree) to 4 (strongly disagree). Spence and Helmreich (1978) reported that the short version of the ATW had a correlation of .91 with the original, longer version of the ATW administered among a sample of 715 students in psychology classes. In addition, the researchers provided strong evidence and support for the scale’s construct validity. The Cronbach’s alpha value of the shorter English version of the ATW was reported to be .89; and the Arabic version of the ATW utilized in previous studies proved to have very good internal reliabilities (e.g., Haj-Yahia 1998a, b, 2003, 2005). Factor analysis carried out for the Arabic version of the ATW used in this study revealed factor loadings that ranged from .41 to .78 for all 14 items. The Cronbach’s alpha internal reliability coefficient for the Arabic version of the questionnaire used in this study was .89.

Marital Role Expectations (MARI)

A short 18-item version of the Marital Role Expectations Inventory (MARI; Dunn and DeBonis 1979) was used in this study to measure the participants’ marital role expectations. Participants were asked to indicate the extent to which they agree or disagree with each item on a 5-point Likert type scale ranging from 1 (strongly agree) to 5 (strongly disagree). Dunn and DeBonis (1979) reported a Spearman-Brown reliability coefficient of .975 for the measure on a split-half correlation analysis. The Arabic version of the MARI utilized in previous studies proved to have very good internal reliabilities (e.g., Haj-Yahia 1998a, b, 2003, 2005). Factor analysis carried out for the Arabic version of the MARI used in this study revealed factor loadings ranging from .43 to .82 for all 18 items of the instrument. The Cronbach’s alpha internal reliability coefficient for the Arabic version utilized in this study was .91.

Witnessing and Experiencing Violence in the Family-of-Origin

Arabic versions of four different forms of the Conflict Tactics Scales (CTS; Straus 1979) were used to measure the extent to which the participants had experienced (i.e., by parents) or witnessed (i.e., between parents) psychological aggression and physical violence while they lived with their parents. Participants were requested to report how many times they had experienced and witnessed each of the acts specified in the measures of psychological aggression and physical violence while they lived with their families of origin (mainly before getting married or before age 18). Responses were based on the following scale: 0 (never), 1 (1–2 times), 2 (3–5 times), 3 (6–10 times), and 4 (11 times or more). These tests included calculation of the Cronbach’s alpha coefficient (at least .78 on the violence scale), as well as item analysis, which indicated that the reliability of the measure was adequate. Haj-Yahia (1991) translated the CTS into Arabic and adapted it to the social, political, and cultural contexts of Arab societies. The Arabic version of the CTS utilized in previous studies proved to have excellent internal reliabilities (e.g., Haj-Yahia 1991, Haj-Yahia et al. 2002). The Cronbach’s alpha internal reliability coefficients for all four forms of psychological aggression and four forms of physical violence measured in this study ranged from .87 to .93 and .86 to .94, respectively.

Procedure

In light of the existing political situation in the region and particularly in the West Bank and Gaza, it was difficult to obtain a comprehensive sampling framework that encompassed physicians from every area of the Palestinian Authority. As a result, it was not possible to select a random sample of Palestinian physicians. Even if it had been possible to select a random sample, we would not have been able to reach the selected participants due to restrictions on travel in the Palestinian Authority. Therefore, a convenience sample was derived, which comprised physicians employed at four Palestinian hospitals in the West Bank (i.e., in the cities of Jenin, Nablus, and Ramallah), and in East Jerusalem. Our sample was heterogeneous in terms of age, religion, area of residence, type of locality of residence, and years of work experience. Physicians who were working in the emergency ward or in other wards at the hospitals on the day the questionnaire was distributed were approached by research assistants and asked to participate in a study on physicians’ attitudes toward and responses to wife abuse. The conceptual framework and design of the study, including the procedures and various ethical dimensions, were discussed and approved by a team of health practitioners and women who were activists in the area of IPV.

The physicians were instructed to fill out the questionnaire at their convenience, place the completed forms in an envelope, and put the sealed envelope in a designated locked box in the emergency room at each participating hospital. The research assistants collected completed questionnaires for 3 weeks, on three different days at the end of each week, as follows: 229 questionnaires were collected at Week 1, 110 at Week 2, and 57 at Week 3. The overall response rate was 83 %.

Data Analysis

First, descriptive statistics (i.e., the distribution of the answers by percentages, means, and standard deviations) were examined for each item that measured the dependent variables of the study, i.e., beliefs about wife beating (for sample items, see Table 1). Second, zero-order correlation analyses were computed for the participants’ demographic characteristics (i.e., age, religion, place of residence, and gender), independent variables (i.e., sex-role stereotypes, attitudes toward women’s social involvement and integration, attitudes toward women, marital role expectations, witnessing father-to-mother and mother-to-father violence, and experiencing violence by the father and mother), and the dependent variables (i.e., justifying wife beating, blaming battered women, the belief that women benefit from beating, and helping battered women; see Table 2). Third, regression and hierarchical multiple regression analyses were conducted in order to examine the extent to which the four above-mentioned beliefs about wife-beating could be explained by the main predictors of the study over and above their predictability by the participants’ sociodemographic characteristics.

Table 2 Zero-order correlations among all independent and dependent variables (N = 396)
Full size table

To carry out these analyses, two new independent variables were produced from the above-mentioned predictors. First, due to the very high and significant correlations and potential multicollinearity among all patterns of violence witnessed between parents (r = .53, p < .0001), participants’ scores on witnessing father-to-mother and mother-to-father psychological aggression and physical violence were added to produce one score for the new variable, Witnessing Interparental Violence. Moreover, due to the very high and significant correlations and potential multicollinearity among all patterns of violence that participants experienced by their parents (r = .56, p < .0001), participants’ scores on experiencing psychological aggression and physical violence by the father and mother were added to produce one score for the new variable, Experiencing Parental Violence. Although the correlation between these two aggregated family violence variables was found to be very high and significant (r = .54, p < .0001), we decided to treat both of them as two separate predictors rather than as one variable in order to examinne the relative contribution of each predictor to explaining each of the beliefs about wife beating. Table 3 presents the results of regression and hierarchical multiple regression analyses with regard to each one of the four beliefs about wife beating that were examined in this study. All of the analyses were conducted using SPSS Software, version 19. Missing values were treated by using listwise deletion for all analyses.

Table 3 Regression and hierarchical multiple regression for beliefs about wife-beating
Full size table

Results

The results regarding each of the four beliefs about wife beating among Palestinian physicians are presented below.

Justifying Wife Beating

The results in Table 1 indicate that although 34 % of the Palestinian physicians agreed or strongly agreed with the statement that “There is no excuse for a man to beat his wife” (item 5, M = 3.02, SD = 1.28), between 24 and 47 % of them agreed or strongly agreed that wives deserve to be beaten in different marital circumstances. For example, 47 and 35 % of the participants agreed or strongly agreed that a husband has the right to beat his wife “If she makes fun of his manhood” (item 1, M = 2.79, SD = 1.22); and “If she keeps reminding her husband of his weak points” (item 2, M = 3.03, SD = 1.17), respectively. Furthermore, 31 and 26 % of the participants agreed or strongly agreed with the statements that “A wife who constantly refuses to have sex with her husband is asking to be beaten” (item 4, M = 3.24, SD = 1.26); and “A wife who lies to her husband deserves to be beaten” (item 3, M = 3.20, SD = 1.09), respectively (see Table 1).

The results of correlation analysis presented in Table 2 indicate that the extent to which the participants justified wife beating correlated significantly with all of their demographic characteristics. For example, the results revealed that the older the participants, the more they tended to justify wife beating (r = .19, p < .01); and that male physicians were more likely than their female counterparts to justify wife beating (r = .15, p < .05).

In addition, the results reveal that the extent to which participants justified wife beating correlated significantly with all of the patriarchy-oriented independent variables measured in the study. For example, the more participants maintain sex-role stereotypes and nonegalitarian and patriarchal expectations of marriage, the greater their tendency to justify wife beating (r = .31, p < .0001, and r = .42, p < .0001, respectively). Furthermore, the extent of justifying wife beating correlated significantly with two out of four of the independent variables related to social learning. Specifically, the more the participants witnessed father-to-mother violence and the more they experienced violence by their mothers, the greater their tendency to justify wife beating (r = .20, p < .001, and r = .19, p < .01, respectively).

The results of regression and hierarchical multiple regression analyses presented in Table 3 reveal that 32 % of the variance in the participants’ tendency to justify wife beating can be attributed to all three blocks of predictors considered in those analyses. In essence, 4 and 26 % of the variance in this belief can be attributed to the participants’ witnessing interparental violence (see the second block) and to all four patriarchal ideology-related predictors (see the third block), respectively, over and above the 2 % of the variance that was explained by their sociodemographic characteristics.

Blaming Women for Their Beating

The results reveal that between 21 and 41 % of the physicians agreed or strongly agreed with the items that measured the tendency to blame women for their beating. For example, 27, 32, and 41 % of the participants agreed or strongly agreed with the following statements: “Sometimes the wife’s provocative words cause her husband to beat her” (item 6, M = 3.25, SD = 1.15), “When a woman is beaten, it is caused by her behavior during the weeks beforehand” (item 7, M = 3.13, SD = 1.19), and “If the wife had known her boundaries, her husband wouldn’t have beaten her” (item 8, M = 2.81, SD = 1.23), respectively (see Table 1).

The results in Table 2 reveal that male physicians were more likely than their female counterparts to blame women for being battered (r = .15, p < .05), and that the older the participants were, the greater the likelihood that they would blame women for being battered (r = .13, p < .05). In addition, the results reveal that the extent to which participants blamed women for being battered correlated significantly with all of the patriarchal ideology-related independent variables of the study. For example, the more the participants opposed women’s social involvement, and the more traditional and negative their attitudes toward women, the greater their tendency to blame women for being battered (r = .17, p < .01, and r = .43, p < .0001, respectively). The results also reveal that the participants’ tendency to blame women for being battered correlated significantly with two out of the four variables related to social learning. Specifically, the more the participants witnessed father-to-mother violence and the more they experienced violence by their mothers, the greater their tendency to blame women for being battered (r = .18, p < .01, and r = .12, p < .05, respectively).

The results presented in Table 3 reveal that 35 % of the variance in the extent to which the participants blamed women for being beaten can be attributed to all three blocks of predictors that were considered in the analyses. Most importantly, 31 % of the variance in this belief can be attributed to the four patriarchal ideology-related predictors, which were found to be significant predictors of the participants’ tendency to blame women for being battered, over and above the variance that could be explained by the two previous blocks of predictors. Notably, neither the participants’ sociodemographic characteristics (which explained 2 % of the variance in this belief) nor patterns of exposure to family violence (which explained 2 % of the variance in this belief) contributed significantly to predicting their tendency to blame women for their beating.

Women Benefit From Beating

The results reveal that whereas 62 % of the Palestinian physicians agreed or strongly agreed with the statement that “Women feel pain and no pleasure when they are beaten up by their husbands” (item 10, M = 2.29, SD = 1.24), between 21 and 30 % of the participants expressed some extent of support for the statements that reflect a tendency to believe that women benefit from being beaten. For example, 21 % of the participants agreed or strongly agreed that “Battered wives try to get their partners to beat them in order to gain attention from others” (item 9, M = 3.35, SD = 1.18, see Table 1).

The results in Table 2 indicate that the older the participants were, the greater their tendency to believe that women benefit from battering (r = .15, p < .05); male and Muslim physicians showed a greater tendency to believe that women benefit from beating than did female and Christian physicians (r = .20, p < .01, and r = .16, p < .01, respectively). The results also reveal that the more the participants held sex-role stereotypes and patriarchal and nonegalitarian expectations of marriage, the more they tended to believe that women benefit from beating (r = .17, p < .01, and r = .17, p < .01, respectively). In addition, the results reveal that the more the participants experienced violence by their fathers, the more they tended to believe that women benefit from beating (r = .18, p < .01).

The results in Table 3 indicate that 35 % of the variance in the extent to which the participants believed that women benefit from being beaten can be explained by all three blocks of predictors that were considered in those analyses. In particular, 4 and 27 % of the variance in the participants’ belief that women benefit from beating were explained by their exposure to family violence and patriarchal ideology, respectively, over and above the 4 % of variance in this belief that could be attributed to their sociodemographic characteristics. In essence, the findings reveal that experiencing parental violence (β = .19, p < .05, second block) and each one of the four patriarchal ideology-related predictors (third block) were the most significant predictors that explain the variance in the extent to which the participants believed that women benefit from being beaten. The results also reveal that of all four sociodemographic characteristics, religion (Muslims more than Christians, β = .15, p < .05) and gender (men more than women, β = .17, p < .05, first block) were significant predictors of the participants’ belief that women benefit from beating.

Helping Battered Women

The results indicate that 33 % of the participants agreed or strongly agreed with the statement that “If a wife tells me that her husband abuses her, I prefer not to intervene” (item 12, M = 3.01, SD = 1.12). However, physicians exhibited a noticeable tendency to support different forms of help for battered women. About 56 and 55 % of the participants agreed or strongly agreed that “The Palestinian Authority should give wife abuse very high priority over other social problems” (item 13, M = 2.53, SD = 1.19), and that “Social agencies should do more to help battered women” (item 14, M = 2.50, SD = 1.26), respectively. Even though 59 % of the participants agreed or strongly agreed that “Women should be protected by law if their husbands beat them” (item 15, M = 2.45, SD = 1.21), only 27 % of the participants agreed or strongly agreed that “Husbands who beat their wives should be arrested” (item 11, M = 3.28, SD = 1.17, see Table 1).

The results in Table 2 reveal that older and male physicians were less likely than younger and female physicians to support the provision of assistance to battered wives (r = −.17, p < .01, and r = −.12, p < .05, respectively). In addition, the results reveal that all four patriarchal ideology-related variables correlated significantly with the extent to which the participants tended to support or oppose the provision of assistance to battered women. For example, the more the participants held sex-role stereotypes and patriarchal and nonegalitarian expectations of marriage, the less likely they were to support the provision of assistance to battered women (r = −.28, p < .0001, and r = −.29, p < .0001, respectively). The results also reveal that the more the participants experienced violence from their fathers, the less likely they were to support the provision of assistance to battered women (r = −.14, p < .05).

The results of regression and hierarchical multiple regression analyses presented in Table 3 indicate that 28 % of the variance in the participants’ tendency to support or oppose the provision of assistance to battered women can be explained by all three blocks of predictors that were considered in those analyses. In essence, all four predictors in the third block (i.e., patriarchal ideology-related variables) were found to contribute significantly to explaining 25 % of the variance in the extent to which the participants support or oppose the provision of assistance to battered wives, over and above the variance that could be attributed to the participants’ sociodemographic characteristics (1 %) and to exposure to family violence (2 %). Notably, none of the variables in those two blocks of predictors significantly explained the variance in the participants’ beliefs about helping battered women.

Discussion

Summary and Conclusions

The paper presents a study conducted among Palestinian physicians, which examined their beliefs about wife beating, including their tendencies to justify wife beating and blame women for being beaten, their belief that women benefit from beating, and their beliefs about helping battered women. The results indicate that substantial percentages of the Palestinian physicians showed a tendency to justify wife beating, to blame women for violence against them, and to believe that battered women benefit from beating. Interestingly, despite these negative beliefs, substantial percentages of the physicians also showed a tendency to help battered women. These four beliefs were investigated in an attempt to examine their relationship to the physicians’ patriarchal ideology and exposure to violence in their families of origin.

The findings relating to beliefs about domestic violence among Palestinian physicians are quite similar to the results of studies conducted among other populations of health care providers. For instance, a study in the US among primary care teams (physicians, physicians’ assistants, nurses and medical assistants) found that 15 % of the respondents indicated that patients’ personalities caused them to be abused; 25 % believed that patient’s passive-dependent personality led to abuse; and 19 % indicated that abused persons stay because they get something out of the abusive relationship (Sugg et al. 1999). Another study conducted among medical students and surgical residents in Ontario found that 18 % of the participants thought victims must get something out of abusive relationships; 11 % indicated that they believed victims chose to be victims; and 41 % indicated that something in patients’ personalities cause them to be abused (Sprague et al. 2013).

The present study of Palestinian physicians revealed that after controlling for their sociodemographic characteristics, a significant and large percentage of the variance in their beliefs about wife beating could be attributed to their patriarchal ideology, over and above the variance that could be explained by their exposure to family violence. This finding aligns with earlier studies among the general population in Palestinian society and in other Arab societies, which revealed that these beliefs can be attributed to the negative and traditional attitudes toward women, as well as to nonegalitarian expectations of marriage, and to sex-role stereotypes (Haj-Yahia 1998a, b, 2002a, 2003). These results are consistent with studies conducted among both Western and non-Western populations, which have linked traditional notions about gender roles with more lenient attitudes about domestic violence perpetration and higher rates of victim blaming. Since the first investigations began in the 1970s, studies from around the world have linked patriarchial ideology with more permissive attitudes about violence against women (Flood and Pease 2009), including European American university students (Berkel et al. 2004; Esqueda and Harrison 2005), Chinese public service professionals (Tang et al. 2002), and Arab societies at large (Haj-Yahia 1991, 1997, 1998a, 2003).

This finding is consistent with our previous contention about patriarchal ideology, that in patriarchal societies there is a tendency to perpetuate male prerogatives by approval of wife abuse, and to legitimize such behavior in order to see that women maintain an inferior status to men in society (Dobash and Dobash 1979, 1992; Yllö and Straus 1990). Based on this approach, it is assumed that people with rigid gender norms tend to approve of wife abuse, that they tend to treat violent husbands leniently, and that they attribute wife abuse primarily to the woman and less, if at all, to the violent husband (Haj-Yahia 2010).

Although a relatively small amount of the variance in the physicians’ beliefs about wife beating was explained by their exposure to family violence, after controlling for their sociodemographic characteristics, the findings of our study also provide partial empirical support for social learning theory. As indicated, the findings showed that the more the Palestinian physicians either experienced or witnessed some patterns of violence in their families of origin, the more they tended to justify wife abuse, the more they tended to believe that battered women benefit from beating, the more they tended to blame women for violence against them, and the less they supported helping battered women. Notably, social learning theory is more often used by researchers to explain violent behavior of men against intimate partners than to explain beliefs about and attitudes toward violence against women. Thus, the results of this study add to the scarce empirical knowledge on beliefs about wife beating from a social learning perspective.

Limitations of the Study

The importance of this study lies in its integration of two theoretical perspectives, which are applied toward examining Palestinian physicians’ beliefs about wife beating. Nonetheless, several of its limitations need to be addressed. As mentioned, a self-administered instrument package was utilized for data collection. Participants in the study were asked to respond to questions about personal issues that are considered sensitive in Palestinian society – particularly questions about topics such as sex-role stereotypes, attitudes toward women, expectations of marriage, and violence in the family of origin. Notably, self-reports are subject to response biases and social desirability effects. Although we attempted to ensure the confidentiality and anonymity of responses, in addition to making sure that the participants answered the questionnaire independently, further efforts are required in order to minimize self-report bias. For example, reports by battered women who were treated by the physicians participating in the study would provide an opportunity for cross-validation of the responses. Specifically, the reliability and validity of the results would be enhanced if the battered women were to indicate how they perceive the physicians’ approach toward them, and how they perceive the physicians’ attitudes toward women in general and toward battered women in particular.

Retrospective self-reports about witnessing interparental violence and experiencing parental violence during childhood are also subject to distortion. For example, it cannot be determined whether the reported behavior in childhood was actually abusive when evaluated or measured objectively. In addition, it cannot be determined whether the abusive behavior was defined by the participants as such at the time it was experienced or witnessed, or whether it was defined as such only in retrospect. Undoubtedly, concurrent cognitive labeling of the event may influence the degree or nature of its impact on the individual’s behavior. Thus, it would have been desirable to include reports from siblings about exposure to different patterns of family violence, or to include reports from a sample of the physicians’ parents concerning past and present styles of conflict management in the family.

Additional limitations of the study relate to the sample and the sampling method. In light of the existing political situation in the region, and particularly in the West Bank and Gaza Strip, it was difficult to obtain a comprehensive sampling framework that encompassed physicians from every area of the Palestinian Authority. As a result, it was not possible to select a random sample of Palestinian physicians working in different health settings (i.e., hospitals, public clinics, and private clinics). Even if it had been possible to select a random sample, we would not have been able to reach the selected participants due to restrictions on travel in the Palestinian Authority. Nonetheless, we did reach a convenience sample of physicians who work at hospitals, which was heterogeneous in terms of age, religion, area of residence, and years of work experience. Owing to the heterogeneous nature of the sample, it is assumed that the participants represented Palestinian physicians from various social strata. Nonetheless, it is difficult to guarantee the generalizability of the findings to all Palestinian physicians. Therefore, future studies of Palestinian physicians should attempt to cover a large random sample that includes specialists in various fields (e.g., family practitioners, physicians in different hospital wards, emergency room physicians, physicians in private clinics, physicians in public and community clinics, gynecologists, dentists, etc.) from all areas of the Palestinian Authority who work in various settings.

Implications for Future Research and Theory Development

Our research findings are consistent with patriarchal ideology and to some extent with social learning theory. Although high percentages of the variance in each of the dependent variables were explained by variables deriving from those two theories, the participants’ patriarchal ideology still explained larger amounts of the variance in their beliefs about wife beating than the amounts of variance that could be attributed to their exposure to family violence. At the same time, it is clear that a substantial share of the variance in Palestinian physicians’ beliefs about wife beating was not explained by the predictors of this study.

This result underscores the importance of considering health practitioners’ beliefs about wife beating from a broad, integrative perspective that includes additional theories, besides the two theories that were examined in this study. The incorporation of additional theoretical frameworks will provide a better understanding of the approach of physicians and other professionals toward the problem of wife abuse, including their beliefs about wife beating. For example, variables such as family honor, perceptions of masculinity, and the significance of collectivism in the nuclear and extended family are relevant to patriarchal ideology (Haj-Yahia 2002a, b), and should, therefore, be examined in terms of their impact on the approach of professionals toward wife abuse.

Moreover, future research in the field should incorporate the structural element of patriarchy, which is “the low status women generally hold relative to men in the family and in the economic, educational, political, and legal institutions” (Yllö and Straus 1990, p. 384). Examination of the different dimensions of the structural element of patriarchy as delineated by Yllö and Straus (1990) would contribute toward enhancing knowledge about the approach of professionals toward battered women and violent husbands.

Furthermore, the impact of economic, sociocultural, and political transformations and transitions on the patriarchal character of Palestinian society over the past four decades may provide a strong basis for studying the beliefs of Palestinian professionals about wife abuse and other issues relating to violence against women. Examinations of this sort are important, in that they seriously consider the ways in which patriarchial beliefs and norms both evolve over time and interact with the official structural responses of the government. In this connection, Kandiyoti (1988) dealt extensively with the issue of patriarchal bargaining and the impact of government intervention or lack of intervention on emancipating women and empowering them to change or enhance the patriarchal character of the society. This may provide an additional conceptual framework for future research into the approach of physicians and other professionals toward wife abuse, based on the patriarchal perspective adopted in this study.

As noted, small proportions of the variance in physicians’ beliefs about wife beating could be attributed to their exposure to family violence. Consequently, future research on attitudes about wife abuse among health practitioners should examine additional independent variables that derive from social learning theory in order to learn more about the influence of learning processes and experiences on physicians’ beliefs about wife beating. For example, Bevan and Higgins (2002) suggested that three interrelated theoretical mechanisms are relevant to the problem of wife abuse: (1) identification with the aggressor, whereby people who are exposed to violence in their family of origin will behave violently in the future toward members of their family if they identified with the aggressor in their family of origin; (2) vicarious reinforcement, whereby intergenerational transmission of violence is caused primarily by principles of modeling; and (3) positive reinforcement, whereby people learn violence not only through exposure during childhood, but also through modes of approval for the use of violence. The present study examined the physicians’ exposure to violence during childhood. However, we did not investigate violence in the physicians’ own intimate relationships and whether they identified with aggressors in their families of origin or received approval and positive reinforcement for violence. Nor did we examine the impact of these dimensions on physicians’ beliefs about wife-beating. Hence, it is recommended that future studies conducted among health practitioners deal with these mechanisims and the extent to which they relate to health practitioners’ beliefs about and approaches to wife beating.

It would also be worthwhile to examine additional theories that could be related to the unexplained variance in the approaches of health practitioners toward the problem of wife abuse. For example, resources theory (Bersani and Chen 1988) can provide a basis for examining the relationship between the presence or absence of instrumental resources in the Palestinian Authority and the approaches of health practitioners toward wife abuse. Notably, there is no law for prevention of wife abuse in the Palestinian Authority. Moreover, special services such as battered women’s shelters and centers for treatment of violent men are very scarce, and there are no special allowances to support battered women. In the same vein, it is possible to examine the impact of emotional resources (e.g., fear of revenge by the families of the victim and the perpetrator, anxiety) and cognitive resources (e.g., lack of information about the problem of wife abuse) on physicians’ approaches toward the problem of wife abuse. It would be particularly worthwhile to examine the content of physicians’ education and training during their academic studies at the university as well as after graduation, and the impact of that content on their current beliefs about wife-beating and approaches to intervention in cases of IPV.

The present study examined the Palestinian physicians’ beliefs about wife beating, but did not consider their actual responses to the problem. It is essential to examine these responses, and to focus specifically on how physicians treat battered women under their care. For example, to what extent and how are the physicians involved in identifying and assessing battered women whom they encounter under their care? How do they behave toward those women and their husbands after they identify the wife as abused? To what extent do they show understanding and empathy toward the battered woman on the one hand, and to what extent do they blame her on the other hand? Clearly, examination of the attitudes, perceptions, and approaches of physicians and other health practitioners toward wife abuse is essential for planning and implementation of programs and provision of training and guidance for these professionals. However, without empirical evidence to show how physicians actually behave toward battered women and violent men, it will be difficult to provide them with the most effective guidance for intervention with battered wives and their husbands.

Implications for Professional Training of Physicians

Physicians can play a critical role in identifying battered women among the patients in their care. In particular, together with other professional staff at hospitals, they can plan an active role in ensuring that battered women and their children receive critical assistance such as protection and support, on one hand, and that their violent partners receive treatment, on the other. To ensure Palestinian physicians’ involvement in this intervention process, it is important to debunk their negative beliefs about battered women and violent men, to increase their willingness to help battered women, and to equip them with appropriate knowledge about this problem, so that they will become more involved in the process of identifying, assessing, and protecting battered women. Debunking these negative beliefs on the one hand, and increasing the physicians’ knowledge about the problem of IPV can be achieved through their academic studies as well as through continued provision of training and supervision after graduation from medical school.

It should be noted that although high proportions of physicians in this study expressed negative beliefs about wife-beating, substantial percentages of physicians also expressed support for helping battered women. Hence, even though training and supervision of physicians should aspire to change their negative beliefs about wife-beating, as indicated earlier, it is also important to enhance and build on their willingness to help battered women and encourage them to cooperate with other professionals and organizations that offer appropriate services and interventions in cases of IPV. This training and supervision should also instil professional norms (e.g., accountability) for helping battered women, and interventions should be guided by these norms rather than by personal beliefs such as those examined in this study.