“I can't escape!”: Avoidantly attached individuals' conflict resolution and relationship satisfaction before and during the COVID-19 lockdown

ORIGINAL ARTICLE “I can’t escape!”: Avoidantly attached individuals’conflict resolution and relationship satisfaction before and during the COVID-19 lockdown Ione Bretaña 1 | Itziar Alonso-Arbiol 1 | Kristel Kittel 1 | Silvia Ubillos-Landa 2 1 Department of Clinical and Health Psychology & Research Methods, University of the Basque Country UPV/EHU, Donostia, Spain 2 Department of Health Science, University of Burgos, Burgos, Spain Correspondence Ione Bretaña, Faculty of Psychology, University of the Basque Country UPV/EHU, Av. Tolosa, 70, 20018 San Sebastian, Spain. Email: [email protected] Funding information pre-doctoral grant from the Education Department of the Basque Government, Grant/Award Number: PRE_2016_1_0138; grant awarded by the Regional Government of Castilla y Leon (Spain) to the Social Inclusion and Quality of Life research group, Grant/Award Number: 2019/00184/001; grant by the Basque Government Research Groups (‘Culture, Cognition, and Emotion’Consolidated Group), Grant/Award Number: IT-1598-22; Two grants by the Spanish Ministry of Science and Innovation, Grant/Award Numbers: PID2020-115738GB-I0, PID2020-116658GB-I00 Abstract Increased time spent together and the lockdown resulting from the COVID-19 pandemic may have created new scenarios for marital conflict. We analyzed how home confinement affects avoidantly attached individuals’: (a) resolution strategies to cope with couple conflict, (b) perception of partner’s resolution strategies, and (c) overall relationship satisfaction. The sample comprised 549 individuals, divided into two subsamples: (a) the confined group, individuals confined with their partners (n=275); and (b) the comparison group, coupled individuals from a dataset collected before the pandemic (n=274). Results indicate that the proposed model works in different contexts (non-confinement and confinement situations), but there are some significant differences in the magnitude of some of the relationships between the variables, being stronger in the confinement group than in the comparison group. In the confined group, in individuals with avoidant attachment, withdrawal was associated with lower relationship satisfaction and a higher demand partner perceived to a higher extent than in the comparison group. This might explain the lower satisfaction with the relationship of the confined group. The different conflict resolution strategies of the couple mediated between avoidant attachment and relationship satisfaction in both groups (confined and comparison). It is concluded that individuals’attachment orientation is a key factor in how individuals experienced their close relationships during the confinement. KEYWORDS avoidant attachment, conflict resolution, lockdown, relationship satisfaction INTRODUCTION The whole world suffered terrible consequences and individual confinements during the SARS-CoV-2 (COVID-19) pandemic. Several authors have already written about the psychological impact of confinements, including how the stressinducing context led to increase of some symptoms (i.e., posttraumatic stress, anxiety, depression, and insomnia; e.g., Brooks et al., 2020; Taylor, 2019). During the COVID-19 pandemic, those symptoms were experienced more often in individuals affected by confinement as compared to the general population in normal circumstances (see Cénat et al., 2020, meta-analysis). Extremera (2020) suggested that in stressful situations, such as confinement, it is very likely that relational well-being and relationship satisfaction could be affected by the frequency and intensity of conflict, and by intimacy and closeness in relationships. The interaction in close spaces may lead to conflicts within the couple, especially during periods of increased stress, uncertainty, and anxiety (Lima et al., 2020), as in the case of the COVID-19 pandemic (Ishikawa, 2020). To reinforce this point, Balzarini et al. (2023) observed higher levels of conflict for cohabiting couples from the beginning of the pandemic. COVID-19 lockdown and attachment system In the confinement situation due to the pandemic, there is a clear need to study the effect of behaviors displayed to solve relationship conflicts, considering attachment orientations. Received: 7 March 2022 Accepted: 28 February 2023 DOI: 10.1002/pchj.646 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2023 The Authors. PsyCh Journal published by Institute of Psychology, Chinese Academy of Sciences and John Wiley & Sons Australia, Ltd. 430 PsyCh Journal. 2023;12:430–442. wileyonlinelibrary.com/journal/pchj According to the attachment theory, individuals are equipped with innate behavioral repertoires associated with the attachment system (i.e., secure base, safe haven) that are activated. Thus, when security threats are experienced, the attachment figure provides a secure base from which the individual may explore the world, as well as a comforting safe haven (Bowlby, 1982). The attachment system serves the purpose of keeping proximity between the preferred caregiving figure and the child during infancy to allow for emotional needs to be met. The quality of the interaction determines to what extent attachment security may be acquired based on the perceived own lovability and on the attachment figure’s emotional availability (Bowlby, 1982). Repeated interactions during infancy (and at later development stages) are the base of the formation of internal schemes that guide the interpretation of experience of affective nature. In adulthood, different degrees of (in)security toward the attachment figure are explained into two attachment orientations that individuals display toward their romantic partners: avoidance (of intimacy) and anxiety (about the relationships; for a review, see Mikulincer & Shaver, 2016). The avoidance dimension is described as individuals’need for independence and emotional distance from their partner; the anxiety dimension may be defined as individuals’excessive preoccupation and fear of being abandoned by their partner (Fournier et al., 2011). In fact, an attachment system (and the need for safety from the attachment figure; i.e., the romantic partner in adulthood) would be activated due to the exceptional situation of uncertainty arising from COVID-19 (Vowels & Carnelley, 2021). Yet, important differences due to attachment orientations would apply. Individuals with an avoidance attachment orientation tend to avoid intimacy and conflict with their partners (e.g., Bretaña et al., 2020). In conditions of restrictions derived from confinement (i.e., the impossibility to escape from the situation and to flee physically), avoidantly attached individuals might have suffered a significant decline in their close relationships. Furthermore, in a series of meta-analyses (Candel & Turliuc, 2019; Hadden et al., 2014;Li&Chan,2012), avoidance has been shown to be the attachment orientation most strongly associated with lower relationship satisfaction (or even the only one in some studies: e.g., Brassard et al., 2009;Moleroetal.,2016; Molero et al., 2011). Thus, the present research deepens the study of avoidance orientation. Analyzing these relational dynamics in avoidantly attached individuals is paramount. To this end, a model is proposed that jointly integrates different variables analyzed by the scientific literature. Studying the same dynamics (i.e., models with the same variables) in comparable samples across different contextual stress levels (i.e., during the confinement vs. before the pandemic’s outbreak) would help in elucidating the impact of confinement on relationship satisfaction. For the analysis of such stress-inducing effects, the examination may involve a site where confinement restrictions were more stringent than in other countries, for instance, where the lockdown involved a total ban on going outside without the possibility of exercising (i.e., Spain). Avoidant attachment, conflict, and relationship satisfaction in stressful situations Attachment insecurity is a key factor associated with maladaptive relational dynamics in stressful situations (Mikulincer & Shaver, 2016; Simpson & Rholes, 2017). As mentioned above, avoidance attachment in particular exerts a noteworthy negative effect on relationship satisfaction (e.g., Candel & Turliuc, 2019; Hadden et al., 2014). Relationship satisfaction is defined as an individual’s overall assessment of their romantic relationship (e.g., Kamp Dush et al., 2008); when individuals perceive that their partners satisfy their needs, they are generally satisfied with their relationships (Fincham & Rogge, 2010). The attachment diathesis-stress process model (Simpson & Rholes, 2012) provides the theoretical underpinning to understand the impact of COVID-19 confinement on relationship dynamics. Based on this theory, stress could play a major role in the association between attachment and relationship satisfaction; different stressors or stress conditions could impact directly in the attachment dimension. Under stressful conditions, individuals perceive their partners and their relationship more negatively. In the case of avoidantly attached individuals—who disengage emotionally, cognitively and behaviorally in stressful conditions—it is reasonable to anticipate that they will be more unsatisfied with their relationship. Interactive pattern of conflict resolution is considered an important element in the association between avoidant attachment dimensions and relationship satisfaction. The current study analyzes withdrawal and demand strategies. Withdrawal is one maladaptive resolution strategy in which being silent, leaving the room, and other distancing behaviors are used when dealing with conflict (Eldridge & Christensen, 2002); and demand is characterized by criticizing, nagging, and demanding behavior toward the partner during conflict (Eldridge et al., 2007). Although the pattern of strategies most commonly analyzed in the literature has been the relationship between (an initiating) own (conflict) demand and (the subsequent) withdrawal (from conflict; Eldridge & Christensen, 2002), the immutability of this directionality has also been challenged by recent works. In fact, some authors have adopted a new perspective where the focus is placed on analyzing the inverse relationship between the strategies, with the withdrawal from the conflict being the activating mechanism of the partner’s demand strategy (Bonache et al., 2017,2019). The explanatory mechanism underlying the initiation of this conflict withdrawal could derive from avoidantly attached individuals’own tendency to use conflict avoiding behaviors in close relational contexts. Conflict withdrawal strategy and relationship satisfaction Insecurely attached people tend to experience remarkable difficulties in deploying satisfactory coping approaches and conflict resolution strategies (e.g., Mikulincer & Shaver, 2016); PsyCh JOURNAL 431 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License although other maladaptive conflict resolution strategies have been observed (Collins et al., 2006; Fowler & Dillow, 2011; Sierau & Herzerg, 2012), avoidance attachment orientation has been mainly associated with conflict withdrawal (e.g., Mikulincer & Shaver, 2016). Two recent studies (Bonache et al., 2019; Bretaña et al., 2020) have provided evidence on the association between the avoidance attachment dimension and own conflict withdrawal strategy. Moreover, Bretaña et al. (2020) found that more avoidantly attached individuals tend to use a withdrawal conflict resolution strategy toward the partner to a larger extent, which in turn functioned as a mediating variable, being associated with a decline in their relationship satisfaction. Coping with stressful conditions requires effort and control from individuals in a context where added stressors from the pandemic along with other preexisting contextual vulnerabilities are likely to further degrade these relationship processes (Pietromonaco & Overall, 2021). The disconnection and distancing from the problem that avoidantly attached individuals constantly seek (Mikulincer & Shaver, 2016) would be more difficult during confinement; they have to deploy strategies of withdrawal and flight from the conflict without being able to leave home due to legal prohibition. Using withdrawal in a situation of external stress (i.e., home confinement) would entail higher effort and bring more discomfort than in a situation of only internal stress (i.e., couple conflict). Withdrawal/demand pattern and relationship satisfaction According to a longitudinal study carried out by Neff and Karney (2007), relationship satisfaction declines in stressful situations, which can be explained by partners’conflict resolution skills. Stressful situations accompanied by a higher use of negative conflict resolution strategies result in lower levels of relationship satisfaction across time (Johnson, 2005). The higher emotionality (i.e., behaviors permeated by emotion) in a couple’s conflict leads to attentional biases and splitting (Sillars, 1998). In individuals with high scores in the avoidance dimension of attachment, own withdrawal conflict (more commonly deployed by them) has been associated with the perception that their partners use demand to a higher extent (Bretaña et al., 2020). Schrodt et al.’s(2014) meta-analysis showed that negative conflict resolution patterns (i.e., withdrawal and demand) were more common in those couples with higher levels of stress. In a situation in which running away is simply not possible, the more visible withdrawal behaviors (e.g., being silent, locking oneself in a room) could generate even more frustration in the partner, who would demand to talk about the problem or might even reproach the partner for their passivity (Feeney & Karantzas, 2017; Miga et al., 2010). This conflict pattern—where one person withdraws from the conflict while perceiving that their partner uses demand—has been associated with lower levels of relationship satisfaction (Bretaña et al., 2020). The positive problem-solving strategy in avoidantly attached individuals As we mentioned previously, avoidantly attached individuals tend to perceive others’behaviors more negatively (e.g., Mikulincer & Shaver, 2016). Vowels and Carnelley (2021) observed that individuals with a higher level of attachment avoidance were significantly less likely to perceive their partner as providing support. According to the attachment diathesis-stress process model (Simpson & Rholes, 2012), the association between avoidant attachment and partner support would be explained because under stressful conditions these individuals are not fully aware they are upset, and therefore, they do not seek their partner’s support. They have to reduce their distress by being self-sufficient, and thus, restoring independence, autonomy, and personal control. The way to get self-sufficiency is through the use of withdrawal and distancing behaviors: these strategies are deployed by avoidant individuals so that their partners are less supportive, and they perceive them as more distant and closed. Although this form of greater distance and less support is what an individual with avoidant attachment would prefer from their partners, it is also what they may interpret as rejection by their partners. Avoidant individuals tend to perceive their partners’behavior less benevolently, leading them also to underestimate the support provided by their partners. All of this, in turn, is associated with a decrease in an individual’s own well-being (Simpson & Rholes, 2012), which would refer to a decrease in relationship satisfaction within the close relationship. Although the perception of social support is framed more in the affective domain, the perception of support would also be closely related to positive conflict resolution strategies (Kaur, 2017), in which the individual demonstrates a certain commitment to assertive conflict resolution (Lawrence & Bradbury, 2007). Therefore, in the case of conflict resolution strategies, positive problem-solving strategy—the ability to cope with problems by displaying adaptive behaviors—would be a proxy for partner support. Bretaña et al. (2020) observed that there was a negative association between avoidant attachment and perceived partner’s positive problem-solving strategy. Avoidantly attached individuals tend to perceive that their partners use positive problem-solving to a lesser extent; this perception would be mediating in the association between avoidant attachment dimension and relationship satisfaction. Therefore, taking into account the COVID-19 pandemic, and more specifically, confinement, we assume that the mental models underlying the attachment dimensions and all the attachment behaviors described above will have been exacerbated, and the relationship between these variables will be more evident. In sum, taking into account the previous hypothesis, we propose a model of relationship satisfaction in avoidantly attached individuals to test the differences in two conditions (the confinement group and the comparison group). As we can see in the path model (Figure 1), we expect that the same model will fit for both samples but differ in beta sizes. 432 AVOIDANCE, COUPLE RELATIONSHIP, AND LOCKDOWN 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Hypotheses Hypothesis 1. We expect that in the confined group, avoidant attachment will be associated more negatively with relationship satisfaction than in the comparison group. Hypothesis 2. We expect that the negative association between withdrawal conflict strategy and relationship satisfaction would be higher during confinement than before. Hypothesis 3. Own withdrawal conflict strategy will mediate the association between avoidant attachment dimension and relationship satisfaction to a higher extent in the confinement condition than in a non-confinement situation. Hypothesis 4. Therefore, we propose that individuals who tend to withdraw from conflict will perceive that their partners demand the conflict to a higher extent during confinement as compared to counterparts in a situation prior to confinement. Hypothesis 5. Since the withdraw/demand pattern will be more salient in a confinement situation, the association between own withdrawal and relationship satisfaction will be mediated by perceived partner’s demand to a higher extent in confinement than in non-confinement. Hypothesis 6. Individuals with higher scores in avoidant attachment will perceive that their partner’s use of positive problem-solving strategy will be to a lesser extent during confinement as compared to individuals in the comparison group. Hypothesis 7. Perception that partner uses positive problem-solving resolution strategy to a lesser extent will mediate the association between avoidant attachment dimension and lower relationship satisfaction to a higher extent in a confinement condition than in a non-confinement situation. Hypothesis 8. Individuals with high scores in avoidant attachment and that scored high in withdrawal will perceive higher demanding behaviors and lower positive problem-solving from their partners, which in turn will predict lower scores in relationship satisfaction. Study site The present study takes place in Spain. As of March 15, 2020, Spain was the country with the second highest number of confirmed cases of coronavirus during the first wave of the pandemic. The state of emergency was decreed in Spain and a very restrictive confinement was imposed as compared to other countries. The entire population of 47 million was required to stay in their homes, who mostly (66.2%) live in apartments of smaller size than in the other European countries (Eurostat, 2021). Going out to the streets was only allowed for basic purchases in supermarkets, pharmacies, going to essential workplaces, and the most urgent medical consultations. Going out for a walk or doing physical exercise was not permitted. The police sanctioned anyone who left their home without justification, and sanctions of up to 10,000 Euros could be imposed. At the time of data collection for this study (by April 11, 2020), Spain had 161,852 positive cases and 16,353 deaths, reflecting the seriousness of the situation the country was facing. The harshness of the confinement justifies examining the hypothesis posed in the current study. As can be seen in Figure 2, we wanted to replicate a previous model in this study with a confined and a comparison group. METHODS Participants The complete sample was composed of 549 adult Spanish individuals in an exclusive relationship (of at least 6 months). Own Withdrawal Perceived Partner Demand Perceived Partner Positive Relationship Satisfaction Avoidance Attachment FIGURE 1 Path model of relationship satisfaction and conflict resolution strategies in avoidantly attached individuals. PsyCh JOURNAL 433 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Informed consent was obtained from all individual participants included in the study. We performed a statistical power analysis with G*Power 3.1. (Faul et al., 2009) for sample size estimation. For this analysis, we relied on data from the metaanalysis of Candel and Turliuc (2019) on dimensions of attachment and relationship satisfaction on the one hand, and on Woodin’s( 2011) dealing with conflict resolution strategies and couple satisfaction on the other hand. In both studies, a medium effect size was obtained. For the current study, with an alpha =.05 and a power of =.95, for both estimations, the projected sample size would be N=158. Considering that a multigroup model was proposed, in each group there should be 158 individuals (i.e., a total of 316 individuals). The sample of the study (N=549) was divided into two subsamples with similar characteristics, except for the education level and percentage of women, although the effect size of differences was moderate and low, respectively (see Table 1). The subsample whose data were collected during confinement was made of 275 individuals (74.1% women) confined with their partners in the same household. To test the hypothesis, a sample of 274 Spanish individuals involved in an exclusive relationship who had also completed the same assessment instruments before the pandemic (pre-COVID-19 comparison group, 58% women) was used (for a more complete description of this second subsample, see Bretaña et al., 2020). The sociodemographic characteristics of both subsamples can be seen in Table 1. Some specific data for the confined group were as follows: 15.9% of the sample claimed to have tested positive or to be experiencing some symptom related to the virus 1 and 20.7% of the sample stated that a family member had tested positive. The mean length of the relationship was 11.44 years (SD =9.90 years) and for couples who were living together, the mean length was 8.71 years (SD =9.68 years). As for working status during the confinement, 48.2% were working at home (telework), 13.8% went out to work (outside the home), 17.1% were dismissed due to company closure since the beginning of the health crisis, and 14.5% did not work for various reasons (unemployment, sick leave, etc.). The remaining 6.4% of the sample did not report any information on this question. 2 Instruments Experiences in Close Relationships Scale Although the Experiences in Close Relationships Scale (Brennan et al., 1998; Spanish adaptation by Alonso-Arbiol et al., 2007) measures the two dimensions of insecure attachment (anxiety [about abandonment] and avoidance [of TABLE 1 Descriptive and Mean Differences Between Confined and Comparison Groups. Variables Confined group (n=275) M(SD) Comparison group (n=274) M(SD)tpCohen’sd Gender (women) 74.1% 58% X 2 =16.41 <.001 V=.17 Age (years) 36.88 (10.79) 36.80 (12.60) 0.00 .99 .01 Children (No) 56.2% 56.6% X 2 =.01 .92 V=.01 Relationship length (years) 11.44 (9.90) 12.91 (12.43) 1.56 .11 .13 Cohabiting time (years) 8.71 (9.68) 9.49 (12.41) 0.85 .40 .07 Education level 5.69 (1.48) 4.77 (1.67) 6.78 <.001 .58 Avoidant attachment 2.25 (0.71) 2.26 (0.77) 0.30 .76 .01 Own withdrawal 2.21 (0.99) 2.70 (0.91) 6.07 <.001 .51 Partner demand 1.15 (0.38) 1.30 (0.54) 3.83 <.001 .32 Partner positive 3.80 (1.01) 4.41 (1.16) 6.66 <.001 .56 Relationship satisfaction 6.07 (0.93) 6.21 (0.66) 2.30 .02 .17 Note:V=Cramer’sV, as the effect size for the difference in nominal variables. FIGURE 2 Path model of relationship satisfaction across confined and comparison groups. Values in italics are for pre-COVID-19 comparison (non-confined) group and values in bold are for the confined group. Beta values are standardized. *p< .05, **p< .01. 434 AVOIDANCE, COUPLE RELATIONSHIP, AND LOCKDOWN 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License intimacy]), in the current study, only the dimension of avoidance was thoroughly analyzed. Each dimension is composed of 18 items on a Likert scale from 1 (Strongly Disagree)to7 (Strongly Agree), where the items about avoidance were related to lack of comfort with closeness with others and preference for emotional distance (e.g., “I prefer not to show a partner how I feel deep down”). Anxious dimension was related to the need for excessive intimacy (e.g., “I worry about being abandoned”). Internal consistency for the confined group in avoidance was good (α=.85 and ω=.86), as it was for the comparison group (α=.83 and ω=.88; see Bretaña et al., 2020). Internal consistency in the anxious dimension was good for both the confined group (α=.89 and ω=.89) and the comparison group (α=.84 and ω=.84). Conflict Inventory Revised The Conflict Inventory Revised (Ridley et al., 2001;Spanish version adapted by Bretaña et al., 2019) consists of 13 items that measure the extent to which individuals (CI-Own) use one of the following conflict resolution strategies: withdrawal, including five items (e.g., “Leaving the room”); demand, including four items (e.g., “Name-calling”); and positive problem-solving, including four items (e.g., “Listening carefully”). A second part of the questionnaire refers to how individuals perceive that their partners resolve conflicts and to what extent they use the strategies mentioned above (CI-Partner). Items for each strategy are measured with a Likert frequency scale from 1 (Never)to7 (Always). In the present study, following the theoretical model to be tested, only own withdrawal, demand, and positive problem-solving strategies perceived in the partner were analyzed. All internal consistency values (Cronbach alphas and McDonalds’ omega) were acceptable. The following values were obtained for the confined group (own-withdrawal: α=.71 and ω=.73; partner’s demand: α=.80 and ω=.81; partner’s positive problem-solving: α=.71 and ω=.72). The following values were obtained for the comparison group (own-withdrawal: α=.75 and ω=.72; partner’s demand: α=.73 and ω=.63; partner’s positive problem-solving: α=.66 and ω=.74). Relationship Assessment Scale The Relationship Assessment Scale (Hendrick, 1988; Spanish adaptation by Molero et al., 2016) assesses an individual’s general satisfaction with their partner through 7 items measured with a Likert scale, ranging from 1 (Not satisfied at all)to7 (Very satisfied). An example of an item on this scale is: “Overall, I am satisfied with this relationship.”The internal consistency of the scale was good for both the confined group (α=.88 and ω=.90) and the comparison group (α=.92 and ω=.83). Procedure Prior to data collection, approval for the study was granted by the Ethics Committee of the University of Burgos (IR 12/ 2020) and the University of the Basque Country (M10/2020/ 119R1). After obtaining permission from the ethics committees, the questionnaire was hosted on the Qualtrics platform for online data collection from the confined group. The link was disseminated through email and different social networks between the dates of April 10 and 20, 2020, a period in which the individuals were home confined. Regarding the comparison group sample, permission was obtained for the use of data from a study carried out the previous year in Spain, in which the same instruments had been used, and for which the authors had previously obtained permission from the ethics committee of their university. In this latter case, participants filled in the questionnaires in a paper-and-pencil version. In both cases, participants did not receive compensation and were recruited by a snowball procedure through people with similar sociodemographic characteristics and living in a similar region or context to avoid large differences between the two samples. Data analyses Differences between two groups (comparison group and confined group) were examined through t-test for independence samples, analyzing Cohen’sdto determine the magnitude of the differences. Then we carried out correlation analyses between target study variables. To test the hypothesis and the overall model fit, we estimated using the maximum likelihood (ML) with AMOS 27.0 (Arbuckle, 2020). To compare the fit between the two groups (comparison and confined group), we carried out multigroup analyses with AMOS 27.0. We analyzed the mediator role of conflict resolution strategies (i.e., withdrawal and demand) between avoidant attachment and relationship satisfaction using the bootstrap method (Cheung & Lau, 2008) separately with each mediator. The mediator effects were analyzed using a bootstrap procedure (5000 resamples) with 95% bias-corrected confidence interval (CI). It is considered that if zero is not included in the interval between the lower and the upper bound, the effect is statistically significant at p≤.05. Nevertheless, this statistical significance is understood in a broader context of evaluating the effect size. Finally, to the comparison of the magnitude of β values of regression across groups (Bou & Satorra, 2007), we carried out tdifferences analyses. RESULTS Sample comparison and preliminary analysis As only minor differences were found in the sociodemographic variables of both samples, except for education level, in subsequent analyses this variable was controlled for (Table 1). Preliminary analyses linking sociodemographic variables with relationship satisfaction were also conducted in both samples. No mean differences were observed according to gender or parenting status (i.e., having children or not), and education level was not associated with relationship satisfaction (r=.03, p=.60 for the confined group; r=.02, p=.77 for the comparison group). PsyCh JOURNAL 435 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License As can be seen in Table 1, some differences were observed in relationship satisfaction (t=2.30, p=.02) between the confined group (M=6.07, SD =0.93) and the comparison group (M=6.21, SD =0.66), but the effect size was small (d=0.17). Regarding conflict resolution strategies (i.e., own withdrawal, partner demand, and partner positive), although the frequency of use was higher in the comparison group than in the confined one, the temporality evaluated was not the same in the two groups. 3 Therefore, they will not be taken into consideration as a measure that can be interpreted. Relationship satisfaction model in confined versus comparison (non-COVID-19-confined) individuals Correlations between variables were investigated. Table 2 shows the correlations between avoidant attachment, own and resolution strategies, and relationship satisfaction in the confined and comparison group. According to Hypothesis 1, it was expected that avoidant attachment would be associated negatively to relationship satisfaction to a greater extent in the confined group than in the comparison group. As shown in Figure 2, the association between avoidant attachment and relationship satisfaction was negative in both groups, being β=.41, p< .01 in the confined group and β=.44, p< .01 in the comparison group. Regarding the comparison between both beta values, as shown in Table 3, we did not find significant differences (t=.42, p=.68). Therefore, Hypothesis 1was not confirmed. Regarding Hypothesis 2, it was expected that own withdrawal (of conflict) would be associated with his/her low relationship satisfaction to a greater extent in the confined group than in the comparison group. As shown in Figure 2, there was a negative effect of own withdrawal on relationship satisfaction in both groups. Difference’s test was significant (t=2.20, p=.03), being higher in the confined group (β=.34, p< .01) than in the comparison one (β=.23, p< .01). Therefore, Hypothesis 2is confirmed (see Table 3). Concerning Hypothesis 3, it was expected that own withdrawal mediated between avoidant attachment dimension and relationship satisfaction to a greater extent in the confined than in the comparison group. The results showed that there were significant indirect effects in both groups (confined group =.13, p=.00, 95% CI =[.19, .07]; comparison group =.08, p=.03, 95% CI =[.14, .03]). However, there were no significant differences between the values for beta (t=1.18, p=.24). Therefore, Hypothesis 3was not confirmed. Hypothesis 4stated that the links between own withdrawal with the perception of partner’s demand would be higher in the confined group than in the comparison group (see Figure 2). The results show that there was a positive association between own withdrawal and the perception of partner’s conflict demand, both in the confined (β=.38, p< .001) and in the comparison group (β=.26, p< .001). Analyses showed that there were differences between the magnitude of the beta TABLE 2 Zero-order Correlations. 12345 1. Avoidant attachment –.25** .15** .22** .55** 2. Own withdrawal .31** –.26** .29** .44** 3. Partner demand .13*.33** –.29** .28** 4. Partner positive .37** .28** .26** –.49** 5. Relationship satisfaction .60** .53** .42** .43** – Note: Confinement group’s correlations are displayed below the diagonal, while comparison group’s correlations are displayed above the diagonal. *p< .05; **p< .01. TABLE 3 X 2 Differences Test Between Confined and Comparison Groups’Beta Values. βfor comparison group Standard error βfor confined group Standard error t differences βvalues comparison Avoidance attachment—Relationship satisfaction .41** .04 .44** .06 t=0.42,p=.68 β comparison =β confined Avoidance attachment—Own withdrawal .25** .07 .31** .08 t=0.57,p=.60 β comparison =β confined Avoidance attachment—Perceived partner positive .14*.09 .32** .08 t=1.50,p=.14 β comparison =β confined Own withdrawal—Perceived partner demand .26** .04 .34** .02 t=2.69,p=.01 β comparison <β confined Own withdrawal—Relationship satisfaction .23** .03 .34** .04 t=2.20,p=.03 β comparison <β confined Perceived partner demand—Relationship satisfaction .07 .03 .08 .11 t=0.59,p=.56 β comparison =β confined Own withdrawal—Perceived partner positive .26** .08 .08 .06 t=1.80,p=.07 β comparison =β confined Perceived partner positive—Relationship satisfaction .31** .03 .14*.04 t=3.40,p=.00 β comparison > β confined *p< .05; **p< .01. 436 AVOIDANCE, COUPLE RELATIONSHIP, AND LOCKDOWN 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License of the comparison group and the confinement group (t=2.69, p=.01), being higher in the case of the confinement group. Therefore, Hypothesis 4was confirmed. Considering the entire model (see Figure 2), the effect of the perceived partner’s demand strategy on relationship satisfaction is negative and small, both in the confined (β=12, p=.01) and the comparison (β=.07, p=.12) group. When we examine the mediating effects of perceived partner demand between own withdrawal and relationship satisfaction separately, there were significant indirect effects in the confined group (indirect effect =.06, p=.03, 95% CI =[.14, .01]) and in the comparison group (indirect effect =.04, p=.01, 95% CI =[.07, .01]). Therefore, perceived partner’s demand is a variable mediating between own withdrawal and relationship satisfaction in both groups. 4 No significant differences were found in the values of mediation between the confined and comparison groups (t=.44, p=.66). Therefore, Hypothesis 5was not confirmed. Hypothesis 6stated that the negative association between avoidant attachment dimension and perceived partner’s positive problem-solving would be higher in the confined group than in the comparison group (see Figure 2). The results show that there was a negative association between avoidant attachment and the perception of partner’s positive problem-solving, both in the confined group (β=.32, p=.00) and in the comparison group (β=.14, p=.02). The analyses showed that there were no significant differences between the beta values of the two groups (t=1.50, p=.14). Therefore, Hypothesis 6was not confirmed. With regard to Hypothesis 7, it was expected that perceived partner’s positive problem-solving strategy mediated between avoidant attachment dimension and relationship satisfaction to a greater extent in the confined than in the comparison group. The results showed that there were significant indirect effects in both groups (confined group =.06, p=.02, 95% CI =[.11, .03]; comparison group =.07, p=.03, 95% CI =[.13, .03]). However, there were no significant differences between the values for beta (t =0.28, p=.78). Therefore, Hypothesis 7was not confirmed. As can be seen in Table 4, the unconstrained model showed themostoptimalfit(χ 2 /df =1.45, p=.23, Tucker–Lewis index [TLI] =.98, comparative fit index [CFI] =.99, rootmean-square error of approximation [RMSEA] =.03). 5,6,7 Thus, we can conclude that the two groups fit independently and that no equality restrictions are imposed. This indicates that, although the variables represented in the model would apply to both confined and comparison groups, the way in which they interrelate, and the weight of the regression coefficients, differ between the two groups. Therefore, Hypothesis 8was confirmed. DISCUSSION The present study aimed at examining the effect of COVID19 home confinement on conflict dynamics and relationship satisfaction, highlighting the role of the avoidant attachment dimension. Some relevant findings are presented here. Avoidantly attached individuals’perceptions about the conflict resolution strategies—displayed by them and by their partners— are all factors that explain diminished levels of relationship satisfaction. Avoidantly attached individuals’relationship satisfaction was affected during the confinement, this decrease being explained by strategies used by them and perceived also in their partners during the conflict. Our study has also shed light on the association between some protective factors, like the role of positive problem-solving in relationship satisfaction. Finally, this study provides more in-depth understanding of the withdrawal/demand pattern role in the explanation of relationship satisfaction levels in more stressful context. Confinement versus non-confinement condition Regarding the association between avoidant attachment and relationship satisfaction, our results reveal that the confinement would be associated in some way with the decline of relationship satisfaction, even though this association also occurs in the comparison group. Therefore, confinement itself does not seem to have affected the relationship of these variables in avoidant individuals. It is possible that the period of confinement carried out in Spain was not long enough to produce relevant changes in the partner satisfaction of avoidantly attached individuals. As Overall et al. (2022) observed, it is possible that confinement creates a context in which the distancing strategies applied by avoidantly attached individuals are those that undermine cohesion. Stress could produce greater disconnection and avoidance of their partners (Bodenmann et al., 2007; Repetti et al., 2009), which would predict a decrease in relationship satisfaction (Falconier et al., 2015). Certainly, our results point in this direction, with the conflict withdrawal strategy being what would produce a greater decrease in partner satisfaction in the confined individuals than in the comparison group. TABLE 4 Multigroup Invariance Model of Relationship Satisfaction from Avoidant Attachment and Conflict Resolution Split by Lockdown Status. Model X 2 df X 2 /df p TLI CFI RMSEA LO90 HI90 Nested model X 2 df Significance level 1Unconstrained 2.92 2 1.45 .23 .98 0.99 .03 .00 .09 2 Structural weights 38.15 10 3.81 .00 .89 0.95 .07 .05 .09 2–1 35.23 8 .00 3 Structural covariances 39.24 11 3.57 .00 .90 0.95 .07 .05 .09 3–2 1.09 1 .00 4 Structural residual 133.74 16 8.36 .00 .72 0.78 .11 .09 .14 4–3 94.50 5 .00 Note: The most restrictive model with a good fit in italics. PsyCh JOURNAL 437 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Moreover, as pointed out by some authors, declines in relationship satisfaction during periods of stress, such as confinement, could be explained in individuals with insecure attachment through conflict dynamics (Mikulincer & Shaver, 2016; Simpson & Rholes, 2012,2017). Regarding the mediating effect of own conflict strategies between attachment and relationship satisfaction, the current study showed that withdrawal was a mediator between avoidant attachment and relationship satisfaction. These results follow other studies that found the negative effect of individuals’behavior in relationship satisfaction in response to COVID-19-related stress (e.g., Bar-Shachar et al., 2023). Although our results did not find differences between the association between avoidant attachment, withdrawal, and relationship satisfaction in confined and in non-confined individuals, this lack of difference could be due to the length of the confinement. Certainly, as Bar-Shachar et al. (2023) pointed out, deactivation and distancing strategies deployed by avoidantly attached individuals over a prolonged period of time may negatively impact both objective and subjective stress. All of this, in turn, could end up negatively impacting relationship satisfaction. This study has shown that the links between the conflict strategies used in the “own withdrawal-partner demand”pattern (i.e., because of someone’s withdrawal, his/her partner demands the conflict) increased in confinement. Some authors argue that under higher stress, there is a notable increase in negative conflict dynamics and behaviors (Schrodt et al., 2014; Williamson et al., 2013). In our study, we found that withdrawal/demand pattern was more evident during confinement. Being confined in a situation in which individuals are forced to adapt their routines, this restructuring could lead to higher levels of stress (Jurblum et al., 2020). Withdrawal and aggression/demand conflict resolution strategies could be the maladaptive expression of confinement. As for the association between the “own withdrawal-partner demand”pattern and relationship satisfaction, our results show that this pattern is not associated with low relationship satisfaction to a higher extent in the confinement group than in the comparison group. This could be for two reasons. On the one hand, a more detailed analysis of the data revealed that this result could be due to a gender difference in the magnitude and meaning of the association between perceived partner’s demand strategy and relationship satisfaction; beta value was negative and moderate in size for women, while for men it was positive and of a very small size. This result would be consistent with what other studies have shown about the differential effect of certain violent behaviors on relationship satisfaction based on gender (Katz et al., 2002; Williams & Frieze, 2005). Specifically, Williams and Frieze (2005) found that aggression perpetrated by their partners—both in milder behaviors (i.e., threats) and in more severe responses (i.e., pushing or grabbing)—predicted only women’s low marital satisfaction, not observing such relationship in the case of men. Nevertheless, a replication with a bigger and gender-balanced (also with a more sexual orientation-diverse) sample is necessitated before this tentative explanation may be accepted. On the other hand, it is possible to think that if confinement would have been longer (several months), withdrawing and perceiving a partner as demanding during the conflict could have had a higher impact on relationship satisfaction. As Marjanovica et al. (2007) pointed out, the longer the confinement, the greater the expression of negative conflict behaviors (i.e., withdrawal and demand), and these maladaptive conflict dynamics could impact more negatively in relationship quality and stability (Eldridge & Baucom, 2012). Regarding the association between avoidant attachment and perceived partner’s positive problem-solving, our results showed that there were not differences between the confined group and the comparison group. Although certain studies have found a negative relationship between the avoidant dimension of attachment and the perception that the partner is a source of support during COVID-19 (Vowels & Carnelley, 2021), there are no studies at present that demonstrate that there are differences between a confined sample and a previous sample in that variable, nor in the perception that the partner solves the problem positively. Therefore, we can conclude that regardless of the stress suffered, the mental models operating in individuals with avoidant attachment refer to the greater degree of self-sufficiency and independence from oneself and the perception that others are not able to provide support or display behaviors that benefit the individual in times of need (Mikulincer & Shaver, 2012). The results on the association between avoidant attachment, perceived partner’s positive problem-solving strategy and relationship satisfaction have not been as expected. The absence of differentiation between the confined and comparison group could be explained by the previous results. That is to say, if in avoidantly attached individuals there are no differences between the two groups in the use perceived that their partners make of positive problem-solving strategies, it is also not to be expected that this perception would have a greater impact on relationship satisfaction in the confined group than in the comparison group. Nevertheless, as we comment previously, this result could be explained by the specific duration of the confinement period (Marjanovica et al., 2007). Prolonged confinement might have resulted in the sustained perception that the partner does not resolve conflicts positively, leading to a decrease in relationship satisfaction in avoidantly attached individuals. Finally, the results indicate that, although the model proposed works in different contexts (non-confinement and confinement situations), there are significant differences in the magnitude of some of the relationships between the variables (own withdrawal-relationship satisfaction; own withdrawal-perceived partner demand), being stronger in the confinement group than in the comparison group. In addition, the differences in the magnitude of the other relationships between variables, although not significant, are in the same direction. Therefore, we can conclude that avoidantly attached individuals activate more their working models according to specific types of stressful circumstances (Simpson & Rholes, 2017). This results in avoidant individuals showing a tendency to be unsupportive, withdrawn, or uncooperative in their romantic relationships (Simpson & Rholes, 2017), but these characteristics may be more visible in some stressful conditions than in others. 438 AVOIDANCE, COUPLE RELATIONSHIP, AND LOCKDOWN 20460260, 2023, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pchj.646 by Universidad Del Pais Vasco, Wiley Online Library on [28/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License