Secure Base Narrative Representations and Intimate Partner Violence: A Dyadic Perspective

Abstract

This study aimed to understand the relationship between secure base phenomena and dating violence among couples. Within a relationship, a secure base can be defined as a balancing act of proximity-seeking and exploration at various times and contexts with the assurance of a caregiver’s availability and responsiveness in emotionally distressing situations. Participants were 87 heterosexual couples. The Actor-Partner Interdependence Model was used to examine the relationship between each partner’s scores on secure base representational knowledge and intimate partner violence. Findings demonstrated that women’s secure base representational knowledge had a significant direct negative effect on the victimization of both men and women, while men’s secure base representational knowledge did not have any significant partner or actor effects. Therefore, findings suggest that women with insecure attachments may be more vulnerable to being both the victims and the perpetrators of

The success of the feminist movement in 1960s America drew attention to male violence against women as a serious violation of women’s human rights. The resultant view that intimate partner violence (IPV) is unanimously male-perpetrated has dominated the social sciences for decades. However, this traditional perspective has morphed in recent years, as studies conducted in developed nations, such as the United States, have demonstrated rates of IPV approximating gender symmetry, especially in regards to psychological aggression or emotional abuse (e.g. Archer, 2000; Harned, 2001; Hines & Saudino, 2003; Straus, 2004). A study by Whitaker, Haileyesus, Swahn, and Saltzman (2007) analyzed data from the National Longitudinal Study on Adolescent Health (N= 11,370) and found that 24% of relationships had some level of violence, with 50% of those involving reciprocal violence (i.e. perpetrated by both partners). The context of violence was a strong predictor of injury, with reciprocal IPV more likely to result in greater injury, regardless of gender (Whitaker et al., 2007).

Although female IPV against male partners is less likely to result in death or serious injury than male-perpetrated violence against women (Archer, 2000), it is still a form of abuse requiring concern and intervention (Straus, 2005). In this study, we investigated the relationship between secure base phenomena beyond specific attachment styles and violence among couples by use of couple level data. In this sense, secure base can be defined as a balancing act of proximity-seeking and exploration, at various times and contexts, with the assurance in a caregiver’s availability and responsiveness in difficult times. Furthermore, the current study examined victimization and perpetration by partners of both genders.

Attachment Theory

In order to understand the variants of relationship violence, it is imperative to note attachment theory, which is a useful framework through which IPV prevention and treatment can be viewed. Based on the initial conceptualizations of Bowlby (1969), attachment theory is concerned with how early experiences, especially with the primary caregiver, influence relationships later in life. Attachment is considered an innate behavioral system that enables survival by triggering an infant to seek proximity and contact with a trusted figure, especially in times of distress, and to use that caregiver as a secure base from which to explore. The concept of a secure base is central to attachment theory.

The relationship with the attachment figure, whether secure or insecure, informs all interpersonal relationships and allows the child to develop a working model of relationships for reference later on in life. If the caregiver is irresponsible, unresponsive, or abusive, the child will have a greater likelihood of developing an insecure attachment (anxious-ambivalent, avoidant, or disorganized), developing mental models that other people are not dependable or trustworthy (Ainsworth, Blehar, Waters, and Wall, 1978).

Secure base phenomenon in adult partnerships has similarities with the secure attachment in infant-parent attachment. In both, an open and consistent expression of the need for comfort and contact exists. In adult partnerships, one partner approaches the other directly for help and support in a time of distress. Through a collaborative process, the partner is able to re-establish emotional equilibrium and return to normal activities and exploration. Providing secure base support requires that individuals have an interest in detecting signals, can recognize that their partners have a need or distress, correctly interpret the need, and are willing to provide appropriate responses in a timely fashion (Crowell & Treboux, 2000). Furthermore, secure relationships entail an ability to move unreservedly from the dependent to depended-on positions. Both partners must empathically appreciate each other’s thoughts and feelings. This way, reciprocity and symmetry exist, not only within the individual, but also within the couple system (Fisher & Crandel, 2001). The end result is that both partners feel supported and comforted.

Fonagy (1999) proposes that relationship violence may be seen as evidence of a disorganized attachment system, especially men’s violence against women. Studies investigating the association between attachment and relationship violence have found that, as compared to men who are non-violent, men who are violent report significantly higher levels of relationship anxiety, fear of being abandoned in relationships, and more anxious attachment to their partners (Holzworth-Munroe, Stuart, & Hutchinson, 1997). In a study examining how people with insecure attachment manage conflicts in close relationships, Simpson, Rholes, and Phillips (1996) found that men and women with ambivalent orientations perceive their partner in less positive terms after discussing a problem and are also viewed more negatively by observers. In another study, Creasey (2002) found that in romantic relationships, secure women are more likely to engage in positive conflict behaviors, while insecure males tend to engage in negative conflict behaviors.

Assessing Relationship Security: Secure Base Narrative Representations

H. Waters and colleagues (Waters, Rodrigues, & Ridgeway, 1998) developed a script-like approach to study the mental representations of attachment, particularly to assess secure base representational knowledge. According to script theory, mental representations of events are formed with repeated experiences. For a specific situation, a script describes an appropriate series of events, as well as actors, actions, and props needed to reach a goal. A goal can be something simple, such as going to bed or going out for dinner (Nelson, 1986; Schank & Abelson, 1977). Other researchers view scripts as cognitive representations of routines or ways that individuals typically approach errands and problems. In sum, scripts provide a cognitive tool for individuals to help them navigate their complex world (Nelson, 1986).

Scripts are constructed through experiences in a particular context; i.e., they are learned and organized through information about specific social goals and activities (Nelson, 1986). Structural knowledge about objects and their relationships, along with knowledge of the social world, are all integrated in scripts, which increase in complexity as children develop. These scripts tend to reflect the dynamic characteristics of children’s experiences. According to Bretherton (1991), attachment scripts are the cognitive building blocks of attachment representations. Since scripts are derived from shared social experiences, they are expected to demonstrate a high degree of mutuality. Additionally, since scripts may reveal the underlying cognitive structure of representations, they are also expected to be consistent over time. In light of these expectations, H. Waters and colleagues (e.g., Guttmann-Steinmetz, Elliot, Steiner, & Waters, 2003; Waters & Rodriguez, 2001; Waters, Rodrigues, & Ridgeway, 1998; Waters et al., 2000) developed an assessment of attachment representation involving attachment-relevant scenarios.

The key concept assessed through an individual’s narrative representation is that of the prototypical secure base script. This script consists of the following components: (1) exploring the situation, (2) maintaining contact with the romantic partner or returning back in the time of need, (3) providing help and support when difficulty or threat arises, (4) providing physical contact and comfort when it is needed, (5) helping to deal with difficulties (e.g. getting rid of the difficulty, removal from the situation, providing an explanation, and helping to regulate affect), and (6) helping return to exploration and the previous state. This secure base script provides a framework for understanding what has happened, the process of resolving the circumstances for possible effects, and the process of recovering and returning to normal. Scripted information about secure base relationships better informs about key cognitive features underlying attachment representations and reflects the prototypical experience in the domain of attachment relationships (Waters et al., 1998).

Narrative representations have been used to assess secure base phenomena among diverse populations, including: clinical samples of maltreated children (Hodges, Steele, Hillman, Henderson & Neil, 2000; Toth, Cicchetti, Macfie, Maughan, and Vanmeenen, 2000) and children who witness interparental conflict (Perkoski, 2007), along with non-clinical samples (Matias, O’Connor, Futh, & Scott, 2014), of both adolescents (Dykas, Woodhouse, Cassidy, & Waters, 2006) and adults (Rodrigues, 2000).

Current Study

As discussed previously, the attachment behavior system provides safety and security through the interaction with an attachment figure. This occurs when partners express their needs and seek support. If these expressions of need are recognized and met, then individuals may develop secure representations (Bartholomew, Henderson, & Dutton, 2001; Crowell, Treboux, Gao, et al., 2002). Through recurrent experiences with attachment figures, secure base attachment behaviors become automatic rather than requiring active or conscious effort on each occasion. With these recurring experiences, script-like representational skills emerge (Bretherton, 1985; Main, Kaplan, & Cassidy, 1985; H. Waters, Waters, & Crowell, 2000). When individuals have coherent, informed, and organized representations, these cognitive schemas will be an effective guide for their relationships (Crowell & Treboux, 2000; Fisher & Crandel, 2001).

The aim of the current study is to explore the relationship between secure base phenomena and violence among couples. Based on prior research, security in the relationship is thought to provide protection against violence, and security is associated with other positive aspects of the relationship (e.g. Creasey; 2002; Holzworth-Munroe, Stuart, & Hutchinson, 1997). We hypothesize that individuals with higher secure base representational knowledge will report less involvement in IPV, both as perpetrators and victims. In other words, we expect an actor effect for both men and women. Specifically, we hypothesize that:

H1: Women with higher levels of secure base representational knowledge will demonstrate behaviors that are less frequently physically assaultive, are less frequently psychologically aggressive, and will use degradation less frequently than would those women with lower levels of secure base representational knowledge.

H2: Men with higher levels of secure base representational knowledge will demonstrate behaviors that are less frequently physically assaultive, are less frequently psychologically aggressive, and will use degradation less frequently than would those men with lower levels of secure base representational knowledge.

Another aim is to understand the relationship between secure base representational knowledge and the level of violence in a couple’s relationship by considering the relationship between the two partners’ attachment representations and violence. We expect a partner effect both for men and women. Specifically, we hypothesize that:

H3: Women with higher levels of secure base representational knowledge will engage in less abusive behaviors toward their partners (e.g. physical assault, psychological aggression, and degradation) than would those with lower levels of secure base representational knowledge.

H4: Men with higher levels of secure base representational knowledge will engage in less abusive behaviors toward their partners (e.g. physical assault, psychological aggression, and degradation) than would those with lower levels of secure base representational knowledge.

In order to test these hypotheses, dyadic data was analyzed with the Actor Partner Interdependence Model (APIM; Cook, & Kenny, 2005).

Methods

Participants

Participants of this study were 87 heterosexual undergraduate couples in dating relationships (n = 145). Demographic data was reported for students and their partners. The average age of the participants in the sample was 22.3 years (SD = 4.80). Seventy percent of the participants were European-American, 8% African American, 6% Hispanic, 9% Asian, and 4% reported other ethnicities; the remaining participants did not report their ethnicities. The mean duration of dating was 32 months, ranging from 1 month to 16 years. Most of the couples (n =31) reported being in a relationship with their current partner more than 20 months, 19 of the couples reported being involved with their current partner from 6 to 20 months, and eight of the couples were dating less than six months.

Procedure

Classroom announcements at a large Midwestern university were used to recruit participants. Undergraduate students who were in a dating relationship and interested in participating the study asked their partners to join the study. After consenting, participants completed questionnaires about their aggression (both verbal and physical) and attachment patterns. Following completion of the questionnaires, participants were asked to complete stories based on four word-lists, each containing a story outline with familiar topics that were expected to elicit attachment themes. These stories were used to assess secure base representational knowledge. Story collection lasted approximately 30 minutes, and responses were audio recorded. Following study completion, participants were debriefed and received course extra credit for their participation.

Measures

Demographic Information

This questionnaire included questions about participant demographic characteristics and relationships (e.g. age, sex, race, education level, and socioeconomic status). The remaining questions were related to relationship characteristics of the participants, including duration of the relationship and anticipated length of the relationship.

Secure base scriptedness

In order to assess secure base representational knowledge, participants were asked to complete stories based on attachment-relevant scenarios. These stories were developed by H. Waters and colleagues (e.g., Guttmann-Steinmetz, Elliot, Steiner, & Waters, 2003; Waters & Rodriguez, 2001; Waters, Rodrigues, & Ridgeway, 1998; Waters et al., 2000) as an assessment of attachment representation. In the present study, participants were given four sets of attachment related scenarios. These scenarios had prompt words to guide the participants to develop stories. Stories involved either romantic partners or mother and child/baby dyads. Two stories were about adults: “Jane and Bob’s camping trip” and “The accident.” For mother-child stories, two stories existed: “Baby’s morning” and “The doctor’s office.” These story stems were selected because they included familiar conditions that were expected to elicit attachment themes. Furthermore, since attachment is a developmental concept starting from infancy, stories from across the lifespan were included to have a more complete understanding of the secure base representations with stronger theoretical indicators. Participants had time to review the outlines before they started. Stories generated by the participants were recorded, transcribed, and later used for coding.

The stories were scored ranging from 1 (no secure base script) to 7 (secure base script). Both content elaboration and prototypic scriptedness of the story completions were carefully assessed during the scoring procedures (Waters et al., 1998). The appendix includes examples of stories that received both high and low scores. Two raters scored the stories. Both raters of the stories were trained doctoral students. Trainings took place in two separate occasions. Raters first received a didactic lecture on the narrative representations and attachment theory, then read the detailed coding manual and rated sample stories. They rated the stories independently for the current study. Cohen Kappa for interrater reliability between the two raters was .80. For secure base scripts, principal component analysis indicated that all items loaded on one factor. For female participants, this factor contained 52% of the total variance (eigenvalue = 2.08), with an alpha of .69. For male participants, this scale contained 55% of the total variance (eigenvalue= 2.23), with an alpha of .72.

Revised Conflict Tactics Scale (CTS2)

In order to assess intimate partner violence in this study, two subscales (psychological aggression and physical assault) from the modified version of the Conflict Tactics Scale (CTS2) were used (Straus, Hamby, McCoy, & Sugarman, 1996). Participants were asked how they managed disagreements with their dating partners over the four months preceding the study. The psychological aggression subscale included 8 items, including items such as “I insulted or swore at my partner.” The physical assault subscale consisted of 12 items. Physical assault included items like “I pushed, grabbed, or hit my partner.” Participants were also asked to indicate the frequency of these actions for themselves and their partners. Reliability of these two subscales ranged from .69 to 80. The modified version of the CTS is a widely used instrument with good construct validity (Archer, 1999; Shook, Gerrity, Jurich, & Segrist, 2000; Straus, 1990; Straus, 1995).

In this study, frequency scores were used to create composite variables as suggested by the scoring manual developed by Straus (2004). More specifically, physical assault and verbal aggression measure the aggression toward another for each participant, computed by averaging the self and partner reports on the frequency of each action. Higher scores indicated higher frequency of the aggressive behavior.

Emotional abuse questionnaire (EAQ)

The EAQ is a 66 item questionnaire designed to assess verbal abuse, sexual abuse, and threatening behavior (Jacobson & Gottman, 1998). Each item is rated on a 4-point frequency scale: 1 (never), 2 (rarely), 3 (occasionally), to 4 (very often). The EAQ has four subscales: isolation, degradation, sexual abuse, and property damage. For this study, only the degradation subscale was utilized due to a lack of variance in the isolation, sexual abuse, and the property damage subscales when using a less violent community sample. The degradation subscale has 28 items, including “my partner humiliates me in front of others,” and “my partner ridicules me.” In this study, internal consistency for the degradation subscale is .94. Previous research using the EAQ found higher correlations with the constructs related to emotional abuse as an indication of content validity (Babcock, Jacobson, Gottman, & Yerington, 2000). The degradation score was obtained by averaging the 28 items. Higher scores indicated higher levels of degradation experienced from the partner.

Analysis Plan

Physical assault frequency and psychological aggression variables were found to be slightly positively skewed; they were transformed by taking the natural logarithm of the variables to avoid violating the normality assumptions. The Actor-Partner Interdependence Model (APIM) was chosen to examine the relationship between scores of women and men on secure base representational knowledge and intimate partner violence (Kenny, Kashy and Cook, 2006). Because the unit of analysis in this study is the dyad, the data for men and women partners are not independent; thus, the APIM must be used to analyze the data without violating the assumption of independence. Model fit was assessed by a χ2/df < 5, an RMSEA < .10, and a CFI >.90 (Wheaton, Muthen, & Alwin, 1977).

A structural equation model (SEM) was fit in which men and women secure base latent variables (measured by the scores on the 4 narratives: baby’s morning, accident, doctor’s office, and camping trip) were allowed to predict their own and their partner’s victimization latent variables (measured by the CTS subscales of psychological aggression and physical assault, along with EAQ degradation scores). Full-information maximum likelihood estimation (FIML, also known as “Raw Maximum Likelihood”) was utilized in AMOS to use all the available data at the couple level (Enders & Bandalos, 2001).

Results

Table 1 presents means, standard deviations, and corresponding group t-tests for gender differences among all study variables. Preliminary paired sample t-test analyses indicated men reported more frequent psychological aggression victimization t(51) = −2.21, p < .05, physical assault victimization t(51) = −2.02, p < .05 and degradation victimization t(53) = −2.94, p < .05 than women. The story of the accident received higher secure base representation knowledge scores from women t(53) = −2.20, p < .05 than men. Table 2 presents correlations among study variables. Strong correlations were observed between women’s psychological abuse victimization and women’s physical assault victimization, r (85)=.53, p < .001, men’s psychological abuse victimization and men’s physical assault victimization, r(85)=.63, p < .001, women’s psychological abuse victimization and men’s psychological abuse victimization, r(85)=.71, p < .001, and women’s physical assault victimization and men’s physical assault victimization, r(85)=.53, p < .001.

Table 1.

Descriptive Statistics and Paired Sample T-Tests for the IPV victimization and Narrative Representations

	Female Participants		Male Participants		T-test
	Mean	S.D	Mean	S.D	T	Df
Narrative Representations:
Baby’s morning	4.37	1.67	3.77	1.49	1.62	55
Doctor’s office	3.86	1.71	3.13	1.51	2.80	55
Accident	3.48	1.58	2.76	1.36	2.20*	53
Camping Trip	3.74	1.61	3.41	1.18	1.61	54
IPV Victimization:
Psychological Aggression	8.38	8.57	11.13	11.67	−2.21*	51
Physical Assault	1.55	2.83	3.21	7.29	−2.02*	51
Degradation	31.93	4.80	34.72	8.62	−2.94*	51

^*p<.05

Table 2.

Correlation table with means and standard deviations (N=87)

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. F. Baby’s morning
2. F. Doctor’s office	.49**
3. F. Accident	.32**	.20
4. F. Camping Trip	.36**	.50**	.27
5. F. Psy.agg	−.26	−.23	.08	−.18
6. F. Physical assault	−.14	−.22	.06	−.12	.53**
7. F. Degradation	−.23*	−.12	−.05	−.18	.47*	.45*
8. M. Baby’s morning	.04	−.12	.09	−.11	.19	.00	.05
9. M. Doctor’s office	.14	.04	.11	.06	.08	−.15	−.01	.31*
10. M. Accident	.06	−.08	.05	−.05	.09	.28	.00	.46**	.38**
11. M. Camping Trip	.19	.11	.12	.11	.05	−.11	−.08	.58**	.53**	.31**
12. M. Psy.agg	−.27	.15	−.04	−.22	.71*	.31*	.31*	.06	.05	−.06	−.05
13. M. Physical assault	−.18	−.18	−.06	−.21	.62**	.53**	.29*	.22	−.09	−.10	.13	.63**
14. M. Degradation	.07	.01	.02	.00	.38*	.37*	.21	.02	−.02	−.10	.01	.39**	.66**
Means	4.37	3.86	3.48	3.74	8.38	1.55	31.93	3.77	3.13	2.76	3.41	11.13	3.21	34.72
S.D.	1.67	1.71	1.58	1.61	8.57	2.83	4.80	1.49	1.51	1.36	1.18	11.67	7.29	8.62

^*p<.05,

^**p<.01

Please note: F = female participant, M = male participants

A series of confirmatory factor analyses were conducted prior to fitting the proposed SEM. These analyses were conducted to understand whether the latent variables of women and men for both secure base representational knowledge and relationship violence victimization are significantly associated with the corresponding observed variables. The women’s narratives baby’s morning (β = 1.00, r= .64, p < .001), accident (β = .53, r= 36, p < .01), doctor’s office (β = 1.20, r= .75, p < .001), and camping trip (β = .95, r= .63, p < .001) were significantly loaded to the women’s secure base representational knowledge. The men’s narratives of baby’s morning (β =1.00, r= .71, p < .001), accident (β = .70, r= .53, p < .001), doctor’s office (β = .88, r= 61, p < .001), and camping trip (β =.90, r= .80, p < .001) were significantly loaded to the men’s secure base representational knowledge. Degradation (β = .30, r = .52, p < .001), psychological abuse (β = 1.00, r= .96, p < .001), and assault victimization (β = .20, r= .58, p < .001) were significantly loaded on the women’s relationship violence victimization. Degradation (β = .44, r = .49, p < .001), psychological abuse (β = 1,00, r = .83, p < .001), and assault (β =.58, r = .76, p < .001) were significantly loaded on the men’s relationship violence victimization. These confirmatory factor analyses (measurement model) with the latent factors were fit to the data, and results showed satisfactory fit (χ²(74, N=87) =84.15, p > .05; χ2/df=1.14; RMSEA=.04, CFI=.95). In this study, to have a more comprehensive definition of violence, we used both physical and emotional abuse indicators to form the relationship violence victimization latent variable.

APIM findings indicate that the proposed model provides a good fit to the data (χ²(70, N=87) =76.99, p > .05; χ2/df=1.10; RMSEA=.03, CFI=.97). Figure 1 represents the results of this analysis. Significant actor effects for women exist, indicating that women’s secure base representational knowledge has a significant negative direct effect on women’s victimization (β = −2.51, r= −32, p < .05), controlling for partner effects. Significant partner effects for women also exist, indicating that women’s secure base representational knowledge has a significant direct negative effect on men’s victimization (β = −3.36, r=−37, p < .05). Men’s secure base representational knowledge has no significant partner or actor effects. The fitted model accounted for 12% variance of the women’s victimization and 14% variance of the men’s victimization.

Figure 1.

Fitted model with non-standardized estimates for actor partner interdependence model with correlations in parentheses (*<.05, ** <.01, ***<.001). Please note that F stands for female participants, M stands for male participants.

Discussion

This study explored the relationship between secure base phenomena, assessed via narrative representations, and IPV among heterosexual couples. Considering the extant literature on the gender-symmetry of IPV, this study also sought to examine victimization and perpetration by partners of both genders. In the current study, results indicated that the expression of violence among couples and its association with secure base representational knowledge were gender-specific. Women’s secure base representational knowledge had a significant negative direct effect on women’s victimization (actor effect) and a significant negative direct effect on men’s victimization (partner effect). However, men’s secure base representational knowledge did not have a significant direct effect on either men’s victimization or women’s victimization. Thus, findings suggest that women with insecure attachments may be more vulnerable to being both the victim and the perpetrator of aggressive behavior. Additionally, men reported experiencing more frequent victimization than women.

Insights from attachment and gender role socialization theories can be utilized to explain these findings. Attachment within an intimate relationship can be understood as a dyadic process of mutual affect regulation, and an unsuccessful strategy by one partner can dismantle the couple system and reinforce insecurity and attachment-based fears (Schneider & Brimhall, 2014). Creasey (2002) found that female attachment security predicts the occurrence of positive joint couple behaviors. However, the expression of positive behaviors during conflict is not governed by gender alone. Secure women may have a better ability to appraise what to do during conflict as compared to insecure women (Creasey, 2002). The present study also demonstrates that lower levels of secure base representational knowledge in women predict their use of psychological and physical violence. The use of violence by women may be related to when women’s needs for support are not recognized and met. It is possible that the use of violence by women represents a dysfunctional attempt to receive support from their partner. The violence can also be related to a “mispairing” of attachment-based responses between partners (Doumas et al., 2008).

In general, the study’s findings are supported by previous research showing that attachment insecurity is a vital factor for understanding both victimization and perpetration of IPV (e.g. Doumas et al., 2008; Karakurt, Keiley, and Posada, 2013). However, much evidence has accumulated that the effects of IPV are more deleterious for women than for men in terms of attachment, relationship quality, and the onset of post-traumatic stress disorder. For instance, a longitudinal study by Weston (2008) found direct effects of violence on relationship quality only for insecurely attached women. Emotional abuse was found to increase women’s insecure attachment, and as Weston (2008) notes, “the possibility arises that violent victimization may lead to increased insecure attachment, which then results in perpetration” (p. 491).

According to Godbout, Dutton, Lussier, and Sabourin (2009), a number of variables might be related to the heightened interpersonal problems of women. For example, women may be prone to more conflict due to limited time and energy and competing demands, as few studies indicate that work/family balance may be harder to obtain for women generally. This higher emotional demand on women might cause them to be more reactive to conflict compared to men. Findings also demonstrate that male secure base representational knowledge did not have any significant effect on men’s victimization and perpetration. Men higher in secure base representational knowledge of their partner may perceive their partner as both emotionally and physically available, as well as responsive to bids for proximity.

Individuals, who have consistent negative interactions with their partners and perceive their partners as unavailable, report feeling disengaged and detached from their partners. The attachment figure serves as a ‘safe haven’ whenever protection, support, and relief are needed. However, it is possible that for men, gender role socialization may intervene in the process of receiving support from their women partners. Men are more likely to be socialized to be independent and to conceal their need for closeness as compared to women; it is possible that this higher need for the appearance of self-sufficiency may prevent men from utilizing their women partners as a safe haven.

It is also possible that gender differences may be better explained by gender role socialization that often leads men in modern society to see themselves as less responsible for caregiving tasks than their women partners. This would explain why men have a general tendency to have lower mean scores on narrative representational knowledge about baby’s morning, doctor’s office, and camping trip than women and significantly lower scores on the accident narratives. Alternatively, men who use general representational knowledge probably would use this information less in their current relationship with their partners.

Past research linked attachment insecurity with perpetration and victimization of violence. However, studies conducted on attachment insecurity are generally not gender-specific and tend to study men’s violence toward men (Mikulincer & Shaver, 2007; Posada & Pratt, 2008). Overall, these studies indicate insecurely attached (particularly anxiously attached) individuals’ inclination to couple violence as shown in two ways. First, compared to nonviolent individuals, people who engage in violent behaviors tend to score higher on attachment anxiety. Second, violent men with higher anxiety scores tend to report more severe and frequent aggressive behaviors, along with more frequent coercive behaviors at the time of conflict. Furthermore, these associations are mediated by ineffective conflict management (Roberts & Noller, 1998) and attempts to control a partner’s behavior (Follingstad, Bradley, Helff, & Laughlin, 2002). Therefore, more research is needed to elucidate the relationships between these constructs.

Finally, it is also possible that some of the inconsistency in findings with the past research might be the application of an alternative measurement tool to assess attachment insecurity. We utilized narrative representational knowledge to assess attachment security, while much past research utilized self-report questionnaires. Related to this issue, much past research also has more consistent findings for the anxious attachment than the avoidant attachment in the perpetration and victimization of couple violence.

Since secure base narrative representations provide scores on only secure and insecure dimensions, a higher score means higher security, while a lower score means higher insecurity. It does not provide a meaningful anxiety and avoidance comparison. Further investigation is needed to determine whether anxiety or avoidance has a role in understanding couples’ victimization and perpetration of intimate partner violence, as the analysis conducted included secure base representational knowledge rather than traditional anxiety and avoidance dimensions of attachment. This will also help in future research to investigate whether men are more avoidant, and whether this gender difference of actor and partner effect of violence perpetration and victimization is due to avoidance. Finally, it will be important to investigate further how attachment insecurities mediate conflict management.

In the current study, men reported experiencing more frequent victimization than women. Research examining the gender of the perpetrators of intimate partner violence is contentious; while some large reports have found IPV perpetration is roughly gender symmetrical (e.g. Archer, 2000), many have argued that these reports lack nuance and/or neglect the broader cultural and socio-political context where women universally have less rights than men. For instance, a study by Allen, Swan, and Raghaven (2009) found comparable numbers of men and women perpetrate and are victimized in their relationships with intimate partners. Yet, further analysis of the path models suggested that women’s violence tended to be in reaction to male violence; men initiated violence, and their women partners responded in turn.

The field of IPV research is convoluted and controversial. Depending on the origins of the data, vastly different conclusions may be reached: crime victimization studies find IPV is rare, serious, escalates over time, and is primarily perpetrated by men; however, family conflict studies find IPV is common, seldom dangerous, stable over time, and is perpetrated equally by women and men (Kimmel, 2002). Studies demonstrating gender symmetry in IPV victimization consistently find that the consequences of IPV are more severe for women than men, with women suffering greater injury, fear, and other negative physical and psychological outcomes (e.g. Whitaker et al., 2007).

Of particular note is that most family conflict surveys rely on the CTS; this measure retrospectively counts acts of violence but takes no account of the circumstances under which these acts occur. The CTS has been widely criticized for neglecting the context and motivation of relationship violence, including: who initiates the violence, the relative size and strength of the people involved, the nature of the relationship, and the intent to control or dominate a partner (e.g. Dobash, Dobash, Wilson, & Daly, 1992). The present study utilized the CTS-2, which was developed to address some of the original critiques, along with the EAQ.

Insistence on the absolute validity of IPV gender symmetry is nonsensical in light of two statistical anomalies. First, there is a dramatic disproportion of women in domestic violence shelters and hospital emergency care facilities due to their injuries. Second, claims of gender symmetry in relationship violence “must be squared with the empirical certainty that in every single other arena of social life, men are far more disproportionately likely to use violence than are women” (Kimmel, p. 1336, 2002). Alternatively, some believe an examination of the construct of gender and the question of why IPV is a gendered social problem deserve the investigation of scholars much more than the intractable debate over which gender is more violent (Anderson, 2013).

In conclusion, results from the current study fit within the family studies conflict paradigm and should be interpreted with sensitivity, in light of the aforementioned convolutions. Compassion to victims is not a zero-sum game, and violence against both men and women should be addressed and prevented. It is important to recognize the existence and needs of male IPV victims while simultaneously acknowledging that most relationship violence occurs within a cultural context historically rooted in the oppression of women. Future studies must investigate the complex factors that cause and sustain IPV through culturally competent methodology and interpretation.

Strengths and Limitations of the Study

Although there are a fair number of studies examining the association between attachment patterns and violence, these studies are limited to the association between attachment dimensions (anxiety and avoidance), or the attachment style (anxious-ambivalent, avoidant dismissive, avoidant fearful, and secure) and violence. This study is one of the few to examine the relationship between secure base and relationship violence. In fact, to the best of the authors’ knowledge, this is the only study using narrative representations to measure secure base phenomena among couples and examine its link to intimate partner violence. Another strength of this study is its utilization of couple data, which allows researchers to examine the relationship dynamic from a dyadic perspective. Finally, this study utilizes partner report data to prevent self-report bias. The study’s main limitation is that it utilizes a convenience sample of college students, which limits generalizability. It is important to extend this study by collecting additional data that spans a broader range of participants in terms of age, socio-economic status, sexual orientation, and marital status. Also, a larger sample size will improve the statistical power.

In conclusion, findings suggest that women who carry coherent, informed, and organized secure base representational knowledge may have less negative feelings and attitudes toward their partners. This effect was particularly stark for psychological aggression, physical assault, and degradation. Attachment insecurity can be considered a general vulnerability to mental disorders, and the current study contributes to the literature by shedding light on how this dynamic relates to IPV and gender. Women with insecure attachments in the current study were more vulnerable to being both the victim and the perpetrator of aggressive behavior. However, this effect was not observed among men.

When mental health providers are working with women victims and perpetrators of intimate partner violence, it is crucial to provide a secure base for them so that these women can explore healthier choices for their relationships. By providing a supportive and emotionally nurturing environment, therapists can help women to improve their relationship choices and reduce their vulnerability to aggression. The therapeutic framework gives women a solid resource to turn to during difficult times that is always accessible. As explained by Mikulincer and Shaver (2012), situationally heightening people’s sense of attachment security, such as in a therapeutic setting, reduces the likelihood and intensity of psychiatric symptoms and distress. For instance, a client’s positive appraisals of his or her therapist’s sensitivity and supportiveness predicted relief from depression and maintenance of therapeutic benefits over an 18-month period (Zuroff & Blatt, 2006). An understanding of attachment-based fears may also be useful in couples counseling if safety concerns are addressed, such as through an attachment-based safety plan (Schneider & Brimhall, 2014). Therapists can offer comfort and safety, eventually becoming loving and caring attachment figures to clients involved in IPV. As clients develop more positive self-representations and learn to regulate their emotions, problems in interpersonal relationships—including IPV—may be ameliorated.

Appendix

Doctor’s Office: Score 7

It was early in the morning and Tommy had just gotten on his bike to outside to ride around the street. There were kids down the neighborhood down the block playing basketball So he ride there to see what was going on in doing when he did this the basketball had bounced on the street make him run over it and he fell down. And he hurt himself badly enough to make his mother scared when she came by running because she heard him screaming. She saw that he was bleeding from his knees because he had fallen of his bike scratched them up. She picked him up in her arms running back to the house and put him in the car and take him to the doctor. When he was get to the doctor he was crying because his knees were all bloody he thought he is going to have scar for the rest of his life even though his mother was reassure him he was not going to. When he get to the doctor’s office doctor called him to the office and looked at him and hmm looked over and cleaned up the wound and gave him a shot just in case there is an infection in doing so this further made him cry because he did not liked shots and find it quite painful and scary. His mother..hmm.. knowing this brought his favorite toy along so he can hold it while right afterwards and doing so he stopped cry once they get back to the house he immediately get back on his bike and start running around, went down the block to play with to play with the kids down the block to play basketball.

Doctor’s Office: Score: 2

Tommy is riding his bike on a beautiful sunny day outside and he stumbles on a rock and falls down the mother hurries out to help because she hears him crying and then Tommy is complaining that his leg hurt and the mother is worried that Tommy broke his leg so they go to the doctor and the doctor examines his leg and determines that he broke his leg the hands him back the toy so Tommy stops crying Tommy holds the toy and he does stop crying the doctor tells Tommy he is going to give him a shot so Tommy starts crying again and the doctor gives him another toy and gives him a shot and Tommy stops crying.