The Effect of Women’s Empowerment on Intimate Partner Violence and Child Nutrition Outcomes in India,Nepal,and Pakistan

Abstract

Women’s empowerment is often defined to include aspects of agency, autonomy and choice, which in turn has consequences for facing intimate partner violence (IPV) and the ability of a woman to fulfil childcare responsibilities. This suggests that empowerment is directly and indirectly (via IPV) associated with child nutrition outcomes (CNOs), especially in South Asian countries where gendered norms may place the onus of childcare on mothers. We explore the interplay between empowerment, IPV and CNOs using nationally representative datasets from three South Asian countries—India, Nepal and Pakistan. We use a multivariate probit approach to investigate the direct and indirect effect of women’s empowerment and autonomy on child malnourishment (stunting, wasting and underweight). Across all countries, we find a strong statistically significant effect of improvements in decision-making power on increased likelihood of facing certain types of IPV. We also find a strong negative relationship between facing less severe violence in particular and CNOs across all countries, indicating that such violence experienced by mothers was detrimental to CNOs. Increasing women’s decision-making power within the household can help ameliorate adverse CNOs, and in India particularly, this increase in decision-making autonomy reduced the incidence of stunting and underweight children. The study concludes with limitations and directions for future work.

Keywords

Intimate partner violence child nutrition women’s empowerment South Asia

Introduction

The United Nations asserted in 1995 that women’s autonomy is essential to human dignity and must be considered a basic human right (UNDP, 1995). Moreover, various terms, including ‘status’, ‘autonomy’ and ‘empowerment’, have been used in literature to illustrate gender equality in the household and community.

Empowerment has been variously conceptualized across disciplines to encompass aspects of decision-making autonomy, choice, agency and status in contexts such as the household, marriage and political representation, among others. For instance, Kabeer’s (2021) concept of empowerment begins with the idea of choice and defines power as the capacity to make choices with positive consequences. Ibrahim and Alkire (2007) note that agency and empowerment are experienced differently in different settings and contexts. They argue that these concepts can be described and measured with respect to different domains of life. One such domain that we explore in this article examines the ways in which women who are mothers experience agency and empowerment within their households. Specifically, we explore the link between this form of empowerment and outcomes related to intimate partner violence (IPV; Paul, 2016) and downstream implications for child nutrition (Chai et al., 2016).¹

In particular, when studying women who are mothers, the concept of empowerment has also been employed by several studies to capture behaviours such as decision-making and mobility that may or may not be under the mother’s control with specific reference to resources directed towards childcare. Mothers are more likely to use scarce resources for the benefit of their child if they are free to decide to do so (Kabeer, 2021). For example, Mistry et al. (2009) found that mothers make greater use of pregnancy services if they have the ability to make decisions. Decision-making ability is often exercised through higher financial autonomy (Wayack Pambè et al., 2013). Financial autonomy allows for women to have higher negotiation power, which affects purchasing decisions and resource allocation to food and childcare, particularly in resource-poor settings (Cunningham et al., 2015; Jejeebhoy & Cook, 1997).

In various studies, women’s empowerment (WE) has also been reflected in intra-household resource allocation, greater decision-making power, absence of domestic violence or a combination of these factors (Kumar et al., 2018). In this article, we examine the effect of WE on the odds of facing emotional, physical and sexual IPV among married women with children in select South Asian countries (India, Nepal and Pakistan).

International research has provided increasing evidence of the prevalence of violence against women, particularly physical violence perpetrated by intimate male partners (Garcia-Moreno et al., 2006). According to the World Health Organization (WHO), one in three women worldwide have experienced physical and/or sexual IPV in their lifetime (World Health Organization, 2017). IPV affects women’s social and economic development, as well as their capacity for self-determination (Koenig et al., 2003). Recent research has focused on the antecedents of IPV, with poor WE being highlighted as an important cause. Limited access to empowerment factors make women more tolerant of IPV, hence perpetuating a cycle of violence and abuse (Simona et al., 2015). Furthermore, we assess the direct and indirect effects of empowerment on child nutrition outcomes (CNOs) in these countries, as well as the association of empowerment and CNOs in select South Asian countries.

Empowerment can have both direct and indirect links to CNOs. One argument for the direct link is attributed to traditional gender norms of mothers being the primary caregivers in most developing countries. The autonomy of the mother allows her to access, gain control of and make use of resources that provide effective nutrition and thus impact child growth. Several studies have observed specific dimensions of WE impacting child growth and survival (Bhagowalia et al., 2012; Burroway, 2016; Carlson et al., 2015; Shroff et al., 2012). There can also be indirect impacts of empowerment which vary by whether the woman faced any form of IPV. Negative consequences of IPV against women may also affect their children’s health (Ackerson & Subramanian, 2008). IPV can negatively affect a child’s health indirectly by impairing maternal caretaking abilities.

Past literature highlights complex direct and indirect pathways between WE, IPV and CNOs. However, no holistic understanding is presently available of the interlinkages between these variables. Therefore, this study proposes to address the following research questions:

RQ1: What is the effect of WE on the odds of facing emotional, physical and sexual IPV among married women with children in select South Asian countries?

RQ2: How does the experience of facing any form of IPV affect CNOs in terms of stunting, wasting and being underweight?

RQ3: Are there additional effects of WE on CNOs, over and above those via changes in odds of facing IPV?

The remainder of this article is organized as follows. The second section outlines the literature on links between WE, IPV and CNOs in the South Asian context. The third section contains the conceptual model, data and empirical methods employed in the article. The main results are presented in the fourth section. The fifth section provides implications and concluding remarks.

Literature Review

WE and IPV

The association between WE and IPV is a two-way process. Whether a woman has experienced IPV or not provides insight about the ability of her household to promote her empowerment, whereas knowing whether or not she has a say in decision-making tells us something about the extent to which she is already empowered within the household (Wayack Pambè et al., 2013). Studies have also found that greater control over resources by wives was associated with a significantly lower risk of domestic violence (Jejeebhoy & Cook, 1997; Koenig et al., 2003).

Patriarchal social norms and gender relations are important to the South Asian context. Family systems and societies designate men as providers and enforce compliance of women to patriarchal norms (Solotaro & Pande, 2014). However, educated women are less likely to suffer from domestic violence. A study in West Bengal noted that educated women were better able to deal with violent husbands, suggesting that access to higher levels of education may have an important role in contributing to women’s sense of self-esteem and importance (Kabeer, 2005).

Schuler et al. (2017) from their study in four villages in Bangladesh found that there are various factors that discourage IPV. These factors included the recognition and value of women’s potential and actual economic contributions. This factor was also seen as supporting WE in the villages. However, they also found that WE might be a risk to them, especially if their empowerment was social in nature without being economically empowering.

Women earning income was associated significantly with experience of higher physical violence in India and rural Bangladesh. Also, in India, women who were responsible for household expenses, compared to when male partners were responsible for household expenses, experienced higher levels of physical violence, while joint responsibility implied significantly lower violence. In India, a woman’s ownership of a house or land was significantly associated with lower levels of violence experienced by them (Vyas & Watts, 2009).

Panda and Agarwal (2005) in their analysis of 500 households among urban and rural areas of Kerala, India, found that among women who did not own property or land, 49% experienced physical violence and 84% experienced psychological violence. In contrast, women who owned both experienced less physical and psychological violence, and when the women owned either a house or land, the incidence of violence was much lower than if they owned neither.

Using seven dimensions to assess the level of WE, including asset ownership and involvement in household decision-making, Jamal (2017) analyses the relationship between WE and spousal physical violence in Pakistan. The study finds that empowerment dimensions such as women’s involvement in household decision-making may increase the risk of violence. Women’s ownership of land and house is negatively associated with risk of violence.

In a systematic review of literature globally, Waltermaurer (2012) found that in Nepal, the prevalence of IPV justification was below 10%, and, for females, arguing back was one reason. However, a quarter of Nepalese females and males felt that IPV was justified when the issue of child neglect took place. Additionally, in India, infidelity justified IPV equally or less than when a woman goes out without her husband’s permission.

WE and CNOs

In their seminal paper, Ramalingaswami et al. (1996) proposed that the extremely low status of women is a key factor responsible for the high rates of child undernutrition in South Asia. They hypothesized that women’s low status not only comprises maternal physical health, resulting in low infant birth weight, but may also lead to a subsequent decline in quality of care provided during infancy and childhood. The implication is that even a woman with sufficient knowledge or resources, accrued because of her education or socio-economic status, would be unable to employ these skills to her child’s benefit if she is not empowered to do so. The authors suggested that if women’s status (i.e., empowerment status) were improved, there would be subsequent improvements in the nutritional status of their children. They argued that the low quality of care given to girls and women by their families is a major determinant of the high levels of child malnutrition in South Asia as compared to global CNOs (Ramalingaswami et al., 1996).

Similarly, Engle (1993) found that mothers with a higher contribution of money to the family income had children with significantly better nutritional status, while Shroff et al. (2012) noted the strong positive influence of the mother’s financial autonomy on child stunting.

Consequences of IPV on CNO

IPV is more likely in early marriage and pregnancy (Kishor & Johnson, 2004), and young children, who rely on their mothers as primary caretakers, are likely to be exposed. Silverman et al. (2009) in their study on maternal experiences of IPV and their effects on child morbidity in Bangladesh find that maternal experiences of IPV are associated with leading causes of childhood mortality. This suggests that the IPV not only affects women but their ability to look after the well-being of and provide for their children as well.

More recently, Chai et al. (2016) conducted a comprehensive study of domestic violence on CNOs in low- and middle-income countries on which Demographic and Health Survey (DHS) data is available. They found that exposure to IPV in any form increased the odds of child stunting in a sample of 29 countries, including data from India (2005) and Nepal (2011). Notably, this study did not examine the role of WE as a protective factor for IPV. From a study in Pakistan using nationally representative data, it was found that there is a significant positive association between malnutrition in children under five years of age and emotional violence experienced by the mother. This relationship was evident for being underweight and for stunting among children and women in rural areas who had children with higher levels of malnutrition, due to a higher incidence of IPV (Shaukat et al., 2018).

Violence affects a woman’s emotional and physical health, which leads to diminished incentives to pursue appropriate maternal healthcare (Simona et al., 2015). The health consequences of IPV extend beyond mothers and translate into health consequences for their children. Many researchers consider this to be an important causal factor for poor CNOs across the world (Ackerson & Subramanian, 2008; Rico et al., 2011; Ziaei et al., 2014). Additionally, poor maternal autonomy and empowerment have also been seen to have a negative consequence on the child’s health. Women with fewer resources and low autonomy do not have the bargaining power to allocate resources to their child’s development (Engle, 1993; Shroff et al., 2012); this is likely to be compounded in violence-complicated relationships.

Interlinkages Between IPV, CNO and WE

A wide range of physical and psychological problems induced by IPV can make abused mothers physically and emotionally less capable of providing their children’s basic needs (Ziaei et al., 2014). Despite the growing recognition of IPV, there are not enough studies that explore the relationship between IPV and CNO, especially in developing countries (Rico et al., 2011).

Ackerson and Subramanian (2008) found that maternal exposure to multiple incidences of physical family violence in previous years was associated with stunting, wasting and underweight among their 12–35-month-old children. They provided a reason for this association: as domestic violence is strongly associated with a woman’s inability to make decisions for herself and her family, this has an impact on the decisions she makes about the types and quantities of food that she prepares as she cares for herself and her children (Ackerson & Subramanian, 2008). In a cross-country analysis using DHS data, Rico et al. (2011) found that when IPV was considered by type, associations remained with stunting and malnutrition, although there were no consistent trends whereby specific types of violence were more strongly associated than others with the CNOs.

Finally, WE, as measured through justifications of domestic violence, also has significant associations with child mortality levels (Kravdal, 2004). Ackerson and Subramanian (2008) describe IPV as an indication of weak bargaining power of wives against husbands and note the increased prevalence of stunted children in violence-affected families. Thus, a child’s health is linked to the health environment in which the child lives, but ultimately it is dependent on the caregiver’s use of facilities and services to optimize the child’s health. The ability of caretakers, usually their mothers, to provide care to children ultimately rests upon the quality of the care they receive (Smith et al., 2003). In our study, we used measures of malnourishment such as wasting, stunting and underweight for children under five years of age, collectively referred to as CNOs.

This study aims to provide crucial insight into the pathways that influence CNOs through the mother’s empowerment status and IPV faced by her. This study acknowledges the importance of the role that women play in the household as mothers and caregivers for the family and builds a link between mothers’ experience of violence and their children’s resultant health. It is also important to understand the effects of the violence that mothers experience on their children’s nutritional outcomes.

Conceptual Framework, Methods and Data

This study explores the relationship between empowerment, IPV and CNOs using the nationally representative datasets of three South Asian countries—India, Pakistan and Nepal. Lower empowerment or autonomy of women in select South Asian countries affects CNOs directly, and indirectly via an exposure to IPV.

Child malnutrition and gender equality (which include WE and IPV faced by women) are key issues addressed in the United Nations Sustainable Development Goals (UNSDGs) 3 and 5, respectively. To understand and effectively work towards attainment of the UNSDGs, the pathways that connect WE, IPV and CNOs must be better understood. This study tests a new conceptual framework in which both women’s experiences of IPV and their children’s nutritional status are influenced by the level of autonomy or empowerment the women within that household possess.

This importance is echoed by Yount et al. (2011) in their conceptual review of the pathways of influence between domestic violence and its influence on child nutrition: DHS data on women and children could be matched to construct retrospective cohorts in multiple countries. Although the DHS data lack some potential mediators (maternal mental health) and confounders (co-occurring child maltreatment), partial path models could be estimated. Ultimately, large, population-based intergenerational studies of diverse populations are needed to assess fully the pathways by which domestic violence affects the nutrition and growth of children.

A conceptual framework was developed to represent the possible linkages among the different sets of variables included in the study (Figure 1). The main outcomes of interest were child health outcomes, with nutritional status of children as the main indicator. The model conceptualized the health of children to be determined through the interplay of a set of covariates such as household and individual characteristics, maternal empowerment indicators and reporting of IPV. It is important to note that there is a possibility of reverse causality in this framework. For example, exposure to violence may lead to changes in WE indicators. However, recent work examining the links between these two variables, particularly in the Indian context, suggest that women who are more empowered are less likely to be subject to IPV (Ahmad et al., 2019; Panda & Agarwal, 2005), and that exposure to this varies by alcohol consumption by the husband (Parekh et al., 2021). However, this evidence varies by the definition of empowerment, which has been highly variable studies mainly due to differing theoretical foundations. Furthermore, it is also important to acknowledge that CNOs and exposure to IPV may have a common source (e.g., socio-economic status or poverty), and that we are able to control for this using a proxy for wealth as detailed in the next section.

Figure 1.

Source: The authors.

Methods

The study’s empirical analysis aimed to identify and test the pathway of mother’s empowerment indicators and experiences of IPV on the child’s nutritional status, using nationally representative household data from three South Asian countries. We use mother-level data at the household, combined with aggregate child outcomes and other individual and household characteristics.

Data

The study used data from the latest round of the DHS of three developing countries in South Asia—India (2015–2016), Pakistan (2011) and Nepal (2011–2012). The datasets were accessed via DHS data portal. These countries are considered a part of South Asia and satisfy criteria for data availability of relevant variables from the DHS. Other countries in the South Asian region where DHS data is available include Afghanistan, Bangladesh and Maldives. However, these surveys did not have data on either violence faced by women or CNOs and, hence, were not included in this analysis.

The surveys use a standardized multistage sampling procedure with weighting to reproduce nationally representative data on maternal and child health. More information on methodology and data collection protocols is available in the respective DHS report. The analysis is carried out at the level of the household, including only those households where a mother and a child were present and fulfilled the inclusion criteria. Data on the indicators in Figure 1 were obtained from the individual recode, domestic violence and child-level questionnaires. Table 1 indicates the number of observations available for each country.

Table 1.

Number of Observations, by Country, Available in the Latest Rounds of the DHS.

Country	Number of Respondents Interviewed for Domestic Violence Module	Number of Children for Anthropometric Measures
Pakistan	3,475	4,226
Nepal	1,938	2,379
India	34,350	236,455

Source: Country Demographic and Health Surveys.

Explanatory Variables

Factors indicating WE included variables assessing ownership of financial assets (land or property held in the female’s name), justification of violence and decision-making power (i.e., the person who usually decides how to spend respondent’s money—alone and/or with other vs other). This grouping of variables is similar to Kabeer’s (1999) proposed three interrelated dimensions of empowerment, that is, the ability to exercise choice: (a) resources (preconditions), (b) agency (process) and (c) achievements (outcomes). Jones et al. (2019) uses Kabeer’s (1999) framework to analyse DHS in Ethiopia, Kenya, Rwanda, Tanzania and Uganda, which is a different sociocultural context from South Asia.

Participation in household decision-making was measured as an aggregate of variables assessing responses to items such as a person who usually decides on respondent’s healthcare, a person who usually decides on large household purchases and a person who usually decides on visits to family or relatives. The responses were coded as ‘yes’ if the woman had answered ‘self’ or ‘self and other’ for all the items (yes vs no). Justification of IPV was an (inverse) measure of WE and was measured as an aggregate of five variables assessing women’s justification of wife beating in different scenarios. Several studies construct indices that aggregate various indicators of WE to minimize using a large number of variables in analyses. However, it is often challenging to find theoretically cohesive measures of WE that include these variables. Additionally, it is statistically challenging to determine reliability of a composite measure. In our dataset, we attempted to construct a composite index encompassing measures of WE but find a low Cronbach’s α across countries as well as in the pooled sample (α_Ind = 0.08; α_Nep = 0.26; α_Pak = 0.12; α_overall = 0.15), indicating low statistical reliability.² We thus retain the variables as individual measures for our analyses.

The domestic violence module of the DHS is administered to one, randomly selected, woman per household. The modified conflict tactics scale (CTS) is used to collect data on physical, emotional and sexual violence. The respondents are asked questions about their experiences of specific acts of physical, sexual and emotional violence perpetrated by a current or former partner. Questions vary slightly by country but largely cover similar aspects of the different types of violence. Three aggregated variables of incidence of violence were employed in the analysis: experiences of sexual violence, severe physical violence and emotional violence in the preceding 12 months.

Outcome Variables

This article employed three standardized indicators of child nutrition based on anthropometric measures computed for children between 0 and 5 years of age: (a) height for age (stunting), (b) weight for height (wasting) and (c) weight for age (underweight). These scores were standardized to Z-scores and measure standard deviations (SD) from the median of a reference population (WHO Child Growth Standards). Children were considered as ‘severely stunted’ if height-for-age Z-scores were below –2 SD. Similarly, Z-scores below –2 SD for weight-for-height were considered as ‘severely wasted’ and ‘severely underweight’ for weight-for-age. In each household, we computed average Z-scores for each measure (and child characteristics) for all children for the corresponding mother.

Covariates/Confounding Variables

Keeping in line with Ramalingaswami et al.’s (1997) commentary, this study employed the mother’s educational status as well as the husband’s educational status as covariates. Additionally, the mother’s health had been found to be another determinant of child’s nutritional status; children of those mothers who had poor health as expressed by low body mass index (BMI) are more likely to be nutritionally poor (Sethuraman et al., 2006). Thus, maternal BMI was also included as a confounding variable to represent mother’s health status. A priori confounds such as household position on the country’s wealth index,³ number of children under the age of five years in the household, and urban or rural residence were included due to their relation to CNO, based on previous findings (Ahmed et al., 2010; Frongillo et al., 1997; Imai et al., 2014; Pongou et al., 2006; Rico et al., 2011; Shroff et al., 2009; Sinha & Chattopadhyay, 2017; Smith et al., 2003; Underwood, 2002). We also included a variable that captures whether the mother is currently working or not. This is because female labour force participation can affect CNOs through higher incomes as well as financial autonomy (see Sangwan & Kumar, 2021).⁴

The summary statistics for all variables can be found in Table 2.

Table 2.

Summary Statistics Across Countries on Dependent, Explanatory and Control Variables.

	India		Nepal		Pakistan
	Mean	SD	Mean	SD	Mean	SD
Control variables
Age	27.56	5.041	26.63	5.447	29.65	6.084
Household size	5.755	2.211	5.373	2.118	8.027	3.710
Number of children five years of age and under	1.659	0.847	1.518	0.754	2.125	1.192
Educational attainment	2.124	1.684	2.001	1.831	1.634	1.952
Wealth index	2.664	1.373	2.625	1.346	2.733	1.375
Age at first birth	21.30	3.701	19.89	3.441	21.15	3.938
Marriage to first birth interval (months)	43.18	127.3	32.50	78.87	31.60	89.68
Proportion currently working	0.178	0.382	0.550	0.498	0.123	0.329
Husband’s educational attainment (years)	7.85	7.23	6.74	5.63	7.13	6.59
Interpersonal violence
Proportion experienced any emotional violence	0.127	0.333	0.130	0.336	0.341	0.474
Proportion experienced any less severe violence by husband/partner	0.293	0.455	0.220	0.414	0.284	0.451
Proportion experienced any severe violence by husband/partner	0.0822	0.275	0.0996	0.300	0.0723	0.259
Proportion experienced any sexual violence by husband/partner	0.0731	0.260	0.0831	0.276	0.0487	0.215
Empowerment indicators
Proportion owning property alone or jointly	0.316	0.465	0.0748	0.263	0.0167	0.128
Average decision-making domains for wife	0.161	0.395	0.740	0.337	0.289	0.246
Average justification of violence	1.152	1.406	0.531	0.303	1.625	1.031
Child characteristics and nutrition outcomes
Proportion of female children	0.479	0.500	0.486	0.500	0.501	0.500
Average age of children	2.063	1.400	2.067	1.406	2.056	1.411
Proportion stunted (WHO, height-for-age)	0.336	0.472	0.344	0.475	0.335	0.472
Proportion wasted (WHO, weight-for-height)	0.178	0.382	0.0867	0.281	0.0696	0.255
Proportion underweight (WHO, weight-for-age)	0.301	0.459	0.251	0.434	0.207	0.405
Observations	34,350		1,938		3,475

Source: Country Demographic and Health Surveys.

Empirical Framework

The study used a multivariate probit approach to investigate the direct and indirect effects of WE and autonomy on child malnourishment (Figure 1). The estimation framework allowed WE to affect child malnourishment via IPV as well as independent of IPV and assumes implicitly that all types of violence can simultaneously (rather than individually) be associated with adverse CNOs.

The binary discrete variable(s) are given by Violence_i, indicating whether or not a women faced the ith type of violence:

\begin{array}{l} V i o l e n c e_{i w v} = 1 i f γ Z_{i w v} + μ_{i w v} >_{} 0; i = (E, L S, S, S E) \\ V i o l e n c e_{i w v} = 0 i f γ Z_{i w v} + μ_{i w v} \leq 0 \end{array}

where, the ith type of violence (E refers to emotional, LS refers to less severe, S refers to severe and SE refers to sexual) is faced by the wth woman residing in the vth household. We wish to examine the effects of facing violence on CNOs, which we also describe as a vector of binary variable as follows:

\begin{array}{l} C h i l d *_{1 i w v} = β_{1} X_{1 i w v} + \int_{1 i w v} C h i l d_{} 1 i w v = I (C h i l d *_{1 i w v} > 0) \\ C h i l d *_{0 i w v} = β_{1} X_{0 i w v} + \int_{0 i w v} C h i l d_{} 0 i w v = I (C h i l d *_{0 i w v} > 0) \end{array}

Note that we wish to link this to the outcomes associated with IPV. Thus,

\begin{array}{l} C h i l d_{i w v} = C h i l d_{1 i w v} i f V i o l e n c e_{i w v} = 1 \\ C h i l d_{i w v} = C h i l d_{0 i w v} i f V i o l e n c e_{i w v} = 0 \end{array}

In any multivariate probit specification, we assume that the error terms are correlated and follow a specific covariance matrix structure with mean zero and some correlations, typically a multivariate normal distribution (Ramful & Zhao, 2009). The univariate marginal probability (i.e., the chances of facing a particular type of violence) is given by:

P (C h i l d_{k i w v} = 1 | V i o l e n c e_{i w v}) = Φ_{1} (V i o l e n c e'_{i w v} β_{i w v}); (i = E, L S, S, S E)

where, Φ₁ is the cumulative univariate normal distribution function for the standard normal distribution. Similarly, multivariate joint probabilities can be derived from these, and we refer the reader to Ramful and Zhao (2009) for more details. The reduced (simplified) form of the estimating equations is as follows:

V i o l e n c e_{i v} = α + β_{1} W E_{i v} + β_{} 2 X_{i v} + η_{i v}

(1)

C h i l d_{i v} = δ + γ_{1} V i o l e n c e_{i v} + γ_{} 2 K_{i v} + γ_{} 3 W E_{i v} + \int_{i v}

(2)

Here, Violence_iv is a vector of binary variables that indicate if a woman faced physical, sexual or emotional abuse; WE_iv is a vector of variables that proxy for women’s autonomy and empowerment; X_iv is the vector of covariate variables that identify Equation (1) uniquely—these could include household-specific characteristics that do not directly affect Child_iv. As mentioned previously, these are her educational attainment, her partner’s educational attainment, her age, household size, household wealth and location of residence (rural or urban). Child_iv is the vector of child stunting/malnourishment variables, K_iv is the vector of child-specific and mother-specific characteristics that uniquely identify Equation (2). This includes number of other children (under five years of age in the household), whether the mother currently works, mother’s age at the time of birth of first child, time between marriage to birth of first child, and sex and age of children under five years of age for whom CNO data is available. η_iv and ϵ_iv are error terms and assumed to be correlated. Since the effect of WE_iv enters in both equations, it is important to identify the causal pathways through which its effects can be decomposed via Violence_iv and its direct impact on Child_iv. These models were fitted separately to data from each country and then overall for all countries. Part of the identification issue relies on having additional confounders in Equation (1) from those in Equation (2) to ensure robustness of estimates. To check for robustness of the specification, we check for correlation between the errors to justify the use of the multivariate probit model.⁵

Results

The results of the multivariate probit estimations are reported in Tables 3, 4 and 5 for India, Nepal and Pakistan, respectively. The pooled estimate using country fixed effects for the South Asia sample can be found in Table 6. Since these are probit coefficients, the interpretation is not straightforward. Generating marginal effects for these coefficients was computationally intractable and hence could not be generated in a reasonable time frame.

Table 3.

Multivariate Probit Estimation Results of WE, IPV and CNOs in India.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
Variables	Stunting	E	LS	S	SE	Wasting	E	LS	S	SE	Underweight	E	LS	S	SE
Emotional violence	0.319***					0.084					0.249***
	(0.06)					(0.073)					(0.064)
Less severe violence	0.641***					0.235***					0.665***
	(0.054)					(0.078)					(0.054)
Severe violence	0.306***					0.128					0.368***
	(0.069)					(0.112)					(0.07)
Sexual violence	0.367***					0.247**					0.299***
	(0.073)					(0.117)					(0.075)
Decision-making power	–0.102***	0.165***	0.155***	0.121***	0.176***	–0.036	0.165***	0.148***	0.114***	0.168***	–0.075**	0.165***	0.152***	0.117***	0.171***
	(0.031)	(0.037)	(0.032)	(0.041)	(0.04)	(0.042)	(0.038)	(0.033)	(0.042)	(0.041)	(0.035)	(0.038)	(0.033)	(0.041)	(0.041)
Justifying violence	–0.023**	0.112***	0.106***	0.119***	0.099***	–0.041***	0.114***	0.108***	0.122***	0.102***	–0.049***	0.112***	0.106***	0.121***	0.1***
	(0.009)	(0.013)	(0.01)	(0.012)	(0.013)	(0.013)	(0.013)	(0.011)	(0.012)	(0.013)	(0.01)	(0.013)	(0.011)	(0.012)	(0.013)
Owns property	0.076***	–0.02	–0.084***	–0.002	–0.04	–0.05	–0.012	–0.078**	0.005	–0.034	0.092***	–0.015	–0.08**	0.007	–0.031
	(0.028)	(0.043)	(0.032)	(0.042)	(0.045)	(0.033)	(0.043)	(0.033)	(0.042)	(0.046)	(0.029)	(0.043)	(0.032)	(0.042)	(0.045)
Chi-squared	1,862.07***					953.66***					1,587.48***
Observations	30,394

Source: The authors.

Note: E – emotional violence; LS – less severe violence; S – severe violence; SE – sexual violence. Coefficients reported are probit coefficients, with clustered standard errors in parentheses. Violence is a binary variable that takes a value of 1 if the woman faced violence from her partner and zero otherwise; stunting, wasting and underweight are binary variables that take a value of 1 if ‘any’ child in the household was stunted, wasted or underweight by WHO standards. Additional controls for violence included are educational status of partner (husband), state/region of residence, type of place of residence (urban/rural), educational attainment, wealth index, number of household members and age. Additional controls for CNOs included are marriage to first birth interval (months), sex and age of the child, whether respondent was currently working, number of children five years of age and under in household. All regressions are weighted by sample weights provided by DHS for India. ^*** p < .01, ^** p < .05, ^* p < .1.

Table 4.

Multivariate Probit Estimation Results of WE, IPV and CNOs in Nepal.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
Variables	Stunting	E	LS	S	SE	Wasting	E	LS	S	SE	Underweight	E	LS	S	SE
Emotional violence	–0.155					0.474					–0.196
	(0.314)					(0.458)					(0.493)
Less severe violence	0.983***					0.66*					0.108
	(0.358)					(0.363)					(0.792)
Severe violence	0.229					0.224					0.367
	(0.554)					(0.371)					(0.979)
Sexual violence	0.243					0.815*					–0.229
	(0.549)					(0.472)					(2.343)
Decision-making power	0.056	–0.364	–0.586***	–0.895***	–0.767**	–0.281	–0.397*	–0.634***	–0.866***	–0.787**	–0.615*	–0.377	–0.611***	–0.878***	–0.766
	(0.252)	(0.224)	(0.221)	(0.245)	(0.344)	(0.257)	(0.218)	(0.213)	(0.248)	(0.34)	(0.336)	(0.25)	(0.235)	(0.271)	(0.557)
Justifying violence	0.017	0.387*	0.269	0.197	0.289	0.015	0.409*	0.316	0.232	0.327	0.259	0.434*	0.339	0.276	0.355
	(0.232)	(0.22)	(0.247)	(0.268)	(0.312)	(0.311)	(0.214)	(0.236)	(0.253)	(0.292)	(0.338)	(0.238)	(0.275)	(0.307)	(0.339)
Owns property	–0.237	–0.339	–0.356**	–0.08	0.623	–0.334	–0.325	–0.321*	–0.091	0.669	–0.031	–0.347	–0.377	–0.096	0.617
	(0.293)	(0.238)	(0.181)	(0.234)	(0.419)	(0.204)	(0.235)	(0.174)	(0.235)	(0.414)	(0.293)	(0.292)	(0.306)	(0.356)	(0.656)
Chi-squared	351.49***					683.24***					412.95***
Observations	3,646

Source: The authors.

Table 5.

Multivariate Probit Estimation Results of WE, IPV and CNOs in Pakistan.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
Variables	Stunting	E	LS	S	SE	Wasting	E	LS	S	SE	Underweight	E	LS	S	SE
Emotional violence	0.104					–0.133					0.048
	(0.243)					(0.192)					(0.264)
Less severe violence	0.928***					0.086					0.752***
	(0.183)					(0.231)					(0.22)
Severe violence	0.137					–0.113					0.131
	(0.251)					(0.25)					(0.376)
Sexual violence	–0.123					0.304					0.139
	(0.348)					(0.415)					(0.441)
Decision-making power	–0.044	0.444*	0.297	0.174	–0.174	–0.279	0.4*	0.219	0.103	–0.188	–0.253	0.461**	0.307	0.197	–0.123
	(0.183)	(0.233)	(0.222)	(0.198)	(0.251)	(0.186)	(0.235)	(0.229)	(0.199)	(0.243)	(0.248)	(0.234)	(0.231)	(0.207)	(0.246)
Justifying violence	0.093	0.156**	0.021	0.017	–0.024	0.004	0.178**	0.044	0.037	–0.009	0.081	0.158**	0.021	0.019	–0.024
	(0.066)	(0.078)	(0.071)	(0.091)	(0.07)	(0.082)	(0.077)	(0.07)	(0.089)	(0.068)	(0.076)	(0.079)	(0.069)	(0.09)	(0.071)
Owns property	–0.278	–0.041	–1.134***	–1.019***	–0.459	–0.227	–0.081	–1.29***	–1.261***	–0.527	-–0.594	–0.046	–1.17***	–1.116***	–0.467
	(0.349)	(0.505)	(0.392)	(0.321)	(0.457)	(0.501)	(0.489)	(0.387)	(0.324)	(0.452)	(0.391)	(0.498)	(0.395)	(0.343)	(0.454)
Chi-squared	360.43***					199.75***					259.19***
Observations	3,088

Source: The authors.

WE and IPV

In India, we find a strong statistically significant association between improvements in decision-making power and likelihood of facing any type of IPV. This is a backlash effect that is also observable in the pooled estimates (Table 6). However, greater decision-making power for women in Nepal is associated with lower likelihood of facing physical violence (but not emotional violence). There are no statistically significant coefficients of decision-making power on facing violence in Pakistan, with the notable exception of a positive association with facing emotional violence, significant at the 10% level. Notably, property ownership is associated with lower incidence of less severe violence in India, Nepal and Pakistan, but no significant effects exist on the incidence of other types of violence. In Pakistan, owning property also has protective effects associated with facing severe violence. In the pooled estimates for all countries, owning property is associated with lower likelihood of facing less severe violence but no statistically significant effects for any other type of violence.⁶

Table 6.

Multivariate Probit Estimation of Experience of IPV and WE on CNOs (Pooled Estimates).

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)
Variables	Stunting	E	LS	S	SE	Wasting	E	LS	S	SE	Underweight	E	LS	S	SE
Emotional violence	0.268***					0.0549					0.245***
	(0.0627)					(0.0681)					(0.0705)
Less severe violence	0.696***					0.102					0.655***
	(0.0521)					(0.0737)					(0.0616)
Severe violence	0.286***					0.0582					0.369***
	(0.0756)					(0.0978)					(0.0812)
Sexual violence	0.313***					0.0364					0.227**
	(0.0777)					(0.0953)					(0.0887)
Decision-making power	–0.0893***	0.155***	0.140***	0.0897**	0.152***	–0.0476	0.148***	0.128***	0.0858**	0.144***	–0.101***	0.151***	0.134***	0.0858**	0.145***
	(0.0324)	(0.0378)	(0.0331)	(0.0414)	(0.0418)	(0.0418)	(0.0378)	(0.0334)	(0.0420)	(0.0420)	(0.0345)	(0.0379)	(0.0333)	(0.0417)	(0.0423)
Justifying violence	–0.0181*	0.111***	0.100***	0.113***	0.0926***	–0.0261**	0.115***	0.105***	0.118***	0.0977***	–0.0385***	0.112***	0.101***	0.114***	0.0945***
	(0.00958)	(0.0127)	(0.0107)	(0.0120)	(0.0130)	(0.0126)	(0.0127)	(0.0108)	(0.0120)	(0.0130)	(0.0101)	(0.0127)	(0.0108)	(0.0119)	(0.0131)
Owns property	0.0702**	–0.0298	–0.0976***	–0.0115	–0.000335	–0.0630*	–0.0230	–0.0912***	–0.00377	0.00781	0.0889***	–0.0235	–0.0898***	–0.00228	0.0150
	(0.0288)	(0.0424)	(0.0318)	(0.0411)	(0.0580)	(0.0329)	(0.0424)	(0.0319)	(0.0411)	(0.0599)	(0.0293)	(0.0423)	(0.0317)	(0.0407)	(0.0620)
Constant	–0.374***	–0.815***	–0.0350	–1.103***	–1.064***	–0.618***	–0.855***	–0.0690	–1.142***	–1.101***	–0.514***	–0.796***	–0.00608	-1.086***	–1.042***
	(0.108)	(0.162)	(0.148)	(0.197)	(0.205)	(0.138)	(0.163)	(0.153)	(0.196)	(0.207)	(0.109)	(0.162)	(0.146)	(0.198)	(0.200)
Chi-squared	1,844.32***					1,073.79***					1,502.24***
Observations	37,128

Source: The authors.

In India, justification of IPV by the wife is also positively associated with the likelihood of facing all types of violence. In Nepal and Pakistan, however, it is only positively associated with the likelihood of facing emotional violence. In the pooled estimates for South Asia, the justification of IPV by the woman is strongly positively associated with facing all types of violence and is the strongest for severe violence in terms of magnitude of the coefficients.

Finally, we note that the association between owning assets (land or property) and incidence of emotional and less severe violence in India is small and negative. It is possible that this is on account of sparse data on land and asset ownership, due to poor or informal record-keeping in developing South Asian countries (Jamal, 2017).

IPV and CNOs

The results of the effects of IPV on CNOs are unequivocal and suggest a strong and statistically significant and negative relationship across countries, types of violence and all CNOs in Tables 3, 4, 5 and 6. In India, the results of facing all types of violence have statistically significant and positive associations with stunting and a child being underweight. In the case of child wasting in India, facing less severe and sexual violence in particular suggests a threat. Thus, when the mother of a child under five years of age faces violence in India, it is associated with a strong increase in the likelihood that her children may be stunted or underweight and, in some instances, suffers from wasting as well. In contrast, in Nepal, we find ambiguous results; facing less severe violence is associated with an increase in likelihood of child stunting and wasting (the coefficient is significant only at the 10% level in the latter case), but not in terms of a child being underweight. There is an increase in the risk of a child being wasted if the mother faced sexual violence in the past 12 months, and this effect is significant at the 10% level. We find no statistically significant association when all types of violence are simultaneously assumed to explain likelihood of adverse CNOs in Nepal. In the case of Pakistan, we find results similar to Nepal, except that increase in likelihood of facing less severe violence was associated with an increase in the rates of stunting and children under five years of age being underweight. There was no statistically significant association with wasting, as well as no statistically significant associations of facing any other type of violence on other CNOs, suggesting that facing less severe violence is a significant threat to child health in India, Nepal and Pakistan, especially when dealing with child stunting. In our overall pooled results, we find that all types of violence are harmful to child outcomes, except for child wasting. In the case of a child being underweight as well as stunting, the incidence is highest when the mother has faced less severe violence in the past 12 months.⁷

WE and CNOs

Results from the overall South Asia regression as well as country-specific results suggest that improvements in women’s decision-making power within the household are particularly beneficial in reducing the incidence of adverse CNOs (with the exception of child wasting in the pooled estimates). Although we do not find any statistically significant effects in Nepal and Pakistan, women who have greater decision-making power were associated with lower incidence of underweight children in Nepal.

In contrast, justifying violence is also associated with a reduction in CNOs, and the coefficients in the pooled estimates suggest that these are small but statistically significant associations. Interestingly, land ownership is associated with small increases in incidence of stunting and underweight children in India, although the effects are only statistically significant in India and overall South Asia sample. It is less clear what the potential channels for such effects could be.

Discussion

This study used a multivariate probit approach to investigate direct and indirect effects of WE and autonomy on child malnourishment (Figure 1) in South Asia. The estimation framework allowed WE to affect child malnourishment via IPV as well as independent of IPV. We find that the indirect impact of WE is critical in being protective of women at risk of facing different types of violence, but in particular less severe violence. Owning property, for example, can reduce the likelihood of facing less severe violence, which contributes significantly to childhood stunting and children being underweight on average. In contrast, the direct impact of owning property on CNOs is actually negative, since property ownership among women is associated with a higher incidence of stunting and children being underweight, although it is associated with less wasting. As van der Meulen Rodgers and Kassens (2018) suggest, there could be unobserved characteristics (such as preferences) that are driving both land ownership and CNOs at the same time, which cannot be adequately accounted for. As Swaminathan et al. (2012) suggest, ownership may not reduce household chores or change control over household incomes in certain circumstances and, therefore, may detract from caregiving responsibilities.

In India, and to some extent Pakistan, we find a strong statistically significant effect of improvements in decision-making power with an increase in likelihood of facing different types of IPV. This is in line with findings from Schuler et al. (2017), where it was found that a moderate level of WE may provoke IPV, as women begin to behave in ways that are contradictory to traditional gender norms. Similarly, Vyas and Watts (2009) find that women solely responsible for household expenses experience higher levels of physical violence in India. However, the results find an exception in Nepal, where an increase in household decision-making is associated with lower likelihood of facing physical forms of violence (severe and less severe). To corroborate this finding, Atteraya et al. (2015) find that in Nepal, married women are at a greater risk of experiencing IPV when they are illiterate, are raised in families where violence is present and lack decision-making autonomy within their households.

The importance of justification of IPV plays a crucial role in determining whether the victim, often a woman, reports the act (Waltermaurer, 2012). If the woman feels that violence witnessed by her is justified, she is less likely to report it. In our study, we found that justification of IPV by the wife is positively associated with the likelihood of experiencing violence, especially emotional violence in Nepal and Pakistan. In South Asian countries, it is possible that cultural beliefs and social norms around IPV are driving some of these results (Solotaro & Pande, 2014). Women may not disclose or seek care when they have experienced any form of violence even in situations where formal and informal mechanisms exist. For some women, violence is considered an inevitable element of marriage and women also do not disclose instances of violence fearing reprisal or additional violence or due to stigma and shame associated with it. From studies in six districts in Pakistan, 35% of women mentioned that they were not allowed to seek medical care for injuries sustained during IPV. Women may also not reveal or disclose experiencing violence because of apathy and the belief that revealing events might not resolve anything (Solotaro & Pande, 2014). Thus, there lie inherent challenges in obtaining accurate data on violence experienced by women. There are several challenges that lie in variations in definitions and survey methodology (Bishwajit et al., 2016).

Maternal health is impacted by IPV, and thus the mother’s ability to take care of her children is negatively affected, which leads to poor CNOs. In our study, we find a strong negative relationship between IPV and CNOs across all countries. All types of violence experienced by the mothers were found to be harmful to CNOs. In particular, less severe violence is associated with nearly a large increase in incidence of stunting. The incidence of a child being underweight is the highest when the mother has experienced less severe violence in the past year in India as well as in the pooled estimates. Notably, the results are strongest for India and Nepal, and for stunting and child underweight. In Pakistan, we found no statistically significant association between facing violence and child wasting. This evidence is broadly in line with past work that suggests facing ‘physical’ violence in particular can hinder a mother’s childcare abilities and potentially result in poor child outcomes (Ackerson & Subramanian, 2009; Kishor & Johnson, 2004; Silverman et al., 2011, 2009; Simona et al., 2015; Yount et al., 2011; Ziaei et al., 2014).

WE ensures that a woman is able to make better decisions for herself and for her children which leads to positive CNOs. In our study, it was found that women’s decision-making power within the household is particularly beneficial in reducing adverse CNOs (independent of whether the woman faced any IPV), and in India particularly, this increase in decision-making autonomy reduced the incidence of stunting and child underweight. Interestingly, justification of IPV among mothers is also associated with a lower incidence of stunting and underweight children, with results for India suggesting that a household where the mother justifies IPV is less likely to have a stunted child. Although the effects are very small (one-tenth of the positive association between decision-making power and reduced adverse CNOs), it is possible that these are because of justification of IPV among both genders when the child is neglected, especially in Asian cultures where family is valued over the individual (Waltermaurer, 2012). Surprisingly, we find that property-owning and wealthier women are likely to have higher incidence of stunting and underweight children, a finding that is potentially driven by the large Indian sample. This finding runs contrary to the literature found in this domain. For example, Swaminathan et al. (2012) find that property ownership among women in a South Indian state boosts their decision-making power and suggest that it is important for CNOs. Pande (2003) finds that property ownership significantly reduced the odds of stunting for children in India between 1992 and 1993. At the same time, there are null results in South Asian countries such as Bangladesh that suggest that the relationship may not be as straightforward (e.g., Bhagowalia et al., 2012).

Conclusion

This study explores the relationship between WE, IPV and CNOs using the nationally representative datasets of three South Asian countries—India, Pakistan and Nepal. Lower empowerment or autonomy of women in South Asian countries affects CNOs directly and via an increased exposure to IPV. Our study points towards important differences between countries in the relationship between IPV and CNOs. For example, in Nepal, having higher decision-making autonomy and owning wealth for women are associated with lower odds of facing certain types of violence, which could be an important channel through which to ameliorate CNOs.

One of the major limitations of this study is the overrepresentation of an Indian sample in our analysis. As Indian households comprised more than 85% of our sample, extending these findings to South Asia is a challenge. One of the major barriers in this regard was the lack of availability of data from DHS on IPV or WE or CNOs for other South Asia countries such as Bangladesh and Sri Lanka. Adding this data will help improve the generalizability of the findings of the study and accordingly help focus on any differences between countries. Second, we are unable to comment on specific pass-through effects that might accumulate as a result of WE, exposure to IPV and changes in CNOs. Future work in this domain can specify empirical frameworks that are suited to establishing the robustness of these impact pathways. Furthermore, our empirical framework does not account for potential reverse causality between experiencing IPV and WE indicators in South Asia. It is possible that the directional associations are difficult to disentangle using the current data, and future work can look at previous episodes of IPV to account for this. Finally, the study implicitly assumes that gender norms, childcare and other social norms are similar in India, Nepal and Pakistan, which may not necessarily be uniform. The authors acknowledge that data related to gender norms, childcare and other social norms are not present in the DHS data and remain unaccounted for in terms of their role in measuring WE and CNOs in the countries selected for the study. More data on how these vary can be incorporated to explain country-level variegations in IPV as well as CNOs. The role that sanitation, infrastructure and ecosystem in the locality such as the health system play in determining CNOs is not present in DHS data. The authors note this lack of data as a limitation of this article.

Supplementary material

Supplementary material for this article is available online.

Footnotes

Acknowledgements

The authors thank Vergard Iversen, two anonymous referees, Hansika Kapoor and Anchal Khandelwal for helpful comments and suggestions, and the research team at SNEHA for preliminary discussions that informed this manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethics Approval Statement

This study was approved by the Monk Prayogshala Institutional Review Board (#019-024).

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

References

Ackerson

L. K.

, & Subramanian

S. V.

(2008). State gender inequality, socioeconomic status and intimate partner violence (IPV) in India: A multilevel analysis. Australian Journal of Social Issues, 43(1), 81–102. 10.1002/j.1839-4655.2008.tb00091.x

Ackerson

L. K.

, & Subramanian

S. V.

(2009). Intimate partner violence and death among infants and children in India. Pediatrics, 124(5), e878–e889. 10.1542/peds.2009-0524

Ahmad

, Khan

, & Mozumdar

(2019). Spousal violence against women in India: A social–ecological analysis using data from the National Family Health Survey 2015 to 2016. Journal of Interpersonal Violence. 10.1177/0886260519881530

Ahmed

, Creanga

A. A.

, Gillespie

D. G.

, & Tsui

A. O.

(2010). Economic status, education and empowerment: Implications for maternal health service utilization in developing countries. PLoS ONE, 5(6). 10.1371/journal.pone.0011190

Atteraya

M. S.

, Gnawali

, & Song

I. H.

(2015). Factors associated with intimate partner violence against married women in Nepal. Journal of Interpersonal Violence, 30(7), 1226–1246. 10.1177/0886260514539845

Bhagowalia

, Menon

, Quisumbing

A. R.

, & Soundararajan

(2012). What dimensions of women’s empowerment matter most for child nutrition? Evidence using nationally representative data from Bangladesh. IFPRI Discussion Papers (No. 1192). https://ideas.repec.org/p/fpr/ifprid/1192.html

Bishwajit

, Sarker

, & Yaya

(2016). Socio-cultural aspects of gender-based violence and its impacts on women’s health in South Asia. F1000Research, 5, 802. 10.12688/f1000research.8633.1

Bloch

, & Rao

(2002). Terror as a bargaining instrument: A case study of dowry violence in rural India. American Economic Review, 92(4), 1029–1043. 10.1257/00028280260344588

Burroway

(2016). Empowering women, strengthening children: A multi-level analysis of gender inequality and child malnutrition in developing countries. In Advances in gender research (Vol. 22, pp. 117–142). 10.1108/S1529-212620160000022016

10.

Carlson

G. J.

, Kordas

, & Murray-Kolb

L. E.

(2015). Associations between women’s autonomy and child nutritional status: A review of the literature. Maternal & Child Nutrition, 11(4), 452–482. 10.1111/mcn.12113

11.

Chai

, Fink

, Kaaya

, Danaei

, Fawzi

, Ezzati

, Lienert

, & Fawzi

M. C. S.

(2016). Association between intimate partner violence and poor child growth: Results from 42 demographic and health surveys. Bulletin of the World Health Organization, 94(5), 331–339. 10.2471/BLT.15.152462

12.

Cronbach

L. J.

(1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. 10.1007/BF02310555

13.

Cunningham

, Ploubidis

G. B.

, Menon

, Ruel

, Kadiyala

, Uauy

, & Ferguson

(2015). Women’s empowerment in agriculture and child nutritional status in rural Nepal. Public Health Nutrition, 18(17), 3134–3145. 10.1017/S1368980015000683

14.

Engle

P. L.

(1993). Influences of mothers’ and fathers’ income on children’s nutritional status in Guatemala. Social Science and Medicine, 37(11), 1303–1312. 10.1016/0277-9536(93)90160-6

15.

Frongillo

E. A.

, de Onis

, & Hanson

K. M.

(1997). Socioeconomic and demographic factors are associated with worldwide patterns of stunting and wasting of children. The Journal of Nutrition, 127(12), 2302–2309. 10.1093/jn/127.12.2302

16.

Garcia-Moreno

, Jansen

H. A. F. M.

, Ellsberg

, Heise

, Watts

C. H.

, & Study

W. H. O. M.

(2006). Prevalence of intimate partner violence: Findings from the WHO multi-country study on women’s health and domestic. Lancet, 368, 1260–1269.

17.

Ibrahim

, & Alkire

(2007). Agency and empowerment: A proposal for internationally comparable indicators. Oxford Development Studies, 35(4), 379–403. 10.1080/13600810701701897

18.

Imai

K. S.

, Annim

S. K.

, Kulkarni

V. S.

, & Gaiha

(2014, May). Women’s empowerment and prevalence of stunted and underweight children in rural India. University of Manchester.

19.

Jamal

(2017). Explaining spousal physical violence through dimensions of women empowerment: Evidence from Pakistan. https://mpra.ub.uni-muenchen.de/83414/1/MPRA_paper_83414.pdf

20.

Jejeebhoy

S. J.

, & Cook

R. J.

(1997). State accountability for wife-beating: The Indian challenge. The Lancet, 349, 10–12.

21.

Jones

, Haardörfer

, Ramakrishnan

, Yount

K. M.

, Miedema

, & Girard

A. W.

(2019). Women’s empowerment and child nutrition: The role of intrinsic agency. SSM - Population Health, 9(August), 100475. 10.1016/j.ssmph.2019.100475

22.

Kabeer

(1999). Resources, agency, achievements: Reflections on the measurement of women’s empowerment. Development and Change, 30(May), 435–464.

23.

Kabeer

(2005). Gender equality and women’s empowerment: A critical analysis of the third millennium development goal 1. Gender & Development, 13(1), 13–24. 10.1080/13552070512331332273

24.

Kabeer

(2021). Three faces of agency in feminist economics. In Berik

& Kongar

(Eds.), The Routledge handbook of feminist economics (pp. 99–107). 10.4324/9780429020612-12

25.

Kishor

, & Johnson

(2004). Profiling domestic violence: A multi-country study. The DHS Program.

26.

Koenig

M. A.

, Ahmed

, Hossain

M. B.

, & Mozumder

A. B. M.

. (2003). Women’s status and domestic violence in rural Bangladesh: Individual- and community-level effects. Demography, 40(2), 269–288. 10.1353/dem.2003.0014

27.

Kravdal

Ø.

(2004). Child mortality in India: The community-level effect of education. Population Studies, 58(2), 177–192. 10.1080/0032472042000213721

28.

Kumar

, Scott

, Menon

, Kannan

, Cunningham

, Tyagi

, Wable

, Raghunathan

, & Quisumbing

(2018). Pathways from women’s group-based programs to nutrition change in South Asia: A conceptual framework and literature review. Global Food Security, (November), 1–14. 10.1016/j.gfs.2017.11.002

29.

Mistry

, Galal

, & Lu

(2009). Women’s autonomy and pregnancy care in rural India: A contextual analysis. Social Science and Medicine, 69(6), 926–933. 10.1016/j.socscimed.2009.07.008

30.

Mosedale

(2005). Assessing women’s empowerment: Towards a conceptual framework. Journal of International Development, 257, 243–257. 10.1002/jid.1212

31.

Panda

, & Agarwal

(2005). Marital violence, human development and women’s property status in India. World Development, 33(5), 823–850. 10.1016/j.worlddev.2005.01.009

32.

Pande

(2003). Selective gender differences in childhood nutrition and immunization in rural India: The role of siblings. Demography, 40(3), 395–418. 10.1353/dem.2003.0029

33.

Parekh

, Tagat

, Kapoor

, & Nadkarni

(2021). The effects of husbands’ alcohol consumption and women’s empowerment on intimate partner violence in India. Journal of Interpersonal Violence. 10.1177/0886260521991304

34.

Paul

(2016). Intimate partner violence and women’s help-seeking behaviour: Evidence from India. Journal of Interdisciplinary Economics, 28(1), 53–82.

35.

Pongou

, Ezzati

, & Salomon

J. A.

(2006). Household and community socioeconomic and environmental determinants of child nutritional status in Cameroon. BMC Public Health, 6. 10.1186/1471-2458-6-98

36.

Ramalingaswami

, Jonsson

, & Rohde

(1996). The Asian enigma. https://www.unicef.org/pon96/nuenigma.htm

37.

Ramalingaswami

, Jonsson

, & Rhode

(1997). Malnutrition: A south Asian enigma. Malnutrition in south Asia: A regional profile. UNICEF Regional Office for South Asia, 11–22.

38.

Ramful

, & Zhao

(2009). Participation in marijuana, cocaine and heroin consumption in Australia: A multivariate probit approach. Applied Economics, 41(4), 481–496. 10.1080/00036840701522853

39.

Rico

, Fenn

, Abramsky

, & Watts

(2011). Associations between maternal experiences of intimate partner violence and child nutrition and mortality: Findings from demographic and health surveys in Egypt, Honduras, Kenya, Malawi and Rwanda. Journal of Epidemiology and Community Health, 65(4), 360–367. 10.1136/jech.2008.081810

40.

Sangwan

, & Kumar

(2021). Labor force participation of rural women and the household’s nutrition: Panel data evidence from SAT India. Food Policy, 102, 102117. 10.1016/j.foodpol.2021.102117

41.

Schuler

S. R.

, Lenzi

, Badal

S. H.

, & Bates

L. M.

(2017). Women’s empowerment as a protective factor against intimate partner violence in Bangladesh: A qualitative exploration of the process and limitations of its influence. Violence Against Women, 23(9), 1100–1121. 10.1177/1077801216654576

42.

Sethuraman

, Lansdown

, & Sullivan

(2006). Women’s empowerment and domestic violence: The role of sociocultural determinants in maternal and child undernutrition in tribal and rural communities in South India. Food and Nutrition Bulletin, 27(2), 128–143. 10.1177/156482650602700204

43.

Shaukat

, Iqbal

, & Khan

M. A.

(2018). Detrimental effects of intimate partner violence on the nutritional status of children: Insights from PDHS 2012-2013. International Journal of Community Medicine and Public Health, 5(5), 1742. 10.18203/2394-6040.ijcmph20181680

44.

Shroff

, Griffiths

, Suchindran

, & Bentley

(2009). Maternal autonomy is inversely related to child stunting in Andhra Pradesh, India. Maternal & Child Nutrition, 361(2), 435–446. 10.1109/TMI.2012.2196707.Separate

45.

Shroff

M. R.

, Griffiths

P. L.

, Suchindran

, Nagalla

, Vazir

, & Bentley

M. E.

(2012). Does maternal autonomy influence feeding practices and infant growth in rural India? Social Science & Medicine, 73(3), 447–455. 10.1016/j.socscimed.2011.05.040.Does

46.

Silverman

J. G.

, Decker

M. R.

, Cheng

D. M.

, Wirth

, Saggurti

, McCauley

H. L.

, Falb

K. L.

, Donta

, & Raj

(2011). Gender-based disparities in infant and child mortality based on maternal exposure to spousal violence: The heavy burden borne by Indian girls. Archives of Pediatrics and Adolescent Medicine, 165(1), 22–27. 10.1001/archpediatrics.2010.261

47.

Silverman

J. G.

, Decker

M. R.

, Gupta

, Kapur

, Raj

, & Naved

R. T.

(2009). Maternal experiences of intimate partner violence and child morbidity in Bangladesh. Archives of Pediatrics & Adolescent Medicine, 163(8), 700–705. 10.1001/archpediatrics.2009.115

48.

Simona

S. J.

, Muchindu

, & Ntalasha

(2015). Intimate partner violence (IPV) in Zambia: Sociodemographic determinants and association with use of maternal health care. https://dhsprogram.com/pubs/pdf/WP121/WP121.pdf

49.

Sinha

, & Chattopadhyay

(2017). Inter-linkages between spousal violence and nutritional status of children: A comparative study of North and South Indian states. Sociology and Criminology—Open Access, 05(02). 10.4172/2375-4435.1000171

50.

Smith

L. C.

, Ramakrishnan

, Ndiaye

, Haddad

, & Martorell

(2003). The importance of women’s status for child nutrition in developing countries. Food and Nutrition Bulletin, 24. 10.1177/156482650302400308

51.

Solotaro

J. L.

, & Pande

R. P.

(2014). Violence against women and girls: Lessons from South Asia. Choice Reviews, 52. 10.5860/choice.188354

52.

Swaminathan

, Lahoti

, & Suchitra

J. Y.

(2012). Women’s property, mobility, and decisionmaking: Evidence from rural Karnataka, India. 10.1016/0584-8539(91)80037-J

53.

Underwood

B. A.

(2002). Health and nutrition in women, infants, and children: Overview of the global situation and the Asian enigma. Nutrition Reviews, 60(5), S7–S13.

54.

UNDP. (1995). Human Development Report 1995: Gender and human development. In Human Development Report (1990 to present). http://hdr.undp.org/sites/default/files/reports/256/hdr_1995_en_complete_nostats.pdf

55.

van der Meulen Rodgers

, & Kassens

A. L.

(2018). Women’s asset ownership and children’s nutritional status: Evidence from Papua New Guinea. Social Science & Medicine, 204, 100–107. 10.1016/j.socscimed.2018.03.026

56.

Vyas

, & Watts

(2009). How does economic empowerment affect women’s risk of intimate partner violence in low and middle income countries? A systematic review of published evidence. Journal of International Development, 21(5), 577–602. 10.1002/jid.1500

57.

Waltermaurer

(2012). Public justification of intimate partner violence: A review of the literature. Trauma, Violence, and Abuse, 13(3), 167–175. 10.1177/1524838012447699

58.

Wayack Pambè

, Gnoumou

, & Kaboré

(2013). Relationship between women’s socioeconomic status and empowerment in Burkina Faso: A focus on participation in decision-making and experience of domestic violence. 10.11564/28-0-563

59.

World Health Organization. (2017). Violence against women. https://www.who.int/news-room/fact-sheets/detail/violence-against-women#:~:text=

60.

Yount

K. M.

, DiGirolamo

A. M.

, & Ramakrishnan

(2011). Impacts of domestic violence on child growth and nutrition: A conceptual review of the pathways of influence. Social Science and Medicine, 72(9), 1534–1554. 10.1016/j.socscimed.2011.02.042

61.

Ziaei

, Naved

R. T.

, & Ekström

E.-C.

(2014). Women’s exposure to intimate partner violence and child malnutrition: Findings from demographic and health surveys in Bangladesh. Maternal & Child Nutrition, 10(3), 347–359. 10.1111/j.1740-8709.2012.00432.x