A key to the process of acceptance and maintenance of violence

The theory of traumatic bonding is based on the theoretical assumptions of the attachment theory (^{Bowlby, 1969}), learned helplessness theory (^{Seligman, 1975}), and cycle of violence theory (^{Walker, 1979}). It consists of psychosocial structures that consider both social and psychological factors, which are important in the process of understanding intimate partner violence (^{Dutton & Painter, 1981}). A social factor is understood as the entire historical trajectory of people's lives (e.g., exposure to violence, especially in the context of COVID-19, educational level), derived from patriarchal societal relations (^{Dutton & Painter, 1981}; ^{Walker, 1979}). A personal factor is understood as all factors related to the person (e.g., age, gender, personality, self-esteem, emotional dependency, among others) (Dutton & Painter, 1981). Thus, the key to understanding the permanence of women in their abusive relationships possibly lies in two fundamental parameters: the power differential and the intermittency of abuse (^{Dutton & Painter, 1981}). The power differential is related to hierarchy in romantic relationships. As the victim perceives these demonstrations of power, operationalized through violence, the more negative her self-assessment will be (i.e., low self-esteem), and the less likely to leave this relationship she will be, increasingly needing a dominator (^{Dutton & Painter, 1993}).

Self-esteem is defined as a set of feelings and thoughts about oneself (^{Sbicigo et al., 2010}). Women victims of violence can express it negatively with their self-esteem (^{Bigizadeh et al., 2021}; ^{Paiva et al., 2017}). Self-esteem may also act as a moderator in the relationship to increase symptoms of depression and anxiety (^{Costa & Gomes, 2018}) and as a mediator in the relation between domestic violence and the symptoms of depression (^{Kim & Kahng, 2011}). The fear of being alone and a need to always please the partner tend to lower self-esteem and increase emotional dependency (^{Urbiola et al., 2017}), which, in turn, may increase beliefs that legitimize intimate partner violence. Dependency strengthens the bond between the person in a lower condition and the person in an upper condition (^{Dutton & Painter, 1993}). In addition, the dependent woman has greater self-denial, that is, reality denial, once she believes the relationship has no problems (^{Moral & Ruiz, 2009}).

As for the intermittency of abuse, it is explained by the phases of abuse: 1) tension build up; 2) acute battering; 3) contrition, described by ^{Walker’s cycle of violence theory (1979}). ^{Dutton and Painter (1981}) explain that these phases serve to connect the victim of violence to her abuser as a “miraculous bonding”. Thus, it becomes an ambivalence between the idealized husband and the abusive husband. Even after the marital separation, it is common for women to return to have contact with their partners, having difficulties in leaving abusive relationships behind (^{Dutton & Painter, 1981}; ¹⁹⁹³). This is partly because they are reinforced by society that their spouses will change, and, for that reason, they always deserve other chances for this change to occur. This phase is conceptualized as “contrition” (^{Walker, 1979}). With this, reinforcement occurs intermittently between negative and positive emotions in the relationship (^{Dutton & Painter, 1981}). In such relationships, in which physical and psychological abuse are present, the demonstrations of power displayed by body marks and threats to her dignity and that of her children render the victim powerless in the face of these situations. Victims, in turn, attach themselves to these relationships as homeostasis, a process of maintaining internal balance (^{Dutton & Painter, 1993}).

In these cases, emotional attachment is nothing more than a cognitive distortion, characterized by the use of behavioral strategies that aim to keep both (victim and abuser), even if unintentionally, in the situation of intimate partner violence (^{De Young & Lowry, 1992}). The dominator may use indirect strategies designed to emphasize the submission and vulnerability of the dependent person. He can also use direct strategies, manipulating and strengthening the bonds, in order to diminish the feeling of abandonment, either assertively or aggressively (e. g., intimidation, keeping the victim in a situation of emotional dependency) (^{Bornstein, 2006}).

Emotional dependency is configured as a pattern of needs not emotionally met, which produces inadequate dysfunctional responses in interpersonal relationships (^{Urbiola et al., 2017}). Affective dependents let other people decide for them, accepting any condition imposed by the other (^{Moral & Ruiz, 2009}). It functions as a need that is only relieved by the presence of the partner (^{Bution & Wechsler, 2016}). Individuals, in general, lose their identities to live according to the other (^{Dutton & Painter, 1981}), and they even become more accepting of abuse in relationships (^{Petruccelli et al., 2014}).

For ^{Rathus and O'Leary (1997}), emotional dependency is divided into three types: anxious attachment, exclusive dependency, and emotional dependency. Anxious attachment is identified as an anxiety out of fear of separation from the spouse, caused by feelings of abandonment and weak affective attachment. Exclusive dependency is identified as an exclusive dedication to the partner, as if there were no other people who mattered more to them than their partner. And emotional dependency is characterized as a need supplied by the partner, which requires the person to stay with the partner at all costs, even if this is the aggressor. Our research focuses precisely on these aspects of emotional dependency as a contributing factor to the maintenance of abusive relationships.

In this regard, we aimed to know the relations between psychological abuse, self-esteem, and emotional dependency in women during the COVID-19 pandemic. In addition, we aimed to analyze the profile and sociodemographic characteristics of the participants. Therefore, this study focused on the psychosocial implications of the pandemic context in the lives of Brazilian women.

The impacts of the pandemic on marital relationships are still unknown. Our hypotheses are guided by theoretical contributions that justify the need for tests to prove the phenomena we seek to answer.

The following hypotheses were tested: (1) We expect that psychological abuse can positively predict emotional dependency and that this process is mediated by self-esteem, based on the theory of traumatic bonding. We justify our hypothesis based on the argument that psychological abuse is a traumatic event that causes a decrease in self-esteem in the victims, thus contributing to a greater attachment, which is characterized as emotional dependency. The more abuse a woman suffers, the more she needs to have a partner, becoming dependent on this abusive relationship (^{Amor & Echeburúa, 2010}). In this regard, a study by ^{Moral et al. (2017}) checked the existing relation between violence in dating, emotional dependency, and self-esteem in adolescents and young adults. As a result, it was shown that there was greater emotional dependency and lower self-esteem in those victimized as compared to those not victimized. Therefore, we specifically aimed that psychological abuse is also a predictor of emotional dependency and self-esteem in adult women.

(2) Given the cyclical nature of the processes of psychological abuse and emotional dependency, we expect that an alternative model is also possible. In this sense, we specifically aim that emotional dependency can also positively predict psychological abuse and that this relation is mediated by self-esteem. This model has not yet been tested by this theory, although the woman victim of abuse remains in this abusive relationship due to emotional dependency created from the traumatic bonding, which can either be derived from exposure to violence of important people (e.g., father, mother) and from the observed experience itself.

(3) We also predict that the relation between psychological abuse and emotional dependency mediated by self-esteem occurs depending on the age of the woman. We specifically hypothesize that this relational process occurs in younger women. Study such as that of ^{Moral et al. (2017}) showed that women who were in high school were more victimized and were more emotionally dependent on their partners as compared to women who were in higher education. The ^{WHO (2018)} stated that about one-third of girls and women aged 15-49 have already suffered some form of domestic violence. Therefore, age can be an important variable in the relations between these constructs.

(4) And finally, we specifically aim for significant differences in emotional dependency - in the specific factors anxious attachment, exclusive dependency, and emotional dependency - and in psychological abuse in women who were living full-time with their partners, as well as those who were living with their partner during the social isolation recommended by health agencies during the COVID-19 pandemic.

Method

Results

We performed descriptive statistics to analyze the mean and standard deviations of the scales in the sample. We also analyzed the Shapiro-Wilk normality test (^{Ghasemi & Zahediasl, 2012}) to assess the distribution of the data in the sample. The more asymmetry, the further away from zero the factor will be, and the smaller the kurtosis, the less flattened the frequency curve of responses will be. We observed that emotional dependency (in all its factors) and self-esteem had asymmetry close to 0 and the most flattened curve. As for psychological abuse, it had an asymmetry more distant from 0 and more elongated kurtosis. Thus, we consider emotional dependency and self-esteem scores to have approximately normal distributions and psychological abuse to have an asymmetric distribution (Table 1). With this, all the techniques implemented below were conducted with bootstrapping of 1000 simulations for normality correction.

Table 1: Descriptive statistics 

^Note: M: mean; SD: standard deviation; p _S-W: p-value of Shapiro-Wilk test.

To find out if the variables have a relation between themselves, we performed a two-tailed Pearson correlation (Table 2). Psychological abuse was correlated with all factors of emotional dependency, as well as correlated with self-esteem, age of the participant and time of cohabitation with the partner. Thus, we infer that the greater the psychological abuse experienced, the lower the self-esteem and the greater the exclusive dependency and anxious attachment in the participants. The emotional dependency factor correlated negatively with psychological abuse, indicating that the greater the psychological abuse, the lower the emotional dependency. The time of cohabitation also has a positive correlation with psychological abuse, that is, the longer the time spent with the partner, the more psychological abuse the participant may suffer.

Table 2: Correlations between the constructs 

^Notes: PA: Psychological Abuse. ** p > 0,001; * p > 0,05.

For a more robust analysis, we tested hypotheses 1 and 2 through simple mediation models. In this model we grouped the factors of emotional dependency, as these formed the second-order factor in order to assess the phenomenon of emotional dependency in general. The results are shown in Figure 1. In model 1, psychological abuse (PA) significantly predicted emotional dependency (ED) mediated by self-esteem (R² = 0.14). The direct effect (b= 0.47; SE= 0.11; p< 0.001; CI 95%: 0.25; 0.68) of PA on ED was lower than the total effect (b= 0.55; SE= 0.20; p< 0.001; CI 95%: 0.34; 0.75). The significant indirect effect (b= 0.08; SE= 0.03; 95% CI 0.02; 0.15) showed that the indirect relation also exists, through a partial mediation. In the alternative model 2, ED predicted PA mediated by self-esteem (R² = 0.15). The direct effect (b= 0.17; SE= 0.04; p< 0.001; CI 95%: 0.09; 0.25) was lower than the total effect (b= 0.20; SE= 0.04; p< 0.001; CI 95%: 0.13; 0.28). The indirect effect was significant (b= 0.03; SE= 0.01; CI 95%: 0.01; 0.06), also indicating a partial mediation. Thus, hypotheses 1 and 2 were confirmed. There is an indirect relation between the constructs through self-esteem.

Figure 1: Simple mediation models 

An age-moderate mediation analysis was performed with model 1 to test hypothesis 3. The results showed a statistically significant interaction between age and self-esteem (b= 0.02; SE= 0.01), all at level p= 0.04, so that the indirect effect of mediation occurred only for women aged up to 33 years (b= 0.13; SE= 0.05; CI 95%: 0.05; 0.22), but not for women over 34 (b= 0.04; SE= 0.04; IC 95%: -0.03; 0.11). Thus, in women over 34 years of age, emotional dependency was more related to psychological abuse. Whereas in women below the age of 34, the reduction in self-esteem indirectly led to a stronger prediction of psychological abuse over emotional dependency.

We assessed whether there were statistically significant differences in the isolated factors of emotional dependency, regarding the time of cohabitation with the partner, and whether the participants were living with their partners during the pandemic period. We found statistically significant differences regarding the time of cohabitation during social isolation in psychological abuse (F(2.222) = 3.93; p= 0.021) and in exclusive dependency (F(2.219) = 4.86; p= 0.009). Multiple comparisons with Tukey's correction showed that women who live full-time with their spouse (M= 1.20; SD= 1.36) reported greater psychological abuse when compared to those who did not live together (M= 0.55; SD= 0.49), all at level p= 0.018. In addition, a lower exclusive attachment was observed in women who did not live together with the partner (M= 2.48; SD= 0.83) when compared to those who live full-time (M= 2.97; SD= 0.90), all at p= 0.023 and partial (M = 2.94; SD = 1.02), all at level p= 0.009. There was also a difference in relation to living with a partner (F(1. 221) = 5.59; p= 0.019) in anxious attachment. In women who were not living with their partner during social isolation, greater anxious attachment occurred (M= 3.87; SD= 1.23) when compared to those who were living with them (M= 3.46; SD= 1.27). Thus, there is evidence that the time of cohabitation and the fact of living with a partner during social isolation positively related to the psychological abuse suffered, exclusive dependency, and anxious attachment of women.

Discussion

Our goal was to apply the Theory of Traumatic Bonding (^{Dutton & Painter, 1981}) to understand the process of victims of violence in remaining in abusive relationships during the COVID-19 pandemic. Our first hypothesis rested on the fact that women who suffer from psychological abuse decreased their self-esteem and consequently increased the emotional dependency on their partner. Our first hypothesis was corroborated through model 1, by showing that traumatic bonding - represented in the present study by psychological abuse - had a direct relation with emotional dependency, as well as being mediated by negative self-esteem. That is, women suffering from psychological abuse can either maintain an emotional dependency without needing a mediator, or they can legitimize the acceptance of psychological abuse through negative self-esteem.

This data is still little explored, although there are studies that focus on the relation between emotional dependency and violence from the perspective of the investment model (^{Drigotas & Rusbult, 1992}), which show that people remain in the relationship, even dissatisfied, for they have a need to fulfill that cannot be gratified in other ways. This need is called emotional dependency, which tries to fill in a distorted way what is not gratified in any way if the person is not with the partner exclusively (^{Bornstein, 2006}; ^{Bution & Wechsler, 2016}). And this can be an important step to understand the process of remaining in abusive relationships (^{Amor & Echeburúa, 2010}; ^{Rathus & O’Leary, 1997}).

However, when assessing the preliminary correlations between the constructs, we observed that only anxious attachment and exclusive dependency correlated positively and more strongly with psychological abuse. The emotional dependency factor, contrary to expectations, correlated negatively and with a coefficient of low magnitude. This result may have occurred due to issues related to the content of the items of the Spouse-Specific Dependency Scale for Women, validated by ^{Paiva et al. (2021}), which was used in this study. It shows that most of its items expose a more positive and even romanticized connotation of emotional dependency, such as items 18 (“My partner is the only one who really understands me”); 21 (“I like my partner because he/she is understanding”); 24 (“I prefer to face the adversities of life with my partner by my side”). Thus, the suggestion is that evidence of more positive attachments is negatively correlated with psychological abuse for masking dependency as a form of romantic love.

Regarding attachment, this is a predictor for violence (^{Rathus & O'Leary, 1997}). So, the traumatic bonding becomes stronger as there is attachment in a cyclical or feedback process. And it was precisely in this argument that we established the alternative model (model 2). Thus, we found that dependency also has a direct relation with psychological abuse, and that this relation is also mediated by self-esteem. This indicates that the more dependent women are, the greater the negative impact on self-esteem, and the more vulnerable they are to psychological abuse. This process confirmed our hypothesis 2, being important to understand that the maintenance of psychological abuse also lies on the dependency that women have on their partner, as exposed by the Theory of Traumatic Bonding. With this, we show that the process is cyclical: as the woman suffers violence, she decreases self-esteem and becomes emotionally dependent on her aggressor; as she is emotionally dependent, she continues to accept her abusive suffering by remaining in the relationship. This whole process can represent a chain that is justified by the negative self-esteem for some women, because when we legitimize self-esteem as a mediator, this represents one of the essential elements to be worked on in psychotherapeutic processes.

Women who are in aggressive relationships report more non-voluntary dependency (i.e., they avoid ending the relationship because they think they might experience a worse situation in another relationship; ^{Thibaut & Kelley, 1959}) than women who have not experienced abusive relationships (^{Tan et al., 2018}). This is because, depending on the level of commitment of these women in their relationships, this can become a tool for maintaining abusive relationships (^{Bornstein, 2006}; ^{Bution & Wechsler, 2016}). It is important to highlight that the literature points out that low satisfaction in marital relations in couples with IPV is recurrent and, therefore, it is not necessarily what influences the maintenance or disruption of the violent cycle (Tan et al., 2018). The more committed to the relationship this woman is, the more predisposed she will be to maintain this relationship and the more difficult it will be to get out of it, because she might blame herself for the ending of the relationship. In this sense, emotional dependency manifests itself in different forms: it can be the commitment to the relationship that leads to devotion of the partner (Tan et al., 2018), or it can also lead to anxious attachment of the partner, such as the fear of losing their love object (^{Bowlby, 1969}; ^{Dutton & Painter, 1993}; ^{Dutton & White, 2012}). This can also vary due to the age groups of women (^{Petruccelli et al., 2014}).

Our study showed that women aged between 18 and 33 had a stronger emotional dependence than women aged over 34. Furthermore, we observed that, in this group of women, the psychological process tested was stronger, that is, a relation between psychological abuse and emotional dependency showed higher coefficients. This data corroborates with our hypothesis 3 and is within the confidence interval proven by WHO (2018), as was also similar to the study conducted by ^{Moral et al. (2017}), in which girls up to the age of 18 (high school students) were more victimized, causing them to lower their self-esteem and raising their level of emotional dependency on their partner. In addition, ^{Petruccelli et al. (2014}) found in their studies that women aged 32-34 years had higher levels of exclusive dependency and anxious attachment, that is, the more anxious attachment, the greater their fear of losing their ideal partner. In a way, these studies are coherent, considering that one of the possible factors of this dependency is the traumatic bonding and the level of commitment these women have regarding their relationships.

A study by ^{Christman (2009}) noted that traumatic bonding mediates the relation between attachment and the intention to return to the abuser. The victims also forgive more easily for the purpose of remaining in the relationship. More exclusive dependency is demonstrated in relationships in which women have been engaged for more than 4 years and where there has already been a traumatic bonding, and/or where they have already been left by their partners (^{Petruccelli et al., 2014}).

Our study also showed that women who were living with their partner during the isolation period had more exclusive dependency on their partner. That is, the more time women spend inside their homes, the more they become dependent on their partners. We obtained evidence that confirmed our hypothesis 4, showing that those women who are living with their partners may have higher levels of anxious attachment, being more afraid to stay away from their partners for fear of being abandoned. In addition, time of cohabitation showed significant differences in psychological abuse, such that women who live with their spouses full-time reported facing more psychological abuse.

In this regard, according to ^{Bradbury-Jones and Isham (2020}), the current COVID-19 pandemic presents itself as a unique and distressing paradox for women who are victims of violence in intimate relationships. They either stay at home with their partner at risk of further violence, or they leave home and risk being exposed to a highly infectious and dangerous virus. Coercive control is a hallmark of abusive relationships and according to studies, and, according to studies, abusers use the COVID-19 virus to instigate fear and obedience in their partners (^{Godin, 2020}; ^{Sandler, 2020}), which is part of the repertoire of psychological abuse.

Despite the results, the present study is not without limitations. One of the limitations found is the lack of empirical studies that prove the hypotheses listed in this study, so that it can be compared cross-culturally. Most studies are about the theory itself (^{Amor & Echeburúa, 2010}; ^{Ramos, 2005}) and do not focus on predictive or experimental models. One of the suggestions for further studies would be to explore more research to expand this theoretical framework, which should address social factors. Another limitation refers to the sampling characteristics, since it is non-probabilistic (by convenience), and a large part of the sample is composed of women with higher education, therefore, considering the generalization of the results for the entire population of women. A minority of 6.3% reported suffering psychological abuse or other types of abuse (physical, sexual, patrimonial). It is recommended that future studies reach participants of specific samples - such as women who are suffering psychological abuse concomitant with other types of abuse - that aim to relate the recurrence of the various forms of violence to emotional dependency and self-esteem.

Conclusion

This study was important to test a psychological process of the relation between psychological abuse, self-esteem, and emotional dependency, under the domain of the Theory of Traumatic Bonding (^{Dutton & Painter, 1981}). That said, we comprehend that the present study is necessary and relevant, especially in the current context of the COVID-19 pandemic. That said, it is understood that the present study is necessary and relevant, especially in the current context of the COVID-19 pandemic, which, in turn, hinder escape routes, search for help, and ways of coping. In addition, we hope this study contributes to psychotherapeutic processes by applying motivational interviews and self-management of emotions to victims of violence (^{Sussman, 2010}). We also hope this study contributes to the proposition of educational policies regarding the impact of violence on self-esteem - consequently on dependency -, as well as on how dependency plays a role in maintaining and accepting violence against women. By using scientific approaches (^{Bowlby, 1969}; Dutton & Painter, 1993; Dutton & White, 2012; ^{Thibaut & Kelley, 1959}; ^{Walker, 1979}), our study demonstrated how psychological mechanisms can act on the acceptance of psychological abuse as well as on the acceptance of emotional dependency. Such approaches helped to identify the mechanisms of personal factors that underlie the risk of lethality of psychological abuse against women.