Socio-Demographic Predictors of Intimate Partner Violence in a Population Sample of Iranian Women

1. Background

Intimate partner violence is a global problem and a threat for women’s health (1) that have numerous physical, psychological, and social harms (2, 3). It refers to a pattern of assaultive and coercive behavior by an individual against her/his partner that may include physical injury, psychological abuse, sexual assault, progressive isolation, stalking, deprivation, intimidation, and reproductive coercion (4). Domestic violence accounts for about 5% of the total burden of disease among women aged 15 to 44 in the developing countries, and about 19% in the developed countries (5). Several studies with varying reports on the prevalence of domestic violence indicate the impact of socio-cultural settings (3, 6).

According to the World Health Organization (WHO) census, concerning all types of violence against women 15 to 49 years old by their husbands, the prevalence of psychological (21% to 90%), physical (13% to 61%), sexual (6% to 59%), and violence and violence during pregnancy (1% to 28%) varied in different countries under study (10 countries) (3). Each year, nearly 5.3 million cases of domestic violence among women 18 years of age and older is happening in the U.S., leading to about 2 million injuries and 1,400 deaths (7). Violence can set the ground for countless harms and consequences at the individual, family, and society level, including the reduction of women’s energy, endangerment of physical health, deterioration of women’s self-esteem, excessive social and economic costs on the society, and women’s fear and insecurity (8).

Studies conducted on the domestic violence in Iran have reported different prevalence rates for each type of violence. In a study on 251 women in the eastern part of Iran, the prevalence of overall violence was 78.1% and the psychological violence (66.5%) was the most prevalent of domestic type (9). In another study in Ahvaz, Iran, the prevalence of violence was 63.8%, with 58.8% of women experiencing psychological abuse (10). Such a difference could be associated with individual, social, ethnic, and cultural differences across Iran. Thus, it is necessary to measure the prevalence of domestic violence in Tabriz (a city in northwest of Iran with a different ethnicity and culture compared with other cities of Iran).

According to the Statistical Center of Iran, women of the reproductive age account for a large number of female populations (approximately 60% of the total female population) in the Iranian population pyramid (11). Considering the fact that women’s health is essential for preserving the familial dignity and childbearing, this study was aimed at determining the socio-demographic predictors of domestic violence in women of the reproductive age to screen the individuals at a greater risk.

2. Methods

2.1. Study Design and Participants

This was a cross-sectional study carried out in 2015 on a total of 558 married women covered by health centers in Tabriz, Iran. The inclusion criteria were as follows: age ranging from 15 to 49 years, living with a spouse or sex partner, and having a secondary school level education or higher. The exclusion criteria were as follows: any stressful events during the last year for each of the couples such as the death of immediate family members or an accident causing serious physical or psychological damage, pregnancy, postpartum period (from childbirth up to one year after delivery), one of the couples suffering from psychological disease (based on self-report), one of the couples suffering from infertility, and withdrawal from continuing cooperation while completing the questionnaire.

2.2. Sampling and Data Collection

Sampling began when the study proposal was approved and an ethics code was obtained from the Ethics Committee of Tabriz University of Medical Sciences (ethical code: 5.4.3820, 13 July 2014). In this study, the sampling was conducted through a two-stage cluster method. A total of 10 out of 39 active health centers and 11 out of 42 active health posts were randomly selected in Tabriz. By referring to the health centers, the researcher prepared a list of reproductive age women. Then, based on the population data (the number of married women covered in each center/post), a proper sample size was proportionally calculated according to the study sample size and the women in each health center were randomly selected from the list of covered women. At the next stage, the researcher contacted the selected women and assessed the inclusion and exclusion criteria by which those women having eligible criteria were invited to cooperate in the study. In the meeting, the purpose and procedures of the study were clearly explained to the participants and then all of them filled a written consent form. Then, the participants completed the self-administered questionnaires.

The basic sample size information was obtained from a study on domestic violence conducted in Tehran, the capital of Iran (59% prevalence of psychological subscale) (12). The sample size was determined as 372, with considering α = 0.05 and precision (d) = 0.05. Given that we used cluster sampling, the design effect equal with 1.5 was considered in the sample size and the required sample size was considered to be 558 (372 × 1.5).

2.3. Data Collecting Instruments

Data were collected by socio-demographic characteristics and revised conflict tactics scales (CTS2) questionnaires.

The socio-demographic characteristics questionnaire included questions on the age of participants and their spouses, duration of marriage, the number of previous marriages of the participants and their husbands, the number of living children, children’s gender, educational level of the participants and their husbands, job of the participants and their husbands, participants’ satisfaction from their own jobs and their husbands’ job, adequacy of monthly family income, spouse’s cigarette smoking, spouse’s substance abuse, spouse’s alcohol use, consensual marriage, and individuals who live with them.

Revised conflict tactics scales (CTS2) was designed in 1995 (13). CTS2 assess prevalence and chronicity of IPA with five sub-scales including negotiation, psychological aggression, physical assault, sexual coercion, and injury. CTS2 has 39 paired items and five sub-scales including negotiation, psychological, physical, sexual, and injuries. CTS2 has 8 response categories (0 to 7) for each items; the category 0 corresponds with “never” and category 7 corresponds with “not in the referent period but it did happen before”, the 1 to 6 response categories correspond with “once”; “twice”; “3 to 5 times”; “6 to 10 times”; “11 to 20 times”; and “more than 20 times” in the referent period, respectively. The approximate midpoints of the frequency response categories are used for scale scoring purposes (i.e., “once” is scored as 1; “twice” as 2; “3 to 5 times” as 4; “6 to 10 times” as 8; “11 to 20 times” as 15; and “more than 20 times” as 25). CTS2 items are presented in pair questions. The first question in the pair asks respondents to indicate how often they carried out each item and second asks how often the partner carried out each behavior. Selecting options 1 to 6 for any item of each subscale or scale was considered as presence, and selecting options 0 or 7 for all items of the subscale or scale was considered as absence of that type of IPA. To assess the prevalence of abuse in ones lifetime, the option 7 was also considered as presence of violence. Chronicity of minor and severe abuse in each type among women with a positive experience of that type of violence was determined by adding up the midpoint scores of their items. CTS2 has high internal consistency. The reliability of the instrument was confirmed in Tehran, Iran with Cronbach’s alpha ≥ 0.9 (14). In this study, the questionnaire was completed as self-report.

In this study, the validity of socio-demographic characteristics questionnaire was confirmed by content and face validity. The reliability of domestic violence questionnaire was determined by intra-class correlation coefficient (ICC) and internal consistency (Cronbach’s alpha). Cronbach’s alpha and ICC (confidence interval 95 %) were 0.71 and 0.71 (0.47 - 0.85) for negotiation sub-scale, 0.73 and 0.86 (0.74 - 0.93) for psychological aggression sub-scale, 0.75 and 0.81 (0.63 - 0.91) for physical assault sub-scale, 0.76 and 0.72 (0.49 - 0.86) for sexual coercion sub-scale, and 0.71 and 0.87 (0.75 - 0.94) for injury sub-scale of domestic violence.

2.4. Data Analysis

The data were analyzed through SPSS 21. The socio-demographic characteristics were described through several statistical measures such as frequency, percentage, mean, and standard deviation. The relationship between the socio-demographic characteristics and each dimension of violence was determined through the Chi-square test. Then, the variables with P values of lower than 0.05, based on the univariate analysis (Chi-square test), were imported together into the multivariate logistic regression model with a backward strategy. A P value < 0.05 was considered significant.

3. Results

The participants in this study were 558 married 15 - 49 year old women. Table 1 indicates the socio-demographic characteristics of the participants.

3.1. Prevalence of Domestic Violence Against Women and Its Relationship with Socio-Demographic Characteristics

The prevalence of male domestic violence against women in the negotiation, psychological aggression, physical assault, sexual coercion, and injury subscales was 97%, 83.2%, 44.6%, 54.3%, and 22.6%, respectively.

Based on the results of the univariate analysis, the variables of women’s age, number of children, women’s job, spouse’s smoking, children’s gender, women’s satisfaction from their own jobs and their husbands’ job, familial income adequacy, consensual marriage and the way couples first met each other, and first marriage of men had a significant relationship with the dimensions of domestic violence against women (P < 0.05).

The results of multivariate logistic regression revealed that the risk of violence in the negotiation dimension within the age group of 25 - 34 years (odds ratio = 5.2; confidence interval 95%: 1.8 to 15.2, P = 0.002) and the age group of 15 - 24 years (5.1; 1.1 to 23.9; P = 0.043) was 5 times higher than the age group of 35 years and older. Moreover, women who were housewives were lower than the employed women (0.11; 0.1 to 0.4; P = 0.026). The risk of psychological aggression in women with smoking husbands was 2.5 times higher than those with non-smoking husbands (2.6; 1.5 to 4.6; P = 0.001). One of the variables contributing to physical assault was women’s age, where the risk of violence against women in the age group of 25 - 34 (1.9; 1.3 to 3.1; P = 0.002) and 15 - 24 years (1.9; 1.1 to 3.4; P = 0.003) was 2 times higher than women in the age group of 35 years and older. Moreover, the women whose family monthly income was either completely (0.3; 0.2 to 0.5; P < 0.001) or partially sufficient (0.2; 0.2 to 0.8; P = 0.007) faced a significantly lower risk of physical assault as compared to women whose family monthly income was not at all covering the living expenses. Furthermore, women who were married on their own consent faced the lower risk of physical assault than others (0.5; 0.3 to 0.8; P = 0.006). The risk of sexual coercion in women whose husbands experienced their first marriage (0.5; 0.3 to 0.9; P = 0.043) was 50% lower than that in women whose husbands had been married before. Moreover, the risk of injury in women who married on their own consent (0.5; 0.4 to 0.9; P = 0.017) was 50% lower than those who were forced into a marriage (Table 2).

3.2. Prevalence of Domestic Violence Against Men and Its Relationship with Socio-Demographic Characteristics

The prevalence of female domestic violence against men in the negotiation, psychological aggression, physical assault, sexual coercion, and injury subscales was 98.2%, 86.2%, 35.5%, 32.8%, and 44.6%, respectively.

Based on the results of the univariate analysis, the variables of household individuals with the couples, women’s and husband’s age, number of children, spouse’s smoking, children gender, women’s satisfaction from their own jobs and their husbands’ job, familial income adequacy, consensual marriage, first marriage, husband’s substance abuse, and alcohol use had a significant relationship with the dimensions of domestic violence against men (P < 0.05).

The results of multivariate logistic regression demonstrated that the risk of violence in the negotiation dimension for couples living together (without their family) was significantly lower than those living with the family of their spouses (men’s family) (OR = 0.1; CI 95%: 0.1 to 0.4; P = 0.001). The risk of violence for the psychological aggression in smoking individuals (3.1; 1.1 to 5.8; P = 0.001) was 3 times higher than that in non-smokers. The risk of physical assault in the age group of 25 - 34 (3.1; 1.8 to 5.2; P < 0.001) and 15 - 24 years (3.2; 1.7 to 6.2; P < 0.001) were 3 times higher than in the age group of 35 years and older. Moreover, the risk of physical assault in individuals with sufficient monthly income was significantly lower than in those who had an insufficient income (0.4; 0.3 to 0.7; P = 0.005). The risk of physical assault in individuals whose wives had married on their own consent was half of those whose wives had been forced into a marriage (0.5; 0.3 to 0.8; P = 0.004). The risk of physical assault by men who were living together (without their family) (0.4; 0.3 to 0.8; P= 0.017) was significantly lower than men who were living with their family. As for the sexual coercion, the risk of violence in men experiencing their first marriage (0.5; 0.2 to 0.9; P = 0.004) was half of those who had married before. The risk of violence in men whose wives’ age ranged from 15 - 24 years (2.2; 1.2 to 3.9; P = 0.009) and 25 - 34 years (1.8; 1.2 to 3.1; P = 0.008) was almost doubled as compared to those whose wives were 35 years and older. The risk of violence in men whose wives had married on their own consent (0.5; 0.3 to 0.8; P = 0.022) was half of that in men whose wives were forced into a marriage. The risk of violence in individuals with sufficient household monthly incomes (0.5; 0.3 to 0.8; P = 0.009) was half of those with insufficient household monthly incomes. The risk of violence in substance-addicted men was 3 times higher than that in men without substance abuse (3.4; 1.1 to 10.8; P = 0.035). The risk of violence in men who used alcohol (1.8; 1.1 to 2.9; P = 0.018) was 2 times higher than that in non-alcoholic men. The risk of violence in men whose wives were satisfied with their jobs (0.6; 0.4 to 0.8; P = 0.007) was almost half of those whose wives were dissatisfied with their jobs (Table 3).

4. Discussion

This study intended to examine the socio-demographic predictors of domestic violence in women of the reproductive age. In the present study, the male domestic violence against women and vice versa scored the highest rate in the negotiation dimension. The lowest male domestic violence against woman was observed in the injury dimension (22.6%), whereas the lowest female domestic violence against men was observed in the sexual coercion (32.8%). In a study in Sweden, male domestic violence against women was reported in physical dimension, and female domestic violence against men was more frequent in the sexual dimension (15). The results of this study are inconsistent with the findings of the study conducted in Sweden, which may be due to variations in sample inclusion criteria, difference in studied ages (15 - 49 in the present study and 18 - 65 in the study conducted in Sweden) and cultural differences. In a study in Iran, Mohammadkhani et al. (12) reported high levels of male domestic violence against woman (72.2%) and vice versa (77.8%) in the negotiation dimension. The low levels of male violence against women was observed in the sexual dimension (48.7%), whereas the lowest level for female domestic violence against men was observed in the injury dimension (13.5%). The results of the study of Mohammadkhani et al. were also consistent with the present study results. In another study in Tehran, Iran, researchers reported the prevalence of psychological, physical, and sexual violence equal with 85.4%, 31.5%, and 30.4%, respectively (16). The consistency of results between this study and other studies conducted in Iran may be due to similar context and culture.

The women’s age was a predictive variable of domestic violence contributing to the negotiation and physical dimensions of male domestic violence against women, and to the physical and sexual dimensions of female domestic violence against women. The results of a study in America reported a significant relationship between women’s age and sexual abuse (17). In Cyprus, researchers found a significant relationship between women’s age and violence against them (18). An epidemiological study from Sweden reported a significant relationship between the couple’s age and domestic violence (15). The results of a study in Ahvaz, Iran, indicated that age was a factor contributing to the intensity of domestic violence (10). It can be argued that the possible cause of violence in younger ages is the poor understanding of life, a sense of motivation and lack of marital experience and partnership skills, as well as inability to resolve marital disputes.

The household income adequacy was another predictor of domestic violence in the physical and injury dimensions concerning both couples. The results of studies in this regard conducted outside of Iran have suggested a significant relationship between family income adequacy and domestic violence in the psychological, physical, and sexual dimensions concerning both couples (15, 19). Another study in Kazeroon, Iran, reported a significant relationship between family income adequacy and domestic violence (20). In families that are unable to control the conditions arising from financial pressure on life and financial insecurity, low income may lead to stress and tension in couples, thus, setting the ground for disputes and domestic violence.

The consensual marriage of women was another predictor of domestic violence in the physical and injury dimensions concerning both couples. In a study, a significant relationship was reported between forced marriages and domestic violence in the physical, psychological, and sexual dimensions (14). The results of other studies found a significant relationship between women who were forced to get married and domestic violence (21). According to a study in Tehran, Iran, there was a significant relationship between forced marriages and domestic violence against women (22). The results of the study of Rasoulian et al. (23) indicated a significant relationship between marital status and domestic violence in men and women. Marriage is the most important stage of life, where the right choice, based on personal consent, would enhance friendship and passion in any marital relationship, curtailing the domestic violence despite the obstacles and difficulties of a married life.

The men’s substance and alcohol abuse was a predictive variable of female domestic violence against men in the injury dimension, whereas men’s smoking was a predictor of domestic psychological violence against both couples. The results of the study of Coker et al. (24) and the Tlapek study (25) pointed to a significant relationship between alcoholism of one of the partners and experience of violence in both couples. Another study indicated the relationship between smoking and violence among both partners (26). In contrast, the Taherkhani et al. (27) study did not find a significant relationship between substance and alcohol abuse by one of the partners and domestic violence, while a significant relationship was reported between smoking and violence. Other studies have revealed a relationship between substance addiction/alcoholism/smoking and domestic violence (22, 28). One potential reason behind the risk of violence can be associated with the fear of loneliness and rejection, women’s spiritual, psychological, and social insecurity aggravating the women’s stress, thus, increasing the chances of dispute concerning the money spent on substance and alcohol.

Given the high prevalence of domestic violence in Iranian reproductive age women, it is suggested that health policymakers make serious efforts to develop programs to reduce the problem using multifaceted methods including empowering health care providers to identify and take the necessary strategies for reducing intimate domestic violence. In addition, since the couples, particularly women, are more likely to cope with an aggressive lifestyle to protect the family foundations, it is suggested that routine screening be conducted on women admitted to health and medical care centers, thus, to examine the effect of cultural factors, awareness level, attitude, and different perception of individuals toward violence. Moreover, it is recommended that high-risk groups be identified, trained, and empowered to become familiar with their individual rights, mutual respect for others’ rights, and the communication skills in a marital life and how to deal with difficult aggressive situations so as to reinforce self-efficacy, self-esteem, and ultimately help improve their health and curtail the underlying grounds of domestic violence among couples. Conducting studies for identifying effective interventions in order to reduce intimate domestic violence seems to be a research priority.

This study was limited due to the fact that it had a cross-sectional design, i.e. the relationship between domestic violence and socio-demographic characteristics do not necessarily imply causality. Another limitation involved the possibly inaccurate reports given on the violence cases due to women’s tendency to keep marital details in private. In addition, another limitation of this study related to the inclusion criteria of this study was that the women with secondary school education were included in this study and also pregnant and infertile women were excluded of this study. Furthermore, the partners’ psychological problems were considered as exclusion criteria. Thus, the findings of the present study have a low potential for generalization to the women with the low education level or illiterate women, pregnant or infertile women, as well as couples with psychological problems. It is suggested that further researches are conducted in these population groups.

4.1. Conclusion

Considering the high prevalence of domestic violence and its relationship with certain socio-demographic factors in the present study, it is crucial to conduct further research involving various ethnicities in Iran. Research in this field will further identify the socio-demographic factors contributing to violence, based on which culturally appropriate strategies and interventions can be developed to mitigate the domestic violence, thus, leading to the stability of marital relationships and family solidarity.