as below

The Effects of Forgiveness Therapy on Depression, Anxiety, and Posttraumatic Stress for Women After Spousal Emotional Abuse

Gayle L. Reed and Robert D. Enright University of Wisconsin—Madison

Emotionally abused women experience negative psychological outcomes long after the abusive spousal relationship has ended. This study compares forgiveness therapy (FT) with an alternative treatment (AT; anger validation, assertiveness, interpersonal skill building) for emotionally abused women who had been permanently separated for 2 or more years (M � 5.00 years, SD � 2.61; n � 10 per group). Participants, who were matched, yoked, and randomized to treatment group, met individually with the intervener. Mean intervention time was 7.95 months (SD � 2.61). The relative efficacy of FT and AT was assessed at p � .05. Participants in FT experienced significantly greater improvement than AT participants in depression, trait anxiety, posttraumatic stress symptoms, self-esteem, forgiveness, environmental mas- tery, and finding meaning in suffering, with gains maintained at follow-up (M � 8.35 months, SD � 1.53). FT has implications for the long-term recovery of postrelationship emotionally abused women.

Keywords: spousal emotional abuse, forgiveness therapy, depression, anxiety, posttraumatic stress

Spousal emotional abuse is a significant problem, with approx- imately 35% of women reporting such abuse from a spouse or romantic partner (O’Leary, 1999); in addition, women often dem- onstrate negative psychological outcomes long after this abuse. Despite the frequent calls for efficacious therapies for these women, no empirically validated treatments have been clearly established (Enns, Campbell, & Courtois, 1997; Mancoske, Stan- difer, & Cauley, 1994; Miller, Veltkamp, & Kraus, 1997; Paul, 2004), and the literature still demonstrates a focus on the definition of and screening for spousal emotional abuse rather than empirical testing of therapeutic strategies (Follingstad, 2000; Gondolf, Heck- ert, & Kimmel, 2002; Tjaden, 2004).

Spousal psychological abuse represents a painful betrayal of trust, leading to serious negative psychological outcomes for the abused partner (Dutton & Painter, 1993; Sackett & Saunders, 1999). According to Sackett and Saunders (1999), spousal psycho- logical abuse functions with the purpose of causing emotional pain to the spouse and establishing an unequal distribution of power in the relationship. Sackett and Saunders (1999) have demonstrated negative outcomes of emotional abuse that are distinct from the impact of physical battery.

Follingstad, Rutledge, Berg, Hause, and Polek (1990) and Sack- ett and Saunders (1999) have identified at least seven categories of spousal psychological abuse: criticizing, ridiculing, jealous con- trol, purposeful ignoring, threats of abandonment, threats of harm, and damage to personal property, with ridicule associated most

strongly with negative outcomes of psychological abuse. More- over, Follingstad et al. (1990) found that 72% of participants reported that emotional abuse had a more negative impact than physical abuse. The negative psychological outcomes of spousal psychological abuse include depression (O’Leary, 1999; Pimlott- Kubiak & Cortina, 2003; Sackett & Saunders, 1999), anxiety (Dutton & Painter, 1993), posttraumatic stress disorder (Astin, Lawrence, & Foy, 1993; Enns et al., 1997; Pimlott-Kubiak & Cortina, 2003; Woods, 2000), low self-esteem (Aguilar & Night- ingale, 1994), learned helplessness (Follingstad et al., 1990; Lau- nius & Lindquist, 1988), and an ongoing, debilitating resentment of the abuser (Seagull & Seagull, 1991). A number of researchers (Astin et al., 1993; Dutton & Painter, 1993; Sackett & Saunders, 1999; Woods, 2000) have demonstrated that these negative out- comes last well beyond the end of the abusive relationship.

Considering the serious, enduring impact on the psychological health of the emotionally abused partner, the theoretical and em- pirical literature on efficacious postrelationship, postcrisis treat- ment for spousal psychological abuse is sparse. There is a lack of empirical evidence for the efficacy of one treatment that is cur- rently recommended for these women: brief therapy with a focus on anger validation (with subsequent mourning of associated losses from the abuse), assertive limit-setting, and interpersonal skill building. Neither Mancoske et al. (1994) nor Rubin (1991) provided clear empirical support for the efficacy of this brief therapy for emotionally abused women. A review of the current literature did not produce empirical evidence for the efficacy of other therapeutic approaches for emotionally abused women.

One promising new area of treatment is forgiveness therapy (FT). Research on FT has established a causal relation between forgiving an injustice and both the amelioration of anxiety and depression and an improvement in self-esteem (Al-Mabuk, En- right, & Cardis, 1995; Coyle & Enright, 1997; Freedman & En- right, 1996; Lin, Enright, Mack, Krahn, & Baskin, 2004; Rye et al., 2005). FT directly targets ongoing resentment, which can lead to

Gayle L. Reed and Robert D. Enright, Department of Educational Psychology, University of Wisconsin—Madison.

Correspondence concerning this article should be addressed to Gayle L. Reed, who is now at Educational Outreach, the University of Wisconsin Extension—Madison, 1050 University Avenue, Madison, WI 53706 – 1386, and LLC Forgiveness Recovery Programs, 1715 Carns Drive, Mad- ison, WI 53719; www.forgivenessrecovery.com. E-mail: gaylelreed@ charter.net

Journal of Consulting and Clinical Psychology Copyright 2006 by the American Psychological Association 2006, Vol. 74, No. 5, 920 –929 0022-006X/06/$12.00 DOI: 10.1037/0022-006X.74.5.920

920

depression, anxiety, and other negative psychological outcomes (Enright & Fitzgibbons, 2000), and it has been shown in one study to ameliorate the negative emotional effects of incest (Freedman & Enright, 1996). FT posits that although anger is a justifiable, initial problem-solving response to severe wrongdoing, as in the case of emotional abuse, lingering resentment can compromise a person’s emotional health and decision making.

In helping clients move toward forgiveness, clinicians need to differentiate forgiving (see Enright & Fitzgibbons, 2000, for a discussion of defining forgiveness) from condoning, excusing, pardoning, forgetting, and reconciling. Forgiveness is a decision to give up resentment and to respond with goodwill (benevolence based on a desire for the ultimate welfare of the other person; North, 1987) toward the wrongdoer. Current research (see Rye et al., 2005) operationalizes forgiveness similarly as refraining from negative responses to the wrongdoer and fostering positive re- sponses to him or her while also clearly distinguishing forgiveness from pardon and reconciliation. For women who have experienced spousal emotional abuse, FT promotes the reclamation of valued personal qualities, such as compassion, without neglecting the injustice of the abuse or encouraging interactions with the former partner, which may result in further abuse. FT assists the emotion- ally abused woman to examine the injustice of the abuse, consider forgiveness as an option, make the decision to forgive, do the hard work of forgiving (grieve the pain from the injustice, reframe the wrongdoer, relinquish resentment, and develop goodwill), find meaning in the unjust suffering, and discover psychological release and new purpose. These aspects of FT fall within the four phases of the Enright forgiveness process (Enright & Fitzgibbons, 2000)— uncovering, decision, work, and discovery— currently used in treatment and research.

Robust results have been found in randomized experiments with FT. Lin et al. (2004), in their FT experiment in a drug rehabilita- tion unit, following Hedges and Olkin (1985), reported an effect size of 1.58 across all dependent measures after adjusting for intercorrelations among variables. Similar effect sizes (1.44 and 1.42, respectively) were reported in Freedman and Enright (1996) and in Coyle and Enright (1997). In all three studies, the sample sizes were relatively small, ranging from 10 –14 participants per study. Other randomized experiments with FT have been reported (e.g., Rye et al., 2005). See Baskin and Enright (2004) for a meta-analysis of FT.

Women who have experienced spousal emotional abuse present at least two unique challenges for recovery. First, learned help- lessness (Sackett & Saunders, 1999) develops as a pattern of self-blame in response to the criticism and ridicule by the abusive spouse and often remains well beyond the end of the abusive relationship (Dutton & Painter, 1993). “If only I had done this to please him” quickly deteriorates in the ongoing, unpredictable stress of the abusive relationship to “I am trying to prevent this, but nothing is working” and remains in a residual “Maybe I am worthless and none of my decisions are valid.” Therefore, any treatment for these women should demonstrate outcomes in prac- tical decision making and moral decision making. This study tests these outcomes (environmental mastery and finding meaning in suffering) and suggests that FT ameliorates this problem more successfully than an alternative treatment (AT).

Second, Seagull and Seagull (1991) described an obstacle to recovery for emotionally abused women labeled accusatory suf-

fering, which entails maintaining resentment and victim status. The assumption in accusatory suffering is that healing the wounds of the abuse will somehow let the perpetrator off the hook. At a deeper level, accusatory suffering may be seen as a defense against the fear that the woman is somehow responsible for her own victimization, a fear that is often inculcated by the victimizer (Sackett & Saunders, 1999). Seagull and Seagull (1991) argued that although accusatory suffering (resentment and victim status) may function as a temporary strategy to help the woman adapt to the extreme experience of spousal emotional abuse, it seriously hinders substantial postrelationship, postcrisis recovery. There- fore, any treatment for these women should demonstrate a change in victim status. This study tests this outcome (story measure) and suggests that FT is more successful than AT.

One current therapeutic approach recommended and tested in the literature (Enns et al., 1997; Mancoske et al., 1994; Miller et al., 1997) for postrelationship, postcrisis emotionally abused women includes anger validation about the wrongdoing of abuse, assertiveness, and interpersonal skill building (AT). During an early period of separation from the abusive relationship (perhaps 1 to 2 years), anger validation likely helps emotionally abused women confirm the injustice of the abusive spouse’s behavior and thus provides support for the woman’s choice to escape the abu- sive relationship. Moreover, as validation of anger after such a deep, personal injustice can be an important step toward helping women uncover and mourn the pain (Mancoske et al., 1994) from this unjust injury, the uncovering phase of FT addresses this as well. However, we suggest that anger validation (even with the subsequent mourning) over time without the inclusion of work toward forgiveness may inadvertently promote the accusatory suf- fering described by Seagull and Seagull (1991) and thus contribute to the debilitating resentment that maintains and likely even in- creases the negative psychological outcomes of the abuse.

FT does have an overall, targeted focus on decreasing this resentment toward the abusing former partner, which Seagull and Seagull (1991) suggested hinders optimal recovery. Moreover, the mourning work done in FT is for the specific purpose of aiding the recovering women to successfully relinquish resentment and re- venge toward the former abuser and to develop goodwill. It is important, again, to reiterate that FT does not require nor encour- age reconciliation (a critical concern that likely prevents recom- mendation of forgiveness in recovery strategies; Herman, 1997). FT, therefore, makes a safe and distinct contribution to postrela- tionship, postcrisis therapy for emotionally abused women by promoting the practice of a specific moral quality (choosing for- giveness, relinquishing resentment, developing goodwill) as a way of integrating the past traumatic experience of emotional abuse with current positive, empowering moral choices (Astin et al., 1993; Frankl, 1969; Reed, 1998). FT thus likely effectively ame- liorates the negative psychological outcomes of emotional abuse because engagement in the forgiveness process does decrease resentment toward the former abuser (along with concomitant depression, anxiety, and low self-esteem, which are associated with emotional abuse; Sackett & Saunders, 1999) and validates the positive, moral decision to replace resentment with goodwill (thus addressing the learned helplessness associated with emotional abuse; Follingstad et al., 1990).

We therefore hypothesized that individuals who participated in FT would demonstrate less depression, anxiety, and posttraumatic

921FORGIVENESS THERAPY AFTER SPOUSAL EMOTIONAL ABUSE

stress symptoms and more self-esteem, environmental mastery, and finding meaning in suffering than those who engaged in the more standard therapeutic procedure (AT), which does not directly target the amelioration of this resentment. To have as fair a comparison as possible between FT and AT, we addressed meth- odological problems in past AT studies for both treatment condi- tions by having a single presenting problem (psychological abuse), a single living arrangement (permanent separation from the abu- sive partner), postcrisis treatment (all women had been separated at least 2 years from the former abuser), and criterion ending (rather than brief therapy at 4 – 6 weeks).

Method

Participant Sample

The participants were 20 psychologically abused women in a Midwest city who had been divorced or permanently separated for at least 2 years from their spouse or romantic partner. They ranged in age from 32 to 54 years (M � 44.95, SD � 7.01). Regarding ethnicity, 18 (90%) self-reported as European Americans, 1 (5%) was Hispanic American, and 1 (5%) was Native American. Educational levels included 4 (20%) with a high school diploma or general equivalency diploma, 6 (30%) with some college education or an associate’s degree, 4 (20%) who were college graduates, 3 (15%) who had some postgraduate education, and 3 (15%) who had postgraduate degrees. Three (15%) of the participants were unemployed, 5 (25%) had part-time employment, and 12 (60%) had full-time employment. One participant (5%) held a job in service, 7 (35%) worked in clerical jobs, 2 (10%) worked in business or sales, and 7 (35%) had professional careers. Six (30%) participants had no children living with them, whereas 14 (70%) had one to four resident children. Five (25%) of the participants had remarried, and 15 (75%) had not remarried or started a new relationship with a live-in partner. These participants were all self-selecting volunteers; 2 (10%) responded to recruitment flyers (posted in domestic abuse resource centers), and 18 (90%) responded to newspaper advertisements (for women between the ages of 25 and 55 who had experienced spousal psychological abuse but not physical abuse and who had been permanently separated for at least 2 years).

The participants reported the following psychological abuse: Eighteen participants (90%) reported criticizing, 20 participants (100%) reported ridiculing, 15 participants (75%) reported jealous control, 20 participants (100%) reported purposeful ignoring, 5 participants (25%) reported threats of abandonment, 6 participants (30%) reported threats of personal harm, and 4 participants (20%) reported threats of harm to property or pets. Six (30%) participants also disclosed experiences of sexual abuse (5 described ridicule followed by demands for sexual favors, and 1 described threats of physical harm combined with demands for sexual favors).

We set the criterion that women be 2 years postseparation to prevent promoting “false forgiveness” (e.g., “He won’t do it again”), which often occurs in the abuse cycle. Also, working on forgiveness too early in a separation might mistakenly encourage a woman to feel empathy and compassion for her abusive former partner in a way that would foster old patterns of reuniting, including inappropriate dependence on the part of the former partner and subsequent harm from further abuse. Actual time since separation ranged from 2 to 10 years (M � 5.00 years, SD � 2.61).

Design

A matched, yoked, and randomized experimental and control group design was used, with 10 pairs formed after screening interviews and pretest measures. Participants in each pair were matched as closely as possible on age, duration of the abusive relationship, and time since permanent separation or divorce. Correlations for matching variables

within pairs were duration of abuse (r � .91, p � .0001), time since permanent separation (r � .72, p � .02), and age (r � .76, p � .01). Duration of the abusive relationship ranged from 1 to 31 years (M � 16.65 years, SD � 9.01). Contact with the former partner (regarding children) ranged from no contact to more than once per week, with a moderate correlation between matched pairs of .356. Following matching, 1 partic- ipant from each pair was randomly selected for FT, and the other was assigned to AT.

Testing Procedure

Screening. Screening measures included the Psychological Abuse Sur- vey (Follingstad, 2000; Follingstad et al., 1990; Sackett & Saunders, 1999), a posttraumatic stress symptom checklist (PTSS; from the Diagnostic and Statistical Manual of Mental Disorders; 4th ed.; DSM–IV; American Psychiatric Association, 1994), and a psychological screening checklist. A participant was included in the study if she demonstrated psychological abuse in at least three categories with a score of at least four and demon- strated at least three symptoms on the PTSS checklist. A score of 41 or higher on the Psychological Abuse Survey was considered indicative of a present and serious pattern of emotional abuse. All participants reported scores of 41 or above.

A participant was excluded from the study if she demonstrated current involvement in an abusive relationship, described a history of childhood physical abuse, or demonstrated evidence of significant ongoing psychiat- ric illness, such as suicidal ideation or psychosis. We excluded persons with a history of childhood abuse to fairly focus treatment (AT or FT) on only one major wrongdoing (i.e., the spousal emotional abuse). Exclusion for suicidal ideation and psychosis was done as such women would likely be better served with crisis care or counseling with the availability of psychiatric medical treatment. Appropriate referrals were offered.

Dependent variables. Before administration of any pretest measure, all participants read and signed informed consent forms approved by the research program’s human subjects board and consistent with American Psychological Association standards. Then a brief description of the study was offered. The participants were told that the purpose of the study was to promote coping strategies for women who had experienced emotional abuse and that it included weekly 1-hr individualized therapy sessions. Each participant was then invited to fill out the nine pretest measures.

Instruments

All screening and dependent measures were presented to each partici- pant in random order at pretest, posttest, and follow-up. All measures were given after signed informed consent.

Psychological Abuse Survey. This questionnaire (an adaptation from Follingstad, 2000; Follingstad et al., 1990; Sackett and Saunders, 1999) asks, “How often did your partner ____?” with seven categories of abuse: (a) criticizing behavior (e.g., “You don’t do anything well enough”), (b) ridiculing of traits (e.g., “You are worthless”), (c) jealous control (e.g., “You can’t maintain any outside social support”), (d) purposeful ignoring (e.g., “You don’t exist”), (e) threats of abandonment, (f) threats of harm, and (g) threats to damage personal property. Frequency of each abuse category was scored on a Likert scale ranging from daily (8) to never (1). Total scores for participants ranged from 41 to 106 (M � 70.63, SD � 17.58). For the purposes of this study, a total score of 41 or above is considered a high level of abuse (Dutton & Painter, 1993; Sackett & Saunders, 1999).

The Enright Forgiveness Inventory (EFI; Subkoviak et al., 1995). The EFI is a 60-item self-report measure of the degree of interpersonal for- giveness, equally divided in six components: Positive and Negative Affect (e.g., “I feel ____ toward him/her”), Positive and Negative Behavior (e.g., “Regarding the person who hurt me, I do or would ____”), and Positive and Negative Cognition (e.g., “I think she or he is ____”). Range is from 60 to

922 REED AND ENRIGHT

360, with high scores representing high levels of forgiveness. Subkoviak et al. (1995) reported an alpha coefficient of .98. Cronbach’s alpha for this study at pretest (N � 20) was .94.

Coopersmith Self-Esteem Inventory (CSEI; Coopersmith, 1989). The adult form of the CSEI consists of 25 true–false statements (e.g., “This is like me or not like me”) evaluating attitudes toward the self in the following domains: general self, social self, self and peers, and self and parents. Range of scores is 0 (low score) to 100 (high score). Reliability and validity for this scale have been well documented (Coopersmith, 1989). The Cronbach’s alpha for pretest scores (N � 20) for this study was .84.

State–Trait Anxiety Inventory (STAI; Spielberger, 1983). The STAI, a common instrument in clinical work, is composed of two self-report questionnaires (20 items each) that assess state (e.g., “Right now at this moment I feel ____”) and trait (e.g., “Generally I feel ____”) anxiety. The range of scores for each questionnaire is 20 (low anxiety) to 80 (high anxiety). Cronbach’s alphas for this study at pretest (N � 20) were .95 (state) and .92 (trait).

Beck Depression Inventory–II (BDI-II; Beck, Steer, & Brown, 1996). The BDI-II is a self-report measure of 21 items in a multiple choice format (four choices, ranging from “I do feel ____” to “I do not feel ____”). Scores can range from 0 (no depression) to 63 (high depression). This instrument has been used extensively in clinical research and demonstrates construct validity and reliability. Freedman and Enright (1996) reported a reliability of .84 for incest survivors. The Cronbach’s alpha for pretest scores in this study (N � 20) was .90.

Environmental Mastery Scale (Ryff & Singer, 1996). The Environmen- tal Mastery Scale is one of six Scales of Psychological Well-Being (Ryff & Singer, 1996), which represent distinct concepts that are evaluated separately. It has 14 items rated from 1 (strongly disagree) to 6 (strongly agree)—for example, “I am the kind of person that ____” or “I am not the kind of person that ____.” The items contain questions about personal mastery in everyday decisions (high score indicates environmental mas- tery). The scale yields adequate reliability (Ryff & Singer, 1996). The Cronbach’s alpha at pretest for this study (N � 20) was .80.

Reed (1998) Finding Meaning in Suffering. This instrument assesses the participant’s engagement in Viktor Frankl’s (1969) concept of finding meaning in suffering. Thus, items are questions about moral decisions in response to unjust suffering and questions about support (identifying with persons and values) for these moral decisions (Astin et al., 1993; Frankl, 1969; Reed, 1998) Each of 22 items is rated from 1 (not true) to 5 (very true)—for example, “I find this experience ____” or “I see this experience as ____”—producing scores of 22 to 110. Internal consistency is high (Reed, 1998). The Cronbach’s alpha at pretest for this study (N � 20) was .88.

PTSS checklist. This checklist was derived directly from the DSM–IV criteria checklist. The lead question reads, “Are the events (of psycholog- ical abuse) re-experienced in one or more of these ways?” The categories include (a) recurring and intrusive memories, (b) distressing dreams, (c) intense distress on reminders of the abusive events, (d) the avoidance or denial of emotional responses to the abusive events, (e) anxious feelings and thoughts, (f) trouble sleeping, and (g) difficulty concentrating. Partic- ipants were to answer yes (1 point) or no (0 points) to each category if the symptom had occurred in the last month. This DSM–IV symptom checklist was used not only for screening and measurement purposes but also to clearly validate to participants in both treatment groups that the treatment focus was on the effects of a past traumatic relationship (rather than on what is wrong with women who enter and remain in abusive relationships). This was done to promote a sense of interpersonal safety (Herman, 1997) and rapport with the intervener.

Story measure. This measure is a one-page narrative from the partic- ipant’s current perspective about the role that spousal psychological abuse has in her life story. Two raters who were blind as to the identity of the participant rated a score for both the old story (victim status) and the new

story (survivor status). One point was given for the following categories if they were present: for the old story, focuses on the power of abuser; describes self as victim of abuse; describes abuse events but no decisions; abuse memories are resentful, repetitive, or intrusive; for the new story, focuses on her power to choose, puts abuse in context of other life events, describes abuse review as impetus for new decisions, contrasts memories of abuse to ongoing personal growth. The interrater agreement on the story scores was 76.25%.

Intervention Procedure

Following pretesting, the experimental participants engaged in 1-hr, weekly FT sessions based on the Enright forgiveness process model (Enright & Fitzgibbons, 2000). AT participants engaged in 1-hr, weekly participant-initiated discussion of current life concerns (considering the impact of the past abuse) and in intervener-facilitated therapeutic discus- sions about the validity of anger regarding the injustice of the past abuse (and subsequent mourning of the associated pain), present strategies for healthy assertive choices, and interpersonal relationship skills. The inter- vener facilitated therapeutic intervener–participant interactions, with restat- ing, paraphrasing, summarizing, and open-ended questions (Pulvino & Lee, 1995), for both treatment conditions (FT and AT). Participants in FT determined the time spent on each forgiveness topic, and participants in AT determined the time spent on each participant-initiated concern. Each matched pair was equal in time of treatment. The mean treatment time (one session per week) for the pairs was 7.95 months (SD � 2.61), with a minimum of 5 months and a maximum of 12 months.

The FT treatment had a manualized protocol to promote uniformity in the treatment across participants. This protocol was a guide for the treat- ment sessions, which proceeded in a similar therapeutic manner as for the AT group (Pulvino & Lee, 1995). The protocol included (a) defining forgiveness (what it is and is not) and the distinction between forgiveness and reconciliation, (b) examining psychological defenses, (c) understand- ing anger, (d) examining abuser-inculcated shame and self-blame, (e) understanding cognitive rehearsal, (f) making a commitment to the work of forgiving, (g) grieving the pain and losses from the abuse, (h) reframing the former abusive partner (his personal history, fallibility, and culpability; the unfair, unequal power established by his abusive behavior; his inherent worth), (i) exploring empathy and compassion, (j) practicing goodwill (i.e., merciful restraint, or foregoing resentment or revenge; generosity; and moral love), (j) finding meaning in unjust suffering, and (k) considering a new purpose in life of helping others. FT was criterion-based, finishing when each participant reported that she had completed the work of forgiv- ing her former partner.

The AT was designed and delivered (with a written protocol) to match as closely as possible the basic elements of the therapy approach (anger validation with mourning, assertiveness strategies, and interpersonal skills) described and tested in the literature (Enns et al., 1997; Mancoske et al., 1994; Miller et al., 1997). Therefore, for comparison purposes, this was considered a treatment protocol intended to reflect real-world practice (Wampold & Serlin, 2000) and provided a therapeutic intervention (AT) for the control participants during the paired experimental participants’ FT. The current life concerns introduced by the control participants (which included the past abuse, child rearing, child placement interactions with the former partner, family, and work relationships) became part of a facilitated discussion within the parameters of the AT therapeutic approach.

Research Design

We compared FT and AT gain scores from pretest to posttest on all dependent variables with matched-pair t tests. We analyzed FT partici- pants’ maintenance of gains by comparing each participant’s pretest to posttest gain scores with her gains scores from pretest to follow-up.

923FORGIVENESS THERAPY AFTER SPOUSAL EMOTIONAL ABUSE

Qualifications of the Intervener

The intervener had extensive education in the forgiveness process model, experience in delivering FT workshops, training and practice as a psychiatric nurse, and previous experience working with women who had been psychologically abused.

Analysis

We determined sample size by first examining the effect size for for- giveness gains in similar studies. Freedman and Enright (1996) demon- strated a 2.09 Cohen’s effect size for gain in forgiveness scores, which was associated with significant positive psychological outcomes in depression and anxiety (N � 12). Coyle and Enright (1997) reported a Cohen’s effect size of 1.20 for gains in forgiveness scores, which were associated with significant improvements in anger, anxiety, and grief (N � 10). Thus, a minimum projected effect size of 1.20 for forgiveness scores in combina- tion with a sample size of 24 (12 pairs) would yield a power of .80, an acceptable possibility of Type II error. The recruited sample was 20 (10 pairs). The experimental group (FT) demonstrated a within-group Cohen’s effect size of 1.79 for gains in forgiveness scores.

Each participant, treated individually in therapy, was independent of all others, which allowed for the individual rather than the group to be the level of analysis. On the basis of earlier studies (Coyle & Enright, 1997; Freedman & Enright, 1996; Lin et al., 2004), we used one-tailed matched- pairs t tests at an alpha level of .05 to compare the amount of change between the two treatment groups.

Results

Mean scores and standard deviations are reported in Table 1 for all measures at pretest, posttest, and follow-up for the FT group

and at pretest and posttest for the AT group. The gains from pretest to posttest between FT and AT are presented in Table 2. The maintenance of gains at follow-up for FT are presented in Table 3.

We performed one-tailed matched-pairs t tests to compare the amount of change between the two groups from pretest to posttest. Statistical significance was demonstrated on all the dependent variables, with the exception of state anxiety (see Table 2). FT participants demonstrated a statistically significantly greater in- crease in forgiving the former abusive partner, t(9) � 5.80, p � .001; in self-esteem, t(9) � 2.12, p � .05; in environmental mastery (everyday decisions), t(9) � 1.84, p � .05; in finding meaning in suffering (moral decisions), t(9) � 2.34, p � .05; and in new stories (survivor status), t(9) � 3.58, p � .01. The exper- imental group demonstrated a statistically significantly greater reduction in trait anxiety, t(9) � �2.43, p � .05; in depression, t(9) � �1.88, p � .05; in posttraumatic stress symptoms, t(9) � �2.54, p � .05; and in old stories (victim status), t(9) � �5.01, p � .001. There was within-group statistical significance (FT from pretest to posttest) for improvements in state anxiety scores, t(9) � �2.22, p � .05.

We note that if pretest means are compared in a randomized design, 5% of the time, the means will differ. Whether they differ significantly, or whether the difference in standard deviations is small or large, does not alter the fact that the participants were randomized to treatment group. The significance test on gain scores is based on a distribution that, as with all statistical tests, will yield significant differences 5% of the time. This distribution, under the null hypothesis, takes into account those times that yield pretest differences and those times that do not. Therefore, if one

Table 1 Mean and Standard Deviations for Dependent Variables

Variable

Pretest Posttest Follow-up

M SD M SD M SD

Experimental forgiveness therapy group

Forgiveness 155.40 38.24 252.50 55.54 256.50 60.24 Self-esteem 66.00 17.20 82.60 8.54 90.00 4.32 State anxiety 42.50 13.61 31.90 8.99 25.10 4.84 Trait anxiety 45.10 9.65 32.60 8.95 26.40 5.12 Depression 16.20 8.81 5.40 5.77 2.60 4.52 Environmental mastery 56.65 8.19 65.65 10.51 69.60 8.82 Finding meaning 85.70 12.05 100.00 9.52 99.00 11.62 PTSS 9.40 3.41 2.20 2.29 2.40 1.10 Old story (victim) 3.90 0.32 1.20 1.47 0.60 0.84 New story (survivor) 0.30 0.48 2.60 1.71 3.70 0.95

Alternative therapy group

Forgiveness 171.00 48.59 172.90 42.47 Self-esteem 43.20 21.23 49.60 24.31 State anxiety 48.70 14.63 44.80 16.09 Trait anxiety 53.40 11.54 52.40 12.63 Depression 23.40 10.62 21.70 13.68 Environmental mastery 47.40 11.99 48.80 8.68 Finding meaning 74.80 16.11 76.50 12.61 PTSS 11.00 2.87 8.50 3.06 Old story (victim) 3.70 0.67 3.70 0.48 New story (survivor) 0.30 0.67 0.40 0.69

Note. n � 10 per group. PTSS � posttraumatic stress symptoms.

924 REED AND ENRIGHT

has a difference in pretest scores between treatment groups (0.75 standard deviations for depression in this study, e.g.), one cannot conclude that in the population the difference is other than zero (Marascuilo & Serlin, 1988).

A comparison of two sets of gains scores, pretest to posttest versus pretest to follow-up, in the FT group for the EFI, the BDI-II, the Ryff Environmental Mastery Scale, the Reed Finding Meaning in Suffering Measure, the PTSS checklist, and the story measure demonstrated no significant differences. This suggests mainte- nance of gains for the FT group to follow-up. The comparison of gain scores of the experimental group for the CSEI and the STAI State and Trait scales from pretest to posttest versus pretest to follow-up demonstrated significant further changes. This sur- passed the expected outcomes, and these data suggest a continu- ation of gain for the FT group from postFT to follow-up.

Case Study

Marianne (name changed to protect confidentiality), age 38, a participant in the FT treatment group, was married for 6 years and

had been permanently separated from her abusive spouse for 10 years. Marianne described the past psychological abuse as criti- cizing, ridiculing, jealous control (clothing, makeup, telephone use), purposeful ignoring, and threats of harm. In addition, at pretest, Marianne demonstrated negative outcomes of spousal emotional abuse, including anxiety, depression, low self-esteem, difficulty in decision making, and posttraumatic stress symptoms.

Marianne actively engaged in all aspects of FT. She examined her past shame and self-blame and grieved the pain from the undeserved abuse. She moved from hate for her former partner to a genuine desire for his welfare (without excusing him or reuniting with him) as she enthusiastically relinquished resentment and thus cognitive rehearsal of the past abuse. She also found a remarkable increase in energy for her current life (new marriage, career, child rearing) and a surprising new sense of joy.

At posttest, Marianne demonstrated considerable improve- ments: from above the published norms for trait anxiety scores (60) to below the norm (25), from moderate depression (25) to no depression (0), from below normative self-esteem to high self-

Table 2 Comparison of Mean Change From Pretest to Posttest

Variables

Forgiveness therapy group gain score (n � 10)

Alternative therapy group gain score (n � 10)

t(9)M SD M SD

Forgiveness 97.10 54.05 1.90 28.87 5.80*** Self-esteem 16.60 20.39 5.80 13.35 2.12* State anxiety �10.20 15.06 �3.90 9.87 �1.08 Trait anxiety �12.00 14.65 �1.00 5.75 �2.43* Depression �10.80 11.67 �1.70 9.36 �1.88* Environmental mastery 9.00 12.29 1.40 8.67 1.84* Finding meaning 14.30 8.46 1.70 13.02 2.34* PTSS �7.20 3.82 �2.40 4.62 �2.54* Old story (victim) �2.70 1.56 0.00 0.67 �5.01*** New story (survivor) 2.50 1.65 0.50 1.35 3.58**

Note. PTSS � posttraumatic stress symptoms. * p � .05. ** p � .01. *** p � .001, one-tailed.

Table 3 Comparison of Mean Changes: Treatment and Maintenance at Follow-Up

Variable

Forgiveness therapy group

t(9)

Pretreatment to posttreatment gain

score (n � 10)

Pretreatment to follow-up gain score

(n � 10)

M SD M SD

Forgiveness 97.10 54.04 101.10 65.43 0.39 Self-esteem 16.60 20.39 24.00 17.68 2.43* State anxiety �10.20 15.06 �17.40 12.20 �2.59* Trait anxiety �12.00 14.65 �19.00 10.95 �3.13* Depression �10.80 11.67 �13.60 6.64 �1.25 Environmental mastery 9.00 12.29 12.30 12.21 1.05 Finding meaning 14.30 8.46 13.30 6.49 0.59 PTSS �7.20 3.82 �4.40 6.68 1.76 Old story (victim) �2.70 1.56 �3.30 0.95 �2.25 New story (survivor) 2.50 1.65 3.40 0.96 1.86

Note. PTSS � posttraumatic stress symptoms. * p � .05.

925FORGIVENESS THERAPY AFTER SPOUSAL EMOTIONAL ABUSE

esteem, from below normative scores on environmental mastery to above normal scores, from a low score on finding meaning in suffering to a score 1.85 standard deviations above the pretest mean, from frequent cognitive rehearsal of the past abuse (three times per week) to no cognitive rehearsal, and from nine posttrau- matic stress symptoms to one. Marianne maintained these im- provements at follow-up.

Discussion

Studies such as Astin et al. (1993), Dutton and Painter (1993), Paul (2004), Sackett and Saunders (1999), and others have dem- onstrated the significant negative impact of spousal psychological abuse, including low self-esteem, anxiety, depression, learned helplessness, and posttraumatic stress symptoms. Seagull and Seagull (1991) suggested that in postrelationship, postcrisis emo- tionally abused women, these outcomes may be exacerbated by accusatory suffering, which includes a debilitating resentment and victim status. This is the first study to demonstrate that FT is efficacious as a therapeutic strategy for the amelioration of these long-term negative psychological outcomes of spousal psycholog- ical abuse. Moreover, this study demonstrates that FT promotes improvements in psychological health to a significantly greater extent than an AT recommended in the literature for emotionally abused women—that is, a focus on anger validation, assertive limit setting, and interpersonal skills (Enns et al., 1997; Miller et al., 1997; Paul, 2004). This study attempted to control a number of potential confounds, such as comparable treatment lengths for the two conditions, presentation of psychological abuse only (without reported physical abuse), and complete separation from the former spouse for at least 2 years at the time of the therapy.

At pretest, all participants were low in forgiveness toward the abusive former spouse (see Table 1) and well below the published mean for nonclinical samples (Subkoviak et al., 1995). The mean for the FT group at posttest was comparable to the norms for nonabused adult populations reported in Subkoviak et al. (1995), whereas the control group was considerably below that. The gains of the FT group from pretest to posttest and from pretest to follow-up are similar to those reported in Coyle and Enright (1997). The Cohen’s effect size of 1.79 for this FT within-group change is a large effect size (Kirk, 1995) and represents a shift from below normative levels of forgiveness to normative levels that are frequently associated with gains in psychological health (Coyle & Enright, 1997; Lin et al., 2004).

Both FT and AT participants presented at pretest with BDI-II mean scores in the mild to moderate range of depression (Beck et al., 1996). The FT group demonstrated clinically significant im- provement at posttest by shifting into the minimal to nondepressed range (Beck et al., 1996). The within-group Cohen’s effect size of 0.93 is a large effect by Kirk’s (1995) criteria. In contrast, the AT group remained in the moderate range of depression. The FT group change to minimally nondepressed was sustained at follow-up.

State and trait anxiety mean scores for both FT and AT partic- ipants at pretest were well above normative mean scores reported by Spielberger (1983) and approaching psychiatric levels. The reduction in trait anxiety for the FT group was significant (with a Cohen’s effect size of 0.88) and significantly greater than that for the AT group. This within-group change for the FT group repre- sents a shift below published norms and a large effect size (Kirk,

1995). The FT group shifted from substantially above the pub- lished norms for state anxiety to below, whereas the AT group remained above the published average. The FT group maintained improvements in both state and trait anxiety and demonstrated further improvements at follow-up.

The FT and AT participants both presented with high scores on the PTSS checklist at pretest. The FT group demonstrated signif- icantly better improvement at posttest than the control group (AT). FT participants went from a mean of nine symptoms to a mean of two, whereas the mean for AT participants remained at approxi- mately eight.

In self-esteem, both FT and AT groups presented at pretest with mean CSEI scores lower than those reported by Coopersmith (1989) for adult women. The gains of the experimental group were not only maintained at follow-up; further improvement was dem- onstrated. A Cohen’s effect size of 0.81 (for within-group FT change, compared with a 0.45 effect size for the control group) represents a shift from below normative levels of self-esteem to above normative levels. This is a similar improvement to that in Freedman and Enright (1996) and Lin et al. (2004).

FT and AT participants at pretest presented below normative (Ryff, 2002) mean scores for environmental mastery (everyday decision making). Those in the FT group moved to normative levels of environmental mastery (Ryff, 2002) and demonstrated statistically significant improvement at posttest, with a within- group Cohen’s effect size of 0.73, and statistically greater im- provement than AT. This improvement in decision making is uniquely important for women who have experienced spousal emotional abuse (Sackett & Saunders, 1999), as a partial amelio- ration of learned helplessness.

Both FT and AT participants at pretest presented low scores for finding meaning in suffering that were similar to scores reported for untreated adults (Reed, 1998). FT participants improved sta- tistically significantly (with a within-group Cohen’s effect size of 1.69) and in comparison with AT participants. The gains were maintained at follow-up testing. Finding meaning in unjust suffer- ing (Frankl, 1969) entails moral decision making, which counter- acts learned helplessness (Sackett & Saunders, 1999) and promotes recovery from trauma (Astin et al., 1993).

FT and AT participants presented at pretest with high scores for their old story (resentful retelling and victim status) and low scores for their new story (survivor status). FT participants improved statistically significantly compared with AT participants. This demonstrates an important shift from accusatory suffering, as described by Seagull and Seagull (1991).

The findings demonstrate the benefit of FT for women who have experienced spousal psychological abuse (and who have been permanently separated from the abusive partner for at least 2 years). The gains made by the FT group compared with the AT group suggest that FT was more efficacious in reducing anxiety, depression, and posttraumatic stress symptoms for these women. The effect sizes ( p � .05) in the between-treatments analysis were robust: 1.83 for forgiveness, 0.68 for self-esteem, 0.77 for trait anxiety, 0.58 for environmental mastery, 0.59 for depression, and 0.74 for finding meaning in suffering. In their meta-analysis with comparisons between psychotherapeutic treatments, Wampold et al. (1997) found effect sizes for dependent outcomes of 0.20 but, after adjusting for such problems as discontinuity between treat- ments in therapeutic intention, time span of treatment, and partic-

926 REED AND ENRIGHT

ipant’s presenting problem, concluded that true effect sizes for treatment comparisons across studies washed out to zero. The comparison in the present study between FT and AT was made with one presenting problem (spousal psychological abuse) in both treatment groups. Both treatments were designed to be therapeutic, were matched for time span of treatment, retained the same trained intervener, and used matched pairs. Thus, this study addresses concerns that led to a lowering of treatment comparison effect sizes in Wampold et al. (1997). We note that having the same intervener in both treatments in itself is a positive factor (Messer & Wampold, 2002), and all therapeutic gains should be considered to include both the treatment and the therapist’s relationship with the participants.

Psychologically abused women may have difficulty recovering long after the termination of the abusive relationship because of continuing resentment about the unfair harm of the abuse. Thera- pies that support the expression of anger about the wrongdoing of abuse (and subsequent mourning of the pain), assertiveness, and reconnection with improved interpersonal skills may not be ade- quate to address this debilitating resentment. FT, by contrast, although it supports the client in appropriately expressing anger about the abusive relationship and grieving the pain from the abuse, specifically targets the debilitating resentment toward the former abusive partner. The FT client is encouraged to tell her own unique story of the abuse experience, with the purpose of working through this story to a healthy resolution that includes forgiveness. During the forgiveness process, the client does the hard work of uncovering anger and shame, grieving the undeserved pain from the abuse, and reframing the former partner (personal history, fallibility, and culpability, yet inherent human worth), with the purpose of relinquishing debilitating resentment. Most important, FT can then take the paradoxical route of focusing the client’s thoughts, feelings, and behavior on a benevolent response to that former spouse rather than retaining the debilitating resentment. Thus, FT ameliorates the negative outcomes of psychological abuse.

Another therapeutic factor in FT may be the acknowledgement that the client herself is a person of worth. In the process of therapy, the client makes the effort to integrate the reality of the abusive spouse’s wrongdoing and his inherent human worth. This does not negate the reality of the wrongdoing of the abuse itself but rather establishes that what the abuser did, however hurtful and unfair, does not change his worth as a human being (see Table 3 of Enright & the Human Development Study Group, 1994, for an explication of unconditional human worth undiminished by wrongdoing). As the client sees her abuser’s inherent worth, she then may be able to understand that she is also a person of inherent worth. The abuser’s mistaken view of her (as a worthless person who can be mistreated at will) does not have to be her view. This rediscovery of inherent human worth can be enhanced in that the client recognizes that she is a person of courage because she is choosing to relinquish resentment and develop goodwill for the former abusive spouse (in contradistinction to his abusive choices). It seems to us that clients benefit psychologically because of the interaction between understanding inherent human worth in others and in the self and relinquishing resentment (which, if left un- treated, can prolong depression, anxiety, learned helplessness, shame, victim status, and low self-esteem). At this point, the client may see that her courageous response to unfair suffering (i.e.,

forgiveness) is making her emotionally and ethically stronger. Thus, finding meaning in unjust suffering (Frankl, 1969) may strengthen the resolve to continue practicing forgiveness and thus also experiencing further psychological improvement.

Another therapeutic aspect of FT that follows from finding meaning in unjust suffering is finding new purpose in helping others who are in pain or experiencing injustice (Enns et al., 1997; Seagull & Seagull, 1991). FT has a distinct advantage in this aspect of recovery because, after forgiveness, engagement in per- sonal and social causes can proceed with a positive energy that is no longer confused by lingering, debilitating resentment. More- over, this engagement can be more clearly focused on relevant social justice issues, untainted by any subtle motivation for re- venge. FT participants at the end of this study planned to work toward change in conditions for other women in divorce law, social services, and disability benefits. Further study could be done after FT to ascertain progress in and psychological benefits from these pursuits.

The robust findings of this study suggest that forgiveness can have a general effect on emotional regulation, reducing anxiety and depression while also increasing self-esteem and healthy (practical and moral) decision making. Previous studies have sug- gested that such variables are difficult to ameliorate (Sackett & Saunders, 1999; Seagull & Seagull, 1991). Perhaps unsuccessful emotional regulation after spousal emotional abuse is a set of symptoms for what is at the heart of the matter, deep and continu- ing resentment that has no apparent resolution. FT provides such a resolution in a safe environment that allows the abused client to move at her own pace in confronting the injustice. Moreover, FT encourages goodwill toward the former abuser with careful dis- cretion. It is important to note that FT distinguishes carefully between the goodwill of forgiveness and reconciliation. For ex- ample, a woman may have goodwill for her former abusive partner (positive thoughts and feelings) but tell the truth about the situation (the former partner still has a problem); she may be benevolent (e.g., “I hope he obtains effective professional treatment”) yet maintain safety (no direct engagement with the former partner). Thus, FT promotes recovery benefits for the woman while avoid- ing the risk of further harm (therefore, it can also be appropriate for women with a past history of spousal physical or sexual abuse; Freedman & Enright, 1996).

Finally, FT is appropriate as a postrelationship, postcrisis recov- ery treatment for emotionally abused women. FT is best offered after safety issues have been addressed (safety from the former partner and by reconnecting with social support), after some initial time for uncovering of the trauma (Herman, 1997), and as an aspect of long-term recovery, which entails empowering moral choices and integration of these healthy choices with memories of the past abuse.

The research design includes strengths such as individualized therapy, moving at the client’s own pace and terminating therapy at the discretion of the participant, the use of an alternative treatment advocated in the published literature, manualized treat- ment, and careful screening for psychological abuse. Limitations include the sample size and some salient aspects of the participant sample. First, an all-volunteer, self-selecting participant sample may contribute particular characteristics (greater ongoing prob- lems after the abuse, e.g.). Second, the ethnic mix of the sample was quite homogeneous (90% European Americans). Replication

927FORGIVENESS THERAPY AFTER SPOUSAL EMOTIONAL ABUSE

is, therefore, recommended to ascertain generalizability. In addi- tion, because of the exclusion of potential participants who exhib- ited evidence of significant psychiatric illness, this study does not provide information on the benefits of FT for these particular women. Moreover, the study design of one intervener for all participants (although a strength of the study regarding continuity across treatment groups) contributes a limitation in that the out- comes of the study (for both treatment groups) may be due in part to an intervener effect (Messer & Wampold, 2002). A dismantled study design that compares FT and AT delivered with an inter- vener and FT and AT delivered with manuals only (Wampold et al., 1997) might further clarify the contributions of the treatments and the intervener relationship. Finally, a replication with only one or two salient dependent measures, with participants matched on these measures before being randomized to treatment groups and with an analysis of covariance, might lower the possibility of pretest mean differences between groups and thus further elucidate the relative benefits of FT and AT.

This study is the first empirical investigation of the benefits of FT for women who have experienced spousal emotional abuse. As the benefits of FT were tested against one therapy currently spe- cifically recommended for emotionally abused women (AT: anger validation with subsequent mourning, assertiveness, and interper- sonal relationship skills), it would now be helpful to test the efficacy of FT with respect to other recovery strategies for trau- matic relationships, such as Herman’s (1997) three phases of safety, mourning, and reconnection. What would be the relative efficacy of FT, Herman’s (1997) three phases, and the integration of these two approaches? The FT research should also deepen its focus by further study on the contributions of specific aspects of FT to the subsequent psychological improvements for emotionally abused women and other populations who have experienced un- just, traumatic relationships. What are the specific contributions of reframing, practicing goodwill, or finding meaning in unjust suf- fering? Future FT research could follow up participants who benefited from FT for one traumatic relationship (spousal abuse) to determine whether the practice and benefits transferred to another such relationship (e.g., an alcoholic parent). FT research should also explore how the benefits of FT transfer from the forgiver through emotional regulation and goodwill to a larger circle of personal relationships beyond the wrongdoer.

FT holds promise as a postrelationship, postcrisis therapeutic approach for women who have experienced spousal emotional abuse, as it provides relief from negative psychological outcomes and fosters the positive characteristics of courage, competence, and altruism. These women have been given the message in the abuse context that they are worthless and that they cannot make good choices. FT corrects that message.

References

Aguilar, R. J., & Nightingale, N. N. (1994). The impact of specific battering experiences on the self-esteem of abused women. Journal of Family Violence, 9, 35– 45.

Al-Mabuk, R. H., Enright, R. D., & Cardis, P. A. (1995). Forgiveness with parentally love-deprived college students. Journal of Moral Education, 24, 427– 444.

American Psychiatric Association. (1994). Diagnostic and statistical man- ual of mental disorders (4th ed.). Washington, DC: Author.

Astin, M. C., Lawrence, K. J., & Foy, D. W. (1993). Post-traumatic stress

disorder among battered women: Risk and resiliency factors. Violence and Victims, 8(1), 17–28.

Baskin, T. W., & Enright, R. D. (2004). Intervention studies on forgive- ness: A meta-analysis. Journal of Counseling and Development, 82, 79 –90.

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck Depression Inventory II manual. San Antonio, TX: Psychological Corporation.

Coopersmith, S. (1989). Self-Esteem Inventories. Palo Alto, CA: Consult- ing Psychologists Press.

Coyle, C. T., & Enright, R. D. (1997). Forgiveness intervention with postabortion men. Journal of Consulting and Clinical Psychology, 65, 1042–1046.

Dutton, D. G., & Painter, S. (1993). The battered woman syndrome: Effects of severity and intermittency of abuse. American Journal of Orthopsy- chiatry, 63, 614 – 621.

Enns, C. Z., Campbell, J., & Courtois, C. A. (1997). Recommendations for working with domestic violence survivors, with special attention to memory issues and post-traumatic processes. Psychotherapy, 34, 459 – 477.

Enright, R. D., & Fitzgibbons, R. P. (2000). Helping clients forgive: An empirical guide for resolving anger and restoring hope. Washington, DC: American Psychological Association.

Enright, R. D., & the Human Development Study Group. (1994). Piaget on the moral development of forgiveness: Identity or reciprocity? Human Development, 47, 63– 80.

Follingstad, D. R. (2000). Defining psychological abuse of husbands toward wives: Contexts, behaviors, and typologies. Journal of Interper- sonal Violence, 15, 891–911.

Follingstad, D., Rutledge, L., Berg, B., Hause, E., & Polek, D. (1990). The role of emotional abuse in physically abusive relationships. Journal of Family Violence, 5, 107–119.

Frankl, V. (1969). The will to meaning: Foundations and applications of logotherapy. New York: World Publishing House.

Freedman, S. R., & Enright, R. D. (1996). Forgiveness as an intervention goal with incest survivors. Journal of Consulting and Clinical Psychol- ogy, 64, 983–992.

Gondolf, E. W., Heckert, D. A., & Kimmel, C. M. (2002). Nonphysical abuse among batterer program participants. Journal of Family Violence, 17, 293–314.

Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. San Diego, CA: Academic Press.

Herman, J. (1997). Trauma and recovery: The aftermath of violence— From domestic abuse to political terror. New York: Basic Books

Kirk, R. E. (1995). Experimental design: Procedures for the behavioral sciences. Pacific Grove, CA: Brooks/Cole.

Launius, M. H., & Lindquist, C. U. (1988). Learned helplessness, external locus of control, and passivity in battered women. Journal of Interper- sonal Violence, 3, 307–318.

Lin, W. F., Enright, R. D., Mack, D., Krahn, D., & Baskin, T. (2004). Effects of forgiveness therapy on anger, mood, and vulnerability to substance use among inpatient substance-dependent clients. Journal of Consulting and Clinical Psychology, 72, 1114 –1121.

Mancoske, R. J., Standifer, D., & Cauley, C. (1994). The effectiveness of brief counseling services for battered women. Research on Social Work Practice, 4(1), 53– 63.

Marascuilo, L. A., & Serlin, R. C. (1988). Statistical methods for the social and behavioral sciences. New York: Freeman.

Messer, S. B., & Wampold, B. E. (2002). Let’s face facts: Common factors are more potent than specific therapy ingredients. Clinical Psychology: Science and Practice, 9, 21–25.

Miller, T. W., Veltkamp, L. J., & Kraus, R. F. (1997). Clinical pathways for diagnosing and treating victims of domestic violence. Psychother- apy, 34, 425– 432.

North, J. (1987). Wrongdoing and forgiveness. Philosophy, 62, 499 –508.

928 REED AND ENRIGHT

O’Leary, D. K. (1999). Psychological abuse: A variable deserving critical attention in domestic violence. Violence and Victims, 14(1), 3–23.

Paul, M. (2004). Clinical implications in healing from domestic violence: A case study. American Psychologist, 59, 809 – 816.

Pimlott-Kubiak, S., & Cortina, L. M. (2003). Gender, victimization, and outcomes: Reconceptualizing risk. Journal of Consulting and Clinical Psychology, 71, 528 –539.

Pulvino, C. J., & Lee, J. L. (1995). A practical guide to interviewing. Madison, WI: Instructional Enterprises.

Reed, G. (1998). Forgiveness as a function of moral agency in the context of divorce and infidelity. Unpublished master’s thesis, University of Wisconsin—Madison.

Rubin, A. (1991). The effectiveness of outreach counseling and support groups for battered women: A preliminary evaluation. Research on Social Work Practice, 1, 332–357.

Rye, M. S., Pargament, K. I., Pan, W., Yingling, D. W., Shogren, K. A., & Ito, M. (2005). Can group interventions facilitate forgiveness of an ex-spouse? A randomized clinical trial. Journal of Consulting and Clin- ical Psychology, 73, 880 – 892.

Ryff, C. D. (2002). Life Histories and Health in Midlife Study: Normative data. Madison: University of Wisconsin.

Ryff, C. D., & Singer, B. (1996). Psychological well-being, meaning and measurement, and implications for psychotherapy research. Psychother- apy and Psychosomatics, 65, 14 –23.

Sackett, L. A., & Saunders, D. G. (1999). The impact of different forms of psychological abuse on battered women. Violence and Victims, 14(1), 105–117.

Seagull, E. A., & Seagull, A. A. (1991). Healing the wound that must not heal: Psychotherapy with survivors of domestic violence. Psychother- apy, 28, 16 –20.

Spielberger, C. D. (1983). State–Trait Anxiety Inventory (Form Y). Red- wood City, CA: Mind Garden.

Subkoviak, M. J., Enright, R. D., Wu, C. R., Gassin, E. A., Freedman, S., Olson, L. M., & Sarinopoulos, S. (1995). Measuring interpersonal for- giveness in late adolescence and middle adulthood. Journal of Adoles- cence, 18, 641– 655.

Tjaden, P. (2004). What is violence against women? Defining and mea- suring the problem. Journal of Interpersonal Violence, 19, 1244 –1252.

Wampold, B. E., Mondin, G. E., Moody, M., Stich, F., Benson, K., & Ahn, H. (1997). A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, “all must have prizes.” Psychological Bulletin, 122, 203–215.

Wampold, B. E., & Serlin, R. C. (2000). The consequence of ignoring a nested factor on measures of effect size in analysis of variance. Psycho- logical Methods, 5, 425– 433.

Woods, S. J. (2000). Prevalence and patterns of posttraumatic stress disorder in abused and postabused women. Issues in Mental Health Nursing, 21, 309 –324.

Received May 19, 2005 Revision received May 31, 2006

Accepted June 16, 2006 �

929FORGIVENESS THERAPY AFTER SPOUSAL EMOTIONAL ABUSE