“My Partner Will Change”: Cognitive Distortion in Battered Women in Bolivia

Abstract

This study examines the role of cognitive distortion in women’s decision to stay with or leave their violent partner in a sample of Bolivian women. Our study is based on a consistency model: Cognitive distortion is assumed to play an important role in maintaining cognitive consistency under threatening conditions. Eighty victims of partner violence aged 18 to 62 years who sought help in a legal institution were longitudinally assessed three times over a time period of 6 months. Measures were taken from previous studies and culturally adapted through qualitative interviews. Nearly half of the participants decreased their intention to leave the violent partner in the time span of 1 month between the first and second interview. Women who had decreased their leaving intention had concurrently increased their cognitive distortion: They blamed their partner less, were more convinced that they could stop the violence themselves, and were more likely to believe that their partner would change. Cognitive distortion was not observed among women who remained stable in their intention to leave. Women whose intention of leaving decreased and who displayed more cognitive distortion after 1 month were more likely to live with the violent partner 6 months later than women whose leaving intention remained stable or increased. Socio-demographic variables were not related to cognitive distortion or stay–leave decisions in this study. We conclude that cognitive distortion plays a role for women’s decision to stay, enhancing their risk of re-victimization.

Keywords

partner violence stay–leave decision cognitive distortion Bolivia

Introduction

Worldwide, the prevalence of intimate partner violence is high. According to a meta-analysis including 141 studies in 81 countries (Devries, Mak, García-Moreno, et al., 2013), in 2010, 30% (95% confidence interval = [27.8, 32.2]) of women aged 15 years and above had experienced physical and/or sexual violence in their intimate partner relationship. Aside from the global prevalence estimate, this meta-analysis also showed considerable regional variance in the prevalence of physical and sexual partner violence. The Andean countries were among the ones with the highest prevalence (41%), after sub-Saharan Africa (65%) and South Asia (42%). Similar rates were found in the World Health Organization (WHO) multi-country study on women’s health and domestic violence against women (García-Moreno, Jansen, Ellsberg, Heise, & Watts, 2006), where the lifetime prevalence for severe physical violence among ever-partnered women was highest in provincial Peru with 49%. In Bolivia, data from the 2008 census indicated that 47% of women had experienced some form of partner violence (43% psychological violence, 25% physical violence, and 7% sexual violence; Meekers, Pallin, & Hutchinson, 2013).

Partner violence has severe consequences for victims’ mental and physical health (Campbell, 2002; Ellsberg, Jansen, Heise, Watts, & García-Moreno, 2008). Intimate partner homicide is among the leading causes of death among women worldwide (Stöckl et al., 2013). Moreover, partner violence is associated with depression and increased suicide rates (Devries, Mak, Bacchus, et al., 2013; Devries et al., 2011). The negative effect of partner violence on women’s mental health was found in Bolivia, as well (Meekers, Pallin, & Hutchinson, 2014). Risk for repeated abuse and re-victimization is generally high: Goodman, Dutton, Vankos, and Weinfurt (2005); Krause, Kaltman, Goodman, and Dutton (2008); and Diop-Sidibé, Campbell, and Becker (2006) found that approximately 38% of participants in their studies had suffered repeated abuse. On a larger scale, the WHO multi-country study on women’s health and domestic violence against women (García-Moreno et al., 2006) showed that “most acts of physical partner violence were part of a pattern of continuing abuse” (p. 1265). In summary, prevalence of partner violence in Andean countries is high, and robust evidence from all over the world shows devastating health consequences of violence for victims. Aside from epidemiological studies on prevalence, risk factors, and health consequences of partner violence, little evidence exists from Bolivia or other Andean countries on battered women’s cognitions related to partner violence, and on their decision to stay with or leave the violent partner.

The present study focuses on battered women’s cognitions and is based on Grawe’s consistency theory (Grawe, 2004). According to this model, humans strive for the realization of motivational goals (approach and avoidance goals; Elliot, 2006). All intra- and interpersonal psychological processes (perceptions, cognitions, emotions, and behavior) serve those approach and avoidance goals. The concept of consistency refers to the compatibility of simultaneously occurring mental processes and particularly to the congruence between motivational goals and perceptions. Consistency is the highest systemic demand. Inconsistency produces negative emotions such as tension and anxiety (called inconsistency-tension). For this reason, various mechanisms exist to re-establish consistency in case of conflicts between mental processes (Gawronski, 2012) or between motivational goals and perceptions (Bélanger & Kruglanski, 2014).

Cognitions play an important role in regularizing consistency in the mental system. Grawe (2004) assigns three different functions to cognitions: (a) representation of real situations or facts, (b) interpretation of facts in view of approach and avoidance goals, and (c) guidance of behavior. Bélanger and Kruglanski (2014) recently showed that in the presence of a biasing motivation (e.g., the motivation for social recognition), cognitions may enable the formation of a desired judgment. Once behavior is based on such biased, motivation-driven judgments, it may become counterintuitive, or even contrary, to what may seem beneficial for a person (Hill & Durante, 2011).

This conceptual framework can be used to explain mental health and stay–leave decisions in battered women: Partner violence goes against women’s motivational goals, thus, enhancing inconsistency-tension. Two options exist for reducing inconsistency-tension: (a) leaving the violent partner or (b) modifying cognitions (e.g., causal attributions, expectations) in a way that they are consistent with the experience of partner violence. A huge amount of literature exists on cognitive distortion in battered women, its effect on mental health and on re-victimization. One group of authors looked at battered women’s coping strategies (Bauman, Haaga, & Dutton, 2008; Calvete, Corral, & Estévez, 2007; Krause et al., 2008; Lilly & Graham-Bermann, 2010; Solveig & Bjorkly, 2014). A difference exists between problem-focused and emotion-focused coping strategies (Lazarus & Folkman, 1984). Among the emotion-focused strategies are denial and avoidance, as well as positive thinking. Those strategies correspond to the concept of cognitive distortion, as they result in reduction of inconsistency-tension. Bauman and colleagues (2008) showed that the vast majority of a sample of battered women had used emotion-focused strategies to cope with partner violence, but had not perceived these strategies as helpful. Other authors showed that emotion-focused coping strategies were associated with higher levels of depression and posttraumatic stress disorder (Calvete et al., 2007; Krause et al., 2008; Lilly & Graham-Bermann, 2010). Moreover, Iverson and colleagues (2013) showed that emotion-focused coping strategies increased victims’ risk of re-victimization, whereas problem-focused strategies, such as seeking help or confronting the aggressor, decreased this risk.

Another group of researchers directly addressed battered women’s cognitions and their associations with negative emotions and stay–leave decisions (Hendy, Eggen, Gustitus, McLeod, & Ng, 2003; Herbert, Silver, & Ellard, 1991; Pape & Arias, 1995, 2000). Pape and Arias (1995) showed that victim’s perception of control over violence reduced their negative emotions, such as anxiety and tension. The authors argue that overestimating control over violence enhances victims’ risk of future abuse, as they do not adequately appraise their situation and do not seek help. Pape and Arias (2000) showed that causal attribution was related to negative emotions: Women who attributed the causes of violence to the partner expressed more negative emotions than women who did not consider their partner as responsible for the abuse. Battered women also tend to overestimate positive aspects of their relationship (Herbert et al., 1991) and maintain hope that the situation will change (Hendy et al., 2003). These cognitions are related to the stay–leave decisions: Women who maintain negative, partner-blaming attributions, as opposed to exculpatory or self-blaming attributions, more often leave the violent partner (Pape & Arias, 2000; Rhatigan, Street, & Axsom, 2006). Positive appraisal of the partnership and the hope that things will change affect victims’ intentions to stay with the violent partner (Byrne & Arias, 2004; Hendy et al., 2003; Herbert et al., 1991).

Cognitive distortion may change over time. In the “cycle theory of violence” by Walker (2009), during and shortly after a violent episode, victims’ optimistic beliefs about the controllability of the violence or the partner changing his behavior become shattered. However, during the phase of reconciliation—the “honeymoon phase”—victims rebuild their perceptions and cognitions about the partner and the relationship. Most of studies on battered women’s cognitions rely on cross-sectional designs and omit the fact that cognitions may change over time (Solveig & Bjorkly, 2012). Few studies exist that adopted a longitudinal design (Bell, Goodman, & Dutton, 2007; Katerndahl, Burge, Ferrer, Becho, & Wood, 2014a, 2014b). When assessing cognitive distortion, adopting a longitudinal design is essential, because victims’ retrospective answers on cognitive strategies may be shaped by the current state within the cycle of violence for reasons of cognitive consistency. Finally, studies on stay–leave decisions of battered women were conducted in western-industrialized countries so far. To the best of our knowledge, no studies exist on the applicability of the existing evidence on stay–leave decisions in a different socio-cultural context.

The current study aims at investigating (a) cognitive distortion in battered women in Bolivia, (b) the effect of cognitive distortion on inconsistency-tension, and (c) the association between cognitive distortion and victims’ stay–leave decisions. Considering the fact that partner violence is strongly related to socio-demographic factors such as poverty, women’s education, and employment (e.g., Akmatov, Mikolajczyk, Labeeb, Dhaher, & Khan, 2008; Heise & Kotsadam, 2015; Kimuna & Djamba, 2008; Meekers et al., 2013), we included these variables in our study.

Method

Sample and Procedure

Participating women were recruited from the Legal Office for Women in Cochabamba, Bolivia, a non-profit women’s rights organization based in Cochabamba that offers legal and psychological aid to victims of violence. All women who sought help at the Legal Office during this 12-month period (about 500) were invited to participate in the study—134 women ultimately agreed to participate. Interviews were conducted with the women within 1 week of their seeking help at the Legal Office (T1). One month later, participating women were asked to answer the same questions (T2). To keep drop-out rates low, detailed contact information was gathered from all women, and the interviewers visited participants at home when no phone number was available. Of the initial 134 women, 101 could be located and expressed their willingness to participate at the second interview. Six months after the first meeting (T3), participants were contacted to obtain their relationship status and prevalence and frequency of partner violence between T2 and T3. Eighty women participated at the third interview, with a total drop-out rate of 40%. Qualitative interviews were conducted prior to the study to assess the relevance of the selected variables in the socio-cultural context of Bolivia and to develop the items on outcome expectations. After developing and compiling the questionnaire, a pre-test was conducted including 21 women. The questionnaire was modified after this test run. Modifications included small changes to wording and deletion of a duplicate item.

At T1, participants were informed about the aims of the study and were assured that their responses would be treated confidentially; this information was included in a consent form that each of them signed. Because many had a low-education level or were even illiterate, the items of the questionnaire were read to all women. Participants answered the questions with the help of a wooden board that illustrated the scales with either a colored triangle (illustrating the Likert-type scale from low to high) or sad smilies (for negative values) and happy smilies (for positive values). Participants moved a wooden figure along these illustrations. With the help of a transparent stencil laid over these illustrations, interviewers were able to read off the correspondent value of the Likert-type scale. Duration of the interviews was between 1.5 and 3 hr, depending on the additional comments participants made or the time they needed to understand the questions. Five local graduate students in psychology were trained as interviewers. During a 2-day training, they interviewed each other under the supervision of the study coordinator. The first 40 interviews were digitally recorded, and 10 (two per interviewer) were randomly chosen for quality checks. All statistics were done in SPSS 16.

Measures

Socio-demographics

Socio-demographic variables including age, education level, employment, and household income were assessed at T1. Household income was assigned to three categories: from 0 to 999 Pesos Bolivianos (= US$145), from 1,000 to 1,999 Pesos Bolivianos, and 2,000 or more Pesos Bolivianos.

Extent of partner violence

Violence was assessed using the Abusive Behaviour Observation Checklist (Dutton, 1992). From the original scale, 28 items were selected and translated. Seven items assessed physical violence, one item assessed sexual violence, and 20 items dealt with different aspects of psychological violence. At T1, participants responded to the items on a 5-point scale, indicating the frequency with which their current partner had engaged in each behavior during the past 12 months. At T2 and T3, participants were asked about the frequency with which their partner had engaged in each behavior since the previous interview. Internal consistency for this scale was .86 (Cronbach’s α).

Intention to leave/to stay

Two items asked participants about their intention to leave their violent partner, and single item asked them about their intentions to continue their relationship. The scale ranged from 1 = not at all to 4 = absolutely. Responses about the intention to continue the relationship were reverse scored. Internal consistency of this brief scale was .9.

Relationship status

Relationship status (together vs. separated) was measured 6 months after the baseline assessment.

Inconsistency-tension

Seven items were presented to measure negative emotions (anger, impatience, nervousness, fear, tension, sadness, and generally feeling affected by the violence) during the month prior to the interview. The response scale ranged from 1 = not at all to 4 = a lot. Internal consistency of this scale was .63.

Cognitive distortion

Five subscales were used to measure cognitive distortion: Seven items concerned partner attribution. This scale included items on the causality, responsibility, malicious intentions, and stability of violent behavior (Pape & Arias, 2000), with lower values representing more cognitive distortion. Internal consistency for this scale was .76. Perceived control included four items regarding the perceived control over the partner’s violent behavior (Pape & Arias, 1995). Internal consistency for this scale was .73. A third scale included 14 items on relationship appraisal (Herbert et al., 1991). Examples included mutual love, trust, support, and understanding in the relationship. Internal consistency for this scale was .88. Ultimately, a scale for outcome expectations was developed. Twenty possible outcomes associated with staying in the violent relationship were presented, both positive (e.g., living in a stable relationship) and negative (e.g., continuing to experience partner violence). Items were developed for this study following a literature review (Byrne & Arias, 2004; Hendy et al., 2003; Rhatigan et al., 2006) and qualitative interviews with battered women. Participants were asked to indicate the likelihood that staying with the violent partner would result in each of these 20 different outcomes, using a 7-point scale ranging from −3 = extremely unlikely to 3 = extremely likely. Participants were then asked to evaluate how good or bad each of the 20 possible outcomes would be for them personally, using a 7-point scale ranging from −3 = extremely bad to 3 = extremely good. Participants’ evaluation of each of the consequences was multiplied by the likelihood that staying with the violent partner would result in this consequence (ranging from −9 to 9). Higher scores reflected more positive outcome expectations. Internal consistency for this scale was .91.

Results

Participants had an age range from 18 to 62 years. Some 29% had no formal education, 19% had completed primary school, 16% had completed secondary school, 21% had a technical education, and 15% had a university degree. Altogether, 37% were unemployed, 30% were employed, and 32% worked in the informal sector. For the total household income, 63% indicated they had less than 1,000 Pesos Bolivianos (= US$145) a month. Only 15% earned 2,000 Pesos Bolivianos or more, meaning that the middle and higher social strata were hardly represented in this sample. Participants who dropped out after the first interview were more likely to have a low level of education: In the drop-out group, only 18% had a technical education or a university degree, compared with 41% in the remainders. This difference was statistically significant (p < .05, Pearson’s χ²). Participants in the drop-out group did not differ from the others in terms of age, income, or employment.

First, a group assignment was conducted by comparing the intentions to leave the violent partner at T1 and T2. Change in intention was assumed if the mean difference between T1 and T2 was .33 or more (difference of 1 point in at least one of the three questions). Participants were assigned to three different groups: (a) honeymoon group, in which intention to leave had decreased; (b) stable intention group; and (c) increased intention group. For almost half of the women (50/101), their intention to leave the violent partner decreased (M difference = −1.16; SD = 0.73). These women were assigned to the “honeymoon group.” Twenty-five participants remained constant in their intention to leave (M difference = 0.0; SD = 0.0) and were assigned to the “stable intention group.” And for 26 participants, their intention to leave had increased (M difference = 0.77; SD = 0.47).

Changes regarding cognitive distortion (i.e., partner attribution, perceived control, relationship appraisal, and outcome expectations) between T1 and T2 were analyzed for each subgroup using Wilcoxon Test for related samples. As shown in Table 1, alterations in cognitive variables were observed in the honeymoon group, but not in the other groups. Women in the honeymoon group had significantly decreased their partner attributions for violence (Z = −5.2; p < .001), perceived more control over the violence (Z = −4.5; p < .001), appraised their relationship more positively (Z = −3.2; p < .001), and showed significantly more positive outcome expectations (Z = −5.4; p < .001). Their expectations for the future relationship with the partner had even changed from a negative value at T1 to a positive value 1 month later. Moreover, these women showed significantly less inconsistency-tension (Z = −5.8; p < .001). Participants in the stable group showed less inconsistency-tension at T2 than at T1 (Z = −3.9; p < .001). All other differences between variables at T1 and T2 were non-significant for the stable intention and the increased intention group. Table 1 further shows means and standard deviations for participants who had dropped out of the study after the first interview. Mean differences between the drop-out group and the ones who remained in the study resulted non-significant (Mann–Whitney Test for independent samples). The number of participants who dropped out after the second interview was highest in the stable intention group (28%), followed by the honeymoon group (20%), and the increased intention group (15%). This frequency difference was not statistically significant (Pearson’s χ² Test).

Table 1.

Mean Differences Between T1 and T2.

	Honeymoon (n = 50)			Stable Intention (n = 25)			Increased Intention (n = 26)			Drop Out^a (n = 33)
	T1	T2	p	T1	T2	p	T1	T2	p	T1
	M (SD)	M (SD)		M (SD)	M (SD)		M (SD)	M (SD)		M (SD)
Intention to leave	3.41 (0.72)	2.26 (0.81)	<.001	3.72 (0.62)	3.72 (0.62)	—	2.41 (0.88)	3.18 (0.80)	<.001	3.18 (0.92)
Partner-attribution	3.27 (0.61)	2.71 (0.77)	<.001	3.42 (0.60)	3.47 (0.64)	.46	3.11 (0.69)	2.95 (0.74)	.21	3.28 (0.64)
Perceived control	2.01 (0.80)	2.66 (0.78)	<.001	1.52 (0.67)	1.73 (0.66)	.08	2.16 (0.89)	2.21 (0.85)	.65	1.89 (0.84)
Relationship appraisal	2.03 (0.54)	2.35 (0.71)	<.001	1.71 (0.53)	1.58 (0.62)	.14	2.07 (0.62)	1.92 (0.78)	.10	2.00 (0.82)
Outcome expectations	−1.97 (3.27)	1.50 (3.28)	<.001	−2.71 (3.73)	−1.73 (3.67)	.15	−0.14 (3.47)	−0.10 (3.35)	.83	−2.97 (3.79)
Inconsistency-tension	4.40 (0.51)	3.05 (1.10)	<.001	4.36 (0.53)	3.40 (1.15)	<.001	4.15 (0.64)	3.83 (0.94)	.18	4.49 (0.50)

Drop out after first interview.

Table 2 shows the partial correlations between cognitive variables and inconsistency-tension, after controlling for the frequency of physical, psychological, and sexual violence. No significant correlations emerged at T1. However, at T2, a higher level of perceived control over violence, more positive outcome expectations, and a lower level of partner attribution were related to a lower level of inconsistency-tension (p < .05).

Table 2.

Partial Correlation Between Positive Illusions and Negative Emotions After Controlling for Frequency of Physical, Psychological, and Sexual Violence (N = 101).

	Inconsistency-Tension
	T1	T2
Partner-attribution	0.18	0.24*
Perceived control	0.05	−0.22*
Relationship appraisal	−0.12	−0.17
Outcome expectations	−0.03	−0.24*

p < .05.

The percentage of women living apart from their partner was calculated for each group at T1, T2, and T3, and for the drop-out groups (drop out after first and second interview). Results are displayed in Table 3. At T1, difference between groups regarding the percentage of women living apart was not significant. At the subsequent interviews, percentage of women living apart from their partner had increased in the stable group. By contrast, the percentage of women living apart from their partner had decreased in the honeymoon group, with only 13% of women living separated from their partner at T3. Interestingly, proportions of separated and non-separated women remained stable in the increased intention group. Chi-square testing revealed non-significant group differences at T2, but significant group differences at T3. Binary logistic regression analysis showed that the intention to leave at T1 did not predict the relationship status at T3. However, intention to leave at T2 predicted the relationship status at T3, −2LL (log likelihood) = 85.43, χ²(1) = 14.69, p < .001, with 77% correctly classified cases.

Table 3.

Frequencies (%) of Participants Living Apart From Their Violent Partner.

	Honeymoon Group (n = 40)	Stable Intention Group (n = 18)	Increased Intention Group (n = 22)	Drop Out After Second Interview (n = 21)	Drop Out After First Interview (n = 33)	χ² Testing (Row)
T1	13 (33%)	7 (39%)	8 (36%)	6 (29%)	14 (41%)	1.18
T2	10 (25%)	9 (50%)	6 (27%)	9 (43%)		4.69
T3	5 (13%)	11 (61%)	8 (36%)	—		14.92***

***

p < .001.

Finally, differences between the honeymoon group, the stable intention and the increased intention group at T1 with regard to the frequency of partner violence and socio-demographic variables were calculated using Mann−Whitney U-Test and Pearson’s chi-square test. The three groups did not differ regarding the indicated frequency of partner violence at T1 (at category or item level), nor regarding demographics (i.e., age, education level, employment, and household income). Moreover, the number of women who had suffered partner violence between T2 and T3 was compared across groups using Pearson’s chi-square test. Statistically non-significant group differences emerged regarding the number of women who had experienced violence between T2 and T3. In the honeymoon group, 43% had suffered physical violence, 93% psychological violence, and 28% sexual violence between T2 and T3. We further compared women who lived with or separate from the partner at T3 with regard to the frequency of partner violence between T2 and T3: These frequencies were higher among women who lived with their partner (41% physical, 95% psychological, and 32% sexual violence) than among those who lived separate (28%, 68%, and 8%, respectively). The Pearson’s chi-square testing for group differences resulted non-significant (Pearson’s χ² = 1.17 for physical, 5.0 for psychological, and 0.86 for sexual violence, respectively).

Discussion

This study indicated that cognitive distortion may play an important role in women’s decision to stay with or leave a violent partner. In this sample of Bolivian women, almost half of the participants decreased their intention to leave the violent partner during the time span of 1 month after the initial interview. Decrease in intention to leave went along with an increase in cognitive distortion: After 1 month, they blamed their partner to a lesser extent, were more convinced that they could stop their partner’s violent behavior, and were more inclined to believe that their partner would change his behavior in the future than at the first interview. Such considerable changes did not emerge among victims who had remained stable in or had even increased their intention to leave. The study also indicated that women who had increased their cognitive distortion during the month after the first interview were more likely to live with the partner after 6 months than women who had maintained their negative cognitions over time. We considered this group of women as “the honeymoon group,” based on the cycle theory of violence by Walker (2009).

Battered women’s cognitions have been discussed extensively in literature. However, to the best of our knowledge, cognitive distortion and its role in reducing cognitive dissonance have not been used to explain battered women’s decision to stay with a violent partner. Our longitudinal design enabled these findings: A cross-sectional design (e.g., at the first or second time of measurement) would not have depicted the change in cognitions and—alongside—intentions. Such longitudinal studies are still scarce, although evidence shows that leaving a violent partner is a dynamic process that is, at best, investigated over time (Bell et al., 2007, 2009).

In our study, the initial intention of the battered women to leave their violent partners was irrelevant to the decision of living together or apart after 6 months—but the intention at mid-term (1 month after the first interview) predicted this decision. Our interpretation of this finding is that at the moment of seeking help in an institution, participants passed through the stage of disillusion and accordingly high intention to leave. But over time, some of them changed their appraisal of the situation to a more positive outlook and lesser intention to leave. Others remained stable, and those were more likely to eventually leave their violent partner. Thus, the decision to stay with or leave the violent partner was consolidated during the first month after study entry, and cognitive distortion was part of this decisional process. Moreover, we found that cognitive distortion was inversely associated with inconsistency-tension. This result stays in line with consistency theory, according to which cognitive distortion contributes to maintaining the emotional balance in the psychic system. As a positive aspect, it may enable victims to cope with the traumatic situation, as they see their situation in a more positive light. Cognitive distortion may also, however, prevent victims from leaving.

One might argue that the positive appraisal of the relationship is not necessarily based on cognitive distortion, but rather on the real situation and the partner’s changed behavior. There are, however, some arguments against this view. In the first place, women of the honeymoon group did not suffer less partner violence than women of the other groups at the first interview. It was, therefore, not the lower frequency of violence that fostered the positive appraisal. Furthermore, the longitudinal data indicated that the amount of women who had suffered partner violence between the second and third interview was equal in the honeymoon group as among women who had remained stable in their intention. Moreover, participants’ change of mind occurred within a relatively short time period, which gives rise to doubts regarding the adequacy of the evaluation. Although aggressors frequently change their behavior during the so-called honeymoon phase, it is doubtful that such changes are substantial and based on a true regret or willingness to behave differently. In fact, behavioral changes among aggressors are rather to be understood as the other side of the same coin, as different forms of manipulation and exertion of power (Walker, 2009).

Women who lived apart from their partner suffered less psychological, physical, and sexual violence than women who had stayed, although this difference was not statistically significant. The percentage of women who still suffered from partner violence after separation is alarming. This is in line with previous studies that show that change in relationship does not necessarily lead to a drop in violence (Bell et al., 2007, 2009). However, our data do not allow specifying whether these women had suffered violence while still living with him or whether they had left after the last violent episode. Interestingly, socio-demographic variables did not distinguish women who changed their mind from women who remained stable in their intention. This result is highly relevant, as it shows that being poor or uneducated does not necessarily lead women to stay with a violent partner—at least not women who make an initial attempt to find help in a formal institution.

Important conclusions can be drawn from this study about psychological intervention with women who seek help from a formal institution. According to the results, counseling should address cognitive distortion by challenging overly positive beliefs about the partner and the relationship. Cognitive distortion may help women preserve a sense of mastery in the threatening situation. By consequence, abolishing cognitive distortion is often linked with nervous reactions, fear, and hopelessness resulting from inconsistency-tension. Psychological intervention may help women face such negative emotions, and, thus, enable victims to make a decision based on a more conscious and mindful evaluation of their situation. Moreover, counseling should motivate and empower women using more problem-focused coping strategies to improve their lives (Bauman et al., 2008). One could argue that women draw on emotion-focused coping strategies when they see no other choice than staying with the partner. Societal changes toward more equitable gender-role models and against the normalization of partner violence are needed in the long run that enable women to take decisions more benefitting for their own and their children’s well-being.

The most important limitation of our study is the self-selection of the sample: They were originally seeking help and thus do not represent all battered women, because few battered women in Latin America turn to formal institutions (Frías, 2013; Kyriakakis, Dawson, & Edmond, 2012; Quintela Modia, Arandia, & Campos, 2004). Interviews were conducted within 1 week after calling on the Legal Office. Time since the last violent episode was not assessed, which might have influenced group assignment (particularly, regarding the honeymoon group). However, most women came to the institution in the direct aftermath of a violent episode, which, after all, had given them the impetus to seek formal help. Moreover, the response rate was particularly low in this study, possibly introducing bias. Another limitation of the present study is that most of the participants were from lower social class—middle and upper classes were not represented. Drop-out rates were high with 40%, and women with a lower level of education were more likely to drop out of the study. Finally, statistical analyses were limited to group comparisons and abstained from repeated or time series measurement analyses due to power calculation.

To the best of our knowledge, this is the first study that investigated women’s cognitions to explain their decision to stay with or leave a violent partner in a low-income country such as Bolivia. This unique sample complements previous results from high-income countries. Despite the high prevalence of partner violence in low- and middle-income countries (Devries, Mak, García-Moreno, et al., 2013; García-Moreno et al., 2006), empirical studies on the stay–leave decision of battered women from these countries are lacking. Evidence from Peru and Bolivia highlights the difficulties women from these countries face when leaving a violent partner (Alcalde, 2006; Quintela Modia et al., 2004). This study provides evidence on one particular aspect that might influence victims’ decision to stay or leave: their overly positive appraisal of the partner and the relationship. More evidence is needed to better understand how such motivation-based judgments interact with social and cultural factors, to design culturally adapted interventions for those millions of women who suffer from partner violence worldwide.

Footnotes

Acknowledgements

The authors would like to thank the Oficina Jurídica para la Mujer as well as the Research Institute of the Faculty of Humanities and Education Sciences (IIHCE) at the Universidad de San Simón in Cochabamba, Bolivia, for their collaboration in this study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was financed by the Swiss Agency for Development and Cooperation (SDC). SDC had no involvement in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

Author Biographies

Eva Maria Heim is a trained psychotherapist and post-doctoral researcher at University of Zurich’s Department of Psychology. Her main research interests are cultural clinical psychology and e-mental health, with a special focus on low- and middle-income countries.

Laura Trujillo Tapia was an undergraduate student during the time this research was conducted.

Ruth Quintanilla Gonzáles is a psychotherapist and lecturer at Universidad Mayor de San Simón with a special research interest in psychosocial intervention methods, health psychology, qualitative research methods, and forensic psychology.

References

Akmatov

M. K.

Mikolajczyk

R. T.

Labeeb

Dhaher

Khan

M. M.

(2008). Factors associated with wife beating in Egypt: Analysis of two surveys (1995 and 2005). BMC Women’s Health, 8, 15. doi:10.1186/1472-6874-8-15

Alcalde

M. C.

(2006). Migration and class as constraints in battered women’s attempts to escape violence in Lima, Peru. Latin American Perspectives, 33, 147-164. doi:10.1177/0094582x06294140

Bauman

E. M.

Haaga

D. A. E.

Dutton

M. A.

(2008). Coping with intimate partner violence: Battered women’s use and perceived helpfulness of emotion-focused coping strategies. Journal of Aggression, Maltreatment & Trauma, 17, 23-41. doi:10.1080/10926770802250942

Bélanger

J. J.

Kruglanski

A. W.

(2014). Bending perception to desire: Effects of task demands, motivation, and cognitive resources. Motivation and Emotion, 38, 802-814. doi:10.1007/s11031-014-9436-z

Bell

M. E.

Goodman

L. A.

Dutton

M. A.

(2007). The dynamics of staying and leaving: Implications for battered women’s emotional well-being and experiences of violence at the end of a year. Journal of Family Violence, 22, 413-428. doi:10.1007/s10896-007-9096-9

Bell

M. E.

Goodman

L. A.

Dutton

M. A.

(2009). Variations in help-seeking, battered women’s relationship course, emotional well-being, and experiences of abuse over time. Psychology of Women Quarterly, 33, 149-162. doi:10.1111/j.1471-6402.2009.01485.x

Byrne

C. A.

Arias

(2004). Predicting women’s intentions to leave abusive relationships: An application of the theory of planned behavior. Journal of Applied Social Psychology, 24, 2586-2601. doi:10.1111/j.1559-1816.2004.tb01993.x

Calvete

Corral

Estévez

(2007). Cognitive and coping mechanisms in the interplay between intimate partner violence and depression. Anxiety, Stress, & Coping, 20, 369-382. doi:10.1080/10615800701628850

Campbell

(2002). Health consequences of intimate partner violence. The Lancet, 359, 1331-1336. doi:10.1016/S0140-6736%2802%2908336-8

10.

Devries

K. M.

Mak

J. Y.

Bacchus

L. J.

Child

J. C.

Falder

Petzold

. . . Watts

C. H.

(2013). Intimate partner violence and incident depressive symptoms and suicide attempts: A systematic review of longitudinal studies. PLoS Medicine, 10, e1001439-e1001439. doi:10.1371/journal.pmed.1001439

11.

Devries

K. M.

Mak

J. Y.

García-Moreno

Petzold

Child

J. C.

Falder

. . . Watts

C. H.

(2013). The global prevalence of intimate partner violence against women. Science, 340, 1527-1528. doi:10.1126/science.1240937

12.

Devries

K. M.

Watts

Yoshihama

Kiss

Schraiber

L. B.

Deyessa

. . . García-Moreno

(2011). Violence against women is strongly associated with suicide attempts: Evidence from the WHO multi-country study on women’s health and domestic violence against women. Social Science & Medicine, 73, 79-86. doi:10.1016/j.socscimed.2011.05.006

13.

Diop-Sidibé

Campbell

J. C.

Becker

(2006). Domestic violence against women in Egypt—Wife beating and health outcomes. Social Science & Medicine, 62, 1260-1277. doi:10.1016/j.socscimed.2005.07.022

14.

Dutton

M. A.

(1992). Empowering and healing the battered woman. New York, NY: Springer.

15.

Elliot

(2006). The hierarchical model of approach-avoidance motivation. Motivation and Emotion, 30, 111-116. doi:10.1007/s11031-006-9028-7

16.

Ellsberg

Jansen

H. A. F. M.

Heise

Watts

C. H.

García-Moreno

(2008). Intimate partner violence and women’s physical and mental health in the WHO multi-country study on women’s health and domestic violence: An observational study. Lancet, 371, 1165-1172. doi:10.1016/S0140-6736(08)60522-X

17.

Frías

S. M.

(2013). Strategies and help-seeking behavior among Mexican women experiencing partner violence. Violence Against Women, 19, 24-49. doi:10.1177/1077801212475334

18.

García-Moreno

Jansen

Ellsberg

Heise

Watts

C. H.

(2006). Prevalence of intimate partner violence: Findings from the WHO multi-country study on women’s health and domestic violence. Lancet, 368, 1260-1269. doi:10.1016/s0140-6736(06)69523-8

19.

Gawronski

(2012). Back to the future of dissonance theory: Cognitive consistency as a core motive. Social Cognition, 30, 652-668. doi:10.1521/soco.2012.30.6.652

20.

Goodman

Dutton

M. A.

Vankos

Weinfurt

(2005). Women’s resources and use of strategies as risk and protective factors for reabuse over time. Violence Against Women, 11, 311-336. doi:10.1177/1077801204273297

21.

Grawe

(2004). Psychological therapy. Göttingen, Germany: Hogrefe.

22.

Heise

L. L.

Kotsadam

(2015). Cross-national and multilevel correlates of partner violence: An analysis of data from population-based surveys. The Lancet Global Health, 3, e332-340. doi:10.1016/s2214-109x(15)00013-3

23.

Hendy

M. H.

Eggen

Gustitus

McLeod

K. C.

(2003). Decision to Leave Scale: Perceived reasons to stay in or leave violent relationships. Psychology of Women Quarterly, 27, 162-173. doi:10.1111/1471-6402.00096

24.

Herbert

T. B.

Silver

R. C.

Ellard

J. H.

(1991). Coping with an abusive relationship: I. How and why do women stay? Journal of Marriage and Family, 53, 311-325. doi:10.2307/352901

25.

Hill

S. E.

Durante

K. M.

(2011). Courtship, competition, and the pursuit of attractiveness: Mating goals facilitate health-related risk taking and strategic risk suppression in women. Personality and Social Psychology Bulletin, 37, 383-394. doi:10.1177/0146167210395603

26.

Iverson

K. M.

Litwack

S. D.

Pineles

S. L.

Suvak

M. K.

Vaughn

R. A.

Resick

P. A.

(2013). Predictors of intimate partner violence revictimization: The relative impact of distinct PTSD symptoms, dissociation, and coping strategies. Journal of Traumatic Stress, 26, 102-110. doi:10.1002/jts.21781

27.

Katerndahl

Burge

Ferrer

Becho

Wood

(2014a). Do violence dynamics matter? Journal of Evaluation in Clinical Practice, 20, 719-727. doi:10.1111/jep.12216

28.

Katerndahl

Burge

Ferrer

Becho

Wood

(2014b). Dynamics of violence. Journal of Evaluation in Clinical Practice, 20, 695-702. doi:10.1111/jep.12151

29.

Kimuna

S. R.

Djamba

Y. K.

(2008). Gender based violence: Correlates of physical and sexual wife abuse in Kenya. Journal of Family Violence, 23, 333-342. doi:10.1007/s10896-008-9156-9

30.

Krause

E. D.

Kaltman

Goodman

L. A.

Dutton

M. A.

(2008). Avoidant coping and PTSD symptoms related to domestic violence exposure: A longitudinal study. Journal of Traumatic Stress, 21, 83-90. doi:10.1002/jts.20288

31.

Kyriakakis

Dawson

B. A.

Edmond

(2012). Mexican immigrant survivors of intimate partner violence: Conceptualization and descriptions of abuse. Violence and Victims, 27, 548-562.

32.

Lazarus

R. S.

Folkman

(1984). Stress, appraisal and coping. New York, NY: Springer.

33.

Lilly

Graham-Bermann

S. A.

(2010). Intimate partner violence and PTSD: The moderating role of emotion-focused coping. Violence and Victims, 25, 604-616.

34.

Meekers

Pallin

S. C.

Hutchinson

(2013). Prevalence and correlates of physical, psychological, and sexual intimate partner violence in Bolivia. Global Public Health: An International Journal for Research, Policy and Practice, 8, 588-606. doi:10.1080/17441692.2013.776093

35.

Meekers

Pallin

S. C.

Hutchinson

(2014). Intimate partner violence and mental health in Bolivia. BMC Women’s Health, 13, 28. doi:10.1186/1472-6874-13-28

36.

Pape

K. T.

Arias

(1995). Control, coping and victimization in dating relationships. Violence and Victims, 10, 43-54.

37.

Pape

K. T.

Arias

(2000). The role of perceptions and attributions in battered women’s intentions to permanently end their violent relationships. Cognitive Therapy and Research, 24, 201-214. doi:10.1023/A:1005498109271

38.

Quintela Modia

Arandia

M. J.

Campos

P. V

. (2004). De la comunidad al barrio: Violencia de pareja en mujeres migrantes en Sucre [From the community to the neighborhood: Partner violence among migrant women in Sucre]. La Paz, Bolivia: Programa de Investigación Estratégica en Bolivia [PIEB, Strategic research program in Bolivia].

39.

Rhatigan

Street

Axsom

(2006). A critical review of theories to explain violent relationship termination: Implications for research and intervention. Clinical Psychology Review, 26, 321-345. doi:10.1016/j.cpr.2005.09.002

40.

Solveig

Bjorkly

(2012). Does separation or divorce make any difference? An interactional perspective on intimate partner violence with focus on marital status. Journal of Family Violence, 27, 45-54. doi:10.1007/s10896-011-9400-6

41.

Solveig

Bjorkly

(2014). An interactional perspective on coping with intimate partner violence: Counterattack, call for help, or give in and obey him? Journal of Aggression, Maltreatment & Trauma, 23, 881-900. doi:10.1080/10926771.2014.953716

42.

Stöckl

Devries

K. M.

Rotstein

Abrahams

Campbell

Watts

García-Moreno

(2013). The global prevalence of intimate partner homicide: A systematic review. Lancet, 382, 859-865. doi:10.1016/s0140-6736(13)61030-2

43.

Walker

(2009). The battered women syndrome. New York, NY: Springer.