Exploring Insecure Romantic Attachment and Justifications for the Use of Intimate Partner Psychological Aggression in Couples

Abstract

The attachment theory has commonly been used to examine intimate partner psychological aggression (IPPA), but few studies have examined its association with self-reported justifications for one’s own use of IPPA. Behaviors, including the use of IPPA, are influenced, maintained, and function within the context of their justifications, highlighting the importance of investigating these justifications to obtain a clearer picture of IPPA. This study examined whether insecure romantic attachment (i.e., attachment anxiety and attachment avoidance) in both partners of a couple was associated with their justifications for their own use of IPPA. A community sample of 81 mixed-sex couples who reported using IPPA in the last year completed self-reported questionnaires on adult romantic attachment and their justifications for their use of IPPA. Results of a path analysis based on the actor-partner interdependence model revealed moderate positive associations between attachment anxiety and one’s use of internal and external justifications for their IPPA perpetration in men and women. An unexpected dyadic positive association is discussed. These results suggest that the use of justifications for one’s use of IPPA may reflect a strategy of hyperactivation that could contribute to the cycle of psychological aggression. Uncovering the function of these justifications could provide important therapeutic benefits, which are discussed in the study’s implications.

Adult romantic relationships represent one of the most central relationships in adulthood. The pervasiveness of intimate partner aggression (IPA) offers a glimpse into a complex and dynamic experience reflecting the reality of numerous couples. Research has shed some light on the phenomenon of IPA; however, its sustained prevalence and detrimental consequences warrant further investigation of aspects related to IPA that received less empirical attention, such as reasons individuals provide to justify its use. Empirical research has demonstrated that justifications play a crucial role in the maintenance and escalation of negative behaviors (e.g., Mulder & van Dijk, 2020), highlighting its potential importance in the cycle of aggression. While research has demonstrated that insecure adult romantic attachment is robustly associated with the use of aggression (see Spencer et al., 2021), to the extent of our knowledge, no study has investigated whether it is also associated with the justifications for the use of intimate partner psychological aggression (IPPA). The current study examined whether insecure romantic attachment (i.e., conceptualized by dimensions of attachment anxiety and attachment avoidance) is associated with one’s own justifications for using IPPA in both partners of a couple. The contribution of this study to the literature is twofold. First, it contributes to knowledge on dyadic associations between adult romantic attachment and IPPA justifications, shedding some light on the role of these justifications in the cycle of aggression. Second, the findings provide clinical utility in the treatment of individual or couples who use IPPA during conflict.

The current study focused exclusively on IPPA, defined as behaviors intended to cause emotional and/or psychological harm, including verbal (e.g., insults) or non-verbal behaviors (e.g., destroying possessions) to belittle, coerce, isolate, or control the partner (Shorey et al., 2012). The lack of research focusing solely on psychological aggression could suggest a systemic problem of minimizing or normalizing IPPA as a form of abuse (Goldsmith &Freyd, 2005; Hannem et al., 2015). This could explain why victims of nonphysical aggression are less likely to consider themselves as having experienced abuse (Goldsmith &Freyd, 2005). In Canada, physical and sexual violence are considered crimes, while most forms of psychological aggression, other than threats of harm and criminal harassment (e.g., stalking), are not formally recognized as such (Government of Canada, 2019). Yet, IPPA is the most common type of abuse in romantic relationships, and tends to precede or co-occur with other forms of aggression, such as physical or sexual violence (Center for Disease Control and Prevention, 2019). While all forms of IPA can lead to serious impacts, IPPA has been shown to cause severe long-term damaging effects on individuals, which can correspond to, or supersede physical aggression (Jordan et al., 2010; Lawrence et al., 2009).

Justifications for the Use of Intimate Partner Psychological Aggression

Justifications can be defined as one’s explanation, rationalization, or interpretation of attributions for engaging in a behavior (Lowell, 2012). Post hoc justifications for negative behaviors are strategies used to reduce moral dissonance (Festinger, 1957; Mulder & van Dijk, 2020), characterized as conflicting or misaligned behaviors and morals. The moral dissonance of IPA perpetration is reflected in the conflicting notion of harming an individual that one has a loving or intimate relationship with. The more an individual’s actions conflict with their values, the more they experience the dissonance as distressing, feeling a greater sense of urgency to reduce it (Lowell, 2012; Tsang, 2002). Reducing the dissonance can be achieved by either changing one’s behavior or, more commonly, by changing the narrative or interpretation of the situation using justifications, often occurring without explicit awareness. Re-interpreting a situation using justifications relieves the dissonance more rapidly and less effortfully relative to behavioral changes and serves to protect or buffer from the guilt, shame, and anxiety of having violated moral standards (Lowell, 2012; Tavris& Aronson, 2020; Tsang, 2002). Theorists have highlighted the potential for justifications to create an “amplifying feedback loop” (p. 1; Lowell, 2012), in which the internalization of the justification leads to an amplification of the negative behavior and unavoidable repetition of the cycle (Lowell, 2012; Tenbrunsel, 1995). As this cycle repeats, the more the perpetrators are to distort their perception of the environment or of others’ behaviors to justify their actions (e.g., “they provoked me”; Tenbrunsel, 1995). Empirical research has supported the role of justifications in the maintenance and escalation of negative behaviors (Martens et al., 2010; Mulder & van Dijk, 2020). This suggests that IPPA is likely highly influenced, maintained, and functions within the context of their justification, highlighting the importance of investigating justifications as part of the cycle of aggression to obtain a clearer picture of IPPA.

Considering that individuals will provide justifications for a behavior when asked, the current study is particularly focused on exploring what type of justifications will be provided for the use of IPPA and how these justifications are associated with romantic attachment. While there are many ways to classify justifications, the current study conceptually classifies justifications as internal or external. Internal justifications for one’s use of IPPA include those that refer to an individual’s own attributes (i.e., malevolent intentions, fear of abandonment, depressive presentation, and personal traits). External justifications include those that pertain to others or relationships with others (i.e., disagreement with one’s partner, parental modeling, or in reaction to the partner’s behavior).

Attachment Theory

Formulated by Bowlby (1982), attachment theory suggests that children develop a behavioral system that orchestrates a series of responses to one’s caretaker to maximize the likelihood of survival. This attachment system evolves based on the caretaker’s ability to meet the child’s needs for protection, comfort, support, and relief. This ultimately shapes the child’s working models of self (e.g., as deserving love), and working models of others (e.g., as available, responsive; Mikulincer& Shaver, 2005). Analogous to the attachment system in children, adults typically turn to their romantic partner for support, protection, validation, and comfort (Brennan et al., 1998). The same system that fosters the bond between a child and their caretakers is responsible for the emotional and intimate bond found in adult’s romantic relationships, referred to as adult romantic attachment in adulthood.

Romantic attachment is conceptualized using a two-dimensional continuum of attachment anxiety and attachment avoidance (Brennan et al., 1998; Mikulincer& Shaver, 2016). Adults with high attachment anxiety and/or attachment avoidance are considered to have an insecure attachment to their partner, depicted by negative working models of self and/or others (Bartholomew & Horowitz, 1991). Adults with high attachment anxiety tend to have a negative view of themselves (e.g., feelings of unworthiness) and fear rejection and/or abandonment from their romantic partner in times of need (Mikulincer& Shaver, 2016). These individuals are likely to use hyperactivating strategies in response to these fears, such as drawing attention to intensified emotional reactions, a desire for excessive proximity, constant need for reassurance, and hypervigilance of any sign of their partner’s emotional distancing (Johnson, 2019; Mikulincer& Shaver, 2016). Adults who display elevated attachment avoidance experience discomfort with intimacy, struggle to rely on others, and consequently, repress their natural desire to seek support when in distress (Bartholomew & Horowitz, 1991; Mikulincer& Shaver, 2016). These individuals tend to use deactivating strategies to regulate their emotions, which aim to minimize distress and hinder emotions that would activate their attachment system. These strategies are used to divert attention away from emotion-eliciting information and promote excessive over-reliance to avoid being hurt by their partner (Bartholomew & Horowitz, 1991; Goncy& van Dulmen, 2016). Low levels of both dimensions of insecure attachment constitute a secure attachment and a positive model of self and others (Mikulincer& Shaver, 2016).

Using the Attachment Theory to Understand IPPA and its Justifications

Using the lens of attachment theory, IPPA can be understood as a maladaptive and unsuccessful strategy to cope with distress and desperate attempt to maintain the emotional bond with one’s partner (Bartholomew & Allison, 2006). The use of aggression to respond to unmet attachment needs has also been empirically corroborated (Mikulincer& Shaver, 2011; Spencer et al., 2020; Velotti et al., 2020). However, no studies, to the extent of our knowledge, have investigated the associations between attachment insecurity and the justifications for using IPPA exclusively, rather than combining it to other forms of partner abuse. Nonetheless, the attachment theory allows for preliminary hypotheses for how justifications may vary depending on an individual’s attachment orientation.

Attachment Anxiety. Based on attachment theory, adults with higher attachment anxiety would use aggression as an ineffective and desperate attempt to forcefully obtain greater proximity and restore closeness with their partner (Mikulincer& Shaver, 2011). The aggression likely results in the victimized partner distancing themselves, causing great distress in the perpetrator and fueling a pursuit of the partner and escalation of the aggression (Follingstad et al., 2002; Mikulincer& Shaver, 2011). Although both dimensions of insecure attachment have been associated with IPA perpetration, including IPPA, attachment anxiety has shown a stronger, more consistent association with IPA relative to attachment avoidance (Follingstad et al., 2002; Goncy& van Dulmen, 2016; Spencer et al., 2020; Velotti et al., 2020). Despite the lack of empirical support, the attachment theory could suggest an association between attachment anxiety and the use of internal justifications, in which the perpetrator justifies their use of IPPA as a result of internal factors rather than external factors.

Attachment anxiety has been empirically associated with negative models of the self (e.g., undeserving, defective) and tendencies to adulate others (Mikulincer& Shaver, 2005). This is congruent with the empirical literature supporting an association between attachment anxiety and the tendency to direct shame, anger, and criticism towards the self when in conflict with their partner (Mikulincer& Shaver, 2005; Nisenbaum& Lopez, 2015). Directing negative emotions (e.g., anger, criticism) towards the self, rather than towards their partner during a conflict allows anxious individuals to channel their natural tendency of hyperactivation while maintaining the connection to their partner, and without threatening their fundamental need for reassurance and validation from their partner (Mikulincer& Shaver, 2005). Moreover, individuals with high attachment anxiety can have greater dependence to their partner that can lead to over-empathizing with their partner’s real or imagined distress, unbalanced self-other differentiation, or difficulties differentiating their partner’s welfare from their own, which can aggravate their natural tendency to direct negative emotions and justifications towards themselves during conflicts (Mikulincer& Shaver, 2011). This can be understood as internal justifications, or justifications related to the self (Merriam-Webster, 2020b).

Attachment Avoidance. For adults with elevated attachment avoidance, aggression can be used to push their partner away when levels of intimacy with their partner have exceeded what they can tolerate (Mayseless, 1991). Contrary to attachment anxiety, preliminary hypotheses based on theory suggest a potential association between attachment avoidance and the use of external justifications, in which the perpetrator justifies their use of IPPA based on external factors (e.g., the partner’s behaviors). Attachment avoidance has been associated with self-preserving positive models of the self and a negative model of others as undependable (Mikulincer& Shaver, 2005). Motivated by the fundamental tendency to preserve their model of self and distance themselves from emotionally activating content, individuals with higher attachment avoidance might be more likely to reject acknowledgment of personal responsibility for their partner’s distress (Mikulincer& Shaver, 2011; Pistole&Arricale, 2003). This mechanism of self-preservation would lead them to re-direct blame towards their partner rather than themselves (Cassidy & Shaver, 2002; Nisenbaum& Lopez, 2015). Corroborating this, studies have shown that higher levels of attachment avoidance were associated with less self-blame during conflicts, less empathy for the partner, and greater hostility, vengefulness, resentment, and repression of anger towards the romantic partner (Brennan et al., 1998; Péloquin et al., 2011; Shaver &Mikulincer, 2002). This can also be understood as external justifications or justifications deriving from sources outside of the self (Merriam-Webster, 2020a).

Romantic Attachment and Partner Associations

Dyadic hypotheses are postulated within the premise that conflicts within a romantic relationship would activate both partner’s attachment system, impacting each partner’s responses to one another (Bartholomew & Allison, 2006). For example, it would be cogent to hypothesize that higher attachment avoidance or anxiety in individuals could be associated with their partner’s internal or external justifications of IPPA. The deactivation and lack of emotional responses during conflict might either heighten their partner’s tendency to blame themselves for their own use of IPPA or heighten their partner’s tendency to blame others. Similarly, an individual’s overt emotional responses could activate their partner’s tendency to blame themselves, or to distance themselves from emotional content could be achieved by deflecting responsibility for their own use of IPPA on external factors (Bartholomew & Allison, 2006). Testing these hypotheses empirically might uncover underlying patterns of justifications for the use of IPPA that could point to their different functions based on attachment representations. It is of utter importance to clarify that investigating IPPA in a dyadic context does not suggest that victims are responsible for the aggression they receive or the justification that their partner uses to justify their aggression. Perpetrators of aggression are responsible for their actions under all circumstances. Rather, dyadic research suggests that insecure attachment can be a risk factor for IPPA because the perception of others as unreliable for those with higher attachment avoidance, and the fear of abandonment for those with higher attachment anxiety, results in additional challenges for seeking help or leaving an abusive relationship (see Velotti et al., 2020).

Objectives and Hypotheses of the Study

Relying on the actor-partner interdependence model (APIM; Kenny et al., 2006), this study aimed to expand the current literature on IPPA by investigating the association between internal and external justifications of its use and insecure romantic attachment (i.e., attachment anxiety and attachment avoidance) in both partners. In line with the theoretical framework of attachment and with empirical research, it was expected that higher attachment anxiety would be associated with one’s greater use of internal justifications and higher attachment avoidance would be associated with one’s greater use of external justification for the use of IPPA (i.e., actor effects). As for partner effects, we hypothesized that higher attachment anxiety in one participant would be associated with greater use of external justification by their partner, and higher attachment avoidance in one participant would be associated with greater use of internal justification by their partner.

Method

Participants

This study included a final sample of 162 Canadian participants (81 mixed-sex couples). Eligibility criteria included (a) being 18 years of age or older, (b) being in a mixed-sex romantic relationship for a minimum of 12 months, and (c) cohabitating with their partner for a minimum of 6 months. Criteria (b) and (c) were included to increase the likelihood that participants were in a stable romantic relationship. Over half of the participants (53%) identified as common law, 36.9% as married, and 10.1% as dating (i.e., in a romantic relationship without legal status). On average, the participants had been involved in their current romantic relationship for approximately 6 years (SD = 6.62) and cohabitating for approximately 4.5 years (SD = 7.72). Approximately 82% of the sample reported having no children, 10% reported one child, 3% reported two children, and 5% reported having three or more children. The average age of participants was 34 (SD = 14.62). Most participants reported a university-level education (58.3%), followed by high school (22.6%), general and vocational college (18.5%), and elementary school (0.6%). Most participants reported their main occupation as white-collar employment (61.9%), 11.9% as a student, 8.3% as blue-collar employment, 3% as stay at home, 2.4% as self-employed, 1.8% as being unemployed, and 10.7% as other (e.g., maternity leave, retired, on disability). Participants reported an average yearly income of CAN$49,671.08 (SD = $32,053.73). As for self-identified racial or ethnic background, 69% identified as White, 1.2% as Black, 1.2% as Latino/Hispanic, 0.6% as Asian, 0.6% as Indigenous, and 27.4% as “other.”

Procedure

The cross-sectional data were gathered in the context of a larger 3-year longitudinal research, consisting of three time points separated by 12 months. This study includes data from the second time-point exclusively because the Justification for Partner Psychological Aggression Scale (JPPAS) was administered at this time point only (Lafontaine et al., 2021). Participants who did not report using psychological aggression in the past year as measured by the screen items on the JPPAS (i.e., seven couples) and those who had completed three items or less from the 24-item JPPAS (i.e., three couples) were excluded from further analyses. Participants were recruited in the general community of a large Canadian city, through advertisements in local newspapers, posters, and brochures posted at a university campus and public facilities (e.g., bookstores, child-care centers, community centers); during university class presentations; on radio advertisements; and at wedding shows. Among the final sample, 16 couples were recruited among participants who completed the first time-point of the study and 68 couples were recruited as part of a new data collection to increase the sample size of the second time-point and larger longitudinal research.

Research assistants informed each participant about the purpose of the study and their right to withdraw from participation confirmed each partner’s individual willingness to participate and offered a list of resources. After informed consent was provided, couples spent approximately two and a half hours at an eastern Canadian university to complete a paper copy of the questionnaire package. The questionnaires used in this study took participants approximately 10–15 minutes to complete. The order of the questionnaires in the questionnaire package was counterbalanced to control for the potential effect of previously administered measures. Each partner was provided with a separate workspace to independently complete the questionnaire package. To maximize participation rate, couples who were unable to participate in person were provided the option of completing the questionnaire package at home by mail or via an encrypted web-based link to the questionnaire package. Participants who completed the questionnaire package at home were instructed to complete this package alone, without consulting their partner. The study was approved by the university’s Research Ethics Committee.

Measures

Socio-demographic questionnaire. The sociodemographic questionnaire included demographic information (e.g., age, gender, ethnicity, education) and information pertaining to the participant’s romantic relationship with their partner (e.g., length of relationship, cohabitation, marital status).

Romantic attachment. The Experiences in Close Relationships Scale-12 (ECR-12; Lafontaine et al., 2016) is a 12-item version of the original ECR (Brennan et al., 1998), which is a validated self-report measure of romantic attachment. The ECR-12 assesses attachment anxiety (6 items; e.g., “I worry a fair amount about losing my partners”) and attachment avoidance (6 items; e.g., “I do not feel comfortable opening up to romantic partners”). Each item is rated on a 7-point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree). Higher averaged scores are indicative of greater attachment anxiety and attachment avoidance. Convergent and predictive validity of this measure have been shown by Lafontaine et al. (2016) in a diversity of samples, including mixed-sex couples and same-sex individuals, clinical and community samples, and French- and English-speaking participants. These studies have yielded Cronbach’s alpha coefficients that range from .78 to .87 for the attachment anxiety subscale and .74 to .83 for the attachment avoidance subscale, indicating good internal consistency. Similarly, our reliability analysis yielded Cronbach’s alpha coefficients of .87 and .85 for attachment anxiety in women and men, respectively, and .86 and .83 for attachment avoidance in women and men, respectively.

Justification for the use of IPPA. The Justification for Partner Psychological Aggression Scale (JPPAS; Lafontaine et al., 2021) was used to assess IPPA justifications. It includes a list of 18 psychologically aggressive behaviors (i.e., screen items) and 24 possible reasons for using these behaviors(i.e., justifications). Participants were asked to report whether they used each of the 18 behaviors in the past year according to a dichotomous scale (0 = no, 1 =yes). This was summed for a total score of diversity of IPPA acts in the past year, ranging from a minimum score of 0 and maximum score of 18 different acts; higher scores are indicative of greater diversity of IPPA acts. If at least one of these behaviors was reported in the past year, participants were asked to complete a list of 24 items regarding their reason for using the aggressive behavior(s). The JPPAS assesses seven types of justifications for IPPA acts: disagreement with the partner (4 items), in reaction to the partner’s behavior (4 items), parental modeling (3 items), malevolent intentions (4 items), fear of abandonment (3 items), depressive presentation (3 items), and personal traits (3 items). Items are evaluated on a 9-point Likert-type scale, ranging from 1 (not at all true) to 9 (completely true), with higher scores indicating a greater likelihood of using that justification. The seven subscales’ scores were obtained by averaging their respective items. This measure demonstrated adequate internal consistency and construct validity in clinical and community samples as well as in French- and English-speaking participants (Lafontaine et al., 2021). In the current study, the Cronbach’s alpha coefficients produced by each scale for women and men, respectively, were: .80 and .85 for malevolent intentions, .63 and .72 for disagreement with partner, .95 and .88 for fear of abandonment, .82 and .76 for parental modeling, .76 and .84 for in reaction to partner’s behavior, .76 and .79 for depressive presentation, and .67 and .79 for personal trait. For this study, the four justifications subscales that referred to the participant’s own attributes (i.e., malevolent intentions, fear of abandonment, depressive presentation, and personal traits) were averaged to create an index of internal justifications (α = .86 for women, .86 for men). The three justifications subscales referring to external factors (i.e., disagreement with one’s partner, parental modeling, in reaction to the partner’s behavior), were averaged to create an index of external justifications (α = .77 for women, .82 for men).

Analytic Strategy

Preliminary analyses were conducted using SPSS 24 software. Using path analysis (Olobatuyi, 2006) based on the APIM model allows to test the theoretical model while simultaneously assessing the associations between insecure romantic attachment (i.e., attachment anxiety and avoidance) and justifications for IPPA (i.e., internal and external justifications) in one individual (i.e., actor effect) and their partner (i.e., partner effect). Using this data analysis technique allows for the control of non-independence in couples to reduce both Type I and Type II errors (Kenny et al., 2006). The hypothesized model was estimated using the maximum likelihood method with robust standard errors with Mplus version 7 (Muthén&Muthén, 2012).

Results

Preliminary Analysis and Assumptions

Preliminary analyses were conducted to identify and evaluate missing data, outliers, and normality. The results of a Little’s Missing Completely at Random test conducted on the attachment and the justifications scores revealed that values missing were random, χ2 (107) = 45.235, p = 1.00. Less than 2% of values were missing, and since all scales are calculated using means, missing data was not imputed. The verification of z-scores demonstrated that there were univariate outliers and a Mahalanobis distance analysis revealed there were also multivariate outliers. Outliers were anticipated in the sample considering the nature of the justifications for IPPA scale, and were therefore left untransformed in the dataset. Lastly, assumptions of normality of data were verified using z-scores, descriptive statistics (e.g., skewness, kurtosis), and histograms, and revealed that the data was not normally distributed. Non-normality was corrected using the maximum likelihood robust for the standard error against non-normality as integrated in the APIM Mplus7 syntax.

A linear mixed model was utilized to test for potential significant differences on sociodemographic and outcome variables among returning participants from a prior study to the newly recruited ones. No statistically significant differences were found among the two waves of recruitment and participants were combined for all remaining analysis. Correlations were conducted to assess the potential for covariate variables that should be considered in the main analysis. Based on previous literature, variables depicted as relevant and potentially statistically significantly associated with the outcome variables used in the APIM were tested, including age, length of the romantic relationship, number of children with the current partner, and diversity of IPPA acts. The Pearson’s correlations were conducted separately for men and women to control for the non-independence of the data. Results revealed no significant effect of age, length of relationship, nor number of children with the current partner. Diversity of IPPA was significantly correlated with internal (r = .664, p< .001; .510, p< .001) and external justifications (r = .528, p< .001; r = .639, p< .001) for women and men, respectively. Diversity of IPPA acts for women and men were included within the model as covariates (see Figure 1).

Figure 1.

Results of the actor-partner interdependence model of men and Women’s romantic attachment and justifications for their use of intimate partner psychological aggression.Notes. E = Error variance, dotted line = non-significant results, solid line = significant results and gray parameters were removed from the analysis. Standardized coefficients are shown. The correlations among women and men’s attachment and women and men’s justifications can be found in Table 1.

Descriptive Statistics

Potential gender differences were also examined through paired t-tests on each study variable. Independent sample t-tests indicated no significant gender differences on scores of attachment avoidance, t(160) = 1.152, p = .251 and external justifications, t(160) = −1.446, p = .150. Significant gender differences on scores of attachment anxiety t(160) = −3.055, p = .003 and internal justifications, t(160) = −3.223, p = .002 were observed, with women obtaining greater scores in both attachment anxiety and internal justifications. The diversity of IPPA acts was assessed to better depict the context of the IPPA in the sample. On average, men reported having perpetrated 5.11 (SD = 2.91) whereas women reported having perpetrated 6.60 (SD = 3.19) different acts of IPPA in the past year, depicting a statistically significant difference, t (160) = −3.11, p = .002. Means, SDs, and inter-correlations among the main variables for men and women are presented in Table 1.

Table 1.

Intercorrelations, Means, and Standard Deviation Across Outcome Variables for Men and Women.

Measure	1	2	3	4	5	6	7	8
1. W. Attachment avoidance	1	.239 (.032)	.344 (.002)	.275 (.013)	.109 (.334)	.038 (.733)	-.021 (.854)	.049 (.665)
2. W. Attachment anxiety		1	.605 (.000)	.421 (.000)	.232 (.037)	.006 (.955)	.059 (.603)	.250 (.024)
3. W. Internal justifications			1	.671 (.000)	.205 (.066)	-.066 (.556)	.061 (.590)	.221 (.047)
4. W. External justifications				1	.303 (.006)	-.061 (.588)	.296 (.007)	.361 (.001)
5. M. Attachment avoidance					1	.143 (.202)	.181 (.105)	.254 (.022)
6. M. Attachment anxiety						1	.499 (.000)	.299 (.007)
7. M. Internal justifications							1	.630 (.000)
8. M. External justifications								1
M (SD)	2.38 (1.15)	3.52 (1.56)	2.77 (1.51)	3.89 (1.40)	2.58 (1.15)	2.80 (1.44)	2.09 (1.16)	3.58 (1.40)

Note. The p-value for each coefficient is presented in parentheses. M = men; W = women. Table 1 results were produced by SPSS but were consistent with the results obtained in the overall APIM in Mplus.

Actor-Partner Interdependence Model

In APIM, a dyad is considered “distinguishable” when there is a characteristic differentiating its members, such as gender in mixed-sex couples (Kenny et al., 2006). While gender is theoretically distinguishable, distinguishability of dyads was tested empirically to respect principles of parsimony and assess whether gender was needed in the model (Kenny et al., 2006). The test of distinguishability yielded a significant chi-square statistic, χ² (20, N = 162) = 98.93, p< .001, demonstrating that the sample is statistically distinguishable by gender. Considering the small sample size and the fully saturated APIM, the three parameters with the highest p-value and therefore the least significant, were removed to free additional degrees of freedom and produce a better model fit. The three parameters removed from the final analysis can be identified in Figure 1. Fit indices were examined to evaluate the model’s goodness of fit (i.e., the degree to which the data fit the hypothetical model): χ2 (7, N = 162) = 10.660, p = .154, CFI = 0.985, TLI = 0.936, RMSEA = .080, and SMRS = .028 (Weston & Gore, 2006). Controlling for diversity of IPPA, the APIM yielded four statistically significant actor effects and one significant partner effects (see Figure 1). For both men and women, a statistically significant association was observed between attachment anxiety and the use of internal justifications (i.e., moderate effect for both genders), and attachment anxiety and the use of external justifications (i.e., small effect for both genders; Mukaka, 2012). No actor effects were found between attachment avoidance and justifications for either gender. As for partner effects, a small positive association was found between men’s attachment avoidance and women’s use of external justification. No other partner effects were found.

Discussion

This dyadic study examined the association between insecure romantic attachment and internal and external justifications for partners’ own use of IPPA within mixed-sex couples to better understand the underlying ways of thinking associated with psychological aggression. When examining the model’s results, it is useful to keep in mind that the sample consisted of couples characterized with bidirectional aggression, in which both partners were perpetrators and victims of at least one act of IPPA in the last year. Taking into account the diversity of IPPA, the most robust associations in the tested analytical model are consistent with the hypotheses in which men and women with higher levels of attachment anxiety are more likely to justify their acts of IPPA as a result of their own being (e.g., malevolent intentions, fear of abandonment, depressive states, and personal traits). These results are consistent with the theoretical literature, in which directing negative emotions towards the self allows those with greater attachment anxiety to act on the urge to use hyperactivating strategies without compromising their connection to their partner. For the individuals with greater attachment anxiety, directing blame towards their partner would otherwise further distance them from the possibility of their partner meeting their fundamental need for reassurance and love (Mikulincer& Shaver, 2005). It is also plausible that the intense contrast of desperately needing closeness from one’s partner yet using IPPA, which likely pushes them away, would cause greater dissonance in individuals with higher attachment anxiety relative to those with higher attachment avoidance. Research has shown that greater moral dissonance is accompanied by an increased urgency to reduce it via behavioral changes or, more commonly, the use of justifications (Lowell, 2012; Tsang, 2002). As justifications are used and internalized, the maintenance and escalation of the behavior is reinforced (see Spencer et al., 2021 and Velotti et al., 2020, for reviews).

The association between attachment anxiety and external justifications, in addition to the lack of association between attachment avoidance and external justifications in men and women were discordant with the study’s hypotheses. One explanation for these results could be that the use of justifications, internal or external, reflects a strategy of hyperactivation. Based on the attachment theory, individuals with greater attachment anxiety tend to use hyperactivating strategies (e.g., overt expressions of emotions) during conflict with their romantic partner to meet their attachment needs. Justifying one’s actions could be understood as an overt response fundamentally embedded in the acknowledgment of an emotion or emotional need. For those with higher attachment avoidance, the implication of an emotional involvement contradicts their natural tendency to minimize their distress and divert attention away from emotion-eliciting information, which opposes their goal of being completely self-reliant (Cassidy & Shaver, 2002; Mikulincer& Shaver, 2016). As such, they might be less likely to endorse any type of justifications to justify their aggression toward their partner. As observed in Table 1 and Figure 1, some significant bivariate correlations, such as those between attachment avoidance and justification, are no longer meaningful when produced by the APIM. This is explained by the model’s ability to account for interdependence, the variance between, as well as within dyads.

While the size of the association is considered small, the model yielded an interesting partner-effect between men attachment avoidance and women’s use of external justification for their own IPPA. This could indicate that men with higher attachment avoidance are associated with women who uses external justification for their use of aggression, such as attributing blame to men’s use of deactivating strategies (e.g., withdraw from conflict, push partner away). These results are consistent with scientific literature on attachment, conflict perceptions, and couple’s satisfaction in which men’s avoidance predicted women’s experience of conflict and relationship satisfaction (Brassard et al., 2009). This pattern of results could be a response to North American gender norms in which men are socialized to be self-reliant and less emotionally expressive, traits related to attachment avoidance (Mondor et al., 2011).

Limitations and Future Research

The present study has limitations that researchers should address in future research. As dyads in this study have shown to be distinguishable by gender, exploring this model within same-sex couples would be a highly relevant future inquiry that would contribute significantly to the body of empirical literature. Future research may also wish to design recruitment techniques with the specific intention of enlisting a more ethnically diverse and representative sample of participants to bolster the generalizability of research results. The cross-sectional research design also impeded our capacity to draw causal interpretations on the nature of the associations between insecure romantic attachment and justifications for the use of IPPA. Although attachment theory has established that individuals develop attachment bonds prior to developing aggression-related behaviors (i.e., including justifications), it would also be plausible for the justifications for the use of IPPA to influence the quality of attachment. The study’s statistically significant associations can be used as a stepping stone to investigate, with a larger sample, potential partner pairings or causal relationships between adult attachment and justifications for the use of IPPA in future research. The scales disagreement with the partner and personal traits produced lower Cronbach’s alpha and were further investigated. An item reliability analysis did not produce significant enough results to justify removing items to increase the coefficients. Lafontaine et al., (2021) have used reliable techniques (i.e., factor analysis) to evaluate the internal consistency of the scales, the lower alpha can be explained by the short lengths of the scales (e.g., 3–4 items; Tavakol&Dennick, 2011). Lastly, the sample size does not meet the recommended 100 participants, which increased risk of potential type II errors, limiting the study’s statistical power and impacting the validity of the APIM chi-square. Nonetheless, APIM remained the most appropriate analytic strategy to test the presented model as it accounts for both variance between and within dyads, justifying the suitability of its use. A larger sample size would allow for the expansion of the model to investigate the seven subscales of the JPPAS. Future research may also consider investigating the effect of matching partners on dimensions of attachment to uncover potential patterns with use of justifications for IPPA.

Implications and Conclusion

Partner aggression reflects a worldwide public health issue warranting further empirical evidence for its prevention and treatment. The results of this study stress the importance for clinicians to include attachment-based assessment and interventions when working with men and women who have perpetrated IPPA. This study suggests that attachment anxiety may be one of the contributing factors that increase the use of internal or external justifications for the use of IPPA. From an attachment standpoint, justifications, like the use of aggression, potentially reflect a strategy of hyperactivation that unskillfully aims to navigate unmet attachment needs and maintain an emotional bond with a romantic partner. Rather than sooth, facilitate emotional connection, or promote support seeking with their partner, the use of justifications could instead contribute to the escalating cycle of psychological aggression. To intervene with this cycle, treatment should aim to promote greater attachment security by increasing awareness and supporting change in the ways to regulate attachment-related distress from non-optimal to optimal or adaptive strategies. This shift can consequently shape new ways of thinking, new behaviors, and new patterns of interactions and restore the ability to connect with others, an essential pillar of wellbeing (Johnson, 2019; Stevens, 2014). By moving closer to a secure attachment, the individual will experience less urges to use strategies aimed at minimizing the dissonance between one’s values and hyperactivating behaviors, such as justifications. Based on the postulate that justifications foster and nurture negative behaviors through a feedback loop of escalation, the less justifications are used, the less IPPA should be perpetrated within an intimate relationship (Lowell, 2012; Mulder & van Dijk, 2020). Uncovering the function of the justifications and bringing these cognitive processes to light can assist individuals in recognizing their contribution to the cycle of psychological aggression, as this insight can be used by clinicians and their client to facilitate prevention and treatment.

Author Biographies

Marie-Eve Leclerc, MSc, is a MA-PhD candidate in Clinical Psychology, Faculty of Social Sciences, at the University of Ottawa. She has a Master’s degree in Psychiatry, faculty of Medicine from McGill University and a Bachelor of Psychology from Concordia University. Her research interests focus on attachment, couples, and trauma.

Marie-France Lafontaine, PhD, is a Full Professor and the Director of the Couple Research Laboratory in the School of Psychology at the University of Ottawa. She studies couple dynamics using attachment theory. Her research projects aim to understand the factors related to the success of romantic relationships, including relationships with specific problems (e.g., attachment injuries, intimate partner violence, chronic pain).

Audrey Brassard, PhD, is a Full Professor in the Department of Psychology at the Université de Sherbrooke. Her research and clinical interests include interpersonal relationships, romantic attachment, intimate partner violence, conflict, and sexuality.

Katherine Péloquin, PhD, is an Associate Professor in the Department of Psychology at the Université de Montréal. She uses the attachment theory to inform her research on different themes related to couple relationships. Her work primarily focuses on partner support, relationship functioning, intimate partner aggression, dyadic stress management, sexuality, and infertility. She also develops interventions designed to optimize couples’ well-being.

ORCID iD

Marie-Eve Leclerc https://orcid.org/0000-0001-6420-9151