The incidence of intimate partner violence (IPV) among women in the world is estimated to range between 13 and72 % (García-Moreno et al. 2006). The World Health Organization (WHO) reported industrialized countries tend to have lower rates of IPV compared to non-industrialized and newly-industrialized countries, indicating that this social problem is found consistently across cultural settings and countries (García-Moreno et al. 2006). On the other hand, recent multi-country research has found that depression disorders are the second leading cause of disease burden among women between the ages of 15 and 44 years (Ribeiro et al. 2008). Indeed, depression and IPV are linked, as there is abundant evidence that IPV is consistently associated with depression across the world (Breiding et al. 2008; Beydoun et al. 2012; Chowdhary and Patel 2008; Devries et al. 2013a; Ellsberg et al. 2008; Loxton et al. 2006; Nduna et al. 2010; Salazar et al. 2009; Zlotnick et al. 2006). In addition, multiple studies have linked the experience of IPV to increased risk of suicide among women in different cultures and settings (Beydoun et al. 2012; Borges et al. 2007; Devries et al. 2011; Maselko and Patel 2008; Mogga et al. 2006; Naved and Akhtar 2008; Pillai et al. 2008; Vizcarra et al. 2004). However, most of the research exploring the relationship of IPV to depression is conducted in North American and European settings. This study seeks to increase knowledge of the protective and risk factors for depression among women who experience abusive relationships in a nationally representative sample from Mexico.

Depression and Intimate Partner Violence

Depression is a common mental disorder with greater severity and persistence than normal mood variations or sadness (Beckham 2000; Cuijpers et al. 2012). It is a serious, recurrent, and debilitating mental disorder that impacts most individuals at some point in life by means of personal experience or a loved one. Characteristics of depression include persistent sad or depressed mood, changes in appetite and body weight, sleep disturbances, difficulty thinking or concentrating, diminished interest in activities, feelings of hopelessness or worthlessness, and often thoughts of death or suicide (WHO 2008; APA 2013). Depression is a life-long condition in which periods of wellness and illness alternate intermittently, despite being highly treatable. It is estimated that fewer than 50 % of individuals receive treatment globally and this percentage varies, being lower than 30 % in some regions and even lower than 10 % in others (WHO 2008). Significantly, depression and associated symptoms are consistently linked to intimate partner violence (Bonomi et al. 2009; Devries et al. 2011; Deyessa et al. 2009; Ellsberg et al. 2008; Flanagan et al. 2015). Numerous studies across cultures have found that IPV is consistently associated with depression and suicidal behaviors in women in both high- and low-income areas around the world (Borges et al. 2007; Devries et al. 2013b; García-Moreno et al. 2006; Maselko and Patel 2008; Mogga et al. 2006; Naved and Akhtar 2008; Pillai et al. 2008; Vizcarra et al. 2004).

Researchers hypothesize that traumatic stress is the main trigger for development of depression and depressive symptoms among victims of IPV (Costello et al. 2008; Devries et al. 2013a). Experiencing a traumatic event has been associated with increased stress and subsequent development of depression and suicidality (Hyde et al. 2008; Kendler et al. 2001). Data from a recent meta-analysis of studies conducted primarily in highly developed countries provided support for the IPV-depression relationship, as women experiencing IPV were more likely to develop depression compared to women who had never experienced IPV (Beydoun et al. 2012). Another meta-analysis found that female victims of IPV had a mean prevalence of depression of 47.6 %, which is significantly higher than rates found in the general population (Golding 1999). Although there is a clear relationship between depression and IPV, the causal directional impact of these variables is less clear. There are studies arguing IPV leads to depression (Beydoun et al. 2012; Golding 1999; Ortiz-Gomez et al. 2014), while other studies indicate that existing depression may predispose individuals to become involved with partners that will be more likely to perpetrate IPV (Devries et al. 2013b; Lehrer et al. 2006; Khalifeh and Dean 2010; McPherson et al. 2007).

On the other hand, research on IPV in the past 20 years has strived to explore typologies of violence, arguing that not all violence is equal (Johnson 1995; Stark 2006). For example, representative studies in the U.S. have found the violence reported tends to be mild in severity and it is characterized by similar perpetration rates by males and females (Johnson and Leone 2005). This type of violence is identified as Common Couple Violence (CCV), and has been characterized as gender symmetric and resulting in mild consequences to mental and physical health of the victim (Johnson 1995). However, studies including samples from clinical settings and domestic violence shelters have consistently reported the majority of the victims they serve are women whom exhibit significant negative consequences to their mental and physical health due to severe IPV (Lawson 2012). Research with domestic shelter victims found that their partners exhibited a significant pattern of coercion and control that was not found in nationally representative studies (Dobash et al. 1992; Stark 2006). Coercive control (CC) has been defined as a pattern in which the abusive partner asserts power over the victim through use of threats, intimidation, withholding resources, and violence (Dutton and Goodman 2005; Johnson 2006; 1995; Johnson and Leone 2005; Pence and Paymar 1986; Stark 2006; Tanha et al. 2009). Johnson (1995) characterized this type of violence as Intimate Terrorism (IT), and it has been characterized as gender asymmetric and resulting in significant impairments to mental and physical health. Indeed, Dutton and Goodman (2005) assert that “exposure to coercive acts means exposure to threats of harm, including those that would be considered traumatic stressors such as threats of harm to self or others” (p. 752). Although the research in this area is limited and has been conducted primarily in highly industrialized societies like the United States, Canada, Australia, and European countries, some studies have found that women in coercive controlling relationships exhibited adverse mental and physical health problems including depression, anxiety, sleep problems, and persistent headaches (Dutton et al. 1999; Dutton et al. 1997; Dutton et al. 2005a). Through ethnographic research with women in the United States, Dutton et al. (2005b) found that coercive controlling relationships are characterized by specific demands by the partner, who then may threaten the victim and engage in surveillance to ensure compliance with demands. However, there is limited empirical understanding of the concept of nonviolent coercion and control in a relationship in newly or low development countries, as most IPV research has focused on violent and aggressive behaviors perpetrated in the context of a coercive relationship (Dutton et al. 2005a). To date, the literature available suggests it is possible all types of IPV and depression may independently share common risk factors; however, ethical concerns prevent researchers from engaging in experimental manipulation of these independent variables. Thus, correlational research seems most appropriate in this context.

Intimate Partner Violence Across Cultures

Intimate partner violence (IPV) is the most common form of violence experienced by women in the world, as it reported by 15 − 71 % of women during their lifetime, and can occur in the form of physical, sexual, or emotional/psychological abuse by an intimate partner (Heise and García-Moreno 2002; García-Moreno et al. 2006). Although both men and women can experience and perpetrate IPV, there is limited research on the physical and psychological sequelae of IPV on men (Archer 2000; Straus 2005). Most cross-cultural studies conducted to date focus on the impact of IPV on women because violence against women occurs at a higher rate than violence against men (Devries et al. 2013b; García-Moreno et al. 2006; Tjaden and Thoennes 2000).

Additionally, data from several studies suggested a dose–response effect with violence. In other words, when violence becomes increasingly persistent and severe, the impact on the victim’s physical and mental health increases as well (Campbell 2002; Lehrer et al. 2006). Research has documented the association of high severity of abuse with poorer mental and physical health, quality of life, and higher levels of depression, PTSD, and substance abuse across highly developed countries like the U.S. and Canada (Dutton et al. 2005b; Golding 1999; Straus et al. 2009; Wathen and MacMillan 2003; Wuest et al. 2010). A study by Hegarty et al. (2013) found that women experiencing co-occurring physical, sexual, and psychological abuse had poorer mental health and quality of life compared to women who only experienced one type of partner violence.

IPV has also been linked to significant increases in negative health outcomes, including broken bones, traumatic brain injury, sexual dysfunction, chronic pain syndromes, cardiovascular disease, obesity, and gastrointestinal disorders, sexually transmitted infections, and unplanned pregnancies (Black 2011; Breiding et al. 2008; Crofford 2007; Wathen and MacMillan 2003). Indeed, there are multiple short- and long-term negative health effects of IPV and in extreme cases, experiencing IPV may result in death either by the victim’s suicide or homicide by intimate partner (Campbell 2002; Stöckl et al. 2013). In fact, a multi-country study sponsored by the WHO found that victims of IPV are more likely to engage in suicidal ideation and attempts (Coker et al. 2002; Devries et al. 2011; Heise and García-Moreno 2002; Roberts et al. 2003; Warshaw et al. 2009). Thus, the significant and adverse consequences of IPV underscore the importance of a thorough understanding risk and protective factors associated with IPV.

Multiple ideas exist regarding societal, cultural, and individual-level explanations of partner violence. For example, Archer (2006) focused on the idea that cultural variables have the most significance in understanding the societal prevalence of IPV. He analyzed different cultural factors across 16 different nations and found that violence was more prevalent within collectivist countries (e.g., Negy et al. 2013). Moreover, while culture plays a role in the origins of IPV within the individual, it also impacts the style of abuse and the way in which an individual will handle the abuse (Yoshioka and Choi 2005). Clearly, the cultural influences on IPV are complex, yet much of the large-scale research concerning IPV has been conducted in highly developed countries. Johnson and Ferraro (2000) relay the importance of maintaining an awareness of ethnocentrism, or the tendency to impose Western ideas in understanding relationship dynamics in other cultures. Thus, exploring risk factors and profiles of IPV in other countries provides information about whether existing theories of this phenomenon hold consistent in different cultural settings (Terrazas-Carrillo and McWhirter 2015). A greater understanding of cultural contexts will help establish effective interventions for victims of IPV and the sequelae of traumatic stress, depression, and suicide attempts.

The Mexican Context

Mexico is a country with high human development according to recent Human Development Reports (UNDP, 2013). It has the second largest economy in Latin America and has maintained economic and financial stability in spite of the slowdown of the U.S. and European economies (World Bank 2015). However, Mexico is a country of contrasts and wide racial and ethnic diversity, with people experiencing affluence and poverty, long and healthy life as well as violence (UNDP 2014). In spite of advances in health, education, and income, inequality is widespread in Mexico (UNDP 2014). Thus, experiences of Mexican men and women are shaped by demographic, geographical, and cultural factors (Zabludovsky 2001). In spite of these differences, studies across demographic and socioeconomic strata have consistently found that the Mexican culture often pressures women to get married and have children (Brumley 2013; Ruiz Castro 2012; Zabludovsky 2001). Although gender roles are constantly evolving in Mexico, women continue to be seen as primarily mothers and wives expected to place their family’s needs before their own. (Zabludovsky 2001). Indeed, the family is very important in Mexican culture (Flake and Forste 2006; Galanti 2003; Ingoldsby 1991). Familismo, also known as concepto de la familia, values putting the needs of the family before individual needs (Coohey 2001). Within familismo, loyalty, solidarity, cooperation, and maintaining the wellbeing of the family are important in both the immediate and extended family (Ayón et al. 2010; Guilamo-Ramos 2009; Marin and Marin 1991).

Although the larger culture permeates society’s views of gender roles, the experiences of IPV among Mexican women may vary as a result of differences in demographic, geographical, and socioeconomic status. For example, Agoff et al. (2005) observed that Mexican women they interviewed in the impoverished and mostly rural state of Chiapas felt that partner violence was justified if inflicted because they had deviated from social norms regarding their relationships, their manner of dressing, and availability to engage in sexual intercourse. On the other hand, a study of women from the urban city of Durango found that leaving home without asking permission, jealousy, defending their children, and bad housekeeping were reasons that licensed their husbands to engage in partner violence (Alvarado-Zaldívar et al. 1998). These studies suggest that men and women’s gender roles are not perceived as equal, and neither is their decision-making power in their relationships. Indeed, Coleman and Straus (1990) have found that violence is most prevalent in relationships where decision-making is non-egalitarian (e.g. the female or male is dominant in making most of the decisions, but decisions are not made together). Therefore, decision-making may also have an effect on the likelihood of experiencing partner violence given predominant gender roles in Mexico (Flake and Forste 2006). Overall, understanding the dynamics of a culture’s gender roles and values is significant in understanding the context of IPV in Mexico.

The research exploring IPV’s relationship to depression and other negative outcomes in Mexico is sparse and limited by methodological shortcomings (Díaz-Olavarrieta et al. 2002; Díaz-Olavarrieta and Sotelo 1996; Hijar-Medina et al. 1997; Ramírez-Rodríguez and Uribe-Vázquez 1993). For example, a retrospective study of women who sought services from domestic violence shelters in a suburban neighborhood in Mexico City between 1989 and 1991 found that one out of every nine women served had attempted suicide (Valdez and Juárez 1998; Valdez and Shrader 1992). In addition, a recent study among recovering drug users in Mexico found that living in domestically violent situations was one of the major risk factors for developing depression and suicidal thoughts among this population (Ortiz-Gomez et al. 2014).

A study conducted in Mexico City by de Castro and colleagues (2015) found that women with children younger than 5 years who experience IPV are at a significantly higher risk for post-natal depression; this risk is exacerbated when coupled with low socioeconomic status, low social support, and a history of depression. However, many of these studies have limited representativeness and low generalizability due to small sample sizes, which make interpretation and comparisons across studies difficult and inappropriate. A notable exception is the study by de Castro et al. (2014), which used a probabilistic representative sample of women with at least one child aged 5 years or younger who participated in the Encuesta Nacional de Salud y Nutricion (National Survey of Health and Nutrition) in 2012. The authors found that women who reported experiences of IPV presented the highest risk of post-natal depression, which was increased by lack of other resources such as social support, food security, and access to healthcare (de Castro et al. 2014). However, this study had the limitation of only including women who had children younger than 5 years of age. Thus, a greater understanding of the cultural context may illuminate risk factors for women experiencing IPV in Mexico and aid in modifying treatment approaches to fit particular needs.

The Relationship of Employment to IPV and Depression

A growing body of research is committed to exploring the impact of IPV on women’s economic status. Inconclusive results in the literature have led to questions concerning whether employment is beneficial or detrimental to women that experience IPV. Some studies across different countries have shown positive outcomes on women’s status from employment, while others have shown that IPV increases when women’s access to financial resources increases (Bott et al. 2005; Heise and García-Moreno 2002; Koenig et al. 2003; Schuler et al. 2013). However, many abusive partners engage in coercive controlling behaviors that directly and indirectly hinder women’s attempts to acquire and maintain a job (Adams et al. 2008; Swanberg and Logan 2005). In coercive controlling relationships, partners may exert demands on the victim and then engage in surveillance or threatening behaviors in order to ensure compliance (Dutton, Goodman, & Schmidt, 2006). On the other hand, research with women in the U.S. and Canada found that relationships solely characterized by IPV but not CC are characterized by poor problem-solving and emotional regulation skills associated with escalating arguments that end in violence (Johnson 1995; Johnson and Ferraro 2000).

Women involved in relationships where a CC is present may be at risk of increased vulnerability if they are to enter the workforce. On the other hand, a study including a large sample of Mexican women found that women who experience IPV without coercive controlling behaviors may find economic independence to be a protective factor (Terrazas-Carrillo & McWhirter, 2015). For instance, data from a nationally representative sample in Mexico found that women engaged in paid employment outside the home were less likely to report depression symptoms than their homemaker counterparts (Lara et al. 1993; Valdez and Juárez 1998). However, this study included all women irrespective of whether they had experienced IPV. Therefore, variations in abusive experiences, gender roles, and cultural values can impact the array of health outcomes that may result from these experiences, including depression (Davies et al. 2015; Hegarty et al. 2013). Researchers postulate that the relationship between IPV and a woman’s employment status is extremely complex and dependent upon the individual factors, family, culture, and local and global economic conditions (Chronister and McWhirter 2006). It is our hope to contribute to the body of knowledge that will help in understanding women’s experience of IPV in relation to employment and depression.

Method

Sample

Participants included Mexican women (n = 13,053) 15 years of age or older who identified as married or cohabiting with a partner at the time of the interview. Data for this research were collected by the Instituto Nacional de Estadística, Geografía e Informática (Mexican National Institute for Statistics, Geography, and Informatics) as part of their National Survey on Household Relationship Dynamics (Encuesta Nacional sobre la Dinámica de las Relaciones de los Hogares [ENDIREH]) in 2011. The sampling strategy used was stratified and probabilistic by using demographic and cartographic information from the 2002 National Household Registry obtained through data collected during the 2000 National Census in Mexico (INEGI 2011a). The primary sampling units included an estimated 160 to 300 households within a geographic area, which were then stratified by state, population size, and urban status (rural, urban, suburb) to provide a representative sample of the Mexican population. In each of the selected households, all women over 15 years of age, currently living with a partner were interviewed in person by field workers trained by INEGI for data collection (Instituto Nacional de Estadística, Geografía e Informática 2011b).

Hypotheses

Given results from studies suggesting an IPV dose–response effect on mental health (Campbell 2002), the first hypothesis proposes there are significant differences in levels of experienced depressive symptoms between mild and severe IPV. Since the literature suggests coercive controlling behaviors have as a main goal to establish pervasive dominance on the victim’s life (Stark 2006), it is likely that a woman employed outside the home may be subjected to higher levels of surveillance and intimidation from her partner to ensure compliance with his demands. This increase in threatening behaviors is likely to be experienced as a traumatic stressor even in the absence of IPV (Dutton and Goodman 2005). On the other hand, it is plausible that employment would no longer be the main traumatic stressor for women in relationships characterized by both CC and IPV. Therefore, the second hypothesis proposes that a woman’s employment status may be a predictor of depression in the context of a coercive controlling relationship but may not be a predictor of depression in the context of a violent relationship characterized by both IPV and CC.

Measurement

The ENDIREH surveyed a variety of contexts where women may experience violence, specifically school, work, and home. All married or cohabiting women 15 years of age and over answered questions on topics including: Household sociodemographic characteristics, experiences at work and school, family of origin dynamics, characteristics of couple relationship dynamics, tensions and conflicts, intimate partner violence, decision-making, financial and social resources, personal liberty, gender role ideology, and household division of labor.

Depression Symptoms Scale

The year 2013 was the first time ENDIREH asked questions to assess the most common symptoms of depression. The complete list of questions is listed on Table 1. Possible answers to these questions were “Yes” and “No.” The Depression Scale was created by adding one point for each affirmative answer to these questions. The Cronbach’s alpha for this scale was .70.

Table 1 Depression symptom scale questions
Full size table

Sociodemographic Variables

The ENDIREH asks interviewees “How old are you today?” and “What is your higher level of education?” to assess age and level of education. In addition, women interviewed answered the following questions to gather information about children and the age at which women dated, married, and had their first child: “In total, how many live children have you birthed?” “How old were you when you had your first child?” “How many live children does your husband have with other women?” “How old were you when you started dating your current husband or partner?” and “How old were you when you married your husband or started living together with your partner?” The ENDIREH also assessed a woman’s employment status by asking “Do you currently work?” A “Yes” answer was coded as “1” and a “No” answer was coded as “0.”

Intimate Partner Violence Scale

The ENDIREH asks questions regarding specific IPV tactics generally defined as abusive according to national and international standards. Women are asked to rate the frequency of such tactics occurring in the context of intimate partner relationships, which can be “several times,” “a few times,” “one time,” or “never.” The questions included in this scale are included on Table 2. The Abuse scale was created by adding 3 points when participants respond “several times,” 2 points when they answered “a few times,” 1 point when they answered “one time,” and 0 points when they answered “never.” Thus, higher scores on the scale indicate high frequency of abuse, and experiencing more types of abuse such as physical, sexual, and emotional. Cronbach’s alpha for this scale was .916.

Table 2 Intimate partner violence scale questions
Full size table

Coercive Control Scale

Questions used to create this scale come from the section on the ENDIREH asking about who holds decision-making power in the relationship. Women were asked to answer who makes decisions in each instance, and the potential answers were “Only the woman interviewed,” “Only the husband or partner,” “Both,” “Another person,” and “Not applicable.” All questions included on this scale are listed on Table 3. The scale was created by adding one point for each “Only husband or partner,” and zero points for all other answers. Thus, higher scores on this scale would indicate higher levels of coercive control displayed by the husband. Cronbach’s alpha for this scale was .93.

Table 3 Coercive control scale questions
Full size table

Gender Role Attitudes Scale

This scale assesses whether the women interviewed endorse egalitarian gender role expectations. Women answered questions regarding gender role attitudes with “Yes,” and “No” answers. One point was added when the interviewees endorsed statements related to more egalitarian gender roles and zero points were added when interviewees endorsed statements related to less egalitarian attitudes. Therefore, higher scores on this scale represent more egalitarian attitudes about gender role expectations. A complete list of the questions included on this scale is listed on Table 4. Cronbach’s alpha for this scale was .98.

Table 4 Gender role attitudes scale
Full size table

Results

Women in this subsample of the ENDIREH had a mean age of 47.39, and a mean education of 3.82, which means that on average, women interviewed had completed middle school. Since employment was measured as a dichotomous variable, the mean of .42 is interpreted as 42 % of the women surveyed by ENDIREH were employed outside the home. In addition, women interviewed had an average of 2.34 children, and a reported mean age of 20.91 at the time their first child was born. The mean number of children their husbands fathered with other women was 3.33. On average, women in the sample started dating their current spouse or partner at age 20.81 and married their current spouse or partner at a mean age of 22.52. The mean score for the Depression scale was 1.62, and it was 5.39 for the Gender Role Attitudes Scale. The mean score for the Coercive Control Scale was 1.92, and the mean score for the Intimate Partner Violence Scale was 4.61 (See Table 5 for descriptive statistics).

Table 5 Descriptive statistics
Full size table

Before testing the first hypothesis, a factor analysis of the items used to create the Intimate Partner Violence Scale was conducted in order to determine whether different items on the scale loaded on different IPV severity levels. The 30 items of the Intimate Partner Violence Scale were subjected to principal components analysis (PCA) using SPSS version 22. Prior to performing PCA the suitability of data for factor analysis was assessed. The Kaiser-Meyer-Oklin value was .86, exceeding the recommended value of .6 (Kaiser 1970) and the Bartlett’s Test of Sphericity (Bartlett 1954) reached statistical significance, supporting the factorability of the correlation matrix. Principal components analysis revealed the presence of 8 components with eigenvalues exceeding 1. However, an inspection of the screeplot revealed a clear break after the second component and it was decided to retain 2 components for further investigation (Tabachnick and Fidell 2001). To aid in the interpretation of these two components, Varimax rotation was performed. The rotated solution revealed the presence of a simple structure, with both components showing a number of strong loadings and all variables loading substantially on only one component (See Table 6). The two component solution explained a total of 22.9 % of the variance, with Component 1 contributing 11.7 % and Component 2 contributing 11.18 %. Component 1 items seemed to allude to mild forms of IPV, while Component 2 items were indicative of more severe forms of IPV (See Table 6).

Table 6 Varimax rotation of two factor solution for intimate partner viiolence scale items
Full size table

Once the factor structure of the Intimate Partner Violence Scale was determined, the first hypothesis was tested by performing an independent samples t-test assuming unequal variances given the differences in the groups’ sample sizes, was conducted to compare depression scores for Severe and Mild IPV. There was a significant difference (p < .05) in depression scores for Mild IPV (Mean = 1.14, SD = .577) and Severe IPV (Mean = 1.03, SD = .271). See Table 7 for details of this difference. In other words, there is support for the first hypothesis regarding significant differences on depression symptoms between women experiencing mild versus severe IPV.

Table 7 Differences in depression symptom scale scores between mild and severe IPV
Full size table

In order to test the second hypothesis, two regression models were conducted to predict depression. The first model included sociodemographic variables, woman’s employment status, gender role attitudes scale, and coercive control scale in order to find out whether employment in a relationship characterized by nonviolent CC would be predictive of depressive symptomology. In the first model, women’s employment status was a significant predictor of depression [t(13,052) = 4.88, p < .001]. The overall model was statistically significant [F(10, 9927) = 55.37, p < .001] its R 2 was .05, which indicates an estimated 5 % of the variance in depression symptoms was explained by the predictors included in the model. The following predictors were statistically significant: age [t(13,052) = –2.06, p < .05], number of birthed children [t(13,052) = 10.74, p < .001], number of children husband fathered with other women [t(13,052) = 4.15, p < .001], woman’s employment status [t(13,052) = 4.88, p < .001], and coercive control [t(13,052) = 19.78, p < .001]. Therefore, predictors of depression in the context of a nonviolent controlling relationship include: number of children a woman has birthed, number of children the husband fathered with other women, coercive controlling behaviors, employment outside the home, and age.

In order to test the second part of the hypothesis, another multiple regression model included the variable intimate partner violence along with all the predictors included in Model 1 (see table 8). This model was statistically significant F(11, 9926) = 725.49, p < .001, and its R 2  = .44, which indicates an estimated 44 % of the variance in depression scores is explained by the predictors included in the second model. Statistically significant predictors included: gender role attitudes [t(13,052) = –2.29, p < .05], coercive control [t(13,052) = 6.21, p < .001], and intimate partner violence [t(13,052) = 83.81, p < .001]. The predictor intimate partner violence alone accounted for 39 % of the variance on depression, while coercive control accounted for only .2 % of the variance on the criterion variable. The increase of R2 from model 1 to model 2 was statistically significant [F(1, 9926) = .393, p < .001]. Therefore, the following variables predicted depression in the context of a relationship characterized by both CC and IPV: gender role attitudes, coercive control, and intimate partner violence. Once IPV was included in the model, employment status was no longer a significant predictor of depression, supporting both parts of the second hypothesis proposed.

Table 8 Multiple regression model
Full size table

Discussion

Our findings provide evidence supporting the existence of different types of violence with different defining characteristics and impacts on mental health (Johnson 1995). Based on these results, we conclude that risk factors predicting depressive symptoms include having an abusive and controlling partner, young age, engaging in paid employment outside the home, and higher birth rates. Specifically, depressive symptomology is more prevalent in relationships characterized by both controlling behaviors and physical and emotional abuse. Our findings indicate that a woman in a nonviolent coercive controlling relationship will have a higher risk of depression if she is employed. However, employment is not a risk factor for depression symptoms among women in relationships characterized by both coercive control and IPV.

Our findings suggest that women employed outside the home who are in relationships characterized by coercive controlling behaviors may experience increases in depression symptomology. This finding is perhaps a result of increased partner threats and surveillance, which may be experienced as traumatic stressors as suggested by the literature (Dutton et al. 2005a; Stark 2006). However, employment was no longer a predictor of depressive symptomology once IPV was introduced to the model. Although unable to establish based on the results obtained in the multiple regression models, it is possible that the presence of IPV in a coercive controlling relationship may become the focus of traumatic stress and increase the likelihood of experiencing depression regardless of employment status.

Moreover, the absence of a negative or positive impact of employment on depression potentially suggests that seeking employment may be a viable alternative for women in relationships where IPV is the central concern. We speculate that while controlling behaviors have been shown to have detrimental effects for women, their nature is insidious rather than chronically severe in relationships where physical, sexual, and emotional abuse is absent. The impact of this insidious pattern of behavior is perhaps underestimated, as our results show its presence increases the risk of developing depressive symptoms. However, IPV (physical, sexual, and emotional) in the context of a controlling relationship becomes the strongest predictor of depression, which is possibly a result of a systematic pattern of abuse.

Implications

Based on these findings, we recommend tailoring domestic violence interventions to fit the context of the woman’s situation. Specifically, programs designed to increase battered women’s financial stability should consider the nature of the relationship she has with her partner. A woman involved in a controlling relationship may become more depressed upon entering the workforce due to the pattern of control and limitations she experiences. In this case, stress inoculation and/or other preventative interventions for depression may be indicated while phasing in employment, due to the increased risk of depression. However, a woman in a relationship where the central concern is IPV may derive some benefits from seeking out employment. Specifically, women who contribute to the family income have more economic power, which makes them more likely to report abuse or leave the relationship (Harris et al. 2005).

Encouraging women in abusive relationships to work outside the home may also stimulate social change by transforming their partners’ expectations and perspectives (Schuler et al. 2013; Vyas and Watt 2009). However, treatment for women involved in violent relationships characterized by coercive controlling behaviors should exercise caution in recommending that they work outside the home due to the increased risk of depression. Many of these women are experiencing control and intimidation that prevents them from making their own choices, working, and communicating with loved ones (Tanha et al. 2009). Women in these relationships need access to more intensive short-term services and resources, such as shelters, health resources, and mental health interventions that address the long-lasting effects related to living in abusive relationships. It is important that these women are provided access to counseling services that focus on the sequelae of abusive relationships, especially PTSD, depression, anxiety, low self-esteem, and low self-efficacy (McWhirter and Altshuler-Bard 2010; Terrazas-Carrillo and McWhirter 2012). Future studies should further explore the impact of gender role ideology as a protective factor of IPV.

Limitations

This study provides valuable knowledge regarding specific protective and risk factors of depression among women experiencing nonviolent CC and also CC in the context of a violent relationship. However, some limitations should be noted. The first limitation is related to the correlational nature of the study, which precludes inferences regarding causality among the variables analyzed. In addition, the study represented a cross-sectional view of a social and psychological phenomena that is generally characterized by a lifetime prevalence. Although the study included a representative sample of Mexican women interviewed as part of ENDIREH, only a subset of cases were analyzed in the study due to missing data. This raises the question of whether only a particular type of respondent answered the questions related to the variables in the study. Nonetheless, in spite of these limitations, this study contributes to the body of research exploring profiles of depression and its relationship to IPV among women of different cultures.

In sum, findings from this study lend additional support to the salience of this issue for Mexican families. Harris et al. (2005) found that traditional familismo and traditional gender role orientations were reported among Mexican-born women interviewed in their study, which contributed to lower reports of abuse. Holding these traditional views also contributed to women being less likely to define some acts of violence as abuse (Harris et al. 2005). Based on the findings of this study, interventions that aim to augment women’s financial independence should include interventions that target the effect of independence on the family. This should combine education for women and men on the meaning of healthy relationships and aim to broaden men’s perception of identity and value beyond their ability to provide for their family (Ayón et al. 2010; Flake and Forste 2006; Ingoldsby 1991). Individuals may also benefit from learning nonviolent alternatives to reach a conclusion during conflicts and coping skills for nontraditional contexts, such as negotiation, conflict resolution, and assertion skills (Harris et al. 2005). Taken together, these findings contribute to our understanding of international research focusing on the effects of IPV and coercive control on women’s mental health and life situations.