Ione Bretaña, Itziar Alonso-Arbiol, Patricia Recio, Fernando Molero
8 min
This study explores the relationship between attachment style—specifically the avoidance dimension—and the quality of romantic relationships through the lens of conflict resolution strategies. Drawing on data from 175 heterosexual couples, the authors investigate whether a specific pattern, the actor’s withdrawal paired with the partner’s demand-aggression, acts as a mechanism that explains why avoidant individuals often report lower relationship satisfaction. The team used the Actor–Partner Interdependence Model with Mediation Analysis (APIMeM) to test how these strategies correlate both within an individual and across the dyad.
The research confirms that avoidant attachment is a significant predictor of lower relationship satisfaction for both the individual and their partner. The core of the findings reveals a dyadic cycle: an individual's higher levels of avoidant attachment correlate with their increased use of withdrawal strategies during conflict. This withdrawal then triggers a demand/aggression strategy from the partner. This specific interaction—the actor withdrawing while the partner demands—is consistently associated with lower satisfaction for both members of the couple. While withdrawal acts as a mediator for the actor's own satisfaction, the study did not find that it mediates the link between the actor’s attachment and the partner’s satisfaction, suggesting that partners may employ complex compensatory mechanisms to maintain the relationship.
Understanding these cyclical dynamics is critical for clinical practice. By identifying that withdrawal is not just an individual defensive maneuver but a catalyst for partner aggression, therapists can provide couples with better tools to break these maladaptive patterns. Recognizing that avoidant attachment can manifest as both withdrawal and, in some cases, demand-aggression, allows practitioners to address the underlying insecurity driving these conflict behaviors, rather than just the surface-level arguments.
This study was conducted with the purpose of analyzing the combined and mediating effect of actor’s withdrawal–partner’s demand conflict resolution strategies between avoidance attachment dimension and relationship satisfaction. We conducted a dyadic study with 175 heterosexual couples (aged between 18 and 72 years) who filled in the questionnaires. Six hypotheses were tested using the actor–partner interdependence model with mediation analysis (APIMeM). Results showed that the avoidance dimension of attachment was more strongly associated with actor’s withdrawal strategy than with demand/aggression strategy. Furthermore, avoidance attachment was negatively associated with both actor’s and partner’s relationship satisfaction, the actor effect being higher. Withdrawal strategy was a mediator between actor’s avoidance and actor’s relationship satisfaction, but it was not a mediator for partner’s relationship satisfaction. The interactive pattern of actor’s withdrawal–partner’s demand/aggression was associated with low levels of both actor’s and partner’s relationship satisfaction. These results point out to the need of discerning the interactive pattern of conflict-solving strategies as well as their intertwined effect on relationship satisfaction.
Alex: What did the sample look like? That matters if we’re imagining either new relationships or couples already seeking treatment.
Sam: These were community couples in Spain, recruited through municipal courses and personal networks, not a therapy sample. Their relationships averaged about thirteen years. Most participants were Spanish, and roughly sixty percent had higher education.
Alex: So relatively established relationships, with a fairly educated sample. How did the analysis keep track of whose behavior related to whose satisfaction?
Sam: They used an actor–partner interdependence model. It separates associations within a person from associations across partners. An actor association connects my withdrawal with my satisfaction; a partner association connects my withdrawal with your satisfaction.
Alex: And where does attachment enter? Is withdrawal supposed to explain part of its association with satisfaction?
Sam: That is the mediation question: whether an association runs statistically through an intermediate variable. Here, the proposed pathway connects attachment avoidance to withdrawal, then to satisfaction. The researchers also examined pathways through the other partner’s demanding behavior.
Alex: Before we follow those pathways, what is the clearest result in the data?
Sam: One partner’s withdrawal was moderately associated with the other partner’s demand/aggression. For men’s withdrawal and women’s demanding behavior, the correlation was about point five four. Reversing the partners’ roles gave about point five six, so the association was similar in both directions.
Alex: That argues against treating withdrawal as exclusively a male role and demanding behavior as exclusively a female role, at least in this sample.
Sam: The pattern appeared in both directions. And the authors found no significant gender differences in the within-person associations linking either conflict strategy with satisfaction. Men reported slightly higher attachment avoidance, but that difference was small.
Alex: What happened when attachment and the conflict strategies were considered together?
Sam: Greater attachment avoidance was associated with more withdrawal, more strongly than with demand/aggression. Withdrawal partly accounted statistically for the association between avoidance and a person’s own lower satisfaction. The authors did not conclude that withdrawal mediated the association with the partner’s satisfaction.
Alex: That distinction seems central. Did withdrawal simply have no association with how the partner felt?
Sam: It did have an association: more withdrawal was linked to lower partner satisfaction. But that alone does not establish the full mediation pathway from attachment avoidance. The authors’ headline conclusion distinguishes those two claims.
Alex: Is the mediation reporting straightforward enough to take that conclusion at face value?
Sam: There is a reporting tension. Some indirect effects are statistically significant even where the authors label the pathway “no mediation.” For anyone building on those results, I would check the reported estimates and the authors’ mediation criteria rather than relying only on that label.
Alex: What about the connected pattern itself? Does the partner’s demanding behavior add anything beyond withdrawal alone?
Sam: In their analysis, it did. Withdrawal was indirectly associated with lower satisfaction through the partner’s demand/aggression. Those indirect associations appeared for both the withdrawing person’s satisfaction and the demanding partner’s satisfaction.
Alex: So the statistical picture is not simply that withdrawing people dislike their relationships. Their partner’s reported way of handling conflict is part of the pattern too.
Sam: That is the contribution. But the model also needed adjustment. The originally proposed model did not fit the data adequately; adding a path linking avoidance with a person’s own demand/aggression produced a good fit.
Alex: They had a theoretical reason for that addition, but it was made after looking at the data. How does that affect your reading?
Sam: I would treat the final model as an account to test further, not a uniquely established explanation. The authors discuss why avoidant people might sometimes demand rather than withdraw. But their data do not show when someone switches strategies.
Alex: And even a well-fitting model cannot tell us which behavior happened first. Is that the biggest limitation here?
Sam: Yes. All measures came from the same assessment period, so the study is cross-sectional: a snapshot, not a sequence. It cannot establish that withdrawal caused demands, or that either behavior caused dissatisfaction.
Alex: Collecting both partners’ reports improves the perspective, then, without turning their answers into observations of an unfolding argument.
Sam: Both partners still used self-report questionnaires. The authors recommend future work combining questionnaires with interviews or behavioral observation, and following couples over time. Generalization also needs testing beyond these Spanish heterosexual couples of relatively advantaged socioeconomic status.
Alex: Given those limits, who should read the full document, and where should they start?
Sam: Researchers studying attachment, couple conflict, or dyadic mediation should start with the Partner Effects results and the Direct and Indirect Effects analysis. Then read the Discussion’s limitations. Those sections show both the contribution and the reasons to be cautious.
Alex: And for therapists, or researchers outside this area who only retain one idea?
Sam: Therapists may find the pattern useful for recognizing connected conflict behaviors, but this study did not test a treatment. For everyone else: withdrawal and demands belong in the same picture—but this snapshot cannot tell us which came first.
Alex: A useful distinction to carry into the next study you read.