Most states have statutes that allow or mandate family court officials to incorporate the existence of intimate partner abuse (IPA) in child custody determinations (Jaffe et al. 2003). This is not surprising, given that IPA is a pervasive problem with serious physical and mental health consequences for women and their children (Evans et al. 2008; Golding 1999; Wolfe et al. 2003). Previous national surveys estimated that 1.5 to over 3 million women are victims of physical or sexual violence at the hands of their partners or ex-partners each year (Tjaden and Thoennes 2000). According to the most recent National Intimate Partner and Sexual Violence Survey (NISVS) more than 42 million women in the United States have been physically assaulted, sexually assaulted, and/or stalked by an intimate partner in their lifetime, with a 12-month prevalence rate of nearly seven million women (Black et al. 2011). Thirty to sixty percent of children with mothers who are abused are also likely to be abused by their fathers, and an estimated three to ten million children witness parental abuse (Edleson 1999a, b). According to the 2009 U.S. Bureau of Justice Statistics report, women were more likely than men to be victims of both non-fatal and fatal intimate partner violence, and to be stalked by intimate partners (Catalano et al. 2009). Therefore, throughout this paper the terms “he/father” refer to the abuser and “she/mother” refer to the victim/survivor. In addition, IPA is defined as the following:

[IPA is] a pattern of coercive control that may be primarily made up of psychological abuse, sexual coercion, or economic abuse, that is punctuated by one or more acts of frightening physical violence, credible threat of physical harm, or sexual assault (Bancroft 2010).

Along with the damaging effects of physical abuse on women’s health and well-being, emotional abuse appears to be even more deleterious, and also tends to occur more frequently than physical or sexual violence (Panchanadeswaran et al. 2010). Emotional abuse includes harassment, controlling and isolating behaviors, destruction of property, degradation, humiliation, threats, and insults (O’Leary 1999). Survivors of IPA often report that the emotional abuse was worse than the physical violence (Beeble et al. 2009; Theran et al. 2006). Emotional abuse has been found to predict poor mental health outcomes such as depression, higher levels of stress and/or Post Traumatic Stress Disorder symptoms, and low self-esteem over and above physical abuse (Coker et al. 2002; Dutton et al. 1999; Theran et al. 2006). Thus, although non-physical abuse is rarely considered a crime, this form of abuse has a multitude of negative consequences for women that must be considered when conducting research on IPA.

Post-Separation Abuse

Although many victimized women end their relationships with the expectation that the abuse will end as well, batterers often continue, or even escalate, their abuse post-separation (Fleury et al. 2000; Hardesty 2002; Hardesty and Chung 2006; Jaffe et al. 2003; Kurz 1996). Much of this abuse is directed at and/or involves the manipulation of shared children (Beeble et al. 2007; Bemiller 2008; Hardesty 2002; Hardesty and Ganong 2006; Harrison 2008; Kurz 1996; Moe 2009; Slote et al. 2005). Women report that abusers continue to use their children to exert control over them by threatening the children’s lives, mistreating their children to punish mothers, and being physically abusive towards the children (Hardesty and Ganong 2006; Slote et al. 2005). Mothers have also mentioned difficulties in getting their children back from visitations or, in instances where the mother is the noncustodial parent, having the abusive fathers restrict their access to the children (Bemiller 2008; Hardesty and Ganong 2006; Kernic et al. 2005). On the extreme end, post-separation abuse (PSA) results in fathers kidnapping or murdering mothers and/or children (Jaffe et al. 2003; Saunders 2009). Given the seriousness of PSA, it is imperative that family courts provide legal supports necessary to protect mothers and children.

Family Court’s Role

Family court custody arrangements determine physical and legal custody of children, visitation arrangements, and thus, abusers’ level of legal access to women and their children (Hardesty and Ganong 2006; Kernic et al. 2005). The U.S. court system is obligated to determine custody based on what is in the “best interests of the children” (Dalton et al. 2003; Hardesty and Chung 2006; Jaffe et al. 2003). Although the definition of these interests varies by state, family court typically encourages joint custody because of the assumption that frequent and prolonged contact with both parents is always in the best interests of the children (Kelly 2004). This assumption, however, is based on post-divorce impact studies that have failed to account for interparental conflict, leading several scholars to discredit this assumption (Jaffe et al. 2003; Salem and Dunford-Jackson 2008; Zorza 2007).

While joint custody may prove beneficial for children in situations where there has been no prior abuse (Amato 1993; Amato et al. 1995), the greater legal access abusers have to their children, the greater the risk PSA is for mothers and their children (Hart 1990). Shared physical custody and mandated visitation arrangements put mothers in frequent and continued contact with their abusers, allowing batterers opportunities to commit PSA with each child exchange (Hardesty 2002; Hardesty and Ganong 2006). Given the increased dangers related to custody when IPA is involved, most states have some provision(s) for considering relationship abuse when determining custody or visitation (Jaffe et al. 2003). However, the extent to which this occurs varies because each state handles divorce differently and considers different factors when deciding the best interests of the children (Jaffe et al. 2003).

Mediation

Mediation is a widely used process to negotiate child custody among divorcing couples. Proponents of mediation state that it is safer, more efficient, and less costly than traditional litigation (Cohen et al. 1999; Kelly 2004). Mediation is claimed to result in better co-parenting, higher compliance with court mandates, and/or longer involvement from the father in the child’s life (Cohen et al. 1999; Erickson and McKnight 1993; Kelly 2004). It is also purported to increase client satisfaction, decrease re-litigation rates, and/or produce better outcomes for children—likely due to the collaborative nature of mediation, which allows parents to make their own decisions regarding custody and visitation (Cohen et al. 1999; Kelly 2004). Some suggest that mediators have the ability to balance the power inequities present in IPA—especially by making adjustments to the process (Ellis and Stuckless 2006a, b). Others go so far as to say that mediation is an empowering, effective intervention that will end current abuse and reduce future abuse (Erickson and McKnight 1993).

Many IPA scholars, however, argue that mediation is neither effective nor safe when IPA exists (e.g., Beck and Frost 2006; Dalton et al. 2003; Hart 1990; Imbrogno and Imbrogno 2000; Jaffe et al. 2003; Johnson et al. 2005; Pearson 1997; Salem and Dunford-Jackson 2008; Tishler et al. 2004). The concept of mediation assumes that cooperation is attainable, there is little to no abuse among the parties, and each party can adequately argue for his or her needs—false assumptions when IPA is present (Beck and Frost 2006; Johnson et al. 2005; Tishler et al. 2004). Forcing victims to be present with, and argue against, their abusers can be unsafe, and abusers’ patterns of power and control may continue in mediated settings, causing survivors to be less able to negotiate for safe custody arrangements or financial resources (Hart 1990; Imbrogno and Imbrogno 2000; Johnson et al. 2005). Several scholars have argued that mediators are unable to identify IPA, let alone equalize the power difference (e.g., Hart 1990). Subtle threats such as “a look” or a word that appears innocuous to an outsider (e.g., the mediator) can in reality be a threat of future abuse (Stark 2007). Such hidden emotional abuse tactics are more likely to occur than physical violence in mediation (Dalton et al. 2003; Hart 1990; Imbrogno and Imbrogno 2000; Johnson et al. 2005; Pearson 1997; Salem and Dunford-Jackson 2008). Abusers often perform well under observation and manipulate mediators by professing a desire for joint custody and, thus, are viewed more favorably than the victims (Dalton et al. 2003; Hart 1990).

Little research exists on mediation in the context of IPA. Mediators, court staff, and researchers have reported the potential for mediation to be safer for survivors than traditional litigation due to mediators’ ability to promote safety by adjusting the process (Beck et al. 2010; Ellis and Stuckless 2006b; Pearson 1997). However, Johnson and colleagues (2005) found that mediators failed to identify 15 % of physically violent IPA cases. In another study conducted by Beck and colleagues (2010), even when exemplary training and screening procedures were implemented to identify and respond to IPA, mediators in one county only identified IPA in 59 % of the 1,015 IPA cases. Furthermore, only 7 % were screened out of mediation, and 19 % received accommodations. Most importantly, less than 1 % of cases that received accommodations included separate screening days and/or separate mediation session days. Common accommodations included requiring couples to leave separately, providing security escorts, mediation teams, and separate waiting rooms. Understanding how adjustments are made provides insight into the mediation process. However, given that ultimately a custody order is decided, it is how those decisions are influenced by the presence of IPA that has the most pressing and long-term impact on survivors and their children.

Currently, only one study identified by the authors used archival data to compare differences in how cases were handled based on whether or not IPA was present and how it influenced mediators’ child custody and visitation decisions (Johnson et al. 2005). When domestic violence was disclosed to mediators, they were significantly more likely to recommend sole custody to the violent fathers. When mediators had safety concerns, they were seven times more likely to recommend supervised visitation in non-domestic violence cases, but only four times in domestic violence cases. Furthermore, mediators recommended protected child exchanges when the perpetrators threatened the mother for 75 % of the non-domestic violence cases, but only for 32 % of domestic violence cases. Thus, mothers who disclosed the domestic violence were more likely to have their concerns ignored, lose custody, receive unsafe custody exchange recommendations, and were more subject to mediators’ arguably random decision making.

Current Study

Little is known about how and what factors influence custody decisions made by mediators. Prior research on child custody decisions in cases of IPA focuses almost exclusively on physical abuse (Kernic et al. 2005). Since emotional abuse is typically more hidden and more difficult to document, it is important to understand whether women decide to disclose this abuse, and how mediators respond. In addition, previous research on the custody negotiation process and custody outcomes in the context of IPA has relied on primarily quantitative or archival data (Johnson et al. 2005; Kernic et al. 2005; Rosen and O’Sullivan 2005). Survivors’ voices have been largely ignored by the literature and, therefore, a qualitative component within the study was needed to understand this process from women’s perspectives. Results from an analysis of women’s experiences with custody negotiation have been published elsewhere (Rivera et al., in press). Thus, this paper addresses the following research questions: (1) at the time of court-mandated mediation, what are women’s concerns about their own and their children’s safety if they have abusive partners; (1a) how do these concerns and the fathers’ abusive histories get relayed to the court mediators; (2) how do court mediators respond to allegations of the fathers’ abusive behaviors; and (2a) how do those responses impact custody and visitation recommendations?

Method

Setting

The Child Custody Experiences Study (CCES) was conducted in a county located in a Midwestern state that considers 12 factors (one of which is domestic violence) to determine the best interests of the children (Child Custody Act 1970). Similar to many jurisdictions, this county adheres to the “friendly parent” concept, and has explicitly stated that joint custody is preferred and that one parent should encourage frequent contact with the other parent. Mandated child custody negotiation occurs through a court staff official whose role is similar to that of a mediator, and the same assumptions of mediation apply (e.g., equal negotiating power). The main differences are that (a) the mediation session is mandated and run by family court, and (b) the court mediator makes a custody and visitation recommendation regardless of whether the divorcing couple comes to an agreement. Therefore, the terms “mediator” and “court mediator” are used interchangeably.

Should one or both of the divorcing parties object to the court mediator’s recommendation, they must file a motion and attend an additional court hearing. However, the court mediator’s recommendation is weighed heavily by the other family court staff and is difficult to change. At any point after the court mediation session and prior to the final divorce judgment being approved, either or both parties may be mandated for drug testing, custody/psychological evaluations, or private mediation to negotiate assets. Once all aspects of the divorce are agreed upon and a divorce decree is written, a judge must approve the decree in a final divorce judgment hearing. Although family court judges make the final judgment, they typically follow the recommendations of previous family court officials.

Procedures

Selection and Screening

After obtaining IRB approval, study participants were located through the county court clerk’s office. Using publicly available computers, all divorce cases with minor children filed from January 2006 through June 2008 were searched. This timeframe was chosen because we wanted to interview women who had at least a year of experience with custody and visitation after the divorce was final, but it was also important to limit recall bias. Every female plaintiff’s docket was reviewed for custody objections. We chose cases with custody objections because, based on the opinion of the county’s family court director, it was believed that the majority of these cases involved IPA.

For each potentially eligible participant, publicly available court records were requested from the court clerk and were then searched for telephone numbers. Potential participants were called from a university office for screening. English-speaking women were eligible for the study if they met the following criteria: (a) they were at least 18 years old, (b) they went through a divorce with at least one minor child in the county, (c) their divorce case remained in the county, (d) they had experienced physical, psychological, emotional, and/or sexual violence at the hands of their ex-husband, causing them to fear for their safety, and (e) they were willing to be audio taped during the interview.

Women were scheduled to participate in the interview at a location of their choosing that provided privacy. Of the 174 women we attempted to contact, 97 had disconnected numbers and 19 did not answer after repeated attempts. Of the 58 women contacted, 5 refused before they were screened, and 6 could not be reached after the initial contact for screening. Of the 47 women who were screened, 29 (62 %) were eligible for the study. Twenty-three of these women (79 %) were interviewed, and six canceled their scheduled interviews. Four interviewed women were removed from analyses because it was determined post-interview that their case did not involve IPA. Thus, the remaining 19 women comprise the sample for this study. Participants were compensated $40 for their time and given a comprehensive community resource guide after the interview.

There were 38 children across the 19 participants. Participants had a range of one to seven children (M = 2), and the children were between the ages of 2 and 25 years old (M = 12 years old). Participants ranged from age 23 to 52 years (M = 39 years). The majority of participants were White (n = 17; 89.5 %), with 5.3 % (n = 1) participants identifying as Latina, and 5.3 % (n = 1) as African American. All names provided in this paper are pseudonyms.

Measures

A semi-structured interview was used to collect information on women’s abuse history, experiences with and perceptions of the custody negotiation process, current custody and visitation experiences, and ongoing safety concerns. Field notes were taken during the interview, and all interviews were transcribed verbatim.

The Court Mediator Experiences Survey (CMES) was created for this study and includes two scales. The CMES was created to measure secondary victimization during court mediation (Orth 2002; Campbell 2005; Campbell and Raja, 1999, 2005; Orth and Maercker 2004). The first section has 30 items about the actions and behaviors of the mediator. Women responded to these questions on a Likert-type scale from 1 (strongly disagree) to 5 (strongly agree). Since this was a pilot test, a “not applicable” choice was also provided. The second section has 25 items and asks participants about the positive (e.g., “safe”) and negative (e.g., “blamed”) emotions they felt during mediation from 1 (not at all) to 5 (a lot). The CMES was given to participants for completion after the interview. Eighteen women completed the CMES. Because this was a pilot test of the CMES for scale development, and was designed for triangulation with the qualitative data, psychometric analyses were not conducted. Information on women’s secondary victimization experiences during mediation are presented elsewhere (see Rivera, et al., in press).

Analytic Induction

The analytic induction procedure used in this study followed on Robinson’s (1951) model of analytic induction: (a) develop an initial definition of the phenomenon of interest; (b) develop an assertion to explain the phenomenon; (c) examine one participant to determine if the assertion adequately describes and explains the phenomenon; (d) if the assertion does not describe/explain the phenomenon, re-define the phenomenon and/or modify the assertion to fit that case; (e) examine the second case with the new assertion; and (f) continue the process of re-definition and modification until all cases that fit the definition are explained by the assertion. An additional step—data reduction—was added prior to developing assertions. For a detailed description of our analytic induction process, see Rivera et al. (2011). Examining the fit between the data and the assertion includes negative case analysis where the researcher specifically seeks disconfirming evidence. In addition, Erickson (1986) describes guidelines to determine whether or not an assertion is adequate. Each assertion was first viewed in light of the evidence available to support it. If an assertion was determined to be inadequate, it was either dropped entirely or merged into another assertion.

The CMES was designed to both validate and supplement the interview data and assertions. For example, if during the interview a participant described how the mediator allowed the father to dominate the conversation, a low endorsement of the item “the mediator did not let my ex-partner control the conversation” was expected. Systematic validation occurred as followed: (a) based on the interview and assertion data, women’s responses on the CMES were predicted by the first author; (b) predictions were tested; and (c) assertion conclusions were edited and finalized. Three women did not have enough interview data to predict CMES responses (Amelia, Kelly, and Nicole), and one did not complete the CMES (Tanya). For seven women (47 %), their responses were predicted with 100 % accuracy. Remaining accuracy rates were 90 % for four women (27 %), 70–80 % for two women (13 %), and 60 % or less for two other women (13 %). Incorrect predictions were the result of women responding based on their second mediator or averaging their experiences across mediators (Jackie, Ashley, Toni, Rose), and/or because they had much more complex experiences than originally considered (Toni, Rose, Cecilia, Celeste, Chelsea). This process enhanced the validity of the final assertions, as all assertions were confirmed or disconfirmed via triangulation, rather than from just one data source.

Results

Eight assertions emerged through analytic induction. Research questions (RQ) and corresponding assertions are presented in Table 1.

Table 1 Research questions and corresponding assertions
Full size table

RQ 1: At the Time of Court Mediation, What Are Women’s Concerns About Their Own and Their Children’s Safety If They Have Abusive Partners?

Many women expressed concern over their children’s physical safety. For some women, these fears were based on prior threats their partners made about the children during the relationship. For example, some fathers threatened to fight for custody or kidnap the children if their wives left them. Some women also had concerns about physical abuse of the children by fathers, if those fathers had previously physically abused the children. In a few cases, women were concerned about the instability of the fathers and that the children would be left unprotected during visitation. Several women also had concerns about potential danger or indirect physical harm due to fathers’ actions. Others described how fathers’ behaviors became erratic during separation in such a way that placed the children in potential danger or demonstrated a complete disregard for the children’s wellbeing.

Most mothers had serious concerns about the children’s emotional wellbeing. Several women described the fathers’ exceptional ability to manipulate others—including the children. Attempting to turn children against their mothers, some fathers told the children that the mothers were to blame for the divorce, were tearing the family apart, and/or were mentally unstable. Unfortunately, these alienation attempts were either successful or caused the children emotional distress (see quote below). Some women were concerned about the fathers’ inappropriate treatment of the children. For example, men would discuss too many details of the divorce and custody battle with the children. Finally, many women were concerned that joint custody arrangements would disrupt children’s stability and cause the children emotional harm as captured by one woman:

My ex he was in a very depressive state his self when we were goin’ through the divorce. And he would always, you know, be talkin’ to my kid about all this adult stuff. And a lot of times my son would be like, “You left us. You abandoned us, mom.” You know, those are words he wouldn’t say at that age and they were coming, of course, from his dad. (Toni)

Two assertions emerged that corresponded with this research question. The first assertion was at the time of mediation, women are concerned for their children’s physical safety. Thirteen women (68 %) confirmed this assertion (Michelle, Jackie, Ashley, Cindy, Toni, Maria, Rose, Eva, Tanya, Chelsea, Nicole, Brittany, Lissette). The second assertion, at the time of mediation, women are concerned for their children’s emotional wellbeing, was confirmed by 14 women (74 %; Michelle, Jackie, Ashley, Cindy, Toni, Maria, Rose, Lillian, Eva, Cecilia, Melody, Kelly, Nicole, Brittany).

RQ 1a: How Do These Concerns and the Fathers’ Abusive History Get Relayed to the Court Mediator?

Women described the abuse in the divorce complaint and/or verbally described their concerns to the mediator. In addition, all six women who had restraining orders against the children’s fathers either told the mediators or had their attorneys tell the mediators about the orders. Mediators, however, inconsistently asked about abuse. Therefore, sometimes women would volunteer the information unprompted, whereas other women provided this information after being directly asked.

The first assertion for this research question follows: Women relay their concerns about their children’s safety to the mediator directly. Fourteen women (74 %) confirmed this assertion (Michelle, Jackie, Ashley, Cindy, Toni, Maria, Rose, Lillian, Eva, Cecilia, Chelsea, Nicole, Brittany, Lissette). One participant (Tanya) disconfirmed this assertion because she did not relay her concerns to the mediator, as he acted belligerently during mediation and she felt she did not need to describe his behaviors:

[The mediator] basically told [father] that if he didn’t calm himself down he wouldn’t have any rights. Because he was very aggressive. He was getting irritated. You could tell he was getting irritated ‘cause he kept clenching his fists and doin’ one of these numbers with his knee [shaking her knee up and down really fast, showing impatience] and…he was turnin’ red in the face….[Ex-husband] started yelling at the person we were talking to [the mediator] and [mediator]started yellin’ back like, “Look! You can’t yell at me.” “You yell at me, I’m gonna start yellin’ back. And then you’re not gonna get what you want.” And he started getting very agitated. [laughs] You could tell. You could see it in his face that he was getting agitated. (Tanya)

The second assertion is that: Women relay their concerns over the abuse or controlling behavior to the mediator directly. Thirteen women (68 %) confirmed this assertion (Michelle, Jackie, Ashley, Toni, Maria, Rose, Eva, Cecilia, Melody, Chelsea, Celeste, Brittany, Lissette). In sum, most women relayed their concerns to the mediators when they felt it is needed. Therefore, mediators in this sample were aware of abuse in most cases, whether or not they asked about it first.

RQ 2: How Do Court Mediators Respond to Allegations of the Fathers’ Abusive Behaviors?

Mediators responded in a variety of ways, and an understanding of their responses is complex. Some women believed mediators thought IPA was too complex to discuss, and their allegations of IPA were completely dismissed. Women’s testimonies alone, for example, were not considered enough evidence for their concerns to be taken seriously. The presence of physical evidence appeared to play an important role in mothers being taken seriously. For example, one participant (Rose) had her concerns taken seriously only after she showed the mediators letters the father wrote to the child from jail, in which the father made clear attempts to alienate the child from the mother. Other women mentioned the problem of not having “enough” or the “right” kind of evidence. Eight of the nineteen women in this sample (Michelle, Toni, Lillian, Tanya, Lissette, Brittany, Maria, Cecilia) did not experience frequent physical or sexual abuse, but rather the abuse was more controlling or emotional—types of abuse that are extremely difficult, if not impossible, to document. This difficulty is in part because these forms of abuse are not illegal. Therefore, when mediators requested evidence (e.g., police reports), women were unable to provide documentation. Not surprisingly, several women described that the courts view physical abuse as the only legitimate form of abuse:

I would say, “No [physical abuse], but he was controlling,” you know. “Do you have any documentation of that?” That’s what, that’s what I mean. If you don’t call the police, you don’t normally call the police if somebody hasn’t hit you. So if you don’t call the police then it’s, it’s your word against his....If he didn’t hit you, it didn’t happen. He didn’t hit you and you didn’t call the police, it didn’t happen. That’s basically the way I felt it is down there. (Toni)

One assertion was made in an attempt to capture the complexity of mediators’ responses: When mothers allege the father was/is abusive or controlling, mediators will ask the father to respond to the allegations and for independent evidence. Without evidence, mediators will be dismissive of abuse allegations, as though they are (a) she said/he said, (b) mutual violence, and/or (c) irrelevant. Independent evidence includes restraining orders or evidence of child abuse (attempts to turn child against the mother, threatening the mother to the child). However, even with independent evidence, mediators will be dismissive of abuse allegations toward the mother as (a) irrelevant, and/or (b) too difficult/complex. This is especially true for non-physical abuse toward the mother, as physical abuse is seen as the only legitimate form of abuse. Eight women (42 %) confirmed this assertion (Jackie, Ashley, Cindy, Toni, Maria, Rose, Lillian, Nicole). Two women disconfirmed this assertion because the mediator had no response (Eva) or the father acted belligerently during mediation (Melody).

These results indicate that mediators first will ask for evidence when they receive an allegation of abuse. However, what mediators consider as evidence varies. Proof of physical violence was generally taken seriously, but not for every participant. For example, restraining orders were given weight in some cases while in other cases they were not. Women whose abuse was marked by more controlling or emotional abuse than physical abuse were less likely to have evidence and, therefore, were more likely to have their allegations dismissed by the mediators.

RQ 2a: How Do Those Responses Impact the Custody and Visitation Recommendations?

Most women wanted full custody (n = 11, 58 %) or sole physical custody (n = 4, 21 %) of the children. Three women (16 %) wanted a joint custody arrangement, and the remaining mother felt the children were old enough to decide for themselves. From the mediators’ initial custody recommendation, seven women (37 %) were given full physical and legal custody of the children, six (32 %) received sole physical and joint legal custody, four (21 %) had joint physical and legal custody—one mother was primary physical, one father was primary physical, and two had 50/50 split custody—and two fathers (10 %) were given full physical and joint legal custody. In other words, of women who had a preference, ten (53 %) women received the custody arrangement they wanted and eight women (42 %) received less custody than they wanted, indicating that courts do not choose custody arrangements based solely on mothers’ requests and concerns.

Women were asked during the study interview what factors they believed were taken into consideration when the mediators made their custody recommendations. An important pattern emerged regarding a specific subset of women. Some women protected their children in ways the mediators did not understand, and were criticized or punished by the mediators for their actions. For example, one mother (Toni) described how she left her child with the father because he had a gun in the home and, due to his previous behaviors; she was scared that he would harm the child, her, and her family if she took the child at that point in time. While the mother viewed her choice as reasonable, the mediator responded negatively toward her leaving a child in that situation. The assertion developed from these data states: Mediators criticize or punish women for attempting to protect their children in ways the mediator does not understand. Four women (21 %) protected their children in this way, all of whom confirmed the assertion (Ashley, Toni, Rose, Lillian).

Several women described mediators’ strong preference for joint custody and that mediators try to decide custody based on a formula:

They recommended joint custody. And I think it was because at that point in time I think they, you have a… I don’t, I don’t mean this to sound sterile but there’s a formula. And I think that they’re looking at, you know, “Well, obviously this guy,” you know, “he didn’t, there, there’s no physical, they don’t have any pictures, they don’t have any police reports so,” you know, “this is gonna be my recommendation because this is what we normally do now for custody cases.” And so, and I, I think there’s a, a trend, you know, I don’t know, over the past whatever ye--, how many years it’s been but, you know, joint custody seems to be the, you know, the, what they’re recommending. (Jackie)

If joint custody was not awarded, a variety factors interacted in a complex way to inform mediators’ decision making. Some women described that their husbands acted so belligerently during mediation that the mediator was less willing to give the fathers custody. Some believed that the mediators recognized the mothers’ roles as primary caretakers and wanted to uphold that stability in the children’s lives. Similarly, the parent who could better provide basic necessities (e.g., proper hygiene, healthy food) was often favored. Others referenced that mediators favored the employed and housed parents who did not abuse substances or have a violent criminal background. Sometimes, the parents’ behaviors regarding getting the children to school and promoting school performance was considered. What was interesting, however, was how so few women felt that abuse was taken into consideration. Even when abuse was considered, it was often not considered to a significant degree, and as much visitation time or custody was awarded to the father as possible.

In an attempt to account for every woman’s experience and perception of what factors the mediator took into consideration, two detailed assertions emerged. The first assertion states that: Mediators prefer to award joint custody. If sole or primary custody is to be awarded, the mediator will favor the parent who meets one or more of the following criteria (a) does not use substances, (b) is employed, (c) promotes children’s education, (d) ensures basic needs are taken care (e.g., hygiene, housing, food), or (e) acts professionally during the meeting (e.g., does not act belligerently). These preferences exist regardless of who wants custody or of the occurrence of IPA against the mother. However, if sole or primary custody is to be awarded, the mediator will prefer to give as much time with the father as possible (i.e., liberal visitation)—especially if the father requests such time. Eighteen women (95 %) confirmed this assertion (Michelle, Jackie, Ashley, Cindy, Toni, Maria, Rose, Lillian, Eva, Tanya, Cecilia, Amelia, Melody, Chelsea, Nicole, Brittany, Lissette, Celeste). No women disconfirmed this assertion.

The second assertion was: Overall, a mediator will base his/her final decision on the above stated preferences. If mothers attempt to protect the child in ways the mediator does not understand, mediators will favor the father for custody. When the father does not want custody and the mother meets the above criteria, the mediator will follow the mother’s custody request. In general, mediators do not take IPA into account in the custody recommendation. Abuse or violence is only accounted for in the custody recommendation if there is direct and independent evidence of child abuse, but the father will still receive visitation if he requests it. Seventeen women (89 %) confirmed this assertion (Michelle, Jackie, Ashley, Cindy, Toni, Maria, Rose, Lillian, Eva, Tanya, Cecilia, Amelia, Melody, Chelsea, Nicole, Brittany, Lissette). One participant (Celeste) disconfirmed this assertion because the mediator gave sole physical custody to the father because the mother worked two jobs and wanted to exchange the children on a Wednesday, whereas the mediator wanted the children exchanged on a Friday. In this case, this result was shocking to both parents, as the mediator had at first talked the father into wanting joint custody (instead of his original request for sole physical custody), but then gave him sole physical custody for that reason.

Complex Factors Affecting Mediators’ Decision Making

During the analytic induction process, it became clear that mediators’ decisions regarding custody and women’s varying experiences with mediation were complex. In addition to assertion development, we examined the relationships among several factors in order to gain an in-depth understanding of these complexities. Using Miles and Huberman’s (1994) data reduction strategies, we used matrices to group data by various factors to identify common themes and patterns about mediators’ custody recommendations. A multitude of factors influenced how mediators made decisions about custody and visitation. For example, we noticed that only in some cases were women’s restraining orders considered seriously. We also noticed that most women told the mediator about the abuse and/or related concerns (e.g., alcohol or drug use), and yet those concerns were only taken seriously in a few cases. Patterns about the fathers also became apparent. In more than one case, fathers acted destructively during the hearing, which helped women gain their desired custody outcomes. Other characteristics about or actions of the father that appeared to influence mediators’ custody decisions included: (a) his criminal history (violent or non-violent), (b) his drug/alcohol use, and (c) whether he did not want or could not have custody (i.e., because he was incarcerated).

Final results indicate that mediators did not always consider women’s allegations of abuse or restraining orders as significant evidence to award the mothers full or sole custody. Only in cases where the father acted belligerently, the father did not want or could not have custody, or the mother requested joint custody, did the mediator’s recommendation match the mother’s request. These results suggest that IPA is not a deciding, or even important, factor in the mediators’ custody recommendations, even if there is a current restraining order. The more evidence women had (e.g., restraining order, fathers’ belligerent actions, fathers’ criminal histories), the more likely their concerns were considered sincerely by the mediator. However, it appears that the actions of the father during mediation may be one of the most important factors that mediators consider.

Discussion

This study was designed to shed light on the custody negotiation experiences of women survivors of IPA. When mothers are abused by their husbands, they have a variety of concerns regarding their safety and the safety of their children. The findings of this study illuminate the additional struggles endured by women who experience abuse that is more emotional and controlling than it is violent. In family court, physical abuse is often regarded as the only legitimate form of abuse and, therefore, women who experience more emotional abuse are not always viewed as “real” victims. These results replicate findings from previous studies, which found that IPA does not significantly influence family court’s custody decisions (e.g., Slote et al. 2005), nor do mediators make custody decisions in a consistent manner for cases involving IPA (Johnson et al. 2005). This study also contributes to the literature by discovering additional factors that may influence mediators’ custody decisions.

Women had a variety of physical and emotional concerns at the time of mediation. These concerns included: (a) the father’s previous and likely ongoing physical and emotional abuse of the children, (b) losing the children to him due to kidnapping or alienation, and/or (c) the father’s inability to provide a sanitary or stable environment for the children. These results were not surprising, given that previous studies have documented similar concerns or PSA experiences that involve abuse, kidnapping, or poor treatment of the children (e.g., Beeble et al. 2007; Hardesty and Ganong 2006). All of the women in this sample reported that either the father threatened to commit the behavior or he had previously exhibited the behavior in the past. In other words, women had legitimate concerns that deserved to be taken seriously. However, their concerns were rarely considered.

The failure of family court to consistently take abuse seriously has been previously documented (Slote et al. 2005), and the results of this study replicate these findings. The majority of women relayed their concerns by specifying them in the divorce complaint, discussing them with the mediator, or both—regardless of whether the mediator asked them about abuse. However, mediators responded inconsistently to these concerns. The women who had their concerns taken seriously often had concrete evidence to support their allegations. Concrete evidence includes the father’s violent criminal background, a current restraining order, and/or his belligerent behaviors during mediation. However, for cases where the abuse was less physical and more emotional, and when the father was not belligerent during mediation, women did not have such evidence and consequently felt their claims were not taken seriously.

Regardless of whether a case had evidence of abuse, mediators responded consistently—and negatively—when mothers attempted to protect children in ways the mediator did not understand. This finding indicates that mediators do not possess a comprehensive understanding of the multitude of creative strategies survivors of IPA may need to employ to protect their children from abusers. Survivors often have years, if not decades, of experience in dealing with their abusers. They are in a better position to identify which threats to take seriously and are more likely to be able to predict what the father will or will not do in regards to harming the children than court mediators. Therefore, actions such as temporarily leaving the child with an abusive father may appear to be dangerous to an outsider when, in fact, it may have been the safest option the mother felt she had (Bancroft 2010). The lack of this comprehensive understanding can have serious consequences. Mediators were unable to acknowledge that mothers’ efforts may have indeed protected the child, and they failed to recommend the safest custody arrangement. In one case, for example, the mediator awarded the abusive father full physical and legal custody of the child. Thus, it is important to understand all of the factors that contribute to mediators’ custody decisions.

This study illuminated two influential factors in meditational decision-making not previously introduced in the literature—the father’s behavior during mediation and the father’s custody request. First, in nearly every case where the father did not want custody, the mother was awarded full physical and legal custody or sole physical and joint legal custody. Although it is not surprising that men who do not want custody are not given custody, it is surprising that this factor may be more influential than fathers’ abuse histories.

This finding challenges two claims made by fathers’ rights (FR) groups. First, FR groups claim that family court is biased against fathers (Dutton 2005; Jaffe et al. 2003). However, this study indicates that family court actually prefers to award joint custody to all divorcing couples with children and typically gives sole custody only in cases where both parents want this arrangement. Second, FR groups assert that allegations of abuse are vastly overestimated and are invented to help mothers gain full custody (Dutton 2005; Jaffe et al. 2003). This study, in addition to the studies presented in the literature review, do not support such an assertion. Given that abuse allegations often hurt mothers in family court, it is highly unlikely that such allegations are invented by mothers and their attorneys to gain custody (Dutton 2005; Jaffe et al. 2003). Second, fathers’ actions during the meeting helped the mediator believe they had obtained a complete picture of his behavior and the mothers’ concerns. One emergent pattern was women who had only a negative mediation experience also had a father in the meeting who did not act belligerently. This indicates that if the abusive father acts calm, professional, or charming, the woman’s allegations are less likely to be believed by the mediator—even if she has a restraining order against him. It may be that mediators will consider his actions during mediation more heavily than his criminal or abusive past. This finding contributes to previous research that has demonstrated that when abusive men act belligerently or aggressively in front of police officers, women’s concerns are taken significantly more seriously (Finn and Bettis 2006).

Taken in combination, study results suggest that victimized mothers receive full or sole physical custody recommendations when the father (in order of hypothesized importance): (a) does not want custody, (b) acts belligerently during mediation, and/or (c) has a criminal history—especially if this history is a violent offense. Thus, although most women received custody of the children, patterns regarding how such decisions were made are cause for concern. IPA allegations or evidence often did not assist survivors in gaining full or sole physical custody of the children. Furthermore, the factors that appear to be most influential rely on the father’s actions during mediation—another way batterers are able to exert control over their partners. It appears that mediators were following the rule for criminal justice systems—proof beyond a reasonable doubt. However, because family court is a division of civil court, mediators should be following the rule of preponderance of evidence—whether it is more likely than not that the alleged action/s occurred. Therefore, it appears that any physical evidence (e.g., restraining order) should have been taken more seriously, and women’s allegations could have been given more weight in mediators’ custody decisions.

There are several limitations of this study. First, most participants were highly educated white women and the conclusions drawn from this study do not represent the voices and experiences of women of color or those with less education. Second, only women whose cases had an objection to the court mediators’ decisions were contacted. Third, we attempted to contact women by calling them based on the phone number provided in their public files. This likely biased the sample to women who had fewer safety concerns, more consistent living situations, and/or more income. Fourth, this was a cross-sectional study. Future research could utilize a longitudinal design to gain an in-depth understanding of women’s experience at different points in the process. Finally, given the jurisdictional differences in child custody case processes and procedures within family courts, our research has limited ability to be widely generalized to differing jurisdictions.

This study also has strengths worthy of note. This was the first known example of attempting to gain an in-depth understanding of women’s custody negotiation experiences. Using mixed-methods allowed us to obtain a rich, contextualized amount of information to understand women’s experiences. Second, this study included women who experienced more non-physical abuse than physical abuse. Previous literature has focused on women who experienced physical abuse and/or who had restraining orders. A third strength of this study was the rigorous tests of trustworthiness and credibility of the data (Lincoln and Guba 1985; Miles and Huberman 1994).

This study was a first step towards understanding women’s custody negotiation experiences. Results indicate that IPA does not significantly influence mediators’ custody decision making. Rather, father’s actions and custody desires during mediation appear to be more important to mediators. Currently, there is little evidence to indicate that family court staff implements policies that demonstrate an accurate or clear understanding of the dynamics of IPA. Further, emotional abuse and control are not taken seriously, and physical abuse is considered to be the only legitimate form of abuse. A number of policy and training changes need to occur within family courts to maximize the safety of women and children in cases of IPA.