Explaining the Relationship Between Insecure Attachment and Partner Abuse

Abstract

Studies have found that male batterers are more often insecurely attached as compared with nonbatterers. However, it is still not clear how insecure attachment is related to domestic violence. Many studies compared batterers and nonbatterers regarding pathological personality characteristics that are related to attachment (e.g., dependency, jealousy) and generally found that batterers report more personality characteristics. However, these studies did not investigate which role these characteristics played in the relationship between insecure attachment and battering. The first aim of this study is to test which personality characteristics are good candidates to explain the relationship between insecure attachment and domestic violence. The second aim is to test whether personality characteristics are predictive of battering over and above attachment. Seventy-two mainly court-mandated family-only males who were in group treatment for battering are allocated to a securely and an insecurely attached group and compared with 62 nonbatterers. Using questionnaires, self-esteem, dependency, general distrust, distrust in partner, jealousy, lack of empathy, separation anxiety, desire for control, and impulsivity were assessed. This was the first study that examined distrust, separation anxiety, and desire for control in relation to battering. The results show that the relationship between insecure attachment and domestic violence can be explained by separation anxiety and partner distrust. Moreover, only partner distrust increased the risk for battering over and above insecure attachment. The findings suggest the presence of two subtypes among batterers based on attachment style, which has similarities to the family-only and dysphoric-borderline subtypes suggested by Holtzworth-Munroe and Stuart. Implications of the present findings for therapy are discussed.

Keywords

attachment batterers intervention/treatment predicting domestic violence

Introduction

Because numerous empirical studies, literature reviews, and meta-analyses of standard model interventions with perpetrators of intimate partner violence have found little or no positive effect on violent behavior (Dutton & Corvo, 2006), there is a clear need of studies laying the basis for evidence-based practice with perpetrators of intimate partner violence (e.g., Corvo, Dutton, & Chen, 2008). Previous studies have shown that male batterers are not a homogeneous group of persons (e.g., Holtzworth-Munroe & Stuart, 1994). Therefore, it is very probable that one intervention does not fit all. More research into the characteristics of batterers can help improving current batterer interventions and develop, if necessary, different intervention programs for subgroups among batterers.

Several subtypes among batterers have been found, which differ regarding severity of marital violence, generality of the violence (toward the partner only or toward others as well), psychopathology/personality disorders, drug and alcohol use, and attachment style (e.g., Dutton & Golant, 1995; Holtzworth-Munroe, Meehan, Herron, Rehamn, & Stuart, 2000; Holtzworth-Munroe & Stuart, 1994). Several studies, using different measures, found that male batterers are more (often) insecurely attached as compared with nonbatterers. More specifically, male batterers are predominantly preoccupied or fearfully attached (e.g., Dutton, Saunders, Starzomski, & Bartholomew, 1994; Mauricio & Gormley, 2001). It is still not clear how insecure attachment is related to (intimate partner) violence. Tweed and Dutton (1998) propose that attachment style and its related personality characteristics are triggered when the relationship is under stress. Timmerman and Emmelkamp (2005) developed a model that explains how insecure attachment may lead to violence. Insecure attachment is thought to result in personality characteristics like dependency, jealousy, self-esteem, empathy, and impulsivity. More specifically, studies have shown that insecure attachment is related to dependency (Mikulincer & Shaver, 2010), more impulsivity (Scott, Levy, & Pincus, 2009), less empathetic concern (e.g., Feeney & Collins, 2001), and high levels of jealousy (e.g., Buunk, 1997; Guerrero, 1998). The self-esteem of insecurely attached persons is dependent on other people’s approval and the slightest indication of disapproval, criticism, or disinterest can strengthen their low self-esteem (Mikulincer & Shaver, 2010). These personality characteristics in turn are thought to result in inadequate coping styles and social skills. One inadequate way of coping is substance abuse. The relationship between intimate partner violence and alcohol abuse has been shown repeatedly (for a review, see Klostermann & Fals-Stewart, 2006). A recent study clearly showed that intimate partner violence occurred when under the influence of substances (Kraanen, Scholing, & Emmelkamp, 2010).

Knowledge about emotional and behavioral tendencies specific to different attachment styles is valuable for the development of future batterer interventions (see also Mauricio & Gormly, 2001). So far, studies comparing batterers and nonbatterers regarding personality characteristics generally found differences between these two groups. That is, most studies found that compared with nonbatterers, wife batterers score higher on abandonment anxiety (Holtzworth-Munroe, Stuart, & Hutchinson, 1997) and impulsivity (e.g., Cohen et al., 2003) but results with respect to interpersonal and spouse-specific dependency (e.g., Butell & Jones, 2001; Butell, Muldoon, & Carney, 2005) and jealousy (Holtzworth-Munroe et al., 1997; Murphy, Meyer, & O’Leary, 1994) are inconclusive. Furthermore, a number of studies have found that male batterers have lower self-esteem than nonbatterers (e.g., Neidig, Friedman, & Collins, 1986; but Hurlbert, Whittaker, & Munoz, 1991) and that low self-esteem is not only related to violence (e.g., Kesner, Julian, & McKenry, 1997) but also predictive of abusive behaviors (Schwartz, Waldo, & Daniel, 2005). These studies show which factors play a role in intimate partner violence. However, although such personality characteristics may indeed be related to insecure attachment, these studies did not investigate which role these characteristics played in the relationship between insecure attachment and battering. Studies investigating whether insecurely attached batterers score more problematic on these personality factors than securely attached batterers will shed more light on why insecure attachment is related to intimate partner violence and can aid in the development of intervention programs for different subtypes among batterers.

Various studies have tested impulsivity, self-esteem, jealousy, and dependency in male batterers. A few studies focused on lack of empathy among male batterers. According to Wexler (1999), intimate partner violence is due to empathic failure. Without empathy or emotional intelligence, anger can quickly turn into violence (Holtzworth-Munroe & Stuart, 1994; Winters, Clift, & Dutton, 2004). It may also be the other way around—that is, empathy may inhibit aggression. Empathy may regulate anger and in that way inhibit violence.

Three further important aspects of insecure attachment have not previously been studied in relation to wife battering. These are separation anxiety, trust, and desire for control. Separation anxiety refers to feelings of extreme anxiety about separations from major attachment figures, fears that harm would befall those close to them, and an intense yearning to go home (Manicavasagar & Silove, 1997). Separation anxiety can be directly linked to insecurities resulting from the unavailability or insensitivity of an attachment figure (Mikulincer & Shaver, 2010). The second factor trust is significant in the maintenance of intimate relationships. Securely attached people feel trust toward relationship partners (Mikulincer & Shaver, 2010). Rempel, Holmes, and Zanna (1985) describe relational trust as the faith people have that their partner will respond to them with fairness, honesty, and positive feelings. Individuals who trust their partners tend to report more love for their partner and more satisfaction with and commitment to their relationship (Couch & Jones, 1997; Rempel et al., 1985). It has further been suggested that although relational trust may have a dispositional aspect, it is most likely influenced by the specific characteristics and actions of ones partner (Couch & Jones, 1997). Regarding desire for control, insecure individuals are more likely than securely attached individuals to adopt a controlling stance (e.g., Feeney & Collins, 2001; Feeney & Hohaus, 2001). Previous studies showed that batterers who report excessive dependency in their interpersonal relationships tend to use violent and controlling behaviors to achieve physical closeness (e.g., Murphey, Meyer, & O’Leary, 1994). In general, the majority of abusive men engage in a wide array of controlling behaviors (e.g., Murphy & Cascardi, 1993). Mauricio and Gormly (2001) found that men who self-reported a high need for dominance reported greater frequencies of physical violence. Furthermore, those men who were insecurely attached and reported a need to dominate their relationships were at greater risk of becoming violent toward their partner than insecurely attached males without this need.

The aim of this study was to test which personality characteristics (self-esteem, dependency, general distrust, distrust in partner, jealousy, lack of empathy, separation anxiety, desire for control, and impulsivity) are good candidates to explain the relationship between insecure attachment and intimate partner violence. We performed two planned contrasts to address this question. First, we examined which factors have a relationship to battering by contrasting batterers to nonbatterers. Next, we examined which factors are influenced by attachment by comparing insecurely attached batterers with securely attached batterers. If no differences were found between securely and insecurely attached batterers regarding these factors, but solely between batterers in general and nonbatterers, this would imply that these factors play a role in battering, irrespective of attachment problems. In that case, all batterers may profit from therapy that target these personality characteristics. However, if insecurely attached batterers did score differently on a personality characteristic as compared with securely attached batterers, it is likely that this factor plays a role in the relationship between insecure attachment and battering. In that case, securely and insecurely attached batterers may benefit from therapies developed specifically for these two subsamples. A further question then would be whether the personality characteristics need to be targeted individually or whether a more broad therapy directed at attachment problems is sufficient. Therefore, a final aim of the present study was to test whether personality characteristics are predictive of battering over and above attachment.

Method

Participants

Seventy-two men who were in group treatment for battering, and 62 nonbatterers (control participants) participated in this study. The mean age of the batterers was 35.5 (SD = 7.98, range= 19-56). The mean length of their abusive relationship was 9 years (SD = 7.6 years, range = 19 months to 40 years). They had a mean number of 1.9 children (SD = 1.6, range = 0-6), and a mean educational level of 6.4 (SD = 2.6, range = 1-12) where 1 = no education, 12 = completed university, 6 = lower secondary education, and 7 = middle secondary education. Of the 72 participants, 4 reported that they had been violent outside their family as well. At the time the questionnaires were filled in, 22 men followed treatment voluntarily and 42 were court-mandated. Of these, 10 were awaiting whether they would be convicted or whether their case would be dropped, 26 had a conditional sentence or a conditional penalty, 4 had a dismissal with conditions to follow, and 2 had a community duty for 6 hr. Information regarding juridical status was missing in 8 batterers. Thirty-two men were Dutch. The others reported that their ethnic background was Surinam (n = 11), Dutch Antilles (n = 5), Turkish (n = 5), Cape Verdian (n = 3), Moroccan (n = 2), Croatian (n = 1), Asian (n = 1), Rwandese (n = 1), and Senegalese (n = 1). The ethnic background was missing in 10 batterers. At the time the questionnaires were filled in, 3 men lived with family, 9 lived alone, 59 lived with their partner (and children), and 2 marked “other.”

Control participants were recruited by means of flyers, which were spread at the hospital, in shops and at a pharmacy. They received 15 euro for their participation in this study. Questionnaires were sent by mail. Of the 73 control participants who filled in the questionnaires, 11 were not included in further analyses. Ten participants scored more than 1 on one or more of the three subscales, namely, Physical Assault, Sexual Coercion, and Injury of the Revised Conflict Tactics Scale (Straus, Hamby, Bony-McCoy, & Sugarman, 1996), and 1 participant had not filled in this questionnaire. The mean age of the 62 control participants was 39.5 years (SD = 10.1, range = 22-64). The mean length of their relationship was 10.9 years (SD = 10 years, range = 1-42.3 years). They had a mean number of 1.3 children (SD = 1.1, range = 0-4), and a mean educational level of 8 (i.e., higher secondary education; SD = 1.7, range = 5-11). Fifty-nine men were Dutch. One was from the Dutch Antilles, and one referred to himself as Hindustan. Ethnic background was missing in one participant. At the time the questionnaires were filled in, 2 men lived with family, 1 lived with family and his children, 6 men lived alone, 1 men lived alone with his children, 50 lived with their partner (and children), and 2 marked “other.”

Measures

Attachment is assessed using the Relationship Questionnaire (RQ; Bartholomew & Horowitz, 1991). Four prototypical attachment patterns as they apply in close adult peer relationships are given. In the first question, participants are asked to choose a single, best fitting attachment pattern. In the following four questions, they are asked to rate their degree of correspondence to each prototype on a 7-point scale, ranging from 1 (not at all like me) to 7 (very much like me).

Distrust is assessed with the Trust Inventory (TI; Couch, Adams, & Jones, 1996) the TI consist of the two subscales, General Trust and Trust in Partner, which are both assessed with 20 items that are rated on a 5-point scale. Scores range from 1 to 5 with higher scores referring to more trust (less distrust). Internal reliability ranges from .87 to .92, and test-retest reliability is .82 for the subscale Trust in Partner and .80 for the subscale General Trust. The two subscales have both concurrent and construct validity (Couch & Jones, 1997).

Desire for control is assessed with the Desirability of Control Scale (DCS; Burger & Cooper, 1979). Twenty items are rated on a 7-point scale. Scores range from 20 to 140 with higher scores referring to more desire for control. Internal reliability is .80, and test-retest reliability is .75. The scale further has discriminant and construct validity.

Separation anxiety is assessed with the Adult Separation Anxiety Self-Report Checklist (ASA; Manicavasagar, Silove, Wagner, & Drobny, 2003), which consists of 27 items that are rated on a 4-point scale. Scores range from 0 to 81 with higher scores referring to more separation anxiety. Internal reliability is .95, and test-retest reliability is .86.

Lack of empathy is assessed with the Basic Empathy Scale (BES; Jolliffe & Farrington, 2004). For the present study, only the subscale Affective Empathy was administered, which consists of 11 items that are rated on a 5-point scale. Scores range from 11 to 55 with higher scores referring to more affective empathy. The scale has construct, convergent, and divergent validity.

Dependency is assessed with the Spouse-Specific Dependency Scale (SSDS; Rathus & O’Leary, 1997), which consists of 30 items that are rated on a 6-point scale. The SSDS consists of the subscales anxious attachment, exclusive dependency and emotional dependency with each 10 items. The total score ranges from 30 to 180 with higher scores referring to more spouse-specific dependency. Internal reliability ranges from .84 to .89.

Jealousy is assessed using the Interpersonal Jealousy Scale (IJS; Mathes & Severa, 1981), which consists of 28 items that are rated on a 9-point scale. Scores range from 28 to 252 with higher scores referring to more jealousy. Internal reliability is .92. The scale has construct and convergent validity (Mathes, Phillips, Skowran, & Dick, 1982).

Self-esteem is assessed using the Rosenberg’s Self-Esteem Scale (RSES; Rosenberg, 1965), which consists of 10 items that are rated on a 4-point scale. Mean scores range from 0 to 3 with higher scores referring to more self-esteem. Internal reliability for various samples are in the range of .77 to .88, and test-retest correlations are typically in the range of .82 to .88 (Blascovich & Tomaka, 1993; Rosenberg, 1986).

Impulsivity is assessed with the Barratt Impulsiveness Scale version 11 (BIS; Patton, Stanford, & Barratt, 1995). For the present study, only the subscale Motor Impulsivity was administered, which consists of 10 items that are rated on a 4-point scale. Scores range from 10 to 40 with higher scores referring to more impulsivity. The internal reliability of the subscale Motor Impulsivity is .59, and the test-retest reliability is .67 (Stanford et al., 2009).

Partner abuse is assessed with the Revised Conflict Tactics Scale (Straus et al., 1996), which consists of 39 items that are rated on an 8-point scale. Internal reliability ranges from .79 to .95. The scale has construct and discriminant validity. Control participants filled in this questionnaire to be able to exclude individuals who are not real controls in the sense that they scored ≥1 on one or more of the subscales Physical Assault, Sexual Coercion, and Injury. This would indicate that they had more than once committed an act of violence toward their partner in the past year. The responses of batterers on this questionnaire are not reported here for their responses were very unreliable. Batterers reported that they merely pushed their partner or called her names, whereas in reality police had come to their house in at least 30 occasions.

Procedure

Men following group therapy for battering at the forensic outpatient clinic where administered the above questionnaires as part of a larger battery of questionnaires. During the first group session, the study was introduced. During the following two to three sessions, a part of the questionnaires was filled out during the first part of the session. Participants filled in the questionnaires individually, in the presence of the therapist. Control participants received the questionnaires by mail and received €15 for their participation.

Statistical Analyses and Data Reduction

SPSS (15) estimated the few missing values by means of regression with residual estimation adjustment. Some measures are missing completely in a number of batterers for some had too many missing items to estimate a total score using SPSS whereas others did not fill in the questionnaire at all. Therefore, the ASA is missing in 5 participants, the BES in 9, the BIS in 3, the IJS in 14, the RSES in 5, the SSDS in 11, and the TI in 11 batterers.

Using an ANOVA with planned contrasts, we tested whether the three groups (i.e., securely attached batterers, insecurely attached batterers, and controls) differed with respect to the dependent variables. First, an overall group difference will be presented. Then, two main planned contrasts were performed. With the first main planned contrast, the two batterer groups taken together (securely attached batterers and insecurely attached batterers) are contrasted with the control group to select factors related to battering. If differences were found, we tested (using two further planned contrasts) whether this difference can be attributed to a difference between insecurely attached batterers and controls, between securely attached batterers and controls, or between both batterer groups and controls.

Finally, stepwise logistic regression (Backward:LR) was used to test whether personality characteristics predicted battering over and above attachment style (secure vs. insecure). Those variables were included as predictors that differed between batterers and controls using the planned contrast analyses described above. In order not to delete valid predictors from the regression analyses (because they did not reach significance due to lack of power in the planned contrast analyses), the p value for the selection of variables was set to .10.

Results

Attachment Style

Batterers were allocated to one of two groups based on their attachment style (secure vs. insecure). Attachment style was based on the Relationship Questionnaire (RQ; Bartholomew & Horowitz, 1991; for a description of this measure see the section Measures). The three insecure attachment styles were collapsed to form the insecurely attached group (see also Mauricio & Gormly, 2001). Seven batterers did not fill in the first general question about the main attachment style but did rate the other 4 questions in which they indicated to what degree each attachment style applied to them. Based on these four questions, for 3 participants we were able to classify their main attachment style. However, 4 batterers rated two or more attachment styles similarly which made it impossible to know whether they were securely or insecurely attached. Three more participants rated two insecure attachment styles similarly which allowed us to allocate them to the insecurely attached group but made it impossible to know which was their main insecure attachment style. The RQ was completely missing in 2 participants. This resulted in 26 batterers (39.4%) who were securely attached, and 40 batterers (60.6%) who were insecurely attached. Of the latter group, 9 were fearfully attached, 11 were preoccupied, and 17 had a dismissing attachment style. Of the control participants, 46 (74.2%) were securely attached, and 16 (25.8%) were insecurely attached. Of the latter, 5 were fearfully attached, 2 were preoccupied, and 9 had a dismissing attachment style.

Overall Group Differences

For means and standard deviations of the three groups, see Table 1. Overall differences between the three groups (securely attached batterers, insecurely attached batterers, and controls) were found with respect to separation anxiety, F(2, 120) = 10.7, p < .001, distrust in partner, F(2, 116) = 23.8, p < .001, general distrust, F(2, 116) = 3.7, p = .03, total dependency score, F(2, 116) = 4.2, p = .02, and the dependency subscale Anxious Attachment, F(2, 116) = 5.9, p = .004. A trend was observed regarding the dependency subscale Emotional Dependency, F(2, 116) = 2.6, p = .08, and Self-Esteem, F(2, 120) = 2.8, p = .07.

Table 1.

Means, Standard Deviations, and Population Size on Personality Characteristics for the Three Groups

	Insecurely Attached Batterers			Securely Attached Batterers			Nonbatterers
Variable	M	SD	N	M	SD	N	M	SD	N
Separation anxiety	22.6^a,c	13.7	37	15.3^c	9.3	24	12.0^a	9.6	62
Distrust partner	33.2^a,c	10.8	35	26.2^b,c	11.3	22	17.9^a,b	10.4	62
General distrust	32.8^a,c	8.4	35	26.8^c	10.1	22	27.4^a	11.1	62
Self-esteem	19.4^a	4.8	37	21.5	4.8	24	21.8^a	5.3	62
Dependency total score	106.9^c	18.0	34	91.8^b,c	24.5	23	103.3^b	19.1	62
Dependency “anxious attachment”	38.4^a,c	8.0	34	31.1^b,c	8.9	23	34.7^a,b	7.5	62
Dependency “exclusive dependecy”	31.6^c	6.4	34	28.0^b,c	7.1	23	31.4^b	8.4	62
Dependency “emotional dependency”	37.0^c	7.8	34	32.7^b,c	11.1	23	37.2^b	7.6	62
Jealousy	125.8	32.1	33	118.5	36.4	21	119.1	31.7	62
Desire for control	92.3	10.2	36	93.4	14.4	24	90.3	12.1	62
Empathy “affective”	33.3	4.6	36	33.4	5.6	23	33.9	4.4	62
Impulsivity “motor”	20.1	3.3	38	18.8	3.9	25	18.9	3.8	62

Note: Within each row, numbers with equal superscripts are significantly different from each other (p < .05).

Comparing Batterers to Controls

Compared with controls, batterers scored higher on separation anxiety, t(120) = 3.4, p < .001, and distrust in partner, t(116) = 6.0, p < .001. Trends were observed with respect to self-esteem, t(120) = −1.5, p = .07, and emotional dependency, t(116) = −1.6, p = .06. On both variables, batterers scored lower.

Two further planned contrasts were performed to test whether these group differences could be attributed to differences between insecurely attached batterers and controls, between securely attached batterers and controls, or between both securely and insecurely attached batterers versus controls. As compared with the controls, the insecurely attached batterers scored higher on separation anxiety, t(120) = 4.6, p < .001, whereas the securely attached batterers did not, t(120) = 1.2, p = ns. Furthermore, both the securely and the insecurely attached batterers scored higher on distrust in partner than controls, t(116) = 6.8, p < .001 and t(116) = 3.1, p = .001, respectively. The difference (trend) between batterers and controls on self-esteem can be attributed to the insecurely attached batterers. They scored lower on this variable as compared with controls, t(120) = −2.3, p = .012, whereas the securely attached batterers did not, t(120) = −0.3, p = ns. The observed differences (trend) regarding dependency can be attributed to the securely attached batterers as well. They scored lower than the controls on the total dependency score as well as on all subscales, indicating less dependency than nonabusive males—total dependency: t(116) = −2.4, p = .01; Anxious Attachment: t(116) = −1.8, p = .034; Exclusive Dependency: t(116) = −1.8, p = .038; Emotional Dependency: t(116) = −2.2, p = .03. Insecurely attached batterers only differed from controls on the dependency subscale Anxious Attachment. They reported more dependency than controls, t(116) = 2.2, p = .015. Finally, although the batterers as a group did not differ from the controls on general distrust, the planned contrasts showed that the insecurely attached batterers did score higher on general distrust than the controls, t(116) = 2.2, p = .007, and a trend was observed that insecurely attached batterers report more impulsivity than controls, t(122) = 1.6, p = .056. The findings are summarized in Table 1.

Comparing Securely Attached Batterers to Insecurely Attached Batterers

Compared with the securely attached batterers, insecurely attached batterers scored higher on separation anxiety, t(120) = 2.5, p = .006, distrust in partner, t(116) = 2.4, p = .009, general distrust, t(116) = 2.2, p = .02, dependency total score, t(116) = 2.8, p = .003, dependency subscale Anxious Attachment, t(116) = 3.4, p < .001, dependency subscale Exclusive Dependency, t(116) = 1.7, p = .045, and dependency subscale Emotional Dependency, t(116) = 1.9, p = .03. A trend was observed regarding self-esteem, t(120) = −1.6, p = .06, and impulsivity, t(122) = 1.4, p = .09. Insecurely attached batterers reported lower self-esteem and more impulsivity than the securely attached batterers. The findings are summarized in Table 1.

Predicting Battering

A final question was whether personality characteristics (which differed between batterers and nonbatterrers using p < .10) are able to predict battering over and above attachment style (secure vs. insecure). An insecure attachment style explained 16% (Nagelkerke) of the variance of battering. Only distrust in partner increased the explained variance to 36% (see Table 2). In the final model, individuals with an insecure attachment style are 2.5 times more likely to be a batterer than those who are securely attached. And with every 1-point increase on a scale of 20 to 100 assessing partner distrust, the risk to be a batterer as compared to a nonbatterer increased with 1.1.

Table 2.

Backward Logistic Regression Predicting Battering Over and Above Attachment Style

		Predicting Battering
	Variable	B	SE	Wald	Odds Ratio	95% Confidence Interval
Step 1	Insecure attachment	1.5	.40	14.0***	4.4	[2.0, 9.7]
Step 4	Insecure attachment	0.93	.45	4.3*	2.5	[1.1, 6.1]
	Partner distrust	0.09	.02	17.3***	1.1	[1.0, 1.1]
	Separation anxiety	—^a
	Emotional dependency	—^a
	Self-esteem	—^a

Note: Nagelkerke R² of the final model is .37.

The variable was not retained in the regression equation in the final step of the stepwise regression analyses.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

Discussion

The main aim of the present study was to examine which personality characteristics can explain why insecure attachment is related to intimate partner violence. The results show that this relationship can be explained by separation anxiety and partner distrust. Both these factors showed to be related to battering (in that they differed between batterers and nonbatterers) as well as to attachment (in that they differed between securely and insecurely attached batterers). Possibly low self-esteem is a third factor that may explain the relationship between insecure attachment and battering, for the difference between batterers and nonbatterers, and between insecurely and securely attached batterers was marginally significant. A second aim was to test whether personality characteristics increased the risk for battering over and above the predictive risk of an insecure attachment style. Only partner distrust increased the risk over and above insecure attachment. Note that partner distrust is a strong predictor of battering.

This was the first study that examined distrust, separation anxiety, and desire for control in relation to battering. Distrust and separation anxiety showed to be important factors in relation to battering. We did not find group differences regarding desire for control. Possibly, our measure was not relevant to battering. We assessed general desire for control including items such as “I would prefer to be a leader than a follower” and “I enjoy making my own decisions.” A more appropriate measure of desire for control may include items reflecting desire to control the behavior of one’s partner. The controlling behaviors of abusive men are directed at “the partner’s autonomy by limiting her social network, reducing her perceived relationship alternatives, restricting her activities outside the home, and limiting her access to finances, education or employment” (Murphy et al., 1994, p. 729). It would further be highly informative to examine the reasons for the controlling behaviors. Does the need to control the partner result from anxiety (e.g., fear of loosing their partner) or from the desire to dominate?

Most previous studies found that batterers report more interpersonal dependency than controls (Buttell, Muldoon, & Carney, 2005; Murphy et al., 1994; but Butell & Jones, 2001), whereas in the present study only a trend was observed that batterers scored higher on emotional dependency than controls. Note that previous studies used a more general measure of dependency (the interpersonal dependency inventory; Hirschfeld et al., 1977), whereas we used a measure that specifically measured dependency on one’s partner. Only one previous study used this measure to assess dependency in an abusive relationship and found that as compared to controls, violent men scored higher on all three subscales (Holtzworth-Munroe et al., 1997). However, they used a short and older version of the SSDS. In contrast to previous findings, the present results showed that only insecurely attached batterers report more dependency as compared to controls and merely on the subscale anxious attachment. Securely attached batterers report less dependency on all subscales, not only as compared to insecurely attached batterers but also as compared to controls! That is, they report less anxious attachment, less emotional dependency, and less exclusive dependency. This indicates that not excessive dependency but lack of dependency is a detrimental characteristic in relation to battering for securely attached males. This finding is contradictive to previous studies and may be explained by the juridical status of the participants (see also Buttell & Jones, 2001). Studies that did find differences regarding dependency (e.g., Holtzworth-Munroe et al., 1997) used volunteers, that is, men from the community or men who voluntarily sought treatment, whereas Buttel and Jones included court-mandated batterers. They suggested that men seeking treatment voluntarily might do so because their partners threaten to leave them if they do not seek treatment, which they will try to avoid because of their excessive dependency. As the present study also included mainly court-mandated batterers, this may explain why we found no difference between batterers as a group and controls regarding spouse-specific dependency either. Indeed, in the present sample, self-referred males scored higher on emotional and exclusive dependency than court-mandated males—t(52) = 2.3, p = .02, and t(52) = 2.1, p = .04, respectively.

Previous studies which used the same measure for jealousy found significant differences between batterers and nonbatterers, whereas we did not (e.g., Dutton et al., 1994; Dutton, van Ginkel, & Landolt, 1996; Holtzworth-Munroe et al., 1997). This was related neither to judicial status nor to ethnicity; the four largest ethnic groups (Dutch, Surinam, Dutch Antillen, and Turks) did not differ on jealousy.

In the present study, there was a trend that insecurely attached batterers report more impulsivity than both securely attached batterers and controls. Contrary to previous findings (e.g., Cohen et al., 2003), as a group, batterers did not score higher on impulsivity than controls. It should be noted that the first module of therapy for batterers included in this study is aimed at reducing impulsivity. However, it is unlikely that this module was so successful that it resulted in an immediate reduction of impulsivity. Moreover, not all batterers had finished this module when they filled in the questionnaires. Still, it may have influenced their responding to the impulsivity scale. Another explanation may be selection bias. We included mainly family-only batterers whereas high scores on impulsivity are more likely in generally aggressive individuals.

Though in the literature on intimate partner violence, postulations are made about the role of empathy, to our knowledge no previous studies tested the relationship between empathy and battering directly. Based on clinical observations, Holtzworth-Munroe and Stuart (1994) “assume that the more empathy a man feels for others, the less likely he is to engage in aggression” (p. 488). However, based on in-depth interview responses regarding empathy and questionnaire scores assessing abusive behavior, Goodrum, Umberson, and Anderson (2001) observed that some of the most violent men reported cognitive empathy, which they relate to the borderline and schizoidal personality traits of the dysphoric/borderline type of batterer (see Holtzworth-Munroe & Stuart, 1994). In the present study, no differences were observed between batterers and controls regarding affective empathy. The most likely explanation is that we included mainly family-only batterers. According to Holtzworth-Munroe and Stuart (1994), family-only batterers are most remorseful and most empathic toward their wives, whereas the generally violent/antisocial batterers feel the least empathy for their wives.

A remarkable finding of the present study was that the most prevalent insecure attachment style among batterers was the dismissing attachment style. Most studies find that the majority of batterers are preoccupied or fearfully attached (e.g., Mauricio & Gormley, 2001, but see Dutton et al., 1994). Dutton et al. (1994) suggest that those with a dismissing attachment style, which is characterized by a positive self-image and a negative other image, are not prone to insecurity in close relationships, and, therefore, should not be prone to angry protest in intimate relationships. However, Dutton (1988) classifies batterers in three subtypes, one of which is the psychopathic subtype characterized by a dismissing attachment style. Similarly, Holtworth-Munroe and Stuart (1994) suggest that the most violent subtype, the generally violent/antisocial subtype, has a dismissing attachment style. A dismissing attachment style is therefore associated with more cold and severe violence and individuals who are generally violent. However, note that the present study included mainly participants who were only violent within the family.

The results of the present study suggest two subtypes among court-mandated family-only batterers based on attachment style; securely and insecurely attached batterers have different characteristics as compared to controls. Both securely and insecurely attached batterers report partner distrust. The securely attached batterers further score low on dependency. The insecurely attached batterers report separation anxiety, general distrust, and low self-esteem. When comparing these subtypes with the typology of Holtzworth-Munroe and Stuart (1994), the securely attached batterers show resemblance to the family-only batterers, in that both show little or no psychopathology. Holtzworth-Munroe and Stuart further hypothesize that family-only batterers have the fewest attachment problems. Lack of dependency may suggest that the securely attached batterer has similarities with Dutton’s psychopathic batterer. However, the psychopathic batterer is characterized by a dismissing attachment style, violent inside and outside the home, and low empathy, all of which do not apply to the securely attached batterers described here. Next, the insecurely attached batterers show resemblance to the dysphoric-borderline batterers. The dysphoric-borderline type is characterized by borderline personality characteristics such as fear of rejection or abandonment, jealousy, highly dependent on partner, difficulty trusting others, and unstable interpersonal relationships. Similarly, the insecurely attached batterers in the present study show borderline characteristics such as separation anxiety, general distrust, and low self-esteem. The dysphoric-borderline type is also characterized by schizoidal characteristics, which are reflected in the dismissing attachment style. About half of the insecurely attached batterers (46%) were dismissing. However, note that Holtzworth-Munroe and Stuart suggest that the generally violent/antisocial batterers show the dismissive attachment style. Our sample by definition cannot be allocated to this group, for participants in this study were mainly violent within the family; mostly the violence was directed toward the wife. However, in our sample, the insecurely attached group does not score high on jealousy, whereas the dysphoric-borderline group does (Holtzworth-Munroe et al., 2000). Similarly, Dutton (1988) and Hamberger and Hastings (1986) found a group with borderline characteristics as well.

Regarding implications of the present findings for therapy, we suggest that all batterers may benefit from attention directed at attachment style and partner distrust. Furthermore, the securely attached group may benefit from focusing on their low dependency, whereas the insecurely attached group may benefit from focusing on separation anxiety, low self-esteem, and general distrust. Future studies may test whether therapy directed at insecure attachment in general will also reduce attachment-related personality characteristics (e.g., partner distrust and separation anxiety) or whether such factors should be specifically addressed during therapy.

A limitation of the present study is the limited usefulness of the measure used to assess desire of control. However, there was no good alternative to the measure used in the present study. Therefore, it is highly desirable that a measure be developed which assesses more specifically the desire to control the partner, as well as the reason for this desire (i.e., anxiety or dominance needs). Furthermore, because we included mainly family-only batterers, the present results cannot be generalized to batterers who are also violent outside the family.

In sum, this is the first study that tried to explain the relationship between insecure attachment and intimate partner violence. It is also the first study that examined the role of separation anxiety, partner distrust, desire for distrust, and empathy in battering directly. Separation anxiety and partner distrust are good candidates to explain the relationship between insecure attachment and battering. Future studies can test this further by analyzing whether these variables mediate the relationship between insecure attachment and battering. For this analysis, a reliable measure of battering is necessary. The next question then is how separation anxiety and partner distrust result in, or facilitate, battering. According to the model of Timmerman and Emmelkamp (2005) personality characteristics result in inadequate coping and social skills. Therefore, a further line of interest is to explore what inadequate coping and social skills result from separation anxiety and partner distrust. The results of the present study further suggest two subtypes among batterers based on attachment style, which have similarities to the family-only and dysphoric-borderline subtypes suggested by Holtzworth-Munroe and Stuart (1994). Juridical status of batterers seems to be an important factor which should not be neglected in future studies. Next, also ethnic and/or religious background may be an important factor that ought to be taken into account. The abusive and controlling behavior of males from a culture that accepts battering and controlling behaviors may be caused by other factors than the abusive behavior of males from cultures that are generally not accepting of abusive and controlling behaviors. The present results apply to men who are violent within the family. Future studies may examine whether the present findings are generalizable to generally violent batterers. In the present study, we allocated all three insecure attachment styles to one group. However, there are differences between the three insecure attachment styles. Therefore, future studies may test whether partner distrust and separation anxiety play an important role in battering regardless of type of insecure attachment style. Finally, whereas previous studies focused on dependency in relation to insecure attachment, the present study shows the reverse: the problem is lack of dependency in securely attached batterers. Therefore, it is worthwhile to further study this lack of dependency in securely attached batterers in relation to their abusive behavior and personality make-up.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

Bios

Nicole M. L. Buck, PhD, is a postdoctoral researcher at the Department of Forensic Psychiatry at Erasmus Medical Center in Rotterdam, The Netherlands. Her current research includes domestic violence, cognitive processing in relation to aggression and sex offending, and aggression in people with schizophrenia.

Ellie P. E. M. Leenaars, PhD, works as a clinical psychologist at the Outpatient and Day Treatment Clinic Het Dok in Rotterdam, the Netherlands, where she treats mainly sex offenders and perpetrators of domestic violence. She is program manager of different therapy programs for aggression and personality disorders and develops modules aimed at prevention of aggressive behavior for adult aggressive patients and perpetrators of domestic violence. She coordinates research on aggressive behavior and on program evaluation.

Paul M. G. Emmelkamp, PhD, is full professor of clinical psychology at the University of Amsterdam and senior consultant at the forensic psychiatry centre De Waag. He has written many books on research into a variety of clinical subjects, including Evidence-Based Treatment of Alcohol and Drug Abuse (with E. Vedel, Routledge, 2006) which has been translated in Dutch and Japanese and Personality Disorders (with J. Kamphuis, Taylor & Francis, 2007) which has been translated in Swedish, Italian, and Chinese. Furthermore, he has written more than 450 publications in peer-reviewed journals or books. He is coeditor-in-chief of Clinical Psychology & Psychotherapy, editor for the section forensic psychiatry of BMC-Psychiatry and serves on the editorial board of a number of other journals in psychology and psychiatry. In 2006, he was awarded a distinguished professorship (“academy professor”) by the Royal Academy of Arts and Sciences.

Hjalmar J. C. van Marle, PhD, is professor of forensic psychiatry at Erasmus University Medical Center and at the School of Law at Erasmus University, Rotterdam, The Netherlands. His current research includes psychiatric determinants of violence and criminal careers, especially psychopathy.

References

Bartholomew

Horowitz

L. M.

(1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-244.

Blascovich

Tomaka

(1993). Measures of self-esteem. In Robinson

J. P.

Shaver

P. R.

Wrightsman

L. S.

(Eds.), Measures of personality and social psychological attitudes (3rd ed., pp. 115-160). Ann Arbor, MI: Institute for Social Research.

Burger

J. M.

Cooper

H. M.

(1979). The desirability of control. Motivation and Emotion, 3, 381-393.

Buttell

Jones

(2001). Interpersonal dependency among court-ordered domestic violence offenders: A descriptive analysis. Journal of Family Violence, 16, 375-384.

Buttell

Muldoon

Carney

(2005). An application of attachment theory to court-mandated batterers. Journal of Family Violence, 20, 211-217.

Buunk

B. P.

(1997). Personality, birth order and attachment styles as related to various types of jealousy. Personality and individual Differences, 23, 997-1006.

Cohen

R. A.

Brumm

Zawacki

T. M.

Paul

Sweet

L. H.

Rosenbaum

(2003). Impulsivity and verbal deficits associated with domestic violence. Journal of the International Neuropsychological Society, 9, 760-770.

Corvo

Dutton

D. G.

Chen

W. Y.

(2008). Towards evidence-based practice with domestic violence perpetrators. Journal of Aggression, Maltreatment and Trauma, 16(2), 111-130.

Couch

L. L.

Jones

(1997). Measuring levels of trust. Journal of Research in Psychology, 31, 319-336.

10.

Couch

L. L.

Adams

Jones

(1996). The assessment of trust orientation. Journal of Personality Assessment, 67, 305-323.

11.

Dutton

D. G.

Corvo

(2006). Transforming a flawed policy: A call to revive psychology and science in domestic violence research and practice. Aggression and Violent Behavior, 77, 457-483.

12.

Dutton

D. G.

Golant

S. K.

(1995). The batterer: A psychological profile. New York, NY: Basic Books.

13.

Dutton

D. G.

(1988). Profiling wife assaulters: Some evidence for a trimodal analysis. Violence and Victims, 3(1), 5-30.

14.

Dutton

D. G.

Saunders

Starzomski

Bartholomew

(1994). Intimacy-anger and insecure attachment as precursors of abuse in intimate relationships. Journal of Applied Social Psychology, 24, 1367-1386.

15.

Dutton

D. G.

van Ginkel

Landolt

M. A.

(1996). Jealousy, intimate abusiveness, and intrusiveness. Journal of Family Violence, 11, 411-423.

16.

Feeney

B. C.

Collins

N. L.

(2001). Predictors of caregiving in adult intimate relationships: An attachment theoretical perspective. Journal of Personality and Social Psychology, 80, 972-994.

17.

Feeney

J. A.

Hohaus

(2001). Attachment and spousal caregiving. Personal Relationships, 8, 21-39.

18.

Goodrum

Umberson

Anderson

K. L.

(2001). The batterer’s view of the self and others in domestic violence. Sociological Inquiry, 71, 221-240.

19.

Guerrero

L. K.

(1998). Attachment style differences in the experience and expression of romantic jealousy. Personal Relationships, 5, 273-291.

20.

Hamberger

K. L.

Hastings

J. E.

(1986). Characteristics of spouse abusers, predictors of treatment acceptance. Journal of Interpersonal Violence, 1, 363-373.

21.

Hirschfeld

R. M. A.

Klerman

G. L.

Gough

H. G.

Barrett

Korchin

S. J.

Chodoff

(1977). A measure of interpersonal dependency. Journal of Personality Assessment, 41, 610-618.

22.

Holtzworth-Munroe

Stuart

G. L.

(1994). Typologies of male batterers: Three subtypes and the differences among them. Psychological Bulletin, 116, 476-497.

23.

Holtzworth-Munroe

Meehan

J. C.

Herron

Rehman

Stuart

G. L.

(2000). Testing the Holtzworth-Munroe and Stuart (1994) batterer typology. Journal of Consulting and Clinical Psychology, 68, 1000-1019.

24.

Holtzworth-Munroe

Stuart

G. L.

Hutchinson

(1997). Violent versus nonviolent husbands: Differences in attachment patterns, dependency, and jealousy. Journal of Family Psychology, 11, 314-331.

25.

Hurlbert

D. F.

Whittaker

K. E.

Munoz

C. J.

(1991). Etiological characteristics of abusive husbands. Military Medicine, 56, 670-675.

26.

Jolliffe

Farrington

D. P.

(2004). Empathy and offending: A systematic review and meta-analysis. Aggression and Violent Behavior, 9, 441-476.

27.

Kesner

J. E.

Julian

McKenry

P. C.

(1997). Application of attachment theory to male violence toward female intimates. Journal of Family Violence, 12, 211-228.

28.

Klostermann

Fals-Stewart

(2006). Intimate partner violence and alcohol use: Exploring the role of drinking in partner violence and its implications for intervention. Aggression and Violent Behavior: A Review Journal, 11, 587-597.

29.

Kraanen

F. L.

Scholing

Emmelkamp

P. M. G.

(2010). Substance use disorders in perpetrators of intimate partner violence in a forensic setting. International Journal of Offender Therapy and Comparative Criminology, 54(3), 430-440.

30.

Manicavasagar

Silove

(1997). Is there an adult form of separation anxiety disorder? A brief clinical report. Australian and New Zealand Journal of Psychiatry, 31, 299-303.

31.

Manicavasagar

Silove

Wagner

Drobny

(2003). A self-report questionnaire for measuring separation anxiety in adulthood. Comprehensive Psychiatry, 44, 146-153.

32.

Mathes

E. W.

Phillips

J. T.

Skowran

Dick

W. E.

III . (1982). Behavioral Correlates of the Interpersonal Jealousy Scale. Educational and Psychological Measurement, 42, 1227-1231.

33.

Mathes

E. W.

Severa

(1981). Jealousy, romantic love, and liking: Theoretical considerations and preliminary scale development. Psychological Reports, 49(1), 23-31.

34.

Mauricio

A. M.

Gormley

(2001). Male perpetration of physical violence against female partners: The interaction of dominance needs and attachment insecurity. Journal of Interpersonal Violence, 16, 1066-1081.

35.

Mikulincer

Shaver

P. R.

(2010). Attachment in adulthood: Structure, dynamics, and change. New York, NY: Guilford.

36.

Murphy

Meyer

O’Leary

(1994). Dependency characteristics of partner assaultive men. Journal of Abnormal Psychology, 103, 729-735.

37.

Murphy

C. M.

Cascardi

(1993). Psychological aggression and abuse in marriage. In Hampton

Gullotta

Adams

Potter

Weissberg

(Eds.), Family violence: Prevention and treatment (pp. 86-112). Newbury Park, CA: SAGE.

38.

Neidig

P. H.

Friedman

D. H.

Collins

B. S.

(1986). Attitudinal characteristics of males who have engaged in spouse abuse. Journal of Family Violence, 1, 223-233.

39.

Patton

J. H.

Stanford

M. S.

Barratt

E. S.

(1995). Factor structure of the Barratt Impulsiveness Scale. Journal of Clinical Psychology, 51, 768-774.

40.

Rathus

J. H.

O’Leary

K. D.

(1997). The Spouse-Specific Dependency Scale: Scale development. Journal of Family Violence, 12(2), 159-168.

41.

Rempel

J. K.

Holmes

J. G.

Zanna

M. P.

(1985). Trust in close relationships. Journal of Personality and Social Psychology, 49, 95-112.

42.

Rosenberg

(1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

43.

Rosenberg

(1986). Conceiving the self. Malabar, FL: Krieger.

44.

Scott

L. N.

Levy

K. N.

Pincus

A. L.

(2009). Adult attachment, personality traits, and borderline personality disorder features in young adults. Journal of Personality Disorders, 23, 258-280.

45.

Schwartz

J. P.

Waldo

Daniel

(2005). Gender-role conflict and self-esteem: Factors associated with partner abuse in court-referred men. Psychology of Men & Masculinity, 6, 109-113.

46.

Stanford

M. S.

Mathias

C. W.

Dougherty

D. M.

Lake

S. L.

Anderson

N. E.

Patton

J. H.

(2009). Fifty years of the Barratt Impulsiveness Scale: An update and review. Personality and Individual Differences, 47, 385-395.

47.

Straus

M. A.

Hamby

S. L.

Bony-McCoy

Sugarman

D. B.

(1996). The revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17, 283-316.

48.

Timmerman

I. G. H.

Emmelkamp

P. M. G.

(2005). An integrated cognitive-behavioural approach to the aetiology and treatment of violence. Clinical Psychology and Psychotherapy, 12, 167-176.

49.

Tweed

R. G.

Dutton

D. G.

(1998). A comparison of impulsive and instrumental subgroups of batterers. Violence Victims, 13, 217-230.

50.

Wexler

(1999). The broken mirror: A self-psychological treatment perspective for relationship violence. Journal of Psychotherapy: Practice and Research, 8, 129-141.

51.

Winters

Clift

R. J.

Dutton

D. G.

(2004). An exploratory study of emotional intelligence and domestic abuse. Journal of Family Violence, 19(5), 255-267.