Hyper-Sensitivity to the Perpetrator and the Likelihood of Returning to Abusive Relationships

Abstract

Study questions: Although most women who are subjected to intimate partner violence attempt to leave their abusive partners, many return, and resultantly are at risk for even greater violence. Research to date has documented relations between several factors (income and economic dependence, frequency of intimate partner violence (IPV), fear of violence escalations, history of childhood abuse, and post-traumatic stress disorder symptoms) and women’s returning to their abusive partners. Nevertheless, the contribution of women’s emotional bonds with their violent partners, known as identification with the aggressor (IWA), in explaining their perceived likelihood of going back to the relationship, has remained unclear. Subjects: The current study, conducted among 258 Israeli women who had left their violent partners, aimed to fill this void. Methods: An online survey was conducted. Demographic variables, history of childhood abuse, frequency of IPV, economic dependence on former partner, fear of future violence escalation, post-traumatic stress disorder symptoms, IWA, and perceived likelihood of returning to the relationship, were assessed via self-report questionnaire. Findings: Results indicated that two aspects of IWA—becoming hyper-sensitive to the perpetrator and adopting the perpetrator’s experience—were related to women’s perceived likelihood of returning to the relationship. Furthermore, a logistic regression analysis indicated that only two factors—income and becoming hyper-sensitive to the perpetrator—uniquely contributed to explaining the likelihood of returning to abusive partners. Major implications: The current findings suggest that women’s tendency to be highly attuned to their partners’ feelings and needs, as a part of IWA, may impede their ability to permanently leave abusive relationships.

Introduction

Intimate partner violence (IPV) against women, defined as physical, psychological, or sexual acts of violence perpetrated by a partner, is a substantial global public health concern. Evidence suggests that around a third of women worldwide (30%) have suffered from IPV during their lifetimes (García-Moreno et al., 2013). Moreover, large-scale crises, such as the current COVID-19 pandemic, are known to escalate IPV. As such, current rates might be even higher (Mazza et al., 2020). The implications of IPV are far-reaching for women’s mental and physical health, and include increased health problems such as injury, chronic pain, and physical difficulties (Campbell, 2002; Coker et al., 2000; Tiwari et al., 2008), posttraumatic stress disorder (PTSD), depression, anxiety, and suicide attempts (Blasco-Ros et al., 2010; Breiding et al., 2014; Caldwell et al., 2012; Campbell, 2002; Dekel et al., 2019; Lahav et al., 2018; Lahav et al., 2019a; Pico-Alfonso et al., 2006; Taft et al., 2011; Tiwari et al., 2008).

Unfortunately, it is still unclear how to end IPV once it begins (Goodman et al., 2005). Evidence has indicated that although most women attempt to leave their abusive partners (Goodman et al., 2003) many return (Strube, 1988). In a study among 104 female residents in a facility for IPV victims, the majority (66.3%) reported that they had separated from and returned to their abusive partners at least once, and of them, 97.1% indicated that they left and returned multiple times (Griffing et al., 2005). Thus, although the decision to leave a violent partner may seem like an obvious choice given the negative consequences of IPV, ending an abusive relationship appears to be extremely complicated (R. Walker et al., 2004) and to require several attempts (Griffing et al., 2002; Lerner & Kennedy, 2000; Rhatigan et al., 2006).

Estrangement and separation are known to increase the risk for an escalation of violence and femicide (McFarlane et al., 2002). Nevertheless, research suggests that returning to an abusive relationship may put women at even greater risk. Specifically, it has been suggested that abusive partners may perceive women’s attempts to leave as disobedience, and so may respond with increased violence when they return, as a way to punish them and reassert power within the relationship (Anderson, 2003). Research has revealed that women who leave their partners temporarily may be subjected to greater violence than women who stay in the relationships in the first place or those who permanently leave (Anderson, 2003; Bell et al., 2007; Fleury et al., 2000). For example, results of a longitudinal study conducted among 206 low-income, primarily Black, battered women indicated that relationship trajectory over 1 year predicted women’s exposure to violence: women who were completely apart (i.e., remained uninvolved with their partners over 1 year) reported the lowest rates of physical abuse, psychological abuse, and stalking; followed by women who stayed in the relationship; women who were together then apart (i.e., were in the relationship at the first 6 months of the study and then were out of the relationship at the last 6 months of the study); and women who returned to their partners (i.e., were involved with their partner at least once after separation; Bell et al., 2007). Hence, identifying the factors that might explain women’s intentions to return to their abusive partners is critical and may promote clinical interventions.

Whereas many studies have explored factors associated with women’s decisions to stay in/leave abusive relationships, only a few have aimed to detect factors that explain women’s returning to their abusive partners. These studies suggest that several factors may explain women’s actual returning to or appraisals regarding the likelihood of returning to these partners (Anderson, 2003; Griffing et al., 2005; 2002; I. M. Johnson, 1992). Specifically, women are more likely to return to abusive relationships if they are economically dependent on their partners (Anderson, 2003; Griffing et al., 2002). Anderson (2003) found that 45.9% of participants returned due to lack of money, and a study conducted by Griffing et al. (2002) revealed that economic need was one reason for women having returned in the past. Frequency or severity of IPV and fear of further violence may also be implicated in women’s decisions to return to these relationships (Foa et al., 2000; Horton & Johnson, 1993; Strube & Barbour, 1983; L. E. Walker & Meloy, 1998). Women’s fear of further violence appears to be warranted, as separating from an abusive partner may result in an escalation of violence (Browne & Bassuk, 1997), and has been found to be a risk factor for femicide and attempted femicide (Campbell et al., 2007; Garcia et al., 2007; McFarlane et al., 2002).

Although the association between a woman being a mother and the likelihood of her returning to an abusive partner has not been explored in detail, studies on women’s decisions about whether to stay in or leave abusive relationships suggest that this factor may also be involved in women’s decisions about reuniting with abusive partners. These studies have shown that being a mother may have contradictory effects on women’s decisions to end abusive relationships (Kelly, 2009; Rhodes et al., 2010; Zink et al., 2003). On the one hand, mothers want to protect their children from the damaging effects of violence and, therefore, might be more inclined to terminate abusive relationships (Gillum, 2008; Kim & Gray, 2008). However, on the other hand, they may not believe they can raise their children by themselves, may fear losing custody to the partner, and may wish to avoid exposing their children to the instability created by involvement with the legal system—all of which may lead them to refrain from leaving their partners (Bui, 2003; Rhodes et al., 2010). These effects may also be implicated in women’s decisions to reunite with their partners, so that being a mother may contribute to women’s likelihood of returning to abusive partners.

History of childhood abuse, posttraumatic stress disorder (PTSD) subsequent to IPV, as well as the unique relational dynamics between women and their batterers, may also contribute to women’s actual returning or perceived likelihood of doing so. Childhood abuse, which tends to shape attachment patterns and interpersonal schemas (Crawford & Wright, 2007; Lahav & Elklit, 2016; Stovall-McClough & Cloitre, 2006), may lead women to endure partner violence as a way to preserve the relationship (Young & Gerson, 1991), increasing the likelihood of their returning to said relationship. Findings of a study conducted among 90 battered women support this notion, indicating that women who had a history of childhood sexual abuse were more likely to have previously returned to the batterers (Griffing et al., 2005).

Posttraumatic stress disorder symptomatology post-IPV may also be associated with women’s returning to the relationship, although the directionality of associations is unclear. On the one hand, hyperarousal symptoms and re-experiencing of past abuse as part of PTSD symptoms may serve as powerful reminders of the danger of abusive relationships and thus reduce the chances of women’s returning to the relationship. A previous study among women indeed revealed associations between higher levels of PTSD symptoms and the women’s greater confidence in their ability to leave the abusive partners (Lerner & Kennedy, 2000). However, as PTSD symptoms have a debilitating effect on functioning (Kessler, 2000) and may substantially degrade individuals’ psychological, social, and economic resources (Hobfoll, 1989; D. M. Johnson et al., 2007), they may impede women’s ability to cope with the stressors involved in the post-separation period and thus may promote their returning. Indeed, a study conducted among a community sample of 286 Canadian women revealed that higher levels of PTSD symptoms increased women’s risk of being unable to stay away from a former or new abusive partner over a 12-month period (Abdulmohsen Alhalal et al., 2012).

Lastly, another factor that may explain returning to an abusive relationship is women’s strong bonding to their violent partners. The IPV literature has documented this unique form of affiliative emotional connection that abused women often have towards their partners, and several conceptualizations have been offered, such as traumatic bonding (Dutton & Painter, 1981), Stockholm syndrome (Graham, 1995; Graham et al., 1988; Wallace, 2007), and identification with the aggressor (IWA; Frankel, 2002; Lahav et al., 2019b), the last of which is at the heart of the current study.

Identification with the aggressor, a concept which was originally developed by Ferenczi (Ferenczi, 1932; 1933), represents a broad, automatic reaction wherein abuse victims take on their perpetrators’ experience as a way to survive abuse. Although Ferenczi’s theory mainly focused on childhood abuse (Frankel, 2002; Lahav et al., 2019c), IWA appears to be applicable to IPV (Lahav, 2021), and to reflect a mental fusion of victims with their abusive partners that goes beyond the emotional attachment common in intimate relationships. Although the defensive function of IWA implies that it might be more prominent under conditions of severe and chronic abuse, theoretical work suggests that IWA could develop as a result of milder forms of abuse when there is power asymmetry between the victim and perpetrator (Frankel, 2002). Identification with the aggressor entails four intertwined components: losing one’s agency and replacing it with the perpetrator’s; becoming hypersensitive to the perpetrator; adopting the perpetrator’s experience concerning the abuse; and identifying with the perpetrator’s aggression (Lahav et al., 2019b).

In order to survive assaults, abuse victims react with mental subordination—they lose their connection with their own feelings, urges, and needs, and have their own agency replaced by the perpetrator’s (Frankel, 2002; Mucci, 2017). To anticipate these attacks and decrease the peril, victims become highly attuned to their perpetrators’ inner experience, and learn “from the inside” their perpetrators’ feelings/needs. Furthermore, they adopt their perpetrators’ point of view and internalize their perpetrators’ aggression; as a result they may deny or justify the abuse and carry out both inward and outward aggression (Lahav et al., 2020; Lahav et al., 2019b).

Although IWA may have an important defensive function that enables victims of IPV to survive the unbearable reality of abuse, it may impede their ability to keep their distance from their violent partners after separation. Victims who are characterized by strong identification with their aggressor may feel greater affection and empathy towards their abusive partners, which may propel them back into the relationship. Their loss of their own sense of agency and their mental subordination may hamper their ability to experience psychological independence after separation and to resist their partners’ pressure to resume the relationship. Moreover, victims’ adoption of their partners’ perspective, and their hypersensitivity towards their partners, may lead to minimizing/denying the abuse, or feeling sympathy for/appeasing these partners, who often display remorse for their past violent acts after separating. Empirical evidence has revealed some trends that are in line with this view (Baker, 1997; Griffing et al., 2002, 2005). In a study among women residing in a facility for victims of domestic violence, participants with a history of past separations were significantly more likely to indicate that they might return to the batterer due to their continued emotional attachment (Griffing et al., 2002). Nevertheless, these studies did not explore identification with the aggressor. Hence, the link between this phenomenon and women’s returning to, or appraisals regarding returning to the partners, has remained unclear.

As such, the present study, which was conducted among women who had separated from their abusive partners, explored the link between IWA and women’s appraisals regarding the likelihood of returning to their partners. Specifically, the current investigation strove to explore whether IWA makes a unique contribution to explaining women’s perceived likelihood of returning to their partners, in comparison with demographic variables, as well as in comparison with external and victim-related factors that have been documented in the literature. Being the first, presumably, to address this subject matter, the current study was exploratory in nature. Based on the theory of IWA, two main hypotheses were set:

• (1) IWA will be related to women’s perceived likelihood of returning to their partners.

• (2) IWA will make a unique contribution in explaining women’s perceived likelihood of returning to their partners, above and beyond demographic, external, and victim-related factors (i.e., income, having children, economic dependence upon the partner, frequency of IPV, fear of violence escalations, history of childhood abuse, and PTSD symptoms).

Methods

Participants and Procedure

An online survey was conducted among a convenience sample of Israeli female adults. Participants were recruited through a Facebook advertisement from April 1–25, 2020. Facebook users were eligible for this study if they were female, ≥ 18 years old, and living in Israel. The survey was advertised as a study exploring the implications of stressful life events among women, and was accessible through Qualtrics, a secure web-based survey data collection system. The survey took an average of 30 minutes to complete. It was anonymous, and no data were collected that linked participants to recruitment sources. The Tel Aviv University institutional review board (IRB) approved all procedures and instruments. Clicking on the link to the survey guided potential respondents to a page that provided information about the study’s purpose, the nature of the questions, and a consent form. The first page included researcher contact information as well as contact information for several Israeli organizations that provide IPV support and treatment. Each participant was invited to take part in a lottery that included four $60 gift vouchers.

A total of 983 women answered some of the survey’s questionnaires. Of them, 258 participants (26.2%)—who were previously subjected to IPV, left their abusive partners, and provided data regarding the study variables—were included in the present analyses. Table 1 describes the sample’s demographic characteristics. All participants were Jewish and their ages ranged from 18–73. Most had children, were secular, had some degree of higher education, and were employed. More than 40% had an average or above-average income. Most reported a history of physical, emotional, or sexual abuse in childhood based on the childhood trauma questionnaire (CTQ; Bernstein et al., 2003).

Table 1.

Description of Demographic Characteristics among Participants (n = 258).

	M (SD) or n (%)
Age, M (SD)	41.98 (14.08)
Having children
Yes	155 (60.1)
No	103 (39.9)
History of childhood abuse
Yes	219 (84.9)
No	39 (15.1)
Education, n (%)
High school diploma or less	59 (22.9)
Some higher education	72 (27.9)
Bachelor’s degree	70 (27.1)
Master’s degree and above	57 (22.1)
Religiosity, n (%)
Secular	192 (74.4)
Religious/traditional	51 (19.8)
Other	15 (5.8)
Employment status, n (%)
Working in a full or part-time job	134 (51.9)
Furloughed	43 (16.7)
Unemployed	10 (3.9)
Retired	8 (3.1)
Receives stipend from National Insurance Institute	21 (8.1)
Other	42 (16.3)
Income, n (%)
Below-average income	144 (55.8)
Average income or above	114 (44.2)

Participants reported various types of IPV violence: 171 (66.3%) reported physical violence, 235 (91.1%) reported verbal violence, 207 (80.2%) reported psychological violence, and 143 (55.4%) reported sexual violence. Thus, the vast majority of the sample (n = 235, 91.1%) was classified as having had at least two types of violence inflicted upon them by former partners. The average frequency of IPV, as reflected by the mean score of the questionnaire developed by Eisikovits et al. (2004), was 0.81 (±0.60). The average duration of the relationship with the abusive partner was 24.7 (±17.8) months; the average time since participants had left their abusive partners was 33.5 (±29.9) months.

Measures

Demographic variables

Participants completed a brief demographic questionnaire assessing age, education, income, religiosity, employment status, and whether they had children.

Intimate partner violence

Participants completed a questionnaire developed by Eisikovits et al. (2004) for use in the first Israeli national survey on domestic violence. The original questionnaire included 13 items measuring different types/frequency of violence: verbal assault (e.g., cursing, insulting), psychological or emotional abuse (e.g., threatening, stalking), and physical assault (e.g., punching or kicking, slamming against the wall). For the present study, four items measuring sexual violence were added (e.g., forcing intercourse, coercing sexual interaction). For each of the items, participants were asked to rank the frequency of past abuse on a 5-point Likert scale ranging from 0 (never) to 4 (every day). Frequency of IPV was calculated by averaging scores on all 17 items. In addition, participants were asked to indicate the duration of the relationship with the abusive partner and the time that had passed since they left their abusive partners. Yet, given the high rates of missing data regarding the duration of the abusive relationship (31%) and time since they left their abusive partners (41%), these variables were not included in the current analyses.

Perceived likelihood of returning to abusive partner

Participants were asked to indicate their odds of returning to the violent partners on a 7-point scale ranging from 1 (There is no chance that I will return to the partner) to 7 (I will surely return to the partner). Of the total sample, 215 (83.3%) participants indicated that there was no chance they would return to their abusive partners; 30 (11.7%) indicated that it was very much not probable or not probable; and 13 (5.0%) indicated that returning to their partner was probable, very probable, or certain. To overcome the skewness of the variable, it was transformed into a dummy variable with “0” indicating no chance of returning to the partner (scores of 1 on the original continuous scale), and “1” indicating that there was some chance of returning to the partner (scores of 2 and above on the original continuous scale).

Economic dependence on former partner and fear of future violence escalation

Participants were asked to indicate whether they experienced economic dependence on their former partners and fear of future violence escalation on a 6-point scale ranging from 1 (not true at all) to 6 (very much true).

Childhood abuse

A history of childhood abuse was assessed via the childhood trauma questionnaire (CTQ; Bernstein et al., 2003). For this study, only items measuring physical abuse (e.g., “hit hard enough to see a doctor”), sexual abuse (e.g., “was molested”), and emotional abuse (e.g., “felt that parents wished he or she was never born”) were utilized. The items were rated on a 5-point, Likert-type scale with response options ranging from 1 (never true) to 5 (very often true). Participants were classified as having a history of abuse if they had scores which were higher than one of the cutoff scores suggested by Tietjen et al., (2010): physical abuse ≥8; sexual abuse ≥6; and emotional abuse ≥9. In this study, internal consistency reliabilities for physical abuse (α = 0.83), sexual abuse (α = 0.91), and emotional abuse (α = 0.88) were good.

Posttraumatic stress disorder symptoms

Posttraumatic stress disorder symptoms subsequent to past IPV were measured via the PCL-5 (Weathers et al., 2013). This 20-item scale corresponds to the newly approved PTSD symptom criteria in the Diagnostic and Statistical Manual of Mental Disorders (5th ed., DSM–5; American Psychiatric Association, 2013). Participants were asked to indicate the extent to which they experienced each PTSD symptom on a 5-point Likert scale ranging from 0 (not at all) to 4 (extremely). The original version was adapted so that the index event was the past exposure to IPV (example items: “feeling very upset when something reminded you of your past experience of IPV”; “avoiding memories, thoughts, or feelings related to your past experience of IPV”; and “having strong negative feelings such as fear, horror, anger, guilt, or shame”). A total score of PTSD symptoms was calculated by summing all 20 items. Although not a definitive diagnostic measure, preliminary research suggests a cutoff score of 33 is a useful threshold to indicate symptomatology that may be at clinical levels (Bovin et al., 2016). The PCL-5 demonstrates high internal consistency and test-retest reliability (Bovin et al., 2016). In this study, internal consistency reliability was excellent (α = 0.96).

Identification with the aggressor. Levels of IWA were assessed via the Identification with the Aggressor Scale (IAS),

a 23-item self-report questionnaire (Lahav et al., 2019b). The items were presented to the respondents as reflecting “possible reactions that people may experience as a result of abuse or offense.” Participants were asked to rate on an 11-point Likert-type scale, ranging from 0% (never) to 100% (all the time), the frequency with which they experienced each manifestation of IWA in regard to their violent ex-partners. The scale comprises four subscales: adopting the perpetrator’s experience concerning the abuse (9 items; e.g., “Some people feel that the point of view of their perpetrator is the right one”); identifying with the perpetrator’s aggression (5 items; e.g., “Some people feel that they behave as aggressively as their perpetrator”); replacing one’s agency with the perpetrator’s (5 items; e.g., “Some people do not know what they want in the presence of their perpetrator”); and becoming hypersensitive to the perpetrator (4 items; e.g., “Some people ‘read the thoughts’ of their perpetrator”). Although significant correlations were found between the IAS subscales (r_s ranged from 0.51 to 0.76), they appear to reflect different facets of IWA. Therefore, all four subscales were used in the present analyses. The IAS has been shown to have good psychometric properties, including high construct and criterion validity, as well as high internal reliability (Lahav et al., 2019b). In this study, internal consistency reliabilities for adopting the perpetrator’s experience concerning the abuse (α = 0.91), identifying with the perpetrator’s aggression (α = 0.94), replacing one’s agency with that of the perpetrator (α = 0.82), and becoming hypersensitive to the perpetrator (α = 0.88) were good.

Analytic Strategy

Analyses were conducted via SPSS 27. To explore the relations between IWA and women’s perceived likelihood of returning to their abusive partners, a multivariate analysis of variance was conducted. To explore the relations between frequency of IPV, fear of violence escalation, economic dependence upon the partner, and PTSD symptoms, and women’s perceived likelihood of returning to their abusive partners, one-way analyses of variance were conducted. To explore the relations between having children, income, history of childhood abuse, and women’s perceived likelihood of returning to their abusive partners, chi-square tests were conducted. Lastly, to explore the unique contribution of IWA in explaining the likelihood of returning to the abusive partner, a multiple logistic regression analysis was conducted. Variance inflation factors (VIFs) were examined to assess for issues with multicollinearity among the independent variables (IWA) and covariates (i.e., demographic, external, and victim-related factors) in the regression analysis. Results revealed that all were within the acceptable range (all VIFs were smaller than 2), indicating that multicollinearity was not a problem in the present analysis.

Results

Characteristics of the Current Sample concerning the Study Variables

The vast majority (83.3%) indicated that there was no chance they would return to their abusive partners, with the rest indicating that there was some chance they would return to their abusive partners. More than half of the women reported having children (60.1%). Average levels of fear of violence escalation and economic dependence upon the partner were 1.61 (±1.31) and 1.50 (±1.20), respectively. Average levels of PTSD symptoms were 23.79 (±19.22), and 80 participants (31.0%) had a total score of 33 and above, suggesting that their PTSD symptoms were clinically significant. Participants’ IWA average levels were 30.64 (±24.02) for adopting the perpetrator’s experience, 28.16 (±27.72) for identifying with the perpetrator’s aggression, 36.16 (±27.72) for replacing one’s agency with that of the perpetrator, and 35.84 (±27.95) for becoming hyper-sensitive to the perpetrator.

Associations between the Study Variables and the Likelihood of Returning to the Partner

As shown in Table 2, only some of the variables were found to have significant relations with women’s perceived likelihood of returning to their abusive partners. Having children, income, and a history of childhood abuse were related to women’s perceived likelihood of returning to their abusive partners. The proportions of women who had children were significantly lower among women who indicated that there was some chance they would return to their abusive partners than among women who indicated that there was no chance they would ever return to their abusive partners. Additionally, the proportions of having a below-average income or a history of childhood abuse were significantly higher among women who indicated that there was some chance they would return to their abusive partners than among women who indicated that there was no chance they would ever return to their abusive partners.

Table 2.

The Relations between Identification with the Aggressor, Demographic, External and Victim-Related Factors, and Women’s Perceived Likelihood of Returning to their Abusive Partners (n = 258).

	Participants Who Indicated that there was No Chance of Returning to their Abusive Partners (n = 215)	Participants Who Indicated that there was Some Chance of Returning to their Abusive Partners (n = 43)	F (1,256) or X2 (1)
History of childhood abuse (yes: n, %)	178 (82.8)	41 (95.3)	4.40*
Frequency of IPV (M, SD)	0.82 (0.61)	0.73 (0.51)	0.80
Having children (yes: n, %)	135 (62.8)	20 (46.5)	3.60*
Income (below-average income; n, %)	113 (52.6)	31 (72.1)	5.55*
Fear of violence escalation (M, SD)	1.54 (1.27)	1.93 (1.45)	3.15
Economic dependence upon the partner (M, SD)	1.48 (1.21)	1.58 (1.16)	0.26
PTSD symptoms (M, SD)	22.43 (18.24)	30.56 (22.53)	6.54*
Adopting the perpetrator’s experience (M, SD)	28.76 (23.80)	40.03 (23.16)	8.09**
Identifying with the perpetrator’s aggression (M, SD)	23.41 (27.06)	31.81 (30.14)	3.32
Replacing one’s agency with that of the perpetrator (M, SD)	37.38 (27.84)	42.05 (27.10)	1.02
Becoming hyper-sensitive to the perpetrator (M, SD)	33.70 (27.91)	46.57 (25.4)	7.80**

Note. IPV = intimate partner violence. M = mean, SD = standard deviation, n = number.

* p < 0.05 **p < 0.01 ***p < 0.001.

Post-traumatic stress disorder symptoms were also related to women’s perceived likelihood of returning to their abusive partners: Participants who indicated that there was some chance they would return to their abusive partners had higher levels of PTSD symptoms than did participants who indicated that there was no chance they would ever return to their abusive partners. Lastly, adopting the perpetrator’s experience and becoming hyper-sensitive to the perpetrator were both related to the likelihood of returning to the abusive partner: Participants who indicated that there was some chance they would return to their abusive partners reported higher levels of adopting the perpetrator’s experience and becoming hyper-sensitive to the perpetrator than did participants who indicated that there was no chance they would ever return to their abusive partners.

The Unique Contribution of Identification with the Aggressor in Explaining Women’s Perceived Likelihood of Returning to the Abusive Partner

To explore the unique contribution of IWA in explaining the likelihood of returning to the abusive partner, a multiple hierarchical logistic regression analysis was performed. Only variables that were found to be related to the likelihood of returning to the abusive partner were included in the analysis. Adopting the perpetrator’s experience and becoming hyper-sensitive to the perpetrator served as independent variables; having children, income, history of childhood abuse, and PTSD symptoms served as covariates; and the likelihood of returning to the abusive partner served as the dependent variable.

The analyses included four blocks. The first block consisted of demographic factors (having children and income). The second block consisted of history of childhood abuse. The third block consisted of PTSD symptoms. The fourth block consisted of adopting the perpetrator’s experience and becoming hyper-sensitive to the perpetrator. The first logistic regression analysis was conducted utilizing the enter method; that is, all variables of each block were included in the model. This analysis, however, did not produce significant effects for any of the variables. Thus, a second logistic regression analysis was conducted, utilizing the forward Likelihood Ratio method, so that only variables that made a significant contribution in explaining the likelihood of returning to the abusive partner were included in the model. Results of the analysis are presented in Table 3.

Table 3.

Logistic Hierarchical Regression Explaining the Likelihood of Returning to the Abusive Partner (n = 258).

	Variable	OR (95% CI)
Step 1
	Income	2.33* (1.14–4.78)
Step 2
	Income	2.33* (1.14–4.78)
Step 3
	Income	2.08* (1.01–4.31)
	PTSD symptoms	1.02* (1.01–1.04)
Step 4
	Income	2.08* (1.01–4.31)
	PTSD symptoms	1.00 (0.99–1.03)
	Becoming hyper-sensitive to the perpetrator	1.01* (1.01–1.03)

Note. Step 1 and Step 2 are identical since no variables of block 2 were entered the model. For income, the OR is the change of the odds of returning to the abusive partner in relation to the reference category of average income or above. For identification with the aggressor, OR is the change of the odds for each unit change of the variable.

As shown in Table 3, income, PTSD symptoms and becoming hyper-sensitive to the perpetrator were the only variables that significantly explained the likelihood of returning to the abusive partner and therefore were included in the model. Yet, as can be seen, the contribution of PTSD symptoms became non-significant in the final step, so that only two variables—income and becoming hyper-sensitive to the perpetrator—were found to make a significant contribution in explaining the likelihood of returning to the abusive partner in the final model. Findings concerning the effects of income and becoming hyper-sensitive to the perpetrator indicated that the odds of participants with a below-average income reporting that there was some chance of returning to their abusive partner were more than twice as high compared to the odds of participants with an average income or above to do so. Furthermore, an increase in one unit of becoming hyper-sensitive to the perpetrator was found to increase the odds for indicating that there was some chance of returning to their abusive partners in 1.01.

Discussion

Although many IPV survivors try to extricate themselves from abusive relationships, many return to their partners and suffer from an elevated risk of further and even greater violence. Research has identified several factors that might be linked to women’s returning to their partners, but has not assessed the implications of the strong and complex bonds often formed between women and these partners. The current study, which investigated the contribution of IWA in explaining women’s appraisals of the likelihood of returning to their partner, is thus imperative.

Contrary to former studies (e.g., Anderson, 2003; Griffing et al., 2002), the present findings indicated non-significant relations between women’s economic dependence upon the partner or fear of violence escalation, and likelihood of returning to the partner. These results might be rooted in the characteristics of the present sample. Women who participated in this study reported relatively low average frequency of IPV (0.81), and a considerable average time of 33.5 months since separation. Therefore, it might be that women who agreed to take part in this study may have experienced relatively lower levels of threat concerning future violence inflicted by their former partners, and therefore their appraisals regarding the likelihood of returning to them were not associated with this factor. Additionally, it might be that the relatively long period of time elapsed since their separation was linked with lesser economic dependence, and consequently this factor was also not related to women’s appraisals regarding the likelihood of returning to their partners.

The present findings indicated that having children was related to women’s assessments regarding the likelihood they would return to their partners. The proportions of women who had children were significantly lower among women who indicated that there was some chance they would return to their abusive partners than they were among women who indicated that there was no chance they would ever return to their abusive partners. These findings are in line with those of other studies (Gillum, 2008; Kim & Gray, 2008), which have suggested that having children may serve as a catalyst for leaving the abusive partner. At the same time however, they are inconsistent with other findings, which have documented the opposite trend, wherein having children was associated with a lower likelihood of terminating the relationship (Bui, 2003).

The current findings, then, may reflect only one potential effect of being a mother on women’s decisions about whether to reunite with their partners. Specifically, the contradictory effects of motherhood, which have been found to shape women’s decisions to end abusive relationships, may also be applicable in explaining women’s decisions around reuniting with their partners, so that having children could both increase and decrease the chance that women will return to their partners. Nevertheless, at a given time, one trend may overtake the other, depending on the specific circumstances of the woman in question. The current findings could be rooted in such conditions. It might be that the considerable amount of time passed since separation, which characterized women who participated in this study, was related to their increased sense of security. That is, these women were less preoccupied with fears (e.g., worries about their ability to raise their children alone) that might lead to their returning to their partners, and were more inclined to take into account factors that repelled them from their partners, such as the potential adverse effects of IPV on their children. Therefore, the proportions of women who had children among those who indicated that there was some chance they would return to their abusive partners were lower. Given that this study did not incorporate the time passed since separation in the analyses, and given its cross-sectional design, the offered explanation is speculative only. Future longitudinal studies are needed to explore how mothers’ considerations regarding the renewal of abusive relationships change over time.

A significant relationship was found between childhood abuse history and likelihood of returning to the partner. These findings are in line with a previous study indicating that women who had been sexually abused in childhood were more likely to report having previously returned to their abusive relationship (Griffing et al., 2005). Yet, the results of the present study, which explored different types of abuse, suggest that not only sexual abuse during childhood, but also childhood physical and emotional abuse, may be linked with difficulties in staying away from an abusive partner.

The association between a history of childhood abuse and the likelihood of returning to the abusive partner may reflect the impacts of childhood maltreatment on survivors’ relational schemas. According to Young and Gerson (1991), childhood abuse, which often takes place in early relationships with a caregiver, may lead to the development of relational schemas according to which one should bear emotional or physical pain as a way of maintaining the relationship with an attachment figure. These schemas could, in turn, shape the quality of relationships during adulthood, and may hamper childhood abuse survivors’ capacity to permanently escape from abusive attachment figures.

Post-traumatic stress disorder symptomatology post-IPV also had a significant relationship with likelihood of returning to the partner: Participants who indicated that there was some chance they would return to their abusive partners had higher levels of PTSD symptoms than did participants who indicated that there was no chance they would ever return to their abusive partners. These trends are not consistent with a previous study that found a relationship between elevated PTSD and women’s greater confidence in their ability to leave an abusive partner (Lerner & Kennedy, 2000), but are consistent with a study that found PTSD symptoms to increase women’s risk of being unable to stay away from abusive partners (Abdulmohsen Alhalal et al., 2012). The current findings may reflect the negative implications of PTSD symptoms on women’s coping during the post-separation period. It might be that the debilitating effect of PTSD symptoms on functioning (Kessler, 2000), as well as its link to loss of individuals’ psychological, social, and economic resources (Hobfoll, 1989; D. M. Johnson et al., 2007), limits women’s capacity to handle the challenges entailed in the post-separation period and thus further the chances of their returning to the abusive relationship.

Nevertheless, the present findings revealed that the effects of all three factors—having children, women’s childhood abuse history and PTSD symptoms in explaining the likelihood of returning to their partner—were non-significant in the final regression model: having children and women’s history of childhood abuse were not included in the model in the first place due to their non-significant effects, and the effect of women’s PTSD symptoms became non-significant in the final step of the regression model after the variable of becoming hyper-sensitive to the perpetrator was included. Hence, it appears that the contributions of having children, childhood abuse history, and PTSD symptoms in explaining women’s likelihood of returning to the partner were not as important as the other factors explored herein.

Such factors included women’s current income as well as their levels of becoming hyper-sensitive to the perpetrator. Consistent with former studies (Griffing et al., 2002; R. Walker et al., 2004), participants’ income was found to be related to the likelihood of returning to the partner: The proportion of having a below-average income was significantly higher among women who indicated that there was some chance they would return to their abusive partner than among women who indicated that there was no chance they would ever return to their abusive partner. Furthermore, participants’ income was one of two factors that were found to make a unique contribution in explaining the likelihood of returning to the partner, so that the odds of participants with a below-average income indicating that there was some chance of returning to the abusive partner were more than twice as high compared to participants with an average income or above.

Women with a low income may experience the termination of the relationship with the abusive partner as more stressful and might go through greater difficulties in the post-separation period (R. Walker et al., 2004). Additionally, this group may suffer from low access to physical and mental health services (R. Walker et al., 2004), which are essential for overcoming the implications of their past abuse and for developing psychological independence after separation. These effects, in turn, may increase the chance of returning to the violent relationship.

The innovativeness of this study lies in its discovery of the link between IWA and women’s perceived likelihood of returning to their abusive partners. The results indicated that two aspects of IWA—adopting the perpetrator’s experience and becoming hyper-sensitive to the perpetrator—were related to a greater likelihood of returning to the abusive partner. Nevertheless, results of the final logistic regression analysis indicated that only becoming hyper-sensitive to the perpetrator had a significant effect in explaining women’s likelihood of returning. Furthermore, the final regression model revealed that this aspect of IWA had significant effects above and beyond income, having children, history of childhood abuse, and PTSD symptoms.

The present findings might reflect the adverse effects of women’s mental fusion with their perpetrators on their ability to extricate themselves from abusive relationships. Terminating such relationships is highly stressful and involves varied external and internal struggles (R. Walker et al., 2004). Nevertheless, as many trauma scholars have argued (Ferenczi, 1932; Herman, 1992; Van der Kolk, 1989), one of the main obstacles that abuse victims face is the difficulty of cutting the strong bonds to their perpetrator and freeing themselves from mental captivity. As such, one might suggest several paths through which becoming hyper-sensitive to the perpetrator could advance women’s likelihood of returning to these relationships. Being highly attuned to the abusive partner’s experience may increase positive emotions, such as care, empathy, and compassion towards the partner and may lead to women’s longing for their former partners. Furthermore, becoming hyper-sensitive to the partner’s feelings and needs may limit women’s abilities to maintain clear boundaries between themselves and their former partners, and to reject their former partners’ attempts to bring them back. In this way, women’s hyper-sensitivity to their partners’ inner experience may be a powerful force that negatively colors the separation, and pushes women back into these relationships.

The findings of the final regression model, which indicated that only income and hyper-sensitivity to the perpetrator significantly explained women’s likelihood of returning to their abusive partners, may suggest that the effects of both variables are particularly strong, to the point that they may overshadow other demographic, external as well as victim-related, factors. For example, it might be that the elevated stress and difficulties during the post-separation period that stem from the women’s low incomes are substantial to such an extent that they eclipse the effects of other factors, such as having children. Additionally, women’s becoming hyper-sensitive to their perpetrators may serve as yet another central element that shapes their decision to reunite with their partners, and the effects of this may overshadow the implications of women’s history of childhood abuse and PTSD symptoms. It is plausible that, while the relational schemas of women with a history of childhood abuse may be linked to their likelihood of returning to their abusive partners, women’s emotional bonds to their former abusive partners have a much stronger effect in pulling them back to the relationship. Furthermore, it might be that, although women’s PTSD symptoms are related to higher chances of returning to their abusive partners, it is their over-attunement to their partners that serves as the mechanism prompting them to return to this relationship. Hyperarousal symptoms as part of PTSD are characterized by heightened sensitivity to potential threats which are related to the traumatic experience (APA, 2013). Thus, women who suffer from PTSD symptoms may exhibit hyper-sensitivity to their abusive partners as a result of their hyperarousal symptoms and as a way to detect threats and limit additional abuse, and this tendency, in turn, might then increase their likelihood of returning to their partners. The present findings, which revealed that the effect of PTSD symptoms became non-significant after becoming hyper-sensitive to the perpetrator was entered into the model, provide some support for this explanation. Nonetheless, future longitudinal studies are needed to explore this possibility and to assess whether the relationship between PTSD symptoms and women’s likelihood of returning to their abusive partners is mediated by IWA.

It is imperative to note that the present findings, which illustrate the key role of IWA in explaining women’s appraisals regarding the likelihood of returning to their abusive partners, in no way suggest that IPV survivors who develop strong bonds with their perpetrators, are responsible for being re-abused. On the contrary: Both theory (Dutton & Painter, 1993; Ferenczi, 1932; 1933; Frankel, 2002; Herman, 1992; Van der Kolk, 1989) and evidence from research among mammals (Cantor & Price, 2007) suggest that identifying with the aggressor comprises an automatic, evolutionary-based reaction that relies upon biological processes. This reaction enables the organism to survive attacks, by connecting with, appeasing, and becoming attuned to the perpetrator’s wants and needs (Frankel, 2002; Lahav et al., 2019b). Nevertheless, similar to other responses that are essential for surviving traumatic experiences, IWA has substantial drawbacks as it may impede survivors’ capacity to leave their abusive partners for good and, in this way, perpetuate the abuse.

The current investigation should be considered in light of its limitations. First, the cross-sectional design precludes any conclusions regarding the directionality of associations as well as causal relations between the study variables. Second, using a composite measure of IPV does not allow for detecting the role of IPV type, which has been found to be related to women’s response to abuse (Koepsell et al., 2006), in the relationship between IWA and women’s returning to their partners. Third, this study relied on convenience sampling and was conducted online among Israeli women, most of them highly educated and employed, and having left their abusive partners for an average time of 33.5 months—a relatively long time, which may have enabled them to process their experience of the abusive relationship. Participants’ ethnic backgrounds were not assessed, and all women who participated in the study were in heterosexual relationships. Studies have indicated that there are some aspects of IPV that are unique to the lesbian, gay, bisexual, and transgender experiences (Ard & Makadon, 2011), and that the process of leaving the abusive partner, as well as that of returning to the abusive relationship, may be different for women from different racial and ethnic groups (Lacey et al., 2011; Torres, 1991). Furthermore, with no knowledge that a pandemic was on its way, of course, the data collection of this study took place during the first wave of the COVID-19 pandemic in Israel. Studies have indicated the pandemic to be particularly debilitating for trauma survivors (Hamam et al., 2020; Siegel & Lahav, 2021) and to have negative effects on individuals’ economic, occupational, and functional domains (e.g., Bareket-Bojmel et al., 2020). Altogether, these aspects limit the study’s generalizability and indicate the need for further research among clinical samples of IPV survivors from diverse cultural and social backgrounds, simultaneously taking into account the potential effects of the pandemic. Fourth, although this study included several demographic and external factors, such as whether women have children, it did not take into account variables concerning abusive men’s behaviors, which affect women’s ability to end violent relationships. Fifth, this study utilized self-report data which may be subject to response biases. Data regarding participants’ PTSD diagnoses were lacking, and PTSD symptoms were assessed based on participants’ reports only. Additionally, women’s appraisals regarding the likelihood of returning to their partners could be highly subjective and may have been influenced by the stage of exit or healing from the relationship. Future studies should therefore include additional methods of data collection, such as clinical interviews, in order to overcome this limitation. Finally, this research did not incorporate data concerning potential mechanisms underlying the relationship between IWA and women’s perceived likelihood of returning to their abusive partners, such as difficulties in maintaining healthy boundaries in their relationships. Future longitudinal studies should employ clinical interviews and measure potential mechanisms in order to uncover the processes that are at the basis of the relationships which were found in the present study.

Nevertheless, this study suggests that identifying with the aggressor might pose a substantial hurdle for abused women, weakening their ability to stay away from abusive relationships. Therefore, appropriately applied, the findings of this study suggest that Intervention and Policy Clinicians working with IPV victims should thoroughly assess not only external factors, such as women’s fear of an escalation of violence or economic struggles, but also their unique way of relating to their partners, manifested in IWA, in order to identify, together with the victims, the most effective course of action to help them cope with the post-separation period. The current findings also suggest that survivors of IPV might benefit from clinical interventions that would help them reprocess the complex relational dynamic of their violent intimate relationships and gradually disconnect from their partners. Although this process may be challenging, particularly as IWA may be biologically-based, IPV survivors might be able to re-connect with their own emotions, feeling, and needs; view the violence they endured from their own point of view; and stay away from their abusive relationships.

Author Biographies

Yael Lahav, PhD, is a senior lecturer in the Department of Occupational Therapy, The Stanley Steyer School of Health Professions, Sackler Faculty of Medicine, Tel Aviv University. Her research focuses on the unique associations between the emotional, interpersonal, somatic, physiological, and functional facets of psychological trauma, as well as on the interpersonal processes involved in the victim-perpetrator dynamics.

ORCID iD

Yael Lahav https://orcid.org/0000-0003-1242-9042