Sexual Violence and Women Empowerment in India: Findings from a Nationally Representative Sample Survey

Promoting gender equality and the empowerment of women and elimination of violence against women was recognised as an important component in the United Nations 2030 Agenda for Sustainable Development. Intimate partner violence is one of the most common forms of gender based violence throughout the world. Empowering women is an effective measure required to tackle the problem of domestic violence. There are various parameters that are used to measure women empowerment like education, work force participation, women’s decision making capacity in the family etc. In this paper we have analysed the relationship between women’s experience of spousal sexual violence and women empowerment using the ecological model of domestic violence proposed by Heise. We have used the data of the 4th National Family Health Survey (NFHS-4) conducted in India in 2015–2016. Our results show that common empowerment related factors like education was not significantly associated with women’s experience of sexual abuse. Moreover, the likelihood of facing sexual abuse by husband was found higher among working women. We observe that relational and contextual factors like husband’s assertion of control over wife, cultural norms that condone wife abuse significantly increased women’s likelihood of experiencing sexual violence by husband.


Introduction
Violence against women by intimate male partners is a common practice and an accepted reality throughout the world. It is perhaps the most widespread form of gender based violence denying women's basic human rights, dignity, security and self-esteem [1]. Intimate partner violence (IPV), often referred to as domestic violence, encompasses physical and sexual assaults, stalking and psychological abuse by a current or former intimate partner. IPV also includes threats of harm, coercive tactics or arbitrary deprivation of liberty that may occur in public or in private life [2][3][4]. IPV may occur in the same-sex relationship, can be perpetrated by women, but most frequently, women are abused by their partners [3]. In 2017 around 87,000 women were murdered and 58 per cent of them were killed by intimate partners or identifies relationship factors, e.g., marital discord and male control over decision making in the family. The third level encompasses the community level factors such as poverty, socioeconomic status, isolation of women from the natal family etc. The fourth level is associated with the broad societal factors and cultural norms like rigid gender roles or norms granting male control over female behaviour.
To end the violence against women we need to change the patriarchal mindset of people that view unequal power relations as natural. It is not easy. According to Heise et al., the first step for change is empowering women which can be achieved through increased access to education, economic resources, health information and increased participation in decision making by women and girls [3]. Ruth Dixon-Mueller defined women empowerment as "the capacity of individual women or of women as a group to resist the arbitrary imposition of controls on their behaviour or the denial of their rights, to challenge the power of others if it is deemed illegitimate, and to resolve a situation in their favour" [38]. Kabeer viewed empowerment as the processes by which one could gain the ability to make strategic life choices [39]. One essential feature of female empowerment is 'agency' , i.e., women themselves must be significant actors in the process of change [40].
From the above discussion it can be said that women empowerment and intimate partner violence are negatively related. In other words, in a society where women empowerment is higher, the prevalence of domestic violence will be lower. In this paper we have analysed this relationship in the context of India where domestic violence is high and connected to the large number of dowry related deaths each year. Although the reported rate of sexual violence by intimate partners in India is much lower than physical violence but the impact of sexual violence, particularly of coercive sexual intercourse on women is immense as it leads to physical injuries, reproductive health problems and long term mental trauma [41]. Physical violence and sexual coercion compromise women's reproductive autonomy which in turn increases women's risk of unwanted pregnancies and sexually transmitted infections including HIV/AIDS [1]. Sexual assault also increases women's risk of gynaecological problems, depression and suicide [42]. It has been observed that compared to women who only reported physical assault, women who faced coercive sexual intercourse have a lower self-esteem, more negative self-image and are more fearful of sexuality [43,44]. Besides, since the International Conference on Population and Development (IPCD) held in Cairo (1994) [45], demands for the sexual and reproductive rights irrespective of gender have gained momentum [45]. In this context, we are interested to examine the association between women's experience of spousal sexual violence and women empowerment in India using a nationally representative sample survey. In this study we have selected a number of proxy variables commonly used to measure women empowerment and tried to fit them in the ecological framework of Heise [37]. Previous studies from India that have analysed the determinants of domestic violence did not use the ecological model of Heise to explain the factors associated with women's experience of spousal sexual violence.

Data and method
For our analysis we have used the data from the Women's file of the fourth National Family Health Survey (NFHS-4), 2015-2016. It is a nationally representative sample survey which collected data from all the 29 states and 7 union territories of India. We have used the information on background characteristics of the respondents, reproduction, marriage, women's work, husband's background, women's empowerment, other health issues and domestic violence for our study. NFHS-4 provides information Sexual Abuse -an Interdisciplinary Approach of total 699,686 women aged 15-49. However, for domestic violence module 83,397 women were chosen and 79,729 women completed the interview.
Whether a woman has faced spousal sexual violence has been determined by asking them the following questions: Does (did) your current (last) husband ever do any of the following things to you: Physically force you to have sexual intercourse with him even when you did not want to? (i) (ii) Physically force you to perform any other sexual acts you did not want to? (iii) Force you with threats or in any other way to perform sexual acts you did not want to? Women could answer 'yes' or 'no' to each item; If a woman said 'yes' to one or more of these items, she was considered to have experienced spousal sexual violence. In the women's file of NFHS-4, variable D108 provides information on women's experience of sexual violence ever by husband or partner. As only 0.6 per cent never married women experience sexual violence [8] and many of the empowerment related questions were asked only to the ever-married women, we have restricted our analysis to the ever-married women only.
We have used several background variables that help to understand women's empowerment. These are husband's control on selected issues, women's decision making power and level of education, current work status of women and whether women have bank accounts and mobile phones. In the domestic violence module, respondents were asked if they faced six situations in their marital relationship. These are: (i) He (is/was) jealous or angry if you (talk/talked) to other men; (ii) He frequently (accuses/accused) you of being unfaithful; (iii) He (does/did) not permit you to meet your female friends; (iv) He (tries/tried) to limit your contact with your family; (v) He (insists/insisted) on knowing where you (are/were) at all times; (vi) He (does/did) not trust you with any money. The answers are grouped as 'no' and otherwise (yes or do not know). If the answer is 'no' in all six situations, then it is designated as 'no control' . If the answers are 'otherwise' for 1-2 cases, then it is 'less control' and if the answers are 'otherwise' for 3-6 cases, then it is designated as high marital control over women by their husbands. During the NFHS-4 survey, women were asked if they justify wife beating in the 5 given situations: (i) if wife goes out without telling husband; (ii) if wife neglects the children: (iii) if wife argues with husband; (iv) if wife refuses to have sex with husband; and (v) if wife does not cook food properly. If a woman supports wife beating in any of the given situations, then we conclude that wife beating is justified by that woman.
Cross tabulation and Pearson's Chi square test have been used to examine the bivariate relationship between marital sexual violence and background characteristics of women. Multivariate analysis has been applied to find out the net impact of the variables related to women empowerment on sexual abuse experienced by women. All analyses are done using SPSS version 21.0. Table 1 shows the distribution of women selected for our analysis by background characteristics. The selection of background characteristics were based on extensive literature review [3, 4, 11, 14, 17, 19, 24-26, 29, 31-33, 43, 46-53]. In our study 6.7 women (4372 of total 66,013 samples) had experienced sexual violence by their current/last husbands. Among the respondents, 17.3 percent were below 25 years of age and 45.1 per cent were in the age group 35-49 years. More than two third women reported to have 2 or more living children. Less than one-fifth women completed secondary education. Majority of the women were Hindu and  belonged to non-Scheduled Caste/Scheduled Tribe category (The Scheduled Castes and Scheduled Tribes are officially designated groups of depressed classes in Indian society). More than one third women witnessed beating of their mother by their father. Twenty two per cent women confessed that they were afraid of their husbands and more than one fourth of the respondents experienced physical violence by their husbands. Around half of the women said that they did not face any marital by their husbands on selected issues but 17.5 percent women reported high control. Almost two third women reported that they used to take the following decisions jointly with their husbands: about their own health care, about major household purchase, about visits to their family/relatives and about the use of husband's earning. 57 per cent of women did not support wife beating. Thirty per cent of women reported that their husbands consumed alcohol. At the time of the interview, three fourth of the women said that they were not working and 4.3 per cent women reported that their husbands were not working. Almost half of the women did not possess any mobile phone and 46 per cent of women did not have bank account. As per wealth quintile, more or less 20 per cent women belonged to each quintile. According to NFHS-4, India has been divided into six regions [8]. The highest number of respondents were from Central India (23.4 per cent) followed by North (

Results from bivariate analysis
In this section we have presented the results of bivariate analysis. Also, the variables have been categorised at four levels according to the general ecological model of IPV described by Heise [37]. It is evident from Table 2 that women's experience of spousal sexual violence did not vary much with age. Those who had three or more children have experienced higher rate of sexual violence. Those who completed secondary education, 3.8 per cent of them experienced sexual abuse by their husbands. On the other hand those who did not complete secondary education, 7.4 percent of them complained about spousal sexual violence. Muslim, non-SC/ST women reported lower rate of sexual violence. The percentage of women who experienced sexual violence was 2.6 times higher among those who witnessed intergenerational violence than who did not. Working women complained about higher sexual abuse by husbands than non-working women. The highest rate of sexual violence was reported by the women whose husbands were not working. The percentage of women experiencing spousal sexual violence was more than three times higher for those who reported consumption of alcohol by their partners. The percentage of women who experienced spousal sexual violence was lower among those who had bank account and mobile phone. Table 3 presents the bivariate association between women's experience of spousal sexual violence and other covariates indicating relational aspects between husband and wife. The percentage of women who experienced sexual violence was 2.7 times higher among those who were afraid of their husbands. The percentage of women who faced sexual violence was almost 12 times higher among those who ever experienced spousal physical violence compared to the women who did not. Women who reported higher marital control on specific issues, 19 percent of them complained about sexual violence by their husbands whereas those who did not face any control on those issues, only 2.1 per cent of them reported spousal sexual violence. The percentage of women facing sexual abuse was lower for those who said they jointly took decisions with husbands about their own heath care, household purchase, visit to family/relatives and the use of husband's earning.   From Table 4 it is evident that the women who justified wife beating in certain circumstances experienced higher rate of sexual violence. The prevalence of sexual abuse in marital relationship was found the highest among the poorest group and the lowest among the richest group. The prevalence of sexual violence was the highest in eastern part of India and the lowest in the western part of India.

Results from multivariate analysis
In the above section we have discussed the bivariate association between women's lifetime experience of spousal violence and background characteristics of women. To understand the net effect of women's empowerment related parameters on women's experience of sexual violence we have applied binary logistic regression. On the basis of Pearson's correlation coefficient value (r), we found that some variables are highly correlated such as age group of women and number of living children (r > 0.5). Decision making on women's health, household purchase, visit to women's family and relatives and the use of husband's earning show strong correlation among them (r > 0.5). Also women's level of education and wealth quintile, women's possession of mobile phone and wealth quintile, women's possession of mobile phone and bank account show moderate correlation. Keeping these into account, we have not used all the background variables selected earlier for the multivariate analysis. Table 5 presents the results of binary logistic regression analysis predicting the probability of a woman experiencing spousal sexual violence. Odds ratio greater than one indicates a positive relationship between the independent variables and the probability of experiencing sexual abuse by husband, and odds ratio less than one indicates a negative relationship. From the logistic regression analysis in Table 5, it is observed that age of the women, their level of education and wealth quintile are not significantly associated with women's experience of spousal sexual violence. Muslim women show significantly higher probability of experiencing sexual violence by their husbands compared to the Hindu women (OR 1.228, p < .01). Interestingly, in the bivariate analysis we found the opposite result. Another study from India also found similar result [33]. As in the multivariate analysis, other factors are controlled; we may find such contradictory results. It indicates that any/some of the background factors have more influence on marital sexual abuse than religion in case of Muslim women. Caste of women does not have significant impact on spousal sexual violence. Currently working women show higher odds of being Sexually abused by husbands but occupation of husband was not significantly associated with sexual abuse of wife. Also women who had bank account, their likelihood of facing sexual abuse was lower compared to their counterparts (OR 0.874, p < .001). After covariate adjustment, it was found that the risk of sexual violence was significantly higher for those who reported consumption of alcohol by their husbands (OR 1.689, p < .001). Women who witnessed abuse of their mother by their father, who suffered spousal physical violence, who experienced higher spousal control on several issues, who reported that they were afraid of their husbands, who justified wifebeating were more likely to have experienced marital sexual violence. The results are significant at <0.001 percent level of significance. Compared to the women who alone used to take decision about own health care, those who jointly took decision with husbands were significantly less likely to face spousal sexual violence when other factors are controlled. Finally, compared to the women of North India, the women of Eastern and North-Eastern India showed significantly higher probability of facing spousal sexual violence but the women of Western and Southern India had significantly lower probability of experiencing sexual abuse by their husbands..

Discussion
The ecological model of Heise provides a comprehensive framework to understand the factors associated with IPV at various levels. In the first level biological and personal factors like women's age, level of education, religion, caste, labour force participation, own bank account, husband's occupation, are included in our analysis along with the women's intergenerational experience of violence and the consumption of alcohol by husband. Several studies from different parts of the world have found strong correlation between the last two factors and domestic violence [8,14,17,49,51,[53][54][55]. Young age is regarded as a risk factor of experiencing   56]. Education enhances women's cognitive ability, self esteem, and participation in decision making; therefore, educated women are less likely to suffer from domestic violence [33,39,56]. However, the positive influence of education frequently disappears when factors associated with relationship stressors are controlled [57]. In our study we have found that age and women's education are not significantly related to sexual abuse of wife. Like education, women's participation in work, particularly paid work is also regarded as crucial factor for women empowerment [39]. It is assumed that labour force participation helps women to attend financial independence. Therefore it is expected that working women are less likely to experience domestic violence. Interestingly, our results are just the opposite. Previous studies from India also found similar outcome [33,58]. This finding does not fit into the general notion that women empowerment through work force participation will reduce IPV. Actually IPV is a way of asserting male authority on women. As working women acquire some sort of independence by joining the labour force, it poses a challenge to their partners. To keep women under their control, violent measures are used by men. Therefore, only enhancing women's opportunity to education and work is not enough for battle against domestic violence.
Although owning a bank account is not a good indicator of women empowerment as husband alone can handle the account, nevertheless, after controlling the effects of other factors, it shows negative association with wife's experience of sexual abuse.
In fact, owning a bank account enhances women's sense of financial security.
In the second level, the variables expressing the power relation in the family (women's experience of spousal physical violence, husband's marital control on selected issues, decision making power of women regarding own health and whether women are afraid of their husband) are taken into account. Husband's assertion of control over wife is the manifestation of patriarchal mindset. Women empowerment is closely linked with women's autonomy. Women's autonomy is their ability to determine events in their lives [59] and like the control issues, women's decision making power regarding their health, household purchase, visit to family and relatives, use of husband's earning are closely associated with women empowerment. It has already been mentioned that the variables related to women's autonomy are highly correlated. Therefore, we have used one variable in the logistic regression model, i.e., women's decision making power regarding their health. As health has both intrinsic and instrumental values, decision making power regarding own health is the most important factor compared to other issues mentioned above. Physical violence by husband comes under relationship factors because men use it as an instrument of power by which women are dominated and inequality between men and women are maintained [33]. Besides, literature reveals that sexual abuse often accompanied by physical violence [3]. When women report that they are afraid of their husbands, it indicates substantial lower position of women in gender based power relation.
Under community level factors (third level) we have included attitude of women towards wife beating and wealth index. Previous studies from developing countries have also considered that the community wife beating norms are closely associated with IPV [17,24]. It is an important contextual-level variable affecting spousal violence. In our analysis we have found that 43 per cent of women supported at least one of the wife beating issues. It reflects conservative attitudes of the society towards gender norms that endorses IPV. As a result, women themselves support wife beating for trivial issues like food is not properly cooked. The socioeconomic status of a family influences both personal and relationship factors of domestic violence. Poverty and unemployment fuels marital conflicts which ultimately led to domestic violence. However, we have found that economic status of a family does not have significant impact on women's experience of sexual violence controlling for other factors. The fourth level represents the broad societal factors. We have put region of India under this level. Previous studies have focussed on the sharp contrast between the North and South India regarding social and cultural norms, level of education and fertility rates [59][60][61]. In fact, in India women's status and vulnerability varies from region to region [62]. There is a general notion that women enjoy higher autonomy in North-East India. However, one study found that it is true only for selective indicators [63]. Although in general Indian society is patriarchal by nature, prevalence and manifestation of wife abuse varies across regions due to regionspecific cultural norms and traditions. Our studies found that women from East and North-East are more vulnerable to sexual violence under marital relationship.

Conclusion
In India domestic violence by husband is a common phenomenon. Using the ecological model of Heise we have tried to find out the association between spousal sexual violence and women's empowerment related variables. Education and workforce participation are often used as proxy to women empowerment but we find that controlling for other relational, community level and broad societal determinants of violence, education and work force participation are not found as protective factors against sexual violence by husband in India. Therefore, empowering women through increased access to education and work will not produce desirable outcome in combating domestic violence, unless and until policy measures take into account the broader cultural norm that view unequal power relation as natural and normal. We think this finding has an important policy implication.
India is a diverse country and cultural norms and traditions vary widely from region to region. Most of the researches on domestic violence concentrate on North-South differences, while the highest prevalence of physical violence as well as sexual violence was recorded in East India [8]. We also observed that after covariate adjustment, women of Eastern India was the most vulnerable to experience spousal sexual violence. Therefore, special strategies should be prepared for this region while formulating policies and programmes to end violence against women in India.

Conflict of interest
There is no conflict of interest as there is no co-author.
© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.