Attachment and Breakup Distress: The Mediating Role of Coping Strategies

Abstract

Breakups are common among emerging adults and are associated with elevated depressive and anxiety symptoms, especially in the presence of attachment insecurities. Previous authors have suggested that inadequate coping strategies might explain this association, yet this has not been examined longitudinally. This study examined the mediating role of five coping strategies (self-help, approach, accommodation, avoidance, self-punishment) in the longitudinal associations between attachment insecurities (anxiety, avoidance) and depressive and anxious symptoms in 196 emerging adults experiencing a romantic breakup. Measures of pre-breakup attachment, post-breakup coping strategies (one-month post-breakup), and depressive and anxiety symptoms (one- and three-month post-breakup) were administered. Results from a longitudinal autoregressive cross-lagged model showed that pre-breakup attachment insecurities were related to higher depressive and anxiety post-breakup symptoms through higher use of self-punishment and lower use of accommodation coping strategies. Findings highlight coping strategies as potential intervention targets to promote the recovery of emerging adults experiencing breakup distress.

Introduction

Between the ages of 18 and 25, emerging adults often engage in several consecutive romantic relationships, which eventually allow them to determine what they wish for in a partner (Arnett, 2004; Shulman & Connolly, 2013). The relationship instability characterizing this period translates into a high frequency of romantic breakups (Norona & Olmstead, 2017). In a prospective study, Lantagne et al. (2017) observed that more than a third (36.7%) of emerging adults experienced a breakup over 12 months. A romantic breakup represents one of the most challenging life experiences (Scully et al., 2000) associated with increased psychological distress and a decline in life satisfaction (e.g., Preetz, 2022). Considering that early adulthood is a critical period for developing mental health problems and romantic breakups are known triggers (e.g., Monroe et al., 1999), a better understanding of the risk factors contributing to breakup distress is warranted. Several cross-sectional studies have identified attachment insecurities as a direct risk factor for breakup distress (e.g., Brassard et al., 2018), while some authors (e.g., Leung et al., 2011) have suggested inadequate coping strategies might be involved. Yet, to our knowledge, no longitudinal studies have examined this proposition, and none have collected pre-breakup measures. Several authors thus emphasize the need to better understand the associations between attachment and post-breakup distress to identify avenues of intervention likely to promote the recovery of emerging adults (e.g., Brassard et al., 2018; Davis et al., 2003).

Breakup Distress

Romantic breakups can affect mental health and result in various psychological symptoms, such as depressive symptoms (see Field, 2017, for a review), including dysphoric mood and affects and the loss of motivation and interest in life (Derogatis, 1993). A longitudinal study conducted by Keller et al. (2007) revealed that 19.6% of the 1731 participants whose symptoms met the criteria for major depression in the last year identified that the main cause was a romantic breakup. Verhallen et al. (2019) found that as many as 26.8% of those who experienced a breakup in the past six months reported depressive symptoms. Romantic breakup can also cause or exacerbate anxiety (e.g., Boelen & Reijntjes, 2009), which includes symptoms of nervousness, tension, panic attacks, and feelings of terror (Derogatis, 1993). Carter et al. (2018) found that 29.7% of university students reported anxiety symptoms following a breakup.

Although post-breakup distress is frequent, most declines in mental health following a breakup of non-cohabiting partners would be temporary (i.e., less than a year Preetz, 2022). The intensity of distress would also vary according to specific characteristics, namely relationship duration, betrayal, time since breakup (Field et al., 2009), breakup initiator (Carter et al., 2018; Davis et al., 2003; Field et al., 2009), breakup unexpectedness, and involvement in a new relationship (Field et al., 2009). To identify who would be most at risk for higher or prolonged post-breakup distress, Sbarra (2015) suggested turning to attachment theory.

Attachment and Breakup Distress

Attachment theory postulates that representations of the self and others formed through repeated interaction with attachment figures during childhood remain, for most people, relatively stable in adulthood (Fraley & Brumbaugh, 2004). Brennan et al. (1998) conceptualized adult attachment insecurities as varying along two continuums. Attachment anxiety is characterized by a negative self-image, excessive doubts about one’s self-worth, chronic fear of abandonment, and a constant search for proximity and reassurance. In response to threat, or when the attachment figure cannot be reached, individuals with a high level of attachment anxiety tend to adopt hyperactivating strategies, which involve seeking attention through repeated demands for reassurance and closeness and amplifying their distress (Mikulincer & Shaver, 2016). Attachment avoidance represents a discomfort with emotional intimacy, a preference for self-reliance, and a relational distrust based on a negative view of others. When a threat is perceived, individuals with high levels of avoidance tend to adopt deactivating strategies, which consist of denying their attachment needs, avoiding dependence on others, and detaching themselves from negative emotions or thoughts likely to activate their attachment system.

Several authors have highlighted the relevance of studying attachment in the context of romantic breakups. Indeed, past studies have consistently shown that attachment anxiety is associated with greater breakup distress, whether it is measured as depressive symptoms and suicidal ideation (Brassard et al., 2018), physical and emotional distress (Davis et al., 2003), or depressive symptoms and negative affects (Fagundes, 2012). Results regarding attachment avoidance are less consistent. Some studies have found no associations (e.g., Brassard et al., 2018; Davis et al., 2003), whereas another found that divorcees with attachment avoidance reported a higher distress level, similar to those with attachment anxiety (Birnbaum et al., 1997). Bourassa et al.'s (2019) longitudinal study revealed that attachment avoidance, but not attachment anxiety, predicted long-term distress 4.5 years after a marital separation. The authors suggested that the hyperactivating strategies typical of attachment anxiety may result in short-term distress, whereas the deactivating strategies typical of attachment avoidance may be related to less short-term distress but worse long-term outcomes.

Many researchers have emphasized the need to study the mechanisms underlying the association between attachment insecurities and breakup distress (e.g., Brassard et al., 2018; Davis et al., 2003). In this regard, Simpson and Rholes’ (2012) theoretical attachment-diathesis-stress model suggests that individuals with distinct attachment orientations respond differently to distressing situations such as a breakup due to specific motivations and coping strategies stemming from attachment insecurities. This model postulates that the higher level of distress of individuals with attachment anxiety can be explained by their unsuccessful attempts to reduce their distress through hyperactivating strategies and emotion-focused coping (e.g., rumination). Conversely, the higher distress of individuals with an avoidant attachment may be explained by their attempts to contain their distress individually through deactivation strategies and avoidance coping. This model supports the relevance of studying coping strategies as an explanatory factor for the links between attachment and breakup distress.

Coping With Breakup Distress

Coping strategies are defined as cognitive and behavioral efforts to manage, tolerate, and reduce stress (Lazarus & Folkman, 1984). Carver et al. (1989) defined 16 coping strategies that Zuckerman and Gagne (2003) grouped into five categories: (1) self-help coping (expressing emotion, emotional and instrumental support seeking) aims to maintain emotional well-being under stress; (2) approach coping (active coping, planning, suppression of competing activities) represents problem-solving strategies; (3) accommodation coping (maintaining optimism, acceptance, positive reframing, replacement) denotes attempts to cope with adversity when the problem cannot be resolved; (4) avoidance coping (denial, mental and behavioral disengagement, other-blame) aims to move away from the problem; and (5) self-punishment coping (self-blame, self-focused rumination) orients the individual towards negative thoughts and emotions. Self-help, approach, and accommodation are adaptive coping strategies that negatively correlate with depressive and anxiety symptoms, while avoidance and self-punishment are maladaptive coping strategies that positively correlate with these symptoms (Zuckerman & Gagne, 2003).

Attachment, Coping, and Breakup Distress

Some studies have suggested that breakup distress in individuals with higher attachment anxiety can be explained by maladaptive coping strategies, such as blaming oneself for the breakup or protesting and ruminating about the lost relationship (Choo et al., 1996; Davis et al., 2003). Fagundes’ (2012) longitudinal study reveal that participants with high levels of rumination report smaller emotional improvements one-month post-breakup when they also have a high level of attachment anxiety. In contrast, attachment avoidance is related to the use of self-reliance (e.g., low social coping) and avoidance coping (e.g., avoiding ex-partner and new relationships) after a breakup (Davis et al., 2003), the latter being associated with greater overall post-breakup distress (Wrape et al., 2016). Several studies have also linked attachment avoidance to distancing coping strategies, such as denial, distraction, and behavioral or cognitive disengagement (see Mikulincer & Shaver, 2016, for review). The type of coping used by avoidant individuals may temporarily mask the distress, resulting in low short-term distress but potentially higher long-term distress.

Only a few cross-sectional studies have examined the mediating role of coping strategies in the links between attachment and breakup distress in emerging adults. Saffrey and Ehrenberg (2007) observed, among 231 university students who had experienced a recent breakup, the mediating role of rumination in the links between attachment anxiety and lower breakup adjustment (global distress, depressive and anxiety symptoms). A study by Leung and al. (2011), conducted among 250 Chinese and 144 Australian young adults who recently experienced a breakup, revealed that avoidance and self-punishment coping explained the links between higher attachment anxiety and greater breakup distress, while accommodation coping explained the links between low levels of attachment anxiety and avoidance and lower breakup distress.

Present Study

Given the paucity of prospective breakup studies involving emerging adults and measuring attachment and coping while distinguishing psychological symptoms of depression and anxiety, the current study proposes to examine the mediating role of coping strategies (self-help, approach, accommodation, avoidance, self-punishment) in the prospective links between pre-breakup attachment insecurities (anxiety, avoidance) and post-breakup distress (depressive and anxiety symptoms) in emerging adults. The first hypothesis postulates that higher use of avoidance and self-punitive coping and lower use of accommodation coping at one-month post-breakup will mediate the link between pre-breakup attachment anxiety and higher breakup distress in the short term and medium term (one- and three-month post-breakup). The second hypothesis suggests that higher use of avoidance coping and lower use of accommodation coping at one-month post-breakup will mediate the link between pre-breakup attachment avoidance and lower breakup distress in the short term (one-month post-breakup) but higher breakup distress in the medium term (three-month post-breakup). Self-help and approach coping will be studied through an exploratory approach, although it can be expected that individuals with attachment insecurities will less likely use adaptive coping strategies. All hypotheses and research questions will consider baseline levels of distress (to isolate breakup distress from general level of distress, Preetz, 2022) and four empirically supported potential covariates: gender (gender differences were found in post-breakup distress, Carter et al., 2018; Field, 2017; Preetz, 2022), breakup initiator (being the initiator has been related to lower post-breakup distress, e.g., Carter et al., 2018; Hunt & Chung, 2012), presence of betrayal (related to higher distress, Field et al., 2009), and involvement in a new relationship (related to lower distress, Boelen & Reijntjes, 2009; Carter et al., 2018).

Method

Participants

This study was part of a larger prospective study on relationship trajectories among emerging adults approved by the Institutional Review Boards of two Canadian universities. We recruited a community-based sample of 1385 Canadians aged between 18 and 25 years old, involved in a romantic relationship, who understood French, and had access to the Internet. Participants were followed over 12 months, and the sample for this study comprises those who experienced a romantic breakup (14.2%) during that time. The final sample includes 196 participants (163 women, 27 men, 6 non-binary or transgender individuals) who were, on average, 21 years old (SD = 2.2). Most were born in Canada (92.3%) and 33% completed a high school diploma, 44% a college degree, and 19% a university degree. Most participants were full-time students (71.4%), while the others were employed (24.5%), unemployed (3.1%), or on medical leave (1.0%). The majority identified as heterosexual (74.0%), and a quarter identified either as bisexual (12.2%), pansexual (5.6%), homosexual (4.6%), in questioning (3.1%), or queer (.5%). Table 1 presents information about the breakup context and previous relationship. Before the breakup, most participants were in a relationship without cohabitation. Although most participants reported that they initiated the breakup, the majority said it was unexpected and that it was their first breakup with this partner. Most participants had no hope of re-establishing the relationship, yet they had ongoing contact with their ex-partner. At one- and three-month post-breakup, 24.3% and 40.1% reported being in a new relationship, respectively. Of these, 39.1% reported that it was with their last partner, and 60.9% with someone else. ¹

Table 1.

Characteristics of the participants’ prior relationship.

Variables	n	%
	M (SD)	(Range)
Relationship length at baseline (in months)	18.79 (32.51)	(1.00–108.00)
Baseline relationship status
Not living together	122	62.20
Cohabitating, not married	71	36.20
Polyamory	2	1.00
Married	1	.50
Breakup initiator
Participant	91	48.70
Partner	58	31.00
Both	38	20.30
Expected breakup
No	118	63.10
Yes	69	36.90
Ongoing contacts
No	59	31.60
Yes	128	68.40
Previous breakup(s) with this partner
No	131	70.10
Yes	56	29.90
Presence of betrayal
No	137	73.30
Yes	50	26.70
Hope to re-establish the relationship
No	125	66.80
Yes	62	33.20

Procedure

Participants were recruited during the COVID-19 pandemic from March 2020 to September 2020 using advertisements on distribution lists (e.g., scientific organizations) and social media (e.g., Facebook) in Quebec, Canada. Interested participants were directed to Qualtrics, an online survey system, to complete the eligibility questionnaire, a consent form, and the first online questionnaires. Every two weeks over a year, participants completed a short online survey that determined whether a romantic breakup had occurred and, if so, they completed questions about the breakup context. They were then followed every two weeks for an additional six months, during which relationship status, distress symptoms and coping strategies were assessed. Thus, this study followed a prospective design with three measurement times. Sociodemographic characteristics, initial psychological distress, and attachment insecurities were collected at baseline (pre-breakup T1), coping strategies were collected one-month post-breakup (T2), and distress symptoms were collected at both one- (T2) and three-months (T3) post-breakup. All participants were entered into a draw for a CAN$100 Amazon gift card. Those who completed half or the entire study received an electronic payment of CAN$75 or CAN$150 as compensation for their time.

Measures

A sociodemographic questionnaire (e.g., gender, sexual orientation, age), including questions about the nature of the relationship (e.g., relationship length and status), was administered. When a breakup occurred, information about the breakup (e.g., initiator of the breakup, presence of betrayal) was assessed. All questionnaires were in French and each subscale presented adequate model-based omega composite reliability (see Table 2).

Table 2.

Standardized Parameter Estimates From the Combined Measurement Model.

	Anxiety		Depression		Attachment anxiety		Attachment avoidance		Coping self-help		Coping approach		Coping accommodation		Coping avoidance		Coping S-P
	λ	δ	λ	δ	λ	δ	λ	δ	λ	δ	λ	δ	λ	δ	λ	δ	λ	δ
Item 1	.701**	.509	.808**	.347	.859**	.263	.790**	.376	.878**	.229	.892**	.204	.809**	.345	.882**	.222	.808**	.347
Item 2	.789**	.377	.753**	.433	.780**	.394	.668**	.554	.881**	.224	.435**	.811	.926**	.142	.985**	.029	.832**	.309
Item 3	.855**	.268	.716**	.487	.812**	.343	.776**	.398	.845**	.285	.139	.981	.868**	.246	.860**	.260	.902**	.187
Item 4	.792**	.373	.819**	.329	.685**	.527	.933**	.130	.868**	.247	.834**	.304	.709**	.498	.158	.975	.824**	.321
Item 5	.893**	.202	.905**	.180	.656**	.570	.748**	.440	.887**	.214	.911**	.170	.895**	.199	.541**	.708	.868**	.247
Item 6	.776**	.397	.863**	.255	.615**	.616	.789**	.377	.850**	.278	.894**	.200	.764**	.417	.763**	.417	.889**	.210
Item 7									.928**	.139	.694**	.519	.832**	.307	.743**	.448	.963**	.073
Item 8									.794**	.369	.519**	.730	.601**	.639	.431**	.814	.983**	.034
Definition	.801		.811		.735		.784	.866			.686		.801		.670		.884
ω	.916		.921		.877		.907		.960		.878		.936		.881		.967

Note. **p < .01. λ: factor loadings; δ: uniquenesses; ω: model-based omega composite reliability. S-P: self-punishment.

Attachment

The 12-item short French version of Experiences in Close Relationships (ECR-12: Lafontaine et al., 2016) was used to assess attachment-related anxiety and avoidance. The ECR-12 measures each dimension using six items on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Each score is calculated by averaging the items, with a higher score indicating a higher level of attachment insecurity. The reliability (α = .74 - .84) and factorial validity of the ECR-12 scales (through confirmatory factorial analyses) were established in six samples of French-Canadian adults (Lafontaine et al., 2016).

Coping Strategies

Coping strategies were measured using the revised version of the Coping Orientation to Problems Experienced (R-COPE; Zuckerman & Gagne, 2003), which was translated into French (Gehl & Brassard, 2019) for the current study using the back translation procedure (Vallerand, 1989). The R-COPE includes 40 items measuring five subscales of coping strategies (self-help, approach, accommodation, avoidance, self-punishment). Each subscale includes eight items rated on a four-point scale, ranging from 1 (I don’t do this at all) to 4 (I do this a lot). Respective item ratings are averaged to form five scores, with higher scores indicating more frequent coping strategy use. Zuckerman and Gagne (2003) demonstrated the convergent validity of the R-COPE through positive links with similar constructs (e.g., perceived control, procrastination) and supported its adequate reliability for all subscales (α = .81 - .92).

Depression and Anxiety Symptoms

Depression and anxiety symptoms were assessed by the French version of the Brief Symptom Inventory (BSI; Derogatis, 1993; Gosselin & Bergeron, 1993). Only the 12 items assessing depression and anxiety symptoms were used. Items are rated on a five-point scale describing how much the symptoms were distressing in the past week ranging from 1 (not at all) to 5 (extremely). Items are averaged to form global scores, higher scores indicating a higher level of symptoms. The French version of the BSI (Gosselin & Bergeron, 1993) showed adequate convergent validity with MMPI scales and reliability (depression: α = .85; anxiety: α = .81).

Data Analysis Strategy

The study variables were first checked for normality and extreme values using SPSS 27. Preliminary correlations were conducted to identify potential covariates among demographics, relationship- and breakup-related variables. Then, using MPlus8 (Muthén & Muthén, 2017), the measurement model’s adjustment and invariance through time were tested before conducting the main analyses to support the scales’ psychometric properties. To test our main hypotheses, we conducted an autoregressive cross-lagged (ARCL) model that examined the longitudinal associations between attachment, coping, and depressive and anxiety symptoms while considering the baseline level of these symptoms to account for intra-individual stability. This model allowed the identification of direct and indirect associations (using 1000 bias-corrected bootstrap samples and 95% confidence intervals) to test our mediation hypotheses. A robust weighted least square estimator with mean- and variance adjusted statistics (WLSMV) was used to account for the possible departure from normality in ordinal items in MPlus. The model adjustment was verified with four indices (Kline, 2016): a non-significant chi-square, a Comparative Fit Index (CFI) and Tucker-Lewis index (TLI) of .90 or more, and a Root Mean Square Error of Approximation (RMSEA) below .05. All measurement and predictive models were estimated using the full available information in the sample (Enders, 2010) via the missing data algorithm implemented in Mplus for WLSMV estimation (Asparouhov & Muthén, 2010).

Results

Preliminary Analyses

Preliminary correlations aiming to identify potential covariates revealed that gender (1 = female, 0 = others) was significantly related to post-breakup distress. The breakup initiator, the presence of betrayal, and the involvement in a new relationship were considered in an initial model, but none of these were found to be related to breakup distress (T2, T3). They were thus removed from the final model to maximize statistical power. Preliminary measurement models were tested (see supplementary Table). Tests of longitudinal measurement invariance were conducted on the distress measurement (depressive and anxiety symptoms) model across the three time points to ascertain whether their properties remained unchanged. The factor structure of the non-longitudinal measurement (attachment, coping) model was also tested to ascertain psychometric adequacy. Overall, all measurement models had an adequate fit, and the factors were well-defined and reliable. The distress measurement model was then combined with the non-longitudinal measurement model. This combined model was re-specified as a predictive model by replacing correlations with regressions to test the mediation hypotheses.

Table 3 presents the descriptive statistics and bivariate correlations among all variables (latent factor) and the covariate (gender and baseline levels of depression and anxiety symptoms). Correlations were consistent with expectations, although attachment avoidance was not significantly related to anxiety symptoms or avoidance coping at all three measurement times. Identifying as a woman (gender) was related to higher levels of attachment avoidance, self-help and accommodation coping, supporting its inclusion as a covariate in the final model.

Table 3.

Descriptive Statistics (Based on Manifest Scores) and Pairwise Correlations (Based on Fully Latent Variables) Among the Study Variables.

	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Anxiety T1†
2. Anxiety T2†	.674**
3. Anxiety T3†	.633**	.880**
4. Depression T1†	.808**	.556**	.624**
5. Depression T2†	.528**	.808**	.733**	.582**
6. Depression T3†	.450**	.702**	.808**	.692**	.827**
7. Attachment: Anxiety T1†	.304**	.170*	.232**	.383**	.317**	.254**
8. Attachment: Avoidance T1†	.073	.072	.155	.296**	.130**	.231**	−.009
9. Coping: Self-help T2†	−.118	−.212*	−.303**	−.271**	−.328**	−.353**	−.054	−.304**
10. Coping: Approach T2†	−.130	−.101	−.151	−.241**	−.256**	−.177*	−.138	−.129	.638**
11. Coping: Accommodation T2†	−.272**	−.360**	−.243**	−.243**	−.583**	−.354**	−.207**	−.237**	.410**	.450**
12. Coping: Avoidance T2†	.259**	.447**	.375**	.214**	.534**	.340**	.360**	.137	−.220	−.142	−.592**
13. Coping: Self-punishment T2†	.294**	.495**	.484**	.331**	.617**	.529**	.267**	.104	−.306**	−.134**	−.455**	.564**
14. Gender	.049	.085	.074	.011	−.156	−.015	−.009	.342**	.313**	.170	.263**	.017	−.093
Mean	1.82	1.79	1.59	1.90	1.97	1.83	4.54	2.42	2.78	2.38	3.06	1.68	2.20
Standard deviation	.78	.84	.74	.81	.91	.86	1.34	1.10	.82	.67	.70	.56	.89

Note. *p < .05. **p < .01. †: Fully latent variables. Gender was coded as 1 = female, 0 = others.

Main Analyses

A three-step approach was adopted to test the longitudinal associations between pre-breakup attachment insecurities, post-breakup coping strategies, and depressive and anxiety symptoms. In the first step, relations between variables were estimated freely. In the second step, relations between variables at multiple time points were constrained to equality. In the final step, the covariate was added. The parameter estimates for the final auto-regressive cross-lagged model are reported in Table 4. The fit indices for the final model are adequate (see supplementary table). The significance of indirect effects was tested via the calculation of 1000 bootstraps 95% confidence intervals (CI), and these effects were considered statistically significant when the confidence intervals excluded zero.

Table 4.

Parameter Estimates From the Final Autoregressive Cross-Lagged Model.

Predictor	Outcome	b (S.E.)	β (S.E.)
Anxiety
Anxiety T1	Anxiety T2	.923 (.093)**	.739 (.044)**
Attachment: Anxiety T1	Anxiety T2	−.155 (.111)	−.089 (.061)
Attachment: Avoidance T1	Anxiety T2	−.033 (.084)	−.018 (.046)
Coping: Self-help T2	Anxiety T2	−.110 (.085)	−.068 (.051)
Coping: Approach T2	Anxiety T2	.019 (.087)	.011 (.051)
Coping: Accommodation T2	Anxiety T2	.034 (.093)	.022 (.060)
Coping: Avoidance T2	Anxiety T2	.144 (.097)	.090 (.029)
Coping: Self-punishment T2	Anxiety T2	.245 (.080)**	.158 (.050)**
Anxiety T2	Anxiety T3	.923 (.093)**	.643 (.044)**
Attachment: Anxiety T1	Anxiety T3	−.155 (.111)	−.112 (.078)
Attachment: Avoidance T1	Anxiety T3	−.033 (.084)	−.023 (.059)
Coping: Self-help T2	Anxiety T3	−.110 (.085)	−.085 (.065)
Coping: Approach T2	Anxiety T3	.019 (.087)	.014 (.064)
Coping: Accommodation T2	Anxiety T3	.034 (.093)	.027 (.075)
Coping: Avoidance T2	Anxiety T3	.144 (.097)	.112 (.075)
Coping: Self-punishment T2	Anxiety T3	.245 (.080)**	.197 (.064)**
Depression
Depression T1	Depression T2	.688 (.096)**	.592 (.063)**
Attachment: Depression T1	Depression T2	.041 (.116)	.026 (.074)
Attachment: Avoidance T1	Depression T2	−.007 (.096)	−.004 (.059)
Coping: Self-help T2	Depression T2	−.108 (.086)	−.073 (.058)
Coping: Approach T2	Depression T2	.067 (.078)	.043 (.050)
Coping: Accommodation T2	Depression T2	−.160 (.080)*	−.113 (.057)*
Coping: Avoidance T2	Depression T2	−.003 (.087)	−.002 (.059)
Coping: Self-punishment T2	Depression T2	.267 (.070)**	.189 (.047)**
Depression T2	Depression T3	.688 (.096)**	.488 (.063)**
Attachment: Depression T1	Depression T3	.041 (.116)	.030 (.085)
Attachment: Avoidance T1	Depression T3	−.007 (.096)	−.005 (.068)
Coping: Self-help T2	Depression T3	−.108 (.086)	−.085 (.068)
Coping: Approach T2	Depression T3	.067 (.078)	.050 (.058)
Coping: Accommodation T2	Depression T3	−.160 (.080)*	−.132 (.066)*
Coping: Avoidance T2	Depression T3	−.003 (.087)	−.003 (.069)
Coping: Self-punishment T2	Depression T3	.267 (.070)**	.219 (.057)**
Mediators
Attachment: Anxiety T1	Coping: Self-help T2	−.244 (.090)**	−.227 (.080)**
Attachment: Avoidance T1	Coping: Self-help T2	−.397 (.099)**	−.357 (.078)**
Attachment: Anxiety T1	Coping: Approach T2	−.273 (.103)**	−.269 (.094)**
Attachment: Avoidance T1	Coping: Approach T2	−.201 (.085)*	−.192 (.078)*
Attachment: Anxiety T1	Coping: Accommodation T2	−.476 (.104)**	−.427 (.077)**
Attachment: Avoidance T1	Coping: Accommodation T2	−.377 (.094)**	−.327 (.072)**
Attachment: Anxiety T1	Coping: Avoidance T2	.485 (.107)**	.448 (.081)**
Attachment: Avoidance T1	Coping: Avoidance T2	.257 (.099)**	.230 (.084)**
Attachment: Anxiety T1	Coping: Self-punishment T2	.562 (.109)**	.504 (.076)**
Attachment: Avoidance T1	Coping: Self-punishment T2	.275 (.096)**	.238 (.077)**
Control variables
Gender	Anxiety T1	.158 (.146)	.062 (.058)
Gender	Anxiety T2	.158 (.146)	.043 (.040)
Gender	Anxiety T3	.158 (.146)	.035 (.032)
Gender	Depression T1	.025 (.136)	.010 (.064)
Gender	Depression T2	.025 (.136)	.007 (.038)
Gender	Depression T3	.025 (.136)	.006 (.033)
Gender	Attachment: Anxiety T1	.695 (.229)**	.226 (.082)**
Gender	Attachment: Avoidance T1	−.168 (.190)	−.066 (.075)
Gender	Coping: Self-help T2	.710 (.206)**	.253 (.071)**
Gender	Coping: Approach T2	.425 (.202)*	.160 (.075)*
Gender	Coping: Accommodation T2	.767 (.219)**	.263 (.070)**
Gender	Coping: Avoidance T2	−.160 (.235)	−.057 (.083)
Gender	Coping: Self-punishment T2	−.499 (.245)*	−.171 (.082)*

Note. *p < .05. **p < .01. b: unstandardized regression coefficient; β: standardized regression coefficient; S.E.: standard error. Gender was coded as 1 = female, 0 = others.

Results from the ARCL model revealed that pre-breakup attachment anxiety and avoidance (T1) were directly related to higher use of self-punishment and avoidance coping (T2) and to a lower use of accommodation, self-help, and approach coping (T2). However, in the final multivariate model, only self-punishment coping significantly predicted elevated depressive and anxiety symptoms, while accommodation coping significantly predicted lower depressive symptoms, but not anxiety symptoms (see Figure 1). The model explains 49.7% and 51.7% of depressive and anxiety symptoms at one-month post-breakup, as well as 62.7% and 69% of depressive and anxiety symptoms at three-month post-breakup. Identifying as a woman (gender covariate) was related to higher levels of attachment anxiety, self-help, approach, and accommodation coping but to lower reports of self-punitive coping.

Figure 1.

Autoregressive Cross-Lagged Model. Notes. Only significant path associations are depicted. The model included gender and baseline levels of depressive and anxious symptoms as covariates. Correlations between the two attachment dimensions, the five coping strategies, and the depressive and anxious symptoms at each time point are included in the model but not shown for simplicity.

Mediations for the Attachment Anxiety – Post-Breakup Distress Links

Results showed a mediating role of two coping strategies (self-punishment, accommodation) at one-month post-breakup (T2) in the associations between pre-breakup attachment anxiety (T1) and post-breakup distress (depressive and anxiety symptoms) at one-month (T2) and three-month post-breakup (T3), while accounting for gender and baseline depressive and anxiety symptoms. Overall, nine significant indirect effects were found.

First, more anxiously attached individuals were more likely to report using self-punishment coping following their breakup, which in turn was related to higher depressive symptoms at one-month (β = .110, 95% CI [.031, .212]) and three-month post-breakup (β = .095, 95% CI [.028, .183]). Higher use of self-punishment coping in individuals reporting higher levels of attachment anxiety was also indirectly related to more depressive symptoms at three-month post-breakup, via higher depressive symptoms at one-month post-breakup, β = .073, 95% CI [.036, .152].

Second, anxiously attached individuals reported more post-breakup self-punishment coping, which in turn was related to higher anxiety symptoms at one-month (β = .099, 95% CI [.039, .187]) and three-month post-breakup (β = .079, 95% CI [.037, .152]). The use of self-punishment coping in individuals with higher levels of attachment anxiety was also indirectly related to more anxiety symptoms at three-month post-breakup, via higher anxiety symptoms at one-month post-breakup, β = .065, 95% CI [.011, .141].

Finally, attachment anxiety was linked to less use of accommodation coping, which was related to higher depressive symptoms at one-month (β = .056, 95% CI [.012, .128]) and three-month post-breakup (β = .048, 95% CI [.012, .118]), but not to anxiety symptoms. The use of less accommodation coping in individuals with higher levels of attachment anxiety was also indirectly related to more depressive symptoms at three-month post-breakup, via higher depressive symptoms one-month post-breakup, β = .033, 95% CI [.007, .073].

Mediations for the Attachment Avoidance – Post-Breakup Distress Links

Results from the ARCL model revealed that two coping strategies (self-punishment and accommodation) at one-month post-breakup (T2) also mediated the associations between pre-breakup attachment avoidance (T1) and distress at both one-month (T2) and three-month post-breakup (T3), while controlling for gender and baseline depressive and anxiety symptoms. Overall, five significant indirect effects were found.

First, individuals with higher attachment avoidance reported higher self-punishment coping use, which was related to greater anxiety symptoms, but not depressive symptoms, at three-month post-breakup β = .038, 95% CI [.003, .090]. Higher use of self-punishment coping in individuals with higher levels of attachment avoidance was indirectly related to more anxiety symptoms at three-month post-breakup, via higher anxiety symptoms at one-month post-breakup, β = .035, 95% CI [.002, .085].

Second, high avoidant attachment was linked to less accommodation coping, which in turn was related to greater depressive symptoms at one- (β = .043, 95% CI [.008, .115]) and three-month post-breakup (β = .037, 95% CI [.006, .095]), but not to anxiety symptoms. Lower use of accommodation coping in individuals with higher attachment avoidance was also indirectly related to more depressive symptoms at three-month post-breakup, via higher depressive symptoms at one-month post-breakup, β = .025, 95% CI [.008, .061].

Discussion

Using a longitudinal design, this innovative study examined the prospective associations between pre-breakup attachment insecurities and post-breakup depressive and anxiety symptoms via five coping strategies among a community-based sample of emerging adults. The longitudinal autoregressive cross-lagged model yielded two main original findings: attachment-related anxiety and avoidance were related to more severe depressive and anxiety post-breakup symptoms through (1) a higher use of self-punishment coping and (2) a lower use of accommodation coping. These findings help understand why attachment insecurities increase the risk of experiencing greater distress following a romantic breakup in emerging adults.

Attachment Anxiety, Coping, and Distress

In line with our first hypothesis, pre-breakup attachment anxiety was indirectly related to greater breakup distress (both depressive and anxiety symptoms) in the short term and medium term (one- and three-month post-breakup) through higher use of self-punishment coping. These coping strategies were the most frequently employed by, and most deleterious for, individuals with attachment anxiety, as indicated by larger effect sizes. These results are consistent with past research showing that individuals with high levels of attachment anxiety tend to use maladaptive coping strategies, such as rumination (Fagundes, 2012) and self-blame (Choo et al., 1996), to deal with the loss of their romantic partner. Results also support past cross-sectional studies’ findings revealing that self-punishment coping (Leung et al., 2011) and rumination (Saffrey & Ehrenberg, 2007) explain the link between attachment anxiety and breakup distress. Consistent with the hyperactivation of their attachment system explanation and Simpson and Rholes’ (2012) attachment-diathesis-stress model, the coping strategies employed by anxiously attached individuals appear to increase their breakup distress, explaining its maintenance over time.

In partial support of our first hypothesis, pre-breakup attachment anxiety was indirectly related to more depressive, but not anxiety, symptoms at one- and three-month post-breakup through lower use of accommodation coping. These findings corroborate the results of studies evidencing that individuals with high levels of attachment anxiety have difficulties accepting the end of a romantic relationship (e.g., Sbarra, 2006). Thus, consistent with attachment hyperactivation, these individuals tend to show a persistent preoccupation with the loss and seek to re-establish the relationship (Davis et al., 2003) rather than move toward accepting and reframing the situation. In a study of emotional recovery following a breakup, Sbarra’s (2006) results show that breakup nonacceptance is a significant predictor of poor sadness recovery in individuals with attachment anxiety. This difficulty in accepting the end of a relationship appears related to sadness and depressive symptoms, but not psychological activation and anxiety symptoms. These results might also be understood in light of the negative self-image of anxiously attached individuals. Indeed, the breakup may have confirmed their fears and representations of themselves as not deserving love from others (Mikulincer & Shaver, 2016), which could lead to despair, resignation, and doubts about their chances of finding love again.

Attachment Avoidance, Coping and Distress

Pre-breakup attachment avoidance was indirectly related to more anxiety symptoms at three-month post-breakup through higher use of self-punishment coping. Although this finding may seem surprising, it aligns with Birnbaum et al.’s (1997) results in a divorce context, which showed that self-defeating thoughts explained the link between attachment avoidance and psychological distress. These authors suggested that the avoidance strategies used by avoidant individuals when coping with minor stresses are ineffective when facing the loss of an attachment figure. Thus, in the presence of a destabilizing or difficult breakup, it is possible that their coping strategies become less effective, leading them to resort to unusual strategies of self-blame and rumination, but to a lesser extent than individuals with attachment anxiety. Another possible explanation could be the pandemic context in which this study took place. During their breakup, the avoidantly attached individuals may have been deprived of their usual distancing coping, which includes distraction (Mikulincer & Shaver, 2016), leaving more space for rumination. Considering this is an unusual strategy for them, it may have led to more anxiety symptoms. However, this explanation is tentative and would necessitate further empirical validation.

Results also revealed that attachment avoidance was indirectly related to more depressive, but not anxiety, symptoms at one- and three-month post-breakup via a lower use of accommodation coping (optimism, acceptance, positive reframing, replacement). This result is partially consistent with our second hypothesis and Leung et al.’s (2011) cross-sectional results. Although they found no link between attachment insecurities and accommodation coping, their results revealed that accommodation coping explained the links between secure attachment and lower post-breakup distress, suggesting that secure individuals would more frequently use this type of coping. This result is also consistent with attachment theory, which states that attachment avoidant individuals have more negative and pessimistic attitudes (Mikulincer & Shaver, 2016), which might be exacerbated after the loss of a romantic partner, resulting in low accommodation coping and higher depressive symptoms. The pandemic context may also have exacerbated their negative view of others and the world, as the restrictions limited contacts and opportunities for social interactions after their breakup. Finally, the deactivation of their attachment system may have refrained their acceptance of the loss and their mourning process (Mikulincer & Shaver, 2016), resulting in higher negative affects and depressive symptoms.

Contrary to our hypotheses, avoidance coping did not mediate the prospective association between attachment insecurities and post-breakup distress when all coping strategies were considered together. This finding aligns with Birnbaum et al.’s (1997), who found no association between attachment insecurities, distancing coping and post-divorce distress. Still, it contrasts with Leung et al.’s (2011) cross-sectional results, revealing that attachment styles marked by higher attachment anxiety (preoccupied, fearful) were related to post-breakup distress through avoidance coping. Cultural differences between the samples, the timing of the post-breakup measures, the fact that some of Leung et al.’s participants did not experience a breakup, and the use of different measures of attachment and distress might explain these discrepancies. More precisely, our results show that individuals with higher levels of attachment insecurities reported higher use of avoidant coping following their breakup; however, using these strategies did not explain their level of distress over and above higher self-punitive and lower accommodation coping strategies. This result suggests that anxiously attached individuals use avoidance strategies, potentially to temporarily relieve negative emotions and distress that have become overwhelming due to rumination and self-blame (Leung et al., 2011), while avoidant individuals likely use them as a way to deactivate their attachment system. Yet, these strategies play a smaller role in explaining their distress, which might be explained by the lack of awareness of their distress when using avoidance coping. The fact that the distinct avoidance coping strategies (e.g., mental disengagement, denial, blaming others) were not considered might also explain the lack of associations with distress. Finally, avoidance coping might work for some individuals but not for others or work sometimes but not all the time, depending on another moderating variable. Further research is warranted, especially among avoidantly attached young adults, since our results contrast with the attachment-diathesis-stress model (Simpson & Rholes, 2012).

Limitations and Future Directions

The longitudinal design of this study made it possible to follow emerging adults through a romantic relationship breakup, in both the short term and medium term. Pre-breakup measures of distress and attachment insecurities allowed to accurately measure the breakup-related distress and to prevent elevated levels of attachment insecurities if measured after the breakup (Birnbaum et al., 1997; Fagundes, 2012). This study is among the first to distinguish depressive and anxiety symptoms following a breakup. Our findings, however, rely on a relatively small sample size, which may have limited the detection of small associations due to low statistical power. The breakup rate in the larger study was lower than expected, possibly due to a volunteer bias attracting participants with fairly positive relationships (Demir et al., 2017). The entire study was conducted during the COVID-19 pandemic, which may have favored the maintenance of certain relationships, due to a reluctance to break up by fear of the loneliness, considering few social contacts during this period. The pandemic may also have exacerbated distress symptoms, despite controlling for initial distress, and limited the use of some coping strategies, such as avoidance. Keeping individuals who renewed their relationship with their ex-partner in the sample may have mitigated post-breakup distress ² . Nevertheless, some attachment system’s activation and post-breakup distress may persist despite the renewed presence of the former partner. Although the simultaneous study of five coping categories in multivariate analyses represents a strength, another limitation concerns the groupings of coping strategies, which may have obscured distinct and specific effects of each coping strategy on distress. Finally, the over-representation of women and the grouping of non-women together represent limitations to the generalization of our findings, despite controlling for gender. Future studies should replicate these results in a non-COVID context and on larger samples with varied genders. They could also examine the distinct coping strategies among the larger coping categories, especially for avoidance coping. Future studies are warranted on other potential mediators (e.g., self-esteem, loneliness) or moderators (e.g., social support) of the associations between attachment insecurities and breakup distress.

Implications

Our findings emphasize the role of attachment insecurities and coping strategies in both short term and medium term post-breakup distress. These findings encourage therapists to inquire about attachment insecurities, coping strategies, and depressive and anxiety symptoms when assessing young adults’ post-breakup distress. As for treatment, therapists could target awareness of attachment-related tendencies (hyperactivation or deactivation) and coping strategies. This focus might help individuals recognize the high psychological costs of self-punishment coping when dealing with a loss. Interventions could also help insecure young adults improve their coping strategies by replacing maladaptive coping strategies (e.g., self-punishment) with adaptive coping strategies (e.g., accommodation). In this sense, self-compassion has been shown to counter the self-punitive coping employed by young adults experiencing a breakup (Soltani & Fatehizade, 2020; Zhang & Chen, 2017). Finally, these findings could be used to help develop preventive psychoeducation programs surrounding romantic breakups in young adults, as transitions between relationships in emerging adults are opportune times to reduce risk factors and enhance relational and mental health protective factors (Vennum et al., 2017).

Author Biographies

Kristin Gehl, PsyD, is a Clinical Psychologist who completed her Doctoral Dissertation in the Department of Psychology at the University of Sherbrooke, Sherbrooke (QC), Canada. Her research and clinical interests include romantic attachment and the determinants of relationship and personal distress, especially among young adults experiencing a breakup.

Audrey Brassard, PhD, is a Full Professor in the Department of Psychology at the University of Sherbrooke, Sherbrooke (QC), Canada. Her research and clinical interests include romantic attachment and the determinants of relationship distress, intimate partner violence, and sexual difficulties in adults from diverse populations.

Caroline Dugal, PhD, is an Assistant Professor in the Department of Psychology of the Université du Québec à Trois-Rivières, Trois-Rivières (QC), Canada. She focuses her research and clinical interests on conflicts and violence between romantic partners, romantic attachment, relational ambivalence and separation, sexual difficulties and the repercussions of childhood maltreatment.

Audrey-Ann Lefebvre, BA, is a Ph.D. candidate in the Department of Psychology at the Université de Sherbrooke, Sherbrooke (QC), Canada. Her research interest includes clinical psychology, intimate partner violence, as well as sexual and couple functioning.

Isabelle Daigneault, PhD, is a Clinical Psychologist and Full Professor at the Université de Montréal. Her research has focused on child and adolescent sexual assault, life trajectories, resilience and health of young victims, as well as on the effectiveness of prevention programs for various age groups and diverse populations.

Audrey Francoeur, PhD, is a Clinical Psychologist who completed her Doctoral Dissertation in the Department of Psychology at the University of Montreal, Montreal (QC), Canada. Her research and clinical interests include child maltreatment, social cognition, and psychological distress in young adults.

Tania Lecomte, PhD, is a Full Professor in the Department of Psychology at the University of Montreal, Montreal (QC), Canada. Her research interests include the prevention and treatment of mental disorders.

Ethical Statement

Ethical Approval

This study was part of a larger prospective study on relationship trajectories among emerging adults approved by the Institutional Review Boards of two Canadian universities.

ORCID iD

Audrey Brassard https://orcid.org/0000-0002-2292-1519