Leaving an Abusive Partner: Exploring Boundary Ambiguity Using the Stages of Change Model



The Stages of Change Model has helped researchers and practitioners explore the process of leaving an abusive partner, but it does not account for the changes in relational boundaries unique to this process. In this paper, we integrate the concept of boundary ambiguity into the Stages of Change Model to explore potential sources of boundary ambiguity in the process of leaving. We theorize that Boss’s two types of boundary ambiguity may be present at different points of the leaving process and are perceptual barriers to leaving permanently. We also theorize that potential sources of boundary ambiguity emerge from women and their partners’ decisions and actions in the process of leaving. Implications and future directions for research and practice are discussed.

An important research agenda in the intimate partner violence (IPV) field is to understand how women leave their abusive partners. Research suggests that leaving an abusive partner is a process that involves multiple stages, rather than a single isolated event (Moss, Pitula, Campbell, & Halstead, 1997). One model that has been used to theorize this process is the Stages of Change Model (Prochaska & DiClemente, 1984). The Stages of Change Model was originally developed to explain how individuals change behaviors (e.g., smoking). When applied to the process of leaving, the model effectively outlines how women leave an abusive partner in five stages that are ordered on the basis of women's readiness to leave (Burke, Gielen, McDonnell, O'Campo, & Maman, 2001). Thus, a woman in an earlier stage in the model would be less ready to leave her abusive partner compared to women in a later stage. Further, in the Stages of Change Model, the desired outcome of the process of leaving is the woman's sustained separation from her abusive partner.

Although the Stages of Change Model is a useful theoretical model for understanding behavior change, it has important limitations when applied to the process of leaving an abusive partner. Specifically, the model focuses on individuals’ efforts to change their own behaviors (e.g., a smoker's efforts to cease smoking), and, thus, it fails to account for the relational components that may be unique to the process of leaving (Brown, 1997). Namely, women make decisions to leave within the context of partner and parenting relationships that may influence not only individual family members but also their family structures. For example, women may take their children with them when they leave and then coparent with their former partners across different households, consequently altering family dynamics and boundaries. One potential co-occurring process that has yet to be explored in studies on the process of leaving and the Stages of Change Model is boundary ambiguity.

Boundary ambiguity is a perceptual state in which family members are uncertain about who is in or out of the family system (Boss & Greenberg, 1984). Although known to influence families experiencing separation or divorce (Boss, 2007; Madden-Derdich, Leonard, & Christopher, 1999), boundary ambiguity has not been studied as a mechanism specific to the process of leaving an abusive partner. Because boundary ambiguity is perceptually based, it may be a critical perceptual barrier affecting women's ability to sustain change (i.e., to remain separate from their partners). Therefore, in this article, we sought to integrate these disparate bodies of literature and propose a theoretical expansion of the Stages of Change Model that incorporates the potential role of boundary ambiguity. First, we provide a theoretical foundation by reviewing the Stages of Change Model and its use in research on the process of leaving. Next, we review the relevant literature on boundary ambiguity and theorize about potential sources of boundary ambiguity at different stages of change in the process of leaving. Finally, we discuss several research and practical implications of incorporating boundary ambiguity into the Stages of Change Model.

Theoretical Foundation: The Stages of Change Model and the Process of Leaving

In essence, the Stages of Change Model includes four components of change. First, the stages of change refer to the levels of individuals’ readiness to change a certain behavior, which include precontemplation, contemplation, preparation, action, and maintenance. Individuals in earlier stages (e.g., precontemplation), compared to those in later stages (e.g., action), are less ready to change and thus are less likely to succeed in changing their behaviors (Prochaska & DiClemente, 1984). For abused women, movements across the five stages of change may be bidirectional and include both linear (i.e., from one stage to the next) and nonlinear (i.e., skipping stages) trajectories (Khaw & Hardesty, 2007). Second, the processes of change denote individuals’ experiential and behavioral strategies used to initiate change. These strategies are central to decision making throughout the process of leaving (Brown, 1997). Finally, decisional balance and self-efficacy, respectively, refer to the weighing of the pros and cons of change and the individual's level of confidence in changing behaviors. Existing work demonstrates how the Stages of Change Model and these four components of change can be applied to the process of leaving an abusive partner.

Stages of Change


In precontemplation, individuals have no intention to change in the foreseeable future (Prochaska, DiClemente, & Norcross, 1992). In most IPV research, abused women have been identified as precontemplators on the basis of retrospective descriptions of their perceptions and behaviors throughout the process of leaving (Burke et al., 2001). Two exceptions are Edwards et al.; (2006) and Shurman and Rodriguez (2006), who identified precontemplators on the basis of a list of self-descriptive items from an existing change assessment tool (the University of Rhode Island Change Assessment; McConnaughy, Prochaska, & Velicer, 1983). A precontemplator is often identified as not being psychologically prepared to change, as she perceives her relationship as normal (Brown, 1997). Because she has not labeled her experiences as abuse, she may deny that anything is wrong in her relationship and blames herself for the situation (Zink, Elder, Jacobson, & Klostermann, 2004). The lack of understanding and denial of the abuse are evident in terms of the implications on women's personal health. In one study that compared abused women in various stages of change, precontemplators were found to be less vulnerable to experience symptoms of psychological distress from the abuse and reported lower levels of depression, post-traumatic stress symptoms, and suicidal ideation (Edwards et al., 2006). Edwards and colleagues’ finding supports the notion that precontemplators tend to view their relationships through rose-colored glasses that enable them to minimize the negative effects of the abuse and disengage from the current reality of their situation (Shir, 1999).


Abused women are identified as being in contemplation when they acknowledge that a problem exists and begin to seriously consider change (Burke et al., 2001). Contemplators are not yet committed to leaving, however, and often report ambivalent feelings about their relationship (Brown, 1997). For example, their desire for change may coincide with feelings of loyalty and love toward the abuser. As a result, many women may stay in their relationships for many years despite the abuse (Moss et al., 1997). Indeed, studies indicate that most individuals stay in contemplation for up to 2 years without ever taking any significant action to change (Prochaska & DiClemente, 1984). In the process of leaving, it has been suggested that contemplation can be further divided into early contemplation, where women may see the relationship as abusive but choose not to act upon it, and late contemplation, in which women start to weigh the pros and cons of leaving (Zink et al., 2004).


When a woman commits to change, she moves into preparation, which is “a stage that combines intention and behavioral criteria” (Prochaska & Norcross, 2001, p. 444). Here, she recognizes the abuse as a problem, has the intention to leave, and begins to develop a plan to leave (Burke et al., 2001). For example, she may secretly arrange for housing to go to after she leaves, secure her finances, or discuss safety options with a friend (Brown, 1997). Unlike other behaviors, however, some IPV researchers have found it difficult to delineate the preparation stage because many women actually make quick and spontaneous decisions to leave without much time or need for deliberate preparation (Chang et al., 2006; Khaw & Hardesty, 2007). For example, among 20 abused women who had either left or stayed with the abusers, Chang and colleagues reported that most women had progressed through the Stages of Change Model in a nonlinear sequence, including those who moved from contemplation to action without experiencing a distinct preparation stage.


In comparison to the preparation stage, individuals in action actually modify their behaviors and take other steps to overcome the problem (Brown, 1997). Although there is no single behavioral criterion to indicate that a woman is in action (Brown), most IPV studies have defined action as leaving the abuser or any other active attempts taken toward achieving safety and changing the nature of their situation (Burke et al., 2001). For example, women may speak to a counselor, seek help in the community, or obtain a job (Chang et al., 2006). Unlike other behaviors (e.g., smoking), decisions and actions in the process of leaving are complex and may rely on extenuating factors, such as the availability of community resources and social support (Burke et al., 2001). This stage also appears critical for many women who want to leave but for various reasons (e.g., children) do not move into the next stage of maintenance and instead return to an earlier stage. In the Stages of Change Model, this behavior is known as relapse and is considered to be a normal part of the change process (Prochaska & DiClemente, 1984).


In general, IPV research reveals little about women's experiences after leaving (Anderson & Saunders, 2003). The Stages of Change Model, however, contends that successful change includes maintaining the changes after the action stage for over 6 months. During this time, women are “preventing relapse and consolidating the gains attained during action” (Prochaska & Norcross, 2001, p. 444). Previous studies suggest that a woman is in maintenance when she does not return to the abuser and keeps herself safe by maintaining separation as a part of developing her “ability to move on” (Burke et al., 2001, p. 1156) after leaving. Although the Stages of Change Model assumes that most desired outcome in the process of leaving is to be permanently separated from an abusive partner (Brown, 1997), in reality permanent separation is not the safest or most desired option for all women (Goodkind, Sullivan, & Bybee, 2004). For example, studies have shown that women's risk of homicide may increase ninefold when she separates from a highly controlling partner (Campbell et al., 2003). Moreover, not all women intend to leave their partners (Byrne & Arias, 2004) and instead may employ other strategies (e.g., placating the abuser) to be safe while remaining in the relationship (Goodkind et al.).

Processes of Change

Whereas the stages of change illustrate what women experience in the process of leaving, 10 cognitive and behavioral processes of change explicate how women move through the stages. Cognitive processes range from consciousness raising to social liberation; behavior processes range from utilizing helping relationships to stimulus control (e.g., leaving the room or ignoring the abuser when he shows signs of escalation). The full range of such processes is depicted in Table 1. Only a few studies have focused on women's processes of change in the process of leaving (e.g., Burke, Denison, Gielen, McDonnell, & O'Campo, 2004). Overall, women reportedly use more processes as they move through the stages of change, and their readiness to leave increases over time (Burke et al., 2004). Women also tend to use more cognitive processes in the earlier stages of precontemplation and contemplation versus more behavioral processes in the later stages of preparation, action, and maintenance (Brown, 1997), indicative of an evolving pattern from thinking about leaving to the actual behavior of leaving the abusive partner.

Table 1. Processes of Change
  1. Note: Adapted from “Ending Intimate Partner Violence: An Application of the Transtheoretical Model,” by J. G. Burke, J. A. Denison, A. C. Gielen, K. A. McDonnell, and P. O'Campo, 2004, American Journal of Health Behavior, 28, p. 125.

Cognitive processes 
 Consciousness raisingSeeking new information to understand the problem behavior
 Self-reevaluationEmotional and cognitive reappraising of the problem behavior
 Dramatic reliefExperiencing and expressing feelings about the problem behavior
 Environmental reevaluationConsidering and assessing how the problem behavior affects the individual’s environment
 Social liberationIncreasing awareness, availability, and acceptance by the individual of alternative, problem-free lifestyles
Behavioral processes 
 Helping relationshipsTrusting, accepting, and using social support to change
 Counter-conditioningLearning and practicing alternative behaviors
 Reinforcement managementRewarding oneself or being rewarded by others for making changes
 Self-liberationChoosing and committing to change
 Stimulus controlControlling situations and other causes that trigger the problem

Decisional Balance and Self-Efficacy

In addition to the five stages of change and the 10 processes of change, the model also incorporates the constructs of decisional balance and self-efficacy. Decisional balance is the weighing of pros and cons of change. Because the pros and cons of leaving can be powerful motivators and inhibitors of change (Brown, 1997), decisional balance is found throughout all stages of leaving (Burke et al., 2004). Burke and colleagues (2004) found that women resisted change and stayed in contemplation longer when they perceived the cons of leaving (e.g., having to sever financial and emotional ties to the abuser) as being greater than the pros of leaving (e.g., achieving nonviolence). As women gained more insight into the pros of leaving, they became more ready to leave and were better able to move on into subsequent stages.

Self-efficacy refers to an abused woman's level of confidence to leave and stay free from the abuse (Burke et al., 2004). In general, abused women have greater intentions to leave if they hold more positive attitudes about leaving and feel in control of leaving the relationship (Byrne & Arias, 2004). Not surprisingly, confidence to leave the relationship tends to increase along the stages of change (Brown, 1997) and plays an important role in women's process of reclaiming self in the process of leaving (Wuest & Merritt-Gray, 2001). Reclaiming self is the social process of reaffirming women's identities in an IPV context, and survivors often report becoming more conscious of their personal power and control over their lives after they have left an abusive partner (Wuest & Merritt-Gray).

Taken together, although the Stages of Change Model has helped researchers theorize the process of leaving, the model in its original form is limited in two important ways. First, the model explores how individuals change their own behaviors but fails to consider that, in the process of leaving, women's actions are in response to their partners’ problem (i.e., abusive) behaviors (Frasier, Slatt, Kowlowitz, & Glowa, 2001). Second, the process of leaving occurs in a relational context (Brown, 1997), whereby women are changing within their relationships with the abusive partner and other family members (e.g., children). Unlike other types of change in which the desired outcome is the termination of an individual's problem behavior and therefore the outcome is individually centered (Prochaska & DiClemente, 1984), a permanent separation may not be the desired outcome for some women because of its potential negative impact on the family (e.g., breakup of the family unit; Goodkind et al., 2004). Further, separating from an abusive partner can place some women at an increased risk for violence (Hardesty & Chung, 2006; Kurz, 1996). Thus, for women in the process of leaving, potential solutions are complex and typically affect not only themselves but their partners, children, and other family members. As is, the Stages of Change Model does not adequately account for the complex and systemic nature of leaving an abusive partner. Thus, little is known about how changes in relational boundaries affect women as they move through the stages of change. Drawing from family stress theory, we propose an expansion of the Stages of Change Model that includes the potential influence of boundary ambiguity (Boss, 2002, 2006) on the process of leaving an abusive partner. In the next section, we review the relevant boundary ambiguity literature and explain possible sources of boundary ambiguity emerging at different stages of change.

Boundary Ambiguity

Boundary ambiguity is a maladaptive perceptual state of mind that blocks effective coping and stress management in families (Boss, 2002). When families experience high boundary ambiguity, “the family boundary is no longer maintainable, roles are confused, tasks remain undone, and the structure is immobilized …. [Psychologically,] cognition is blocked by the ambiguity, decisions are delayed, and coping and grieving processes are frozen” (Boss, 2002, p. 95). Boss (2002) has identified two types of boundary ambiguity, known as Type I and Type II boundary ambiguity. Overall, studies show that both types of boundary ambiguity are major barriers to coping and generate negative outcomes in families, such as feelings of anticipatory grief, depression, and negative outlooks on family life (Boss, 2002; Kaplan, 2001). These boundary ambiguity types are briefly described in the following paragraphs.

Type I boundary ambiguity is when one is physically absent from the family but maintains a degree of psychological presence in the family (Boss, 2007, p. 105). For example, among families with missing, at-war fathers (whose living statuses were undetermined), spouses and children often felt emotionally preoccupied with their fathers and reported being in limbo (Boss, 1977). In other words, these families felt stuck in a state of uncertainty over their absent fathers for an indeterminate amount of time and thus were unable to grieve their loss and move on with their lives (Campbell & Demi, 2000). Type I boundary ambiguity may also occur in families experiencing divorce or separation. Specifically, after a marriage dissolves, family members may question if one or the other partner is still a part of the family despite not being physically there (Madden-Derdich et al., 1999). Because of prior shared ties, such as children, boundaries in divorced and separated families can remain unclear for a long time, and the task of creating new family boundaries (e.g., due to remarriage) may become even more challenging (Emery, 1994).

In contrast, Type II boundary ambiguity occurs when a family member is physically present but psychologically absent (i.e., mentally and emotionally unavailable) from the family. Studies on Type II boundary ambiguity have mostly focused on caregivers of family members with chronic illnesses (e.g., Alzheimer's). Although these individuals are physically present, they have lost a degree of mental and emotional abilities to function in the family due to their cognitive impairments (Boss, 2007). In one study, families reported being uncertain about including a psychologically absent family member in family roles and rituals but also felt guilty if the family member was excluded (Garwick, Detzner, & Boss, 1994). In a boundary ambiguity situation, it is normal for families to experience ambivalent feelings toward the family member, or they may simply deny that any problem exists. No doubt, these responses can affect families’ ability to accept and cope with their losses and to move on with their lives (Boss, 2002).

Overall, much of the existing boundary ambiguity research has focused on families experiencing some form of loss of membership and roles (e.g., through death, illness, separation, and divorce). Nevertheless, the breadth of these studies has allowed for the construct of boundary ambiguity to “gain a solid base of empirical support in family situations involving similar yet distinct types of loss” (Carroll, Olson, & Buckmiller, 2007, p. 225). As evident in the special issue on ambiguous loss in the journal Family Relations (Boss, 2007), efforts to test and explore boundary ambiguity in different settings and with new populations are ongoing and further warranted. On the basis of our review of the process of leaving and boundary ambiguity literatures, we posit that the process of leaving an abusive partner has the potential to elicit boundary ambiguity and that boundary ambiguity may act as a perceptual barrier to sustaining change. We also posit that the context of violence presents unique dynamics that have not been considered in studies on boundary ambiguity due to separation and divorce. These notions subsequently led us to identify possible sources of boundary ambiguity at each stage of change in the process of leaving.

Possible Sources of Boundary Ambiguity in the Process of Leaving

Stage of Precontemplation: Women's Denial and Ambivalence

In precontemplation, boundary ambiguity may emerge as women experience contradictions between their expectations of a husband (e.g., love, protection) and their husbands’ actual behavior toward them (e.g., abuse). As these contradictions occur, women are beginning to experience an irretrievable loss in their relationship (e.g., loss of prior trust and expectations of their partners) in which they cannot go back to the way things were (Boss, 1988). Similar to families who have just encountered a psychological loss (e.g., due to Alzheimer's), women have to first make sense of the changes occurring in the relationship. Because women at this stage are not ready to contemplate change, however, they may respond to the initial signs of loss with denial and ambivalence. By deluding themselves and ignoring or redefining their realities, women create an illusion of a perfect relationship (Senter & Caldwell, 2002). Similarly, it is often easier for families with high boundary ambiguity to use denial to distort their realities in order to avoid the painful truth of the situation and to reconstruct it on the basis of wishful thinking (Boss, 1993, 2002). Thus, at this stage, women may use denial to create a temporary and false sense of boundary clarity (or security) in their relationship, ultimately preventing them from recognizing the abuse and taking steps to change (Brown, 1997).

Abused women also experience profound ambivalence, or contradictory perceptions and feelings that affect their decisional balance (i.e., weighing of pros and cons) and self-efficacy (i.e., confidence) to change (Burke et al., 2004). As previously mentioned, this experience is also common in families experiencing boundary ambiguity. For example, family caregivers of Alzheimer's patients have reported feeling angry at the individual patient for being emotionally absent, only to be consumed with guilt for having such thoughts (Boss, 1999). Similarly, an abused woman may feel shame and fear for her safety but still love her partner and thus continue to commit to the relationship and hope for change (Hendy, Eggen, Gustitus, McLeod, & Ng, 2003). By responding to the abuse with denial and ambivalence, a precontemplator will “actively construct a particular meaning of her circumstances that will allow both the continuation of the love that she feels toward her partner and the endurance of the pain from abuse” (Anderson & Saunders, 2003, p. 175). As a result, abused women may get stuck in the early stages of change, explaining why many women stay in their relationships for years despite their experiences of abuse (Burke et al., 2001).

Stages of Contemplation Through Action: Fluctuations of Being In and Out

The stages of contemplation, preparation, and action may have the most potential for boundary ambiguity because women are shifting from cognition to action (i.e., from using cognitive processes to more behavioral ones; Khaw & Hardesty, 2007; Prochaska & Norcross, 2001). Changes resulting from this shift may consequently elicit stress and alter family boundaries. Because the process of leaving occurs within a relational context, we posit that researchers must account for changes elicited by both women engaged in the process of leaving and abusers who may exhibit abusive behaviors in response to women's process of leaving.

Abusers’ fluctuating psychological presence

One possible source of boundary ambiguity is the pattern of fluctuation in abusers’ psychological presence in the family (i.e., they are psychologically absent at some times but psychologically present at others). In terms of psychological absence, research has documented examples of abusers’ psychological absence and emotional detachment from their marital and parental relationships. Comparing marital interactions between violent and nonviolent husbands, Babcock, Jacobson, Gottman, and Yerington (2000) found that violent husbands were more likely to be insecurely attached, resulting in higher emotional withdrawal and distancing from their partners. Abusers were also more unwilling to sacrifice the time and effort to meet certain family duties, particularly those that they viewed as socially prescribed female roles (e.g., housework or child care; Ptacek, 1988).

Findings also suggest that abusers tend to be physically present but underinvolved and neglectful parents. Margolin, John, Ghosh, and Gordis (1996) found that abusers exerted more negative parenting skills than nonabusers. For example, compared to nonabusers, abusers were more likely to use physical punishment and authoritarian parenting styles that often lacked feelings of love or affection for their children. Rather, abusers tend to perceive their own children negatively as a “hindrance or annoyance… [and they] are not reliable in knowing the names of their children's schoolteachers, daycare providers, details of medical conditions… or being able to describe knowledge about their children's interests, strengths, or ambitions” (Bancroft & Silverman, 2002, p. 32). Although some abusers, particularly biological fathers, may perceive themselves to be good fathers and show concern for their children's outcomes due to the violence in the home (Fox, Sayers, & Bruce, 2001), a recent study by Rothman, Mandel, and Silverman (2007) found that fathers’ concerns for their children's well-being did not necessarily result in the intention to cease or refrain from violence against the children's mothers. Thus, we posit that an abuser may be more psychologically absent and emotionally detached from his family because he neglects and violates his partner and fathering roles (Boss, 2002).

Subsequently, abusers’ psychological absence and emotional detachment from the family may trigger women to see that something is wrong in the relationship (e.g., “he is never there for me and the children”). Recognizing abusers’ psychological absence may be a significant turning point of realization (Khaw & Hardesty, 2007) when women gain awareness of their partners’ behaviors and accept the reality of their situations (Burke et al., 2001). Similarly, in a Type II boundary ambiguity situation, families often do not realize that there is a problem with a family member until they are able to see that he or she is psychologically absent (Boss, 1999). Only when families accept their losses and stop denying the situation can they begin to seek solutions and prepare for the future (Boss, 2006).

Once women take explicit steps toward change in the preparation and action stages, abusers’ psychological absence in the family may then fluctuate to being present. Specifically, when she prepares or threatens to leave, an abuser may respond by intentionally creating a false sense of emotional and psychological presence as a partner. For example, he may display more emotional affection toward his partner and romance her as he did in the start of their relationship. This particular response pattern resembles the kindness and contrite loving behavior phase in the cycle of violence (Walker, 1979). At this phase, the abuser often behaves in a charming and loving manner, is usually sorry for his abusive actions, and conveys his remorse by begging for her forgiveness and promises to never hurt her again (Walker). Recent work suggests that these responses are actually abusers’“purposeful, reactive and proactive tactics designed to mitigate their responsibility for violent behaviors” (Cavanagh, Dobash, Dobash, & Lewis, 2001, p. 700). Cavanagh and colleagues labeled these responses as remedial work, or damage control strategies, intended to change the meaning of IPV into something more acceptable in order to sustain the abusive relationship. Thus, when threatened by women's attempts to leave, we posit that abusers use remedial work to intentionally reestablish their psychological presence and create a temporary and false sense of boundary clarity. Because abusers have much to lose if women leave (e.g., perceived loss of control and rights to their partners and children; Mahoney, 1991), the reestablishment of abusers’ psychological presence becomes a necessary extension of control over women's decision making (Cavanagh et al.) As a result, women may find it difficult to clarify the nature of their relationships and to make the decision to leave (Liang, Goodman, Tummala-Narra, & Weintraub, 2005).

When children are involved, mothers’ decisions to leave may be further complicated when an abuser establishes himself as a psychologically available father. Studies have shown that while some abusers are abusive toward their partners, they are able to stay emotionally available to their children at the same time (Sullivan, Juras, Bybee, Nguyen, & Allen, 2000). Many IPV studies have found that abused mothers typically perceived their partners as being mostly good fathers and talked about wanting to preserve father-children relationships as reasons to hold back from leaving (Khaw & Hardesty, 2007). Mothers also particularly wanted their sons to continue having a male figure in the home (Zink, Elder, & Jacobson, 2003). Consequently, they may choose to adopt placating strategies (e.g., follow abusers’ demands) in an attempt to stay safe while remaining in the relationship (Goodkind et al., 2004). Additionally, abused mothers may not realize or they may underestimate the effects of direct or indirect exposure to IPV (Zink et al., 2003). It is normal for mothers to “delude themselves into thinking that the child is unaware of the assaults” (Holden, 2003, p. 156) unless they were directly abused by their fathers (Stephens, 1999). Therefore, the literature suggests that although mothers may acknowledge the violence, they may not consider leaving the relationship as long as they believe their partners to be good fathers and that their children are safe in the home. Once that threshold is violated and the negative impact of IPV on children becomes undeniably apparent, mothers often feel compelled to leave for the safety of their children (Campbell & Lewandowski, 1997; Mohr, Noone Lutz, Fantuzzo, & Perry, 2000).

Women's fluidity of relationship status

Boundary ambiguity may also result from the fluidity of relationship status among women leaving abusive partners (Campbell, Rose, Kub, & Nedd, 1998; Lerner & Kennedy, 2000). The fluidity of relationship status refers to the “ebb and flow of a couple's relationship course over time” (Bell, Goodman, & Dutton, 2007, p. 415). In one study, Campbell et al. (1998) asked abused women to describe their relationship status at three separate time points over 3 years. Although some women indicated that they were clearly “in” their abusive relationships and some were clearly “out,” many women were in an intermediate “in/out” status in which they either had expectations of the relationship ending while they were partnered or continuing while they were separated. On the basis of Campbell and colleagues’ (1998) findings, we theorize two possible sources of boundary ambiguity occurring in the middle stages of change for women: women's psychological absence while they are physically with their partners and their psychological presence while they are separated from their partners.

The first source of boundary ambiguity emerges from women's process of mentally leaving before physically leaving (Merritt-Gray & Wuest, 1995). This process is similar to the “goodbye without leaving” analogy of a Type II boundary ambiguity situation (Boss, 2007). In the middle stages of change, women often emotionally detach themselves from their partners as they begin to weigh the pros and cons of leaving and make preparations to leave (Brown, 1997). Once women move out of precontemplation, they typically experience fewer periods of love and affection toward their partners, subsequently causing previously suppressed feelings of loss, anger, and fear to emerge (Anderson & Saunders, 2003; Shurman & Rodriguez, 2006). Women's psychological absence is evident when their perceptions shift from wanting to maintain the relationship to trying to leave the abusive partner (Anderson & Saunders).

Boundary ambiguity in these stages may also transpire from repeated incidences of leaving and returning that are common in the process of leaving. Each time an abused woman leaves and is physically absent from the relationship, she may experience ambivalence about her relationship and be forced to respond to a variety of forces that collectively push her to return to her partner (Anderson & Saunders, 2003). For example, she may have continued loving feelings toward her partner and desires to reunite her family while simultaneously coping with insufficient economic and emotional resources necessary to sustain her separation. Boundary ambiguity may then emerge because there is uncertainty about whether she is in or out of the relationship or when she is physically absent from the relationship but has not made a final break and remains psychologically attached to her partner (Campbell et al., 1998). This situation is similar to the “leaving without saying goodbye” analogy of Type I boundary ambiguity (Boss, 2007), which is commonly perceived by families who experience separation or divorce. For these families, it is hard to exclude a physically absent member from the family, especially when there is still a lingering emotional connection to this person (Madden-Derdich et al., 1999). In terms of the fluidity of relationship status, it appears that repeated incidences of leaving and returning may transition women from being “out” to “in/out” to eventually back “in” to their relationships (Campbell et al., 1998). Because exits from and entries into relational systems are inherently stressful (Boss, 2006), a situation consisting of multiple exits and entries, as is common in the stages of contemplation through action, likely creates more potential for stress and ambiguity to occur. IPV studies support this notion. For example, in a year-long study tracking women's relationship status, women who were “in/out” reported a lower quality of life (e.g., they felt worse about their current lives and family situation) in comparison to women who were either completely ”in” or completely “out” of their relationships (Bell et al., 2007).

In short, boundary ambiguity may emerge from both abusers’ fluctuating psychological absences and women's fluidity of relationship status in contemplation through action. Within a family context, children also play an important role in facilitating women's decisions to stay/return (e.g., wanting to preserve father-child relationships) or leave permanently (e.g., wanting to keep children safe). As noted in the literature, perceptions of boundary ambiguity can block or delay the process of restructuring and clarifying family boundaries and roles (Boss, 2002). This notion is quite evident among women who seek protective orders when they leave. Specifically, women who retained protective orders often referred to their abusers as former partners (Malecha et al., 2003), whereas women who revoked or ignored their protective orders reported lingering emotional commitments and desires to reconcile with their partners (Logan, Shannon, & Walker, 2005). By using relationship terminology that indicates they are no longer partnered with abusers, the first group of women established a clear boundary separating themselves from their partners. In the latter group, women's attempts to set clear boundaries were not followed through. Because emotional ties to their partners were not clearly severed, some women may remain in a fluid “in/out” status until boundary clarity is achieved (Campbell et al., 1998).

Stage of Maintenance: Boundary Renegotiation and Intrusion

According to the Stages of Change Model, women make continued efforts to remain separated from their abusers in the maintenance stage (Brown, 1997). The process of leaving does not end when women physically leave, however; rather, “leaving for the battered woman is the continuation of a process that begins at the emotional and cognitive level while she is still in the relationship and extends well beyond her physical departure” (Anderson & Saunders, 2003, p. 179). Similar to divorced partners, abused women have to renegotiate their psychological boundaries to minimize the potential for Type I boundary ambiguity to emerge (Madden-Derdich et al., 1999). For example, attachment and perceptions of interpersonal boundaries with former partners must be adjusted in this stage. The process of boundary renegotiation may be particularly challenging when former partners must coparent (Madden-Derdich & Arditti, 1999). Studies have shown that continued coparenting relationships after separation place women at risk of experiencing ongoing control and intrusion by their former partners (Hardesty & Ganong, 2006). In this situation, Type II boundary ambiguity may emerge as former partners purposefully reassert their physical presence in women's lives.

Boundary renegotiation

Boundary renegotiation often involves women's attempts to move on with their lives after leaving an abusive relationship, reframing their identities as an individual (i.e., reclaiming self), and eventually considering or establishing new relationships with others (Wuest & Merritt-Gray, 2001). A similar process of boundary renegotiation was observed among the spouses of Alzheimer's patients in which they engaged in a process of reconstructing their identities from being a part of a couple (i.e., “we”) to being an individual (i.e., “I”; Kaplan, 2001). For abused women, several factors may influence their ability to renegotiate boundaries from a “we” to an “I.” First, any ongoing involvement with former partners for coparenting purposes affects how women renegotiate boundaries and move on with their lives (Ford-Gilboe, Wuest, & Merritt-Gray, 2005). For example, because many family courts still prioritize coparenting in order to serve the best interests of children, mothers are required to develop parenting plans and share custody with their former partners (Hardesty & Chung, 2006). Adherence to these legal pressures consequently takes precedence over mothers’ own desire to avoid further contact with their former partners (Jaffe, Lemon, & Poisson, 2003).

Second, women may want to keep abusers present in their lives even after they have separated. As the emotional workers in the family (Arendell, 1995), women are generally committed to preserving father-child relationships and thus may be willing to facilitate these relationships for their children (Wuest, Ford-Gilboe, Merritt-Gray, & Berman, 2003). Other stressors stemming from financial and emotional dependency on former partners or having to raise children on their own (Levendosky, Lynch, & Graham-Bermann, 2000) may also compel women to continue contact with their former partners in maintenance. Further, efforts to renegotiate boundaries may be hindered by existing extended family relationships (e.g., abusers’ family) that could persist after separation. For example, the divorce literature notes that grandparents tend to maintain relationships with their former in-laws and grandchildren after divorce (Ahrons & Bowman, 1982), particularly if they lived close by and if their adult child still had custody of their grandchildren (Hilton & Macari, 1997). Although not IPV specific, these studies suggest that when children are involved, a complete cutoff from the abusers’ family of origin may not be possible. In fact, we posit that some women may not want to sever ties with the abusers’ family, especially if they had played a helping role in childrearing or in their process of leaving.

Finally, a third factor that affects women's ability to renegotiate their boundaries is their psychological well-being after leaving (Anderson & Saunders, 2003). Although the common assumption is that women will fare better psychologically after leaving, studies indicate that some women may actually experience equal or higher levels of negative psychological symptoms (e.g., depression and post-traumatic stress disorder) as they grieve the loss of their relationship and encounter the challenges of life without their partners (Anderson & Saunders; Anderson, Saunders, Yoshihama, Bybee, & Sullivan, 2003). As a result, women may feel more attached and preoccupied with their former partners after leaving compared to while they were partnered (Shurman & Rodriguez, 2006). This notion is also supported in the divorce literature. For example, Madden-Derdich and Arditti (1999) found divorced spouses to be more attached to each other, especially if they coparented after divorce, if they stayed married longer, and if women perceived more difficulty in parenting alone.

Overall, the literature indicates that Type I boundary ambiguity (i.e., where abusers are physically absent but psychologically present) may emerge in the maintenance stage because women's attempts to renegotiate psychological boundaries and maintain separation are challenged when there is ongoing involvement with former partners (e.g., through coparenting). Importantly, these tasks may become increasingly difficult to accomplish if former partners respond by forcefully intruding into women's lives (Ford-Gilboe et al., 2005).


Intrusion is characterized as an “external control or interference that demands attention, diverts energy from family priorities and limits [women's] choices” (Ford-Gilboe et al., 2005, p. 482). Intrusion includes various forms of harassment and intimidation by a former partner, ranging from making menacing phone calls to stalking and making frequent unwanted visits to women's homes (Anderson & Saunders, 2003; Fleury, Sullivan, & Bybee, 2000; Hardesty & Ganong, 2006). Former partners may also use indirect forms of abuse, such as through child support, visitation, and custody issues, in order to control women and interfere with their ability to make autonomous parenting decisions (Kurz, 1996; Shalansky, Ericksen, & Henderson, 1999). Both direct and indirect forms of abuse in the maintenance stage are indicative of abusers’ tactics to reestablish a physical presence in women's lives, further creating a potential situation for Type II boundary ambiguity (i.e., where abuser is physically present but psychologically absent) to emerge. Indeed, intrusion can be harmful (Fleury et al., 2000) and detrimental to women and children's health promotion because it impedes women's efforts to rebuild a sense of security in the family and to move on with their lives (Ford-Gilboe et al.).

On the whole, our literature review suggests that various sources (from both women's and abusers’ behaviors) and types of boundary ambiguity may be present in all five stages. First, denial and ambivalence are psychological and emotional responses that may be most salient in precontemplation when women have not yet labeled their situation as abuse or are in denial. In the middle stages of change, Types I and II boundary ambiguity (i.e., when abusers’ and women's psychological and physical presences are incongruent) appear most salient as women are considering, planning, and taking the action to leave. Further, denial and ambivalence related to children may persist in these stages until women make the final decision to leave (Frasier et al., 2001). Finally, we posit that Types I and II boundary ambiguity reemerge in the maintenance stage, because women attempt to renegotiate family boundaries after leaving. Such efforts may be met with abusers’ direct and indirect intrusive behaviors that reassert their physical presence in women's lives. In the final section of this paper, we outline several implications of our literature review and propose future directions for researchers and practitioners.

Implications and Future Directions

In this paper, we proposed an expansion of the Stages of Change Model to include the theoretical construct of boundary ambiguity and articulated how boundary ambiguity maps on to the different stages of change. The inclusion of boundary ambiguity into the process of leaving addresses limitations of the original model outlined by several IPV researchers (e.g., Brown, 1997; Shurman & Rodriguez, 2006) for extending boundary ambiguity to new understudied areas in family research. Both the Stages of Change Model and boundary ambiguity have much to offer researchers and practitioners who work with women and their families affected by IPV. We posit, however, that the theoretical and practical utility of the model is greatly enhanced by taking boundary ambiguity into consideration. We discuss three areas that we believe are most fruitful for further theoretical and practical development.

First, incorporating boundary ambiguity enhances the theoretical and practical utility of the Stages of Change Model beyond individual change to understanding the stages of change within relational and systemic contexts. In the contemplation through action stages, abusers’ physical and psychological presence in the relationship fluctuates as a response to women's steps toward change. Likewise, women may fluctuate between being in and out of the relationship as they respond to their abusers' behavior. For example, women may decide to stay in their relationship when their partners are temporarily more psychologically present. The cyclical nature of the individual behaviors (i.e., moving in and out of psychological and physical presences) and the bidirectional influence each individual has on the other's behaviors during these stages are not currently represented in the individual-focused Stages of Change Model. As Anderson and Saunders (2003) noted, previous process of leaving studies, including those that utilize the Stages of Change Model, “seem to suggest that leaving an abusive relationship is primarily contingent on changes in the subjective meaning of the situation to the women” (p. 177) without much regard to the relational nature of the process of leaving. No doubt, further research is necessary to develop our theoretical understanding of the complex processes occurring within the middle stages of the model.

Because the process of leaving occurs over time, longitudinal research using both qualitative (e.g., in-depth interviews) and quantitative methods (e.g., boundary ambiguity scales) would be well suited for advancing the theory of boundary ambiguity in the process of leaving (Boss, 2007). As illustrated by Campbell et al.'s (1998) 3-year study of women's relationship statuses, longitudinal research allows IPV researchers to track relational and systemic change patterns throughout the stages of change, which may help identify possible sources of boundary ambiguity emerging over time.

Longitudinal designs would also allow researchers to identify linear and nonlinear trajectories that may emerge in the process of leaving. A developmental systems perspective may be useful in guiding this task. This perspective considers the dynamic interaction between people and their surrounding context in shaping their developmental trajectories (Lerner, 1998). In the process of leaving, both individual characteristics (e.g., women's readiness to leave) and social context (e.g., availability of social support) influence how women respond to the violence and make decisions in their relationship (Anderson & Saunders, 2003). For example, two women who experience abuse by their partners for the first time may respond very differently to the situation. Whereas one may leave immediately to the local shelter and maintain separation, the other may leave at a later time because there are no available shelters in her area. Although both women experience a similar situation, individual characteristics and social context shape how they move from one stage of change to another, resulting in different trajectories of leaving.

In a previous study (Khaw & Hardesty, 2007) we examined variations in women's trajectories of leaving. Using secondary retrospective data, we visually mapped the movements of 19 abused mothers across the Stages of Change Model. Three trajectories of leaving emerged, each of which represented a pathway formed by mothers’ change processes over time. The trajectories are depicted in Figure 1. The first trajectory is a continuous linear model, in which mothers progressed from one stage of change to the next without physically returning to their partners. These mothers had the most distinct preparation stage before leaving. The second trajectory is a discontinuous cyclical model, in which mothers progressed in a linear form until they reached action. Instead of moving into maintenance after action, however, they returned to an earlier stage of change (e.g., contemplation) before moving forward again at a later time. In this trajectory, the pattern of movement formed the cycle of “backing and forthing” (Khaw & Hardesty, p. 418) as they left and returned multiple times. The third trajectory that emerged is a discontinuous leaping model, in which mothers skipped the preparation stage and moved directly from contemplation to action. Chang et al. (2006) found a similar pattern of skipping stages, which they called “leapfrogging” (p. 333). We found that mothers who left in this trajectory made quick decisions to leave their partners without experiencing a distinct preparation stage.

Figure 1.

Leaving Trajectories in the Process of Leaving Using the Stages of Change Model.

Mapping trajectories of leaving may be a useful method for understanding how boundary ambiguity functions in the process of leaving. In the developmental literature, a continuous model (as illustrated in the first trajectory) indicates predictability and stability, whereas a discontinuous model (as illustrated in the second and third trajectories) indicates unpredictability and change in structures, purpose, or meanings (Lerner, 1986). Perhaps women who leave in a continuous linear model perceive less boundary ambiguity in their relationships, which might explain why they were able to maintain a permanent separation without returning to their partners. In contrast, boundary ambiguity may be the underlying mechanism of the cyclical movement in the discontinuous cyclical model. More research is warranted to delineate the potential role of boundary ambiguity within diverse trajectories of change.

Our second recommendation for theoretical development is for researchers to explicate the role of boundary ambiguity as a perceptual barrier in the Stages of Change Model. Although IPV studies suggest that cognitive processes of change are more prevalent in the earlier versus the later stages of change, we posit that boundary ambiguity may be a perceptual barrier that presents itself throughout the entire model. At the same time, we suggest that boundary ambiguity may serve as an initiator of change (or turning point; Khaw & Hardesty, 2007) by moving some women from precontemplation to contemplation. Specifically, when abusers’ psychological absence from their partner or parental roles elicits high levels of boundary ambiguity, women may then become more cognizant of the problems in the relationship and begin to contemplate change. Existing theoretical work has only conceptualized boundary ambiguity as a barrier to coping rather than a facilitator (Boss, 1993). Given the unique context surrounding the process of leaving, the potential role of boundary ambiguity as a facilitator of change needs further exploration.

Finally, theoretical advances are needed to better understand the process of ongoing change after women leave abusive partners. Currently, the Stages of Change Model defines maintenance as sustaining change for at least 6 months. This definition is limited, especially when applied to women who have children with their abusive former partners. Abuse and control issues often continue after separation, particularly when women share custody with abusers and are involved in ongoing court litigation (Hardesty & Ganong, 2006; Wuest et al., 2003). Further research is needed to understand how sharing custody after separation and intrusion may prevent women from renegotiating safe boundaries after leaving.

From a practical standpoint, if boundary ambiguity is an important perceptual barrier to change, professionals can help women normalize ambiguity and learn effective strategies for managing and tolerating ambiguity throughout the process of leaving. Several recommended strategies that may apply to working with abused women include labeling the problem, gathering information about the situation, and helping women to reconstruct their perceptions in order to help them regain a sense of mastery over their situations (Boss, 1999). Depending on women's readiness to change, exposure to these strategies may help women normalize, understand, and anticipate their feelings and perceptions of ambiguity in their relationships (Boss, 2006). Another point of intervention is to address challenges faced by women in renegotiating boundaries and managing intrusion in maintenance. Here, practitioners may become a source of support for women who must coparent with their former partners and help them establish ground rules in order to maintain healthy boundaries (Wuest et al., 2003).

The application of the Stages of Change Model to theorize women's process of leaving has advanced current knowledge about how and why women make decisions to leave an abusive partner. To address the unique complexities experienced in the process of leaving and to enhance the theoretical and practical utility of the Stages of Change Model, however, we integrated the construct of boundary ambiguity and explored possible sources of boundary ambiguity emerging in the process of leaving. The inclusion of boundary ambiguity in the field of IPV creates new possibilities in research to explore the notion of boundary ambiguity as a perceptual barrier for women to maintain and sustain change. Subsequently, the available resources and interventions for abused women must be adapted in order for practitioners to assist women in identifying and coping with their feelings and perceptions of boundary ambiguity.


An earlier version of this article was presented at the annual meetings of the National Council on Family Relations in Pittsburgh, Pennsylvania (November 2007).