Shame, posttraumatic stress disorder, and intimate partner violence perpetration

Abstract

Drawing from theoretical and empirical literature linking shame to aggression and violence, the authors propose that shame may be an important variable to examine in studies of posttraumatic stress disorder (PTSD) and intimate partner violence (IPV) perpetration. The authors review the literature linking shame, PTSD, and IPV, propose ideas for future research, and suggest that shame and shame regulation may be a useful target of clinical interventions aimed at violence perpetrators.

Introduction

Each year in the United States, an estimated two million women are physically assaulted by intimate partners (Tjaden & Thoennes, 2000). Despite growing recognition of intimate partner violence (IPV) as a public health problem, our conceptual understanding of it remains limited (Bell & Naugle, 2008), and interventions aimed at reducing IPV have produced disappointing results (Babcock, Green, & Robie, 2004). Researchers have called for the elucidation of risk factors for IPV and the development of theory-driven interventions (Bell and Naugle, 2008, Whitaker et al., 2006).

Although a literature linking shame and aggression dates back to the 1970s (e.g., 1; Scheff, 1988), shame has received little attention from IPV researchers. Below we review the theoretical basis and empirical evidence for the link between shame and violence perpetration. We discuss the importance of studying the role of shame and shame regulation in IPV, particularly among individuals with posttraumatic stress disorder (PTSD), review the limitations of the research base, and propose ideas for future work.

Psychologists, philosophers, and anthropologists have attempted to define shame and differentiate it from closely related emotions (e.g., guilt) (Hutchinson, 2008, Lebra, 1971, Sabini and Silver, 1997). There is a considerable overlap among definitions of shame, with most authors conceptualizing it as an emotion characterized by global, negative evaluation of the self (e.g., Tangney, Stuewig, & Mashek, 2007). Consistent with the theories emphasizing the dimensional nature of emotions and psychological disorders (e.g., Brown & Barlow, 2009), Retzinger (1991) has proposed that shame exists along a continuum, ranging from low-intensity, transient self-consciousness to intense, long-lasting, and pathological humiliation. Likewise, (Scheff, 1988) differentiated between the normal, universal experience of periodic shame and “toxic” shame, which may result from repeated instances of humiliation, ridicule, or social exclusion. Although both shame and guilt may occur in response to negative self-referent events, guilt involves a focal concern about a particular behavior or failure (Lewis, 1971, Tangney et al., 1996).

Despite their conceptual overlap, shame and guilt appear to be associated with distinct action tendencies. Whereas guilt is correlated with empathy and reparative actions (e.g., confessing, apologizing) (de Hooge et al., 2007, Tangney, 1994), shame is associated with attempts to deny or escape from shame-inducing situations (Tangney & Dearing, 2002). Most relevant to the study of IPV, attempts to escape from shame may be accompanied by aggression or violence (Stuewig et al., 2010, Thomaes et al., 2008, Thomaes et al., 2011).

The link between shame and aggression has long been recognized by clinicians. Lewis (1971) coined the term “humiliated fury” to describe the expression of rage following the experience of shame. Drawing from his clinical experience with violent offenders, Gilligan, 1997, Gilligan, 2003 theorized that acts of violence are maladaptive attempts to escape shame, and associated feelings of weakness, inferiority, and worthlessness. He argues that shame is likely to lead to violence in individuals who do not feel sufficient levels of guilt and remorse, whose experience of shame is so strong that it threatens their sense of self, and who lack nonviolent means of restoring their self-esteem. Empirical work has provided some support for Lewis (1971) and Gilligan, 1997, Gilligan, 2003 clinical observations, with correlational and quasi-experimental studies demonstrating a link between shame and destructive expressions of anger (Andrews et al., 2000, Harper and Arias, 2004, Tangney and Dearing, 2002).

In most studies linking shame and aggression, shame has been conceptualized as a trait variable, shame-proneness. These studies have demonstrated that shame-proneness is significantly correlated with anger arousal, malevolent intentions, and maladaptive responses to anger (e.g., speaking maliciously behind another person's back, displacing aggression onto someone or something that is not the target of the anger) (Tangney et al., 1996). Guilt-proneness, in contrast, is associated with constructive responses to anger, corrective actions, and nonhostile discussions with the target of the anger (Tangney et al., 2007).

In other studies, shame has been operationalized as the frequency of shaming experiences over a particular timeframe. In a study of aggression among adolescents, Åslund, Starrin, Leppert, and Nilsson (2009) found that the frequency of shaming experiences (e.g., being ridiculed, having one's dignity insulted) over the previous three months was positively associated with verbal and physical aggressions. In fact, girls who reported more shaming experiences were four times more likely to have perpetrated physical aggression than girls who reported a low frequency of shaming experiences. Consistent with Åslund et al.'s (2009) findings, Thomaes et al. (2011) observed that children were viewed by peers as angrier on days during which they had experienced a shame-inducing event (e.g., having a secret revealed, being disregarded by others).

Taken together, the studies mentioned above provide support for clinical and theoretical accounts of the link between shame and aggression. Below we discuss why the relationship between shame and aggression may be particularly relevant to the study of PTSD.

There is an increasing recognition among researchers and clinicians that conceptualizations of PTSD as a fear-based disorder may not adequately describe the experience of trauma survivors. Cognitive models of PTSD posit that inadequate processing of traumatic events may lead to “threats to the self”, or shame-inducing views of oneself as an inadequate or unacceptable person. Accordingly, cognitive processing therapy (CPT; Resick & Schnicke, 1992) addresses threats to self-worth that may occur following exposure to trauma and maladaptive beliefs (e.g., “I am basically damaged or flawed”) that are consistent with the global, negative, self-evaluations that Tangney and Dearing (2002) and Lewis (1971) associate with shame. Recently, Litz et al. (2009) have argued that war veterans with PTSD may experience “moral injury” and consequent shame as a result of perpetrating, failing to prevent, or witnessing acts that violate their moral beliefs.

Empirical studies support the notion that trauma exposure is associated with higher levels of shame. Children who have been abused and neglected report higher levels of shame than children who have not been mistreated (Bennett, Sullivan, & Lewis, 2005). In a study of former prisoners of war, Leskela, Dieperink, and Thuras (2002) found that shame-proneness was positively associated with PTSD symptom severity, whereas the inverse relationship was observed with guilt-proneness. One prospective study demonstrated that shame was the only independent predictor of PTSD symptom severity in violent crime victims six months after initial assessment (Andrews et al., 2000).

Thus, the evidence suggests that individuals with PTSD are at increased risk for experiencing pathological levels of shame. A growing body of research has also demonstrated that PTSD is predictive of both relationship dissatisfaction and physical aggression perpetration (Carroll et al., 1985, Orcutt et al., 2003, Taft et al., 2011). Findings linking PTSD with pathological levels of shame, as well as with relationship dissatisfaction and aggression perpetration, highlight the importance of examining shame and IPV in individuals with PTSD.

Despite the theoretical and empirical works linking shame, PTSD, and aggression, shame has been understudied in the context of IPV. Much of the research on negative emotions as risk factors for IPV has focused on anger and hostility, though the relationship between these emotions and IPV has not been consistently borne out in the literature. In a meta-analysis of studies examining anger, hostility, and perpetration of IPV by men, Norlander and Eckhardt (2005) concluded that, relative to nonviolent men, IPV perpetrators report higher levels of anger and hostility across a variety of measurement approaches. However, the functional relationship between anger/hostility and violence perpetration is less clear, with little data to support the hypothesis that acute anger arousal is a risk factor for discrete episodes of violence.

Of note, one task that has commonly been used to examine the relationship between anger and IPV perpetration is the Articulated Thoughts in Simulated Situations paradigm (ATSS; Davison, Robins, & Johnson, 1983). As an alternative to questionnaire-based assessment, this paradigm requires participants to listen to audiotaped scenarios, imagine that the situations are actually occurring, and articulate their thoughts and feelings during a 30-second pause. These responses are transcribed and coded along with relevant dimensions (e.g., hostile attributional biases, anger-control strategies). Although the ATSS appears to be an effective anger induction procedure (Eckhardt, Barbour, & Davison, 1998), the scenarios, which frequently depict infidelity or betrayal, also appear likely to evoke shame. It is possible that shame may be an important but unmeasured variable in studies examining the association between anger and IPV.

The few studies that have explicitly examined the relationship between shame and IPV have indicated that shame is positively associated with the perpetration of both physical and psychological aggressions in dating relationships (Dutton et al., 1995, Harper, 2005). More recently, Sippel and Marshall (2011), using a modified Stroop task, found that men and women with PTSD exhibited faster color naming of shame-related words than did individuals without PTSD. Furthermore, the relationship between PTSD and IPV perpetration was mediated by facilitated color naming of shame-related words. These results suggest that PTSD may be associated with shame processing biases, and that these biases may serve as a risk factor or partner violence. In the one published study examining shame, PTSD, and IPV perpetration among military veterans, Hundt and Holohan (2012) found that shame, guilt, depression, and PTSD were all significantly associated with the perpetration of IPV. Shame, conceptualized as a trait variable, was the most important of these variables in discriminating between perpetrators and nonperpetrators. Furthermore, shame mediated the relationship between PTSD and IPV (although this relationship was not significant when depression was included in the model).

The studies described above did not examine the functional relationship between shame and episodes of aggression and violence. Thus, there is limited evidence of the role of shame and shame-proneness as antecedents to IPV. Many IPV episodes, however, occur in the context of prototypically shame-inducing events (e.g., arguments regarding jealousy, infidelity, or betrayal) (Fenton & Rathus, 2010), suggesting that, consistent with theories advanced by Lewis (1971) and Gilligan, 1997, Gilligan, 2003, a functional relationship may exist between the experience of shame and the perpetration of IPV.