When the domestic violence awareness movement began in the 1970s, its goal was ending physical violence against women. Though some studies report women are as or more likely than men to engage in physical assaults against their partner (Straus 1990; Straus et al. 1996), it became apparent that women suffer greater injuries from their partner’s use of violence (Gelles 1997; Straus 1993). Not only are women more likely to report injuries as the result of abuse, they are also more likely to suffer emotional and/or psychological effects (Campbell 1998; Dobash and Dobash 2001; Schwartz 1987).

Women often report their partner is controlling and emotionally abusive (Bancroft 2002; Johnson 1995; Stark 2007). In some cases, abusive men only engage in controlling behaviors without resorting to physical abuse. Researchers then began to examine how abusive behavior occurs in relationships by looking beyond incidents of physical violence (Dobash and Dobash 2004; Dobash et al. 1992; Johnson 1995). Even with increased attention on abusers’ controlling behaviors in relationships, much remains unknown about how men can continue to control their partners, especially after the relationship ends.

Abuse may not end when relationships end. When a woman begins the process of separation, the dynamics of the relationship change. Since separation is a challenge to an abuser’s control, he may increase his manipulating and intimidating behaviors to get back the control he is losing over his partner’s behavior (Bancroft 2002; Campbell et al. 2007). The abuser may also no longer have direct access to his former intimate partner. However, in relationships where the abusive partner has children with his victim, the children can serve as tools for the abuser to continue his abusive behavior. Tangential spouse abuse is the term for when an abuser uses the children as tools to control, intimidate or manipulate the mother (Bancroft and Silverman 2002; Stark 2007).The abusive partner can also involve the woman’s family and friends as another way to control, intimidate or manipulate her. Using children, family and friends are indirect ways for an abuser to continue controlling the woman upon separation.

Research on controlling behaviors in relationships prior to separation may shed light on how men change the mechanism they use to control their partner during separation. Prior research has shown that abusive men are likely to display controlling behaviors, especially involving children, during child custody disputes (Bancroft and Silverman 2002; DeKeseredy and Schwartz 2009; Field 1998). The current study examines whether prior controlling and physically abusive behavior influences the male’s propensity to indirectly control the woman through children, family and friends post separation. The sample consists of women in the New York City family court system. This paper concludes with policy recommendations for identifying abusive men who are no longer physically abusive but who still pose a danger to their partner and children.

Controlling Behaviors in Relationships Prior to Separation

Abusive behaviors can take many forms, such as physical violence, sexual violence, psychological/emotional abuse and/or controlling behaviors. While there is conceptual overlap between psychological, emotional, and verbal abuse with controlling behaviors, the terms are often used interchangeably (Follingstad 2007). Psychological abuse includes constant criticism, verbal abuse and/or isolation (O’Leary 1999). Controlling behavior, on the other hand, has a narrower focus by shifting attention to what the abuser does not let the woman do as opposed to what he actually does against her (Stark 2007). Controlling behaviors can include financial control, intimidation, isolating the woman from a support network, or regulation of a woman’s behavior to stereotypical gender roles (Stark 2007). The abuser may not allow the woman to work even though the family needs an additional income, he may require she keep a log of her purchases or dinner must be on the table as soon as he gets home from work. These behaviors are often difficult to detect if one does not consider the ongoing and cumulative impact of the behaviors (Dobash and Dobash 2004; Stark 2007).

Many intimate partner violence victims claim the controlling behaviors are worse than the physical violence (Bancroft 2002; Stark 2007). The abuser’s controlling actions, which may not seem oppressive as individual acts, accumulate over time and entrap the victim in the relationship. Men, even though the number is smaller than females, may experience control in abusive relationships (Simonelli and Ingram 1998). However, this entrapment and systematic use of violence and control is often considered unique to female victims of intimate partner violence because it is based on women’s dependence (e.g. economic) on their partner (Johnson 1995; Stark 2010).

As the domestic violence movement raised awareness about physical violence in intimate relationships, it is possible abusive men had to devise new tactics that would allow the behavior to go undetected. It is often difficult to arrest an abuser for controlling behavior that is not physically abusive because laws focus on physical assaults (Buzawa and Buzawa 2003). Men are also able to use physical characteristics to assert power and control. For instance, it is the male’s physical size and strength that makes men’s violence against women, when weapons are not involved, more severe and threatening to the women’s safety than women’s violence against men (Felson 1996).

Yet, a male’s entitlement and control may begin to change as the woman asserts her autonomy by beginning the process of separation. When a woman makes the decision to end an intimate relationship, she may begin to resist her partner’s control (Field 1998; Johnson and Ferraro 2000). The process of separation becomes complicated as one considers the unique barriers many women face, such as employment or having children in common with their abuser. Prior research has focused on how physical abuse and sexual violence may continue once the relationship ends as opposed to how the dynamics of the relationship change upon separation (DeKeseredy et al. 2006; Fleury et al. 2000). Instead, the abusive partner may retain control over the woman during separation by indirectly engaging in abusive behaviors. One important way for an abuser to indirectly assert control is through the children.

Separation and the Abuser’s Use of Children to Control Former Partners

When the couple separates, the abuser will need to devise new tactics that allow him to monitor and control his former partner. Bancroft and Silverman (2002) argue that the children might become “weapons after separation” (p. 75). When former partners have children together, the parents may need to discuss times for visitation or decisions regarding the children. The parents must still retain some type of relationship, though it may no longer be romantic or intimate. The abusive partner can use this new form of the relationship to contact his partner directly and indirectly by interacting with the children. The children, and matters that involve them, become the tools by which he continues to control his former partner.

A mother will often put her children’s needs before her own (Beeble et al. 2007; Ferraro 2006). Bancroft and Silverman (2002) identify a number of ways in which an abuser can control the woman through the children. As revenge for ending the relationship, the father may threaten to take or hurt the children during visitation. He may allege his former partner was abusive to the children, when she was not, in order to remove the children from her care. The abuser can also manipulate her subtly, by exposing the children to things she does not find appropriate. All these behaviors demonstrate to his former partner that he retains control over her through the children.

Abusive men may also use children to pressure partners into getting back together (Bancroft and Silverman 2002). As Bancroft and Silverman (2002) note, the abusive partner will usually tell the children their parents separated because of their mother, not because of his abusive behavior. Since he tells the children the separation is their mother’s fault, the children then pressure their mother to take him back. He can also have the children report to him on their mother’s behavior as a way to keep surveillance over the woman’s life (Beeble et al. 2007). These behaviors are indirect ways by which the abuser manipulates and monitors the mother even though they are no longer together.

When reconciliation no longer seems like an option to the abuser, child custody disputes and visitation become ways to continue abusing the woman (Jaffe et al. 2008; Sauders 1994; Shepard 1992). Abusive men are frequently at an advantage in custody disputes since they are the ones who retained control of the money or they did not let their partner work (Bancroft and Silverman 2002). The abuser can afford legal counsel while the woman may not be able to. The abuser is also not the one who has been subjected to years of physical abuse and control. The woman’s parenting skills may suffer as a result of the control and abusive behavior she has experienced (Jaffe et al. 2008). He may request sole custody because of the mother’s supposed parenting deficits that are a result of his behavior (Dalton 1999). In addition, the abusive partner can make intimidating statements to the mother during visitation exchanges. He can also threaten the mother indirectly by saying he will take the children during a court ordered visitation. Therefore, visitation and child custody disputes provide a multitude of ways for the abusive partner to control, intimidate and manipulate the mother.

Tolman (1989) was the first to report that over a third of battered women in English and American studies claim their partner threatened to hurt the children or report the mother as abusive. Practitioners, based on their experience, have described how abusive men use the children as tools during separation (Bancroft 2002; Bancroft and Silverman 2002; Stark 2007). Yet, there is little empirical research on how the father actually uses the children as tools to manipulate or control the mother, especially during the process of separation (Beeble et al. 2007).

Use of Family and Friends Post-Separation

Over the course of the relationship, the abuser typically presents a different image of himself to outsiders than he does to his family (Bancroft 2002). The man often plays the role of loving husband and father around others. When the couple separates, it will be unlikely outsiders will believe the woman’s allegations since they are only exposed to the man who cares for his family. However, abuse may not be disclosed until the couple separates because of the stigma associated with telling others one is a victim of intimate partner violence (Dalton 1999). The abuser is able to use the outsider’s unfamiliarity with the abuse to slowly turn them against the woman (Bancroft 2002). Therefore, family and friends may not believe the woman because her version of the relationship does not coincide with the version they are exposed to.

A common theme identified by the women Bancroft (2002) has worked with is that abusers are often able to have the women’s family and friends believe them. The abuser is able to convince outsiders the problems in the relationship stem from the woman’s behavior, not his. He uses the societal pressures like “Children should grow up in an intact home” and “You made your bed, now go lie in it” to further his manipulation of family and friends. There is a discrepancy between the abuser’s actual behavior in the home and his outside persona, which allows him to coax others into pressuring the woman to call off the separation.

Current Study

The current study has three general goals: (1) to determine if physical violence and controlling behavior is less likely to be reported during the follow-up interview (2) to discover if men who were previously physically abusive switch to controlling behaviors and (3) to examine whether prior controlling and physically abusive behavior affect the likelihood that an abusive partner will continue indirectly abusing the woman after the couple separates. An understanding of how the abuser employs his control tactics during separation and divorce are vital to making appropriate custody recommendations (Bancroft 1998). Recommendations for court personal who may encounter men indirectly engaging in abuse will be given.

Aim of Study

The focus on separation in abusive relationships has been on ending physical violence in the relationship. Though earlier studies have shown an increased risk for violence and femicide during separation (Fleury et al. 2000; O’Sullivan 2002; Stark 2007), the abuser may also continue to engage in abusive behavior indirectly. Especially in a sample involved with the courts, it is reasonable to assume that abusive men will need to be more covert about their continued abuse of their partner (Adams 1989). Over the course of the follow-up period, an abuser may decrease his physical abuse while still controlling the woman.

Hypothesis 1

There will be fewer incidents of reported physical abuse during the follow-up interview compared to incidents of control.

Based on their experience with abusive men and their families, Bancroft and Silverman (2002) note that the more controlling abusers are of their partners, the more likely they are to engage in tangential spouse abuse. Furthermore, tangential spouse abuse may be an extension of controlling behaviors that occurred while the couple was together. Instead of manipulating the woman, the abuser may have to be indirect and manipulate the woman through the children.

The abuser may also use the woman’s family and friends to control her and monitor her behavior. An abusive man may contact the woman’s family or friends in order to pressure her into reconciliation. This tactic may be more successful than tangential spouse abuse because family and friends are often unaware of the abuser’s behavior while children may have witnessed it firsthand. The woman’s decision not to expose the abuse until after the couple separates can also be used by the ex-partner to manipulate family and friends (Bancroft 2002; Dalton 1999). Family and friends also become pawns by which the abuser continues to engage in his controlling behavior.

Hypothesis 2

Prior physical violence will be associated with later tangential spouse abuse and contacting the woman’s family and friends.

Hypothesis 3

Prior controlling behavior will be associated with later tangential spouse abuse and contacting the woman’s family and friends.

Data and Method

Sample

Data used for this study was collected between 2002 and 2005 to examine whether prior abuse influenced the court’s visitation decision (O’Sullivan et al. 2009). The sample was drawn from all female victims of intimate partner violence who obtained an order of protection in a New York City family court between 2002 and 2005. In order to be included in the original study, participants had to have at least one child with the male she was seeking an order of protection against. The father also had to be filing for visitation or actively visiting the child(ren). O’Sullivan et al. (2009) were interested in the court’s visitation decision. Therefore, the participants were recruited based on their ex-partner’s visitation arrangement. The original sample was separated into supervised visitation and unsupervised/family supervised/supervised transfer visitation.

The supervised visitation sample was recruited from the Bronx, Brooklyn and Queen’s Safe Horizon visitation centers. Some supervised visitation participants were recruited from the Brooklyn Society for the Prevention of Cruelty to Children that offered supervised visitation. Unfortunately, this program lost funding during the course of the study. The majority of respondents recruited for the supervised visitation sample were from the Queens visitation center. The Brooklyn Family Court Visitation Center referred a number of participants though it closed for 2 months during the course of the study. Only a few participants came from the Bronx Family Court Supervised Visitation program because it lost funding.

The unsupervised or partially supervised visitation sample was recruited from Safe Horizon’s family court program that provides assistance and a waiting room for women filing for an order of protection. Women were asked to participate while waiting in the Reception Center or were referred by a case manager, attorney or receptionist. If the visitation petition was dismissed in the unsupervised or partially supervised sample, the respondent was not included in the study. Therefore, the sample for this study is limited to those women in which the father had some form of visitation.

The majority of the study’s sample had unsupervised or partially supervised visitation and was therefore recruited from the Safe Horizon Reception Centers in family court. The woman was either approached or contacted a toll-free number after seeing a flyer at the Reception Center. The baseline interview was conducted at the family courthouse or visitation center from 2002 to 2004. A total of 242 participants from the five New York City counties completed the baseline interview. The interview was available in Spanish and English. The instrument included questions on the occurrence of intimate partner violence, the children’s exposure to the violence, children’s behavioral issues and court orders.

Attrition in research on intimate partner violence is common (Bennett and Williams 2001). In order to the increase the number of respondents at time two, a stipend of $20 was given for the follow-up phone interview. The researchers also obtained an alternate phone number to contact the woman at and gave the woman a toll-free number to call if her contact information changed or she wanted to schedule an interview.

During follow-up, the interviewer first called the number given by the woman during the first interview. If an individual other than the woman answered the telephone, a specific protocol was established in order to ensure the woman’s safety (see Appendix A for specific protocol). If the woman could not be reached at the given phone number, the interviewer called her alternate contact. If the woman could not be reached at the alternate contact, she was sent a letter with the toll-free number. If the letter was returned, it was sent to the alternate contact’s address.

Follow-up interviews were conducted from 2003 to 2005 over the phone, except in a few instances where the woman asked for an in-person interview. The follow-up interview occurred between 2 1/2 and 18 months after the first interview, with the average interview occurring 6 months after the baseline interview. The follow up portion of the study had 168 women, a retention rate of 69 %.

Measures

Dependent Variables

The dependent variables in this study capture how a woman can be indirectly abused during the process of separation. Indirect abuse was operationalized using variables that represent (1) tangential spouse abuse and (2) contacting the woman’s family and friends without permission. Tangential spouse abuse is conceptually different from other controlling behaviors because it occurs after separation when the abuser’s access to the woman is restricted (Stark 2007). The abuser also involves other people when engaging in indirect abuse (Bancroft 2002), further highlighting the uniqueness of the dependent variables. Much like sexual abuse is disaggregated from other forms of physical abuse (DeKeseredy et al. 2006), tangential spouse abuse should be considered a distinct type of emotional abuse abusers can engage in.

Told Lies to the Child

Participants who completed the follow-up interview were asked if their ex-partner told the children bad things about them, whether the statements were true or false. This variable captures the way an abuser can manipulate both the mother and children during separation. It also represents an attack on the mother’s parenting because he is trying to manipulate the children’s feelings and/or attitudes toward her. This can create tension between the mother and children and hinder their relationship. It is a dichotomous variable with a score of one indicating the children’s father told lies to the children about their mother during the follow-up period (Yes = 26.6 % vs. No = 73.4 %).

Kept the Children Longer

This variable captures the unilateral control a father can maintain over visitation and does not represent a negotiated agreement between equal partners. Visitation becomes a way for the abuser to maintain control over the woman (Field 1998). By identifying factors associated with visitation arrangement, child custody evaluators can identify cases in which the violence may escalate (e.g. the Melanie Edwards case). This is also a dichotomous variable during the follow-up portion of the study in which a score of one indicates the children were kept longer during visitation (Yes = 25.8 % vs. No = 74.2 %).

Contacted Family of Friends

An abuser can also contact family and friends during separation in order to manipulate the woman. Though this behavior does not explicitly involve the children, it is another avenue for the abusive partner to indirectly control and monitor the woman’s behavior during separation. The woman was asked during the follow-up interview if her ex-partner contacted her family or friends that she did not want him to contact. A score of one indicates the man contacted her family or friends without the woman’s permission (Yes = 23.4 % vs. No = 76.6 %).

Independent Variables

Physical Abuse

Physical abuse variables assessed during the first interview were originally classified as (1) minor to moderately severe and (2) severe by O’Sullivan et al. (2006). Given that physical violence often escalates over time, it is likely that women who experienced severe physical abuse would also have experienced lower levels of violence over the course of the relationship. The physical abuse variables from the baseline interview were combined into an additive index, without making a distinction between minor and severe violence. The index ranges from 0 to 10 (Median = 6.00 M = 5.69, SD = 2.48) and is negatively skewed. The majority of respondents reported experiencing multiple instances of physical abuse over the course of their relationship as seen in Table 1.

Table 1 Controlling behavior and physical abuse scales for respondents who completed follow-up interview
Full size table

Controlling Behaviors

Controlling behaviors were also measured during both the baseline and follow-up interview. Variables from the baseline interview were combined into an additive index based on the scale used by O’Sullivan et al. (2006). Each of the controlling behaviors was coded as 0 or 1, with a score of one indicating the children’s father engaged in the behavior at some point during the relationship. The resultant index ranges from 0 to 11 (Median = 6.00 M = 6.25, SD = 2.29), with higher scores indicating the woman experienced more controlling behavior at time one. Again, most respondents reported their partner engaged in multiple types of controlling behaviors over the course of the relationship.

Control Variables

Level of education and employment status for both the abuser and victim have been found to have an association with intimate partner violence (Della-Giustina 2010; Maxwell et al. 2001; Roberts and Roberts 2005). For instance, a man who is not working may continue with his abusive behaviors because he does not have a job to lose if he is arrested. It has also been argued that as women gain status in society by obtaining higher education and employment, they are more likely to be a victims of intimate partner violence or intimate partner violence homicide (Della-Giustina 2010; Stark 2007). Separate variables that measure the employment status of the couple were included in the analysis, with both the male and female not working serving as the reference category.

Though previous research has found an increased rate of intimate partner violence among African Americans and ethnic minorities compared to Whites (Caetano et al. 2000; Hampton et al. 2003; Sokoloff 2008), an earlier analysis revealed the respondent’s or her ex-partner’s race did not impact the occurrence of indirect abuse (not shown here; results available from author upon request). Therefore, race control variables were not included in the final analysis.

The demographic characteristics of the women and their ex-partners can be found in Table 2. Women typically had higher levels of education than their male partners (51.8 % of females had some college education or more compared to 30.5 % of males). Women (M = 31.62, SD = 7.55) were often younger than their ex-partners (M = 33.89, SD = 7.66). On average, women had two children (M = 1.99, SD = 1.27).

Table 2 Demographic characteristics of respondents who completed follow-up interview
Full size table

Data Analysis

Physical abuse and controlling behaviors were examined at time one and compared to indirect abuse at time two. Three separate stepwise logistic regressions were used to estimate the odds ratios associated with the two types of tangential spouse abuse and contacting family or friends based on previous control and physical abuse. The first step contains the physical abuse index in order to assess its’ unique effect on tangential spouse abuse and contacting family and friends at time two. The second step includes controlling behaviors to assess their effect beyond physically abusive behaviors. The final step includes employment status of the couple and level of education for both the respondent and her ex-partner. The sample was limited to respondents who completed both the baseline and follow-up interviews. Listwise deletion was used for cases missing data on variables included in analysis.

Results

Table 1 contains the percent of respondents who reported during the baseline and follow-up interviews that their partner engaged in physical abuse and/or controlling behavior. During the follow-up interview, few respondents reported their partner engaged in any form of physical abuse. The most common form of physical violence reported during the follow-up interview was grabbing, with only 3.7 % of the sample saying they had been grabbed. On the other hand, a larger portion of the sample reported that their partner was still engaging in controlling behavior. For instance, 45.7 % of the sample said they were still blamed for their partner’s problems. Over a fifth of the sample reported in the follow-up interview that their partner threatened to take the children. These findings support the first hypothesis that while physical violence decreased during the separation, the men were still engaging in controlling behaviors.

Table 3 contains the median, mean and standard deviation for the physical abuse and controlling behaviors scales for women who responded affirmatively to each of the dependent variables. During the baseline interview, the majority of these women said they experienced multiple types of physical abuse. The average respondent who experienced each of the dependent variables reported during the baseline interview that her partner engaged in more than half of the physically abusive behaviors. Of the women who reported their partner contacted their family or friends or told lies to their children during the follow-up period, almost 80 % experienced five or more kinds of physical violence prior to the baseline interview.

Table 3 Prior physical abuse and control for women who responded yes on dependent variables
Full size table

Moreover, the women who experienced the dependent variables also reported during the first interview that their partner engaged in a number of controlling behaviors (Median = 7.00). Women who reported during the follow-up interview that their ex-partner told lies to the children or contacted family and friends, said their partner engaged in at least one type of controlling behavior before they took part in the study. Over 75 % of the respondents who responded affirmatively to each of the dependent variables experienced three or more types of controlling behavior during the course of their relationship. The findings in Table 3 highlight that women typically experience multiple victimizations in an abusive relationship.

Logistic regressions were performed for each dependent variable. The first logistic regression (Table 4) analyzes the effects of the physical abuse index, controlling behavior index and control variables on telling lies to the children over the follow-up period. While the indexes are not predictors of lying to the children over the follow-up period, men with higher levels of education (Exp (B) = 1.48, Wald = 3.98, p < 0.05) were more likely to tell the children lies about the mother over the follow-up period. However, as the mother’s level of education increased, the abuser was less likely to tell the children lies (Exp (B) = 0.71, Wald = 3.00, p < 0.10).

Table 4 Logistic regression for telling lies to the children over follow-up period (N = 128)
Full size table

Table 5 presents the results from the logistic regression that examines the effect of the independent and control variables on keeping the children longer during the follow-up period. Consistent with the findings for the other dependent variable that represents tangential spouse abuse, neither of the indexes were predictors of keeping the children longer during the follow-up period. However, as the man’s level of education increased, he was less likely to keep the children longer during a visit (Exp (B) = 0.57, Wald = 5.36, p < 0.05). In contrast, as the woman’s level of education increased, the abuser was more likely to keep the children longer (Exp (B) = 1.54, Wald = 2.93, p < 0.10).

Table 5 Logistic regression for keeping the children longer over follow-up period (N = 96)
Full size table

The results from the logistic regression for contacting the woman’s family and friends can be found in Table 6. Respondents who reported physical abuse during the baseline interview were more likely to have their ex-partner contact their family or friends over the follow-up period (Exp (B) = 1.21, Wald = 3.10, p < 0.10). As a man’s level of education increased, he was more likely to contact the woman’s family and friends over the follow-up period (Exp (B) = 1.54, Wald = 3.78, p < 0.05). However, when the man is the only one working, he is less likely to contact the woman’s family and friends (Exp (B) = .20, Wald = 3.91, p < 0.05).

Table 6 Logistic regression for contacting the woman’s family or friends over the follow up (N = 125)
Full size table

Discussion

Findings presented here depart from earlier work on the increased risk of physical violence during the process of separation. Of particular interest, is the large reduction in reported physical abuse during the follow-up period without the corresponding decline in controlling behaviors. The reduction in physical violence over the follow-up period stands in contrast to the body of literature that finds physical violence increases during the process of separation (DeKeseredy et al. 2006; Fleury et al. 2000; O’Sullivan 2002; Stark 2007). It may be that physically abusive men switch to covert tactics, such as threats or telling lies, if the courts become involved.

The effect of education on keeping the children longer and telling lies about the mother may be one example of how an abuser uses covert tactics. Keeping the children longer is an overt action that would attract attention during child custody disputes. More educated men typically have better jobs and would not want to do something that would jeopardize their job and/or their status. On the other hand, telling lies about the mother is a covert behavior. An abuser with higher levels of education may be able to manipulate his way around allegations that he told lies to the children if it comes to the attention of outsiders, especially if they consider him charming or do not believe the woman’s allegations. However, abusers are more likely to keep the children as the women’s education increases. Future research can identify if educational levels and employment status are more appropriate indicators of later indirect abuse during the process of separation than prior abuse.

Furthermore, it may be difficult for the man to engage in physical violence during separation since he may only see his partner during child visitation. The new arrangement, especially if visitation is supervised, may not present the abuser the opportunity to force sex or use a knife or gun. Separation may be considered a success for this sample because it reduced the occurrence of physical violence. However, one must also consider the covert controlling behaviors the abuser continues to engage in.

The results from this study also show that women who experienced indirect control over the follow-up period had partners who engaged in multiple types of physical violence and controlling behaviors. Women in this sample did not experience one type of abuse in isolation. It is important for future research to keep in mind that women often experience ongoing abuses that accumulates over time instead of focusing on isolated incidents (Dobash and Dobash 2004; Stark 2007).

Nevertheless, variables used to measure controlling behaviors do not capture the wide range of behaviors abusive men can engage in. The measures used in this analysis are imperfect indicators of controlling behavior and were originally used by O’Sullivan et al. (2006) to assess psychological abuse. Though the terms psychological abuse and control are often used interchangeably (Follingstad 2007), it is important for future studies to distinguish between the various types of controlling behaviors. Future studies should also include more variables that explicitly measure infringements on the woman’s autonomy. Researchers should also allow the respondent to provide examples of controlling behavior she has experienced in order to capture the dynamics of each relationship.

Findings that show limited effects of previous abuse on tangential spouse abuse are also important. The dependent variables representing tangential spouse abuse are also weak indicators and only captured two behaviors that represent this dynamic. Tangential spouse abuse encompasses a wide range of behaviors and is not limited to telling lies to the children or keeping the children longer. Future studies should look at multiple indicators of tangential spouse abuse to capture the various dimensions.

Furthermore, little empirical research has been conducted on how abusive men use children to control and manipulate their mother upon separation (Beeble et al. 2007). Bancroft and Silverman (2002) theorized the ideas underlying tangential spouse abuse based on their experience with women in a domestic violence shelter. However, some scholars have argued that women in domestic violence shelters are not representative of women involved in child custody disputes (Dutton 2005; Gould et al. 2007). The current study utilized a different sample but one in which researchers have discussed the likelihood of tangential spouse abuse occurring (Bancroft and Silverman 2002; Sauders 1994; Shepard 1992). The finding that prior physical abuse and controlling behavior did not predict later tangential spouse abuse in this sample is not consistent with the idea that abusive men may use children as tools to control the mother.

However, limitations associated with small sample size and limited measures of tangential behaviors could have also produced the non-significant findings in this study. Future studies should be designed that specifically analyze the occurrence of tangential spouse abuse among multiple sample types. Although O’Sullivan et al. (2006, 2009) notes that the sample used for this analysis may not be representative; there is no reason to assume it is extremely biased. Though the researchers had various incentives and methods to limit attrition, future studies should focus on having a larger follow-up sample size. Studies should be replicated in other cities with larger samples and updated measures in order to improve generalizibility. Nonetheless, men may continue to control their ex-partners through other relationships the women have.

Continued control of the mother can be found in the data showing the abuser contacted the woman’s family or friends without her permission over the follow-up period. This finding supports the idea that an abusive man will try to retain control over the woman; however, he may do it in a way that does not involve the children. Furthermore, the physical abuse index was a predictor of contacting the woman’s family or friends but the prior controlling behaviors index was not. This finding may be because men who engaged in physical violence did so to reinforce their controlling behaviors, increasing the lethality associated with physical abuse. Future studies should tease out the possible interaction effect of physical abuse and controlling behaviors on later indirect abuse.

Nevertheless in light of the findings of continued control over the follow-up period, improvements must be made to child custody evaluations. Currently, the focus of many evaluations is on risk factors and physical violence, which include previously using a knife/gun, threatening to kill or to commit suicide (Campbell et al. 2003). The problems of using the Conflict Tactics Scale, which was used to assess physical abuse in this study, have been documented (Dobash and Dobash 2004; Dobash et al. 1992). As indicated by the lack of respondents who reported their partner used a knife or gun during the follow-up period, the Conflict Tactics Scale may be an inappropriate measurement tool in this setting.

Assessments that identify abusive men should also include items to assess tangential spouse abuse. Women’s experiences should be incorporated in these measurement tools. Other assessment tools, such as the Women Experience with Battering Scale, can be used to identify victims who experience controlling behaviors as opposed to physical abuse (Stark 2007). Caseworkers, researchers and intimate partner violence victims need to work together to create tools that identify the types of abusive behaviors women find most damaging.

As part of the reform of child custody evaluations, allegations of intimate partner violence in child custody cases need to be taken seriously (Bancroft and Silverman 2002). One of the reasons the woman may have not have disclosed the abuse before separating from her abuser is because she was ashamed or afraid (Dalton 1999). Therefore, allegations may not arise until after the couple separates as the woman challenges her partner’s control. When these allegations come to the attention of evaluators, certain steps should be taken to ensure the safety of both the woman and her children. One way to improve safety is to use supervised visitations.

Even though the abusive partner can still engage in manipulating behaviors in supervised visitations, the risks of manipulation are greater as the structure with supervision is decreased (Bancroft and Silverman 2002). The abuser will face more obstacles should he try to take the children during a supervised visitation as opposed to an unsupervised visitation. There will also be no reason for the children’s father to contact the woman’s family or friends when he has supervised visitation since the location of visitation is already arranged. If the abuser’s behavior is being monitored, there should be less of a chance he will be able to indirectly control the woman during separation.

Though this study reports modest statistical results, these findings highlight the need to continue research on this topic. Many researchers have theorized the dynamics of what happens during separation yet the empirical findings to support these conclusions are absent. Existing evidence suggests men are using children, which reinforces the need for more research on this topic (Bancroft and Silverman 2002; Beeble et al. 2007). Future studies should increase the sample size, as the final sample in this study was much smaller than the total number of women who completed the first interview.

The results from this study represent an interesting departure on how men can continue to indirectly abuse their partner upon separation. Though the occurrence of physical abuse decreased from the baseline interview to follow up, many women still reported their partner engaged in controlling behavior over the follow-up period. Furthermore, women who reported their ex-partners contacted family or friends during the follow-up period reported multiple victimizations over the course of the relationship. Future research should be directed at analyzing how abusive men continue to manipulate and control their partner throughout the process of separation. We must begin to unravel the different types of control and physical violence that occur during the process of separation.