![[Adobe Stock]](https://sp-ao.shortpixel.ai/client/to_webp,q_glossy,ret_img,w_750,h_375/https://www.psypost.org/wp-content/uploads/2026/03/upset-couple-750x375.jpg)
When a romantic partner refuses to apologize after hurting them, people who worry about rejection tend to report holding more grudges, pulling away, and withholding forgiveness, according to new research published in The Journal of Social Psychology. Across three studies, the findings suggest that a partner’s response after a conflict may shape how anxiously attached people react, and an experiment hinted that being ignored altogether may be especially hard on them.
Every romantic relationship involves occasional transgressions, which happen when one partner breaks an explicit or unspoken relationship rule. How couples navigate the aftermath of these missteps is shaped in part by their adult attachment style, a psychological pattern of expectations and insecurities regarding closeness and dependability in intimate bonds. Psychologists generally measure attachment insecurity along two dimensions: avoidant attachment, which involves discomfort with intimacy and a preference for emotional independence, and anxious attachment, which combines a strong craving for closeness with an intense fear of being rejected or abandoned.
Because people high in attachment anxiety often doubt their own worth and worry about whether their partner truly cares for them, relationship conflicts can quickly set off their deepest insecurities. In a 2005 diary and observation study of dating couples, individuals with higher attachment anxiety perceived more daily relationship conflicts, felt more distressed when discussing major problems, and tended to escalate arguments rather than calm them down. When a partner commits a wrongdoing, victims can react with constructive behaviors like forgiveness, which helps restore trust, or destructive behaviors like avoiding the partner, nursing a grudge, and refusing to forgive, which can erode the relationship over time.
Past studies looking at whether anxious attachment is linked to destructive reactions after a wrongdoing have turned up mixed results. For instance, a 2019 meta-analysis of 26 studies indicated that people with higher attachment anxiety generally have a harder time forgiving others who hurt them. However, a 2015 study co-authored by C. Ward Struthers, one of the authors of the new paper, found that anxiously attached adults reported more grudge-holding and avoidance after a past wrongdoing in an initial survey, but those links were not consistently replicated in a second study in which attachment styles were induced experimentally.
Don’t miss the next big psychology finding
Every morning, PsyPost sends a short briefing on the most interesting new peer-reviewed research in psychology and neuroscience.
The new research, led by Brianna Davison, a PhD candidate in the University of Calgary Love Lab who conducted the research while at York University, alongside Struthers, Jingyuan Sophie Li, and Amy Muise, examined whether the offending partner’s behavior after a transgression helps explain when anxiously attached victims turn to destructive responses.
Rather than offering a sincere apology, offending partners often try to protect their own image by giving a “nonapology.” In the framework the researchers draw on, a nonapology is not merely the absence of an apology, but an active, defensive reaction that pairs refusing to say sorry with behaviors like making excuses, denying fault, blaming the victim, or lashing out. Because anxiously attached people tend to be especially attentive to signs of whether their partner is emotionally available and responsive to their needs, Davison and her colleagues hypothesized that a partner’s nonapology would signal unavailability and prompt anxiously attached victims to react destructively.
“Previous research tells us that adult attachment plays an important role in how individuals initially respond to and manage relational wrongdoing in their romantic relationships; however, far less is known about how adult attachment shapes individuals’ long-term responses after their romantic partner wrongs them,” Davison told PsyPost. “Our research fills the gap by exploring whether anxiously attached individuals are more likely to hold grudges, avoid their partner, or struggle to forgive following a relational wrongdoing and whether these long-term responses are shaped by a partner’s refusal to apologize. Understanding how individuals with anxious attachment respond to relational wrongdoing matters because these responses can have lasting consequences for relationship quality, potentially making it harder for them to maintain the close, secure relationships they desire.”
To test these ideas, the researchers conducted three preregistered studies and made their data, survey materials, and analysis code publicly available, though the main analysis in Study 1 departed from its preregistered plan, which called only for simple correlations. In the first study, the team surveyed 242 adults comprising 121 romantic couples who had been together for at least six months, with an average age of 36 and an average relationship length of roughly 12.6 years.
Each participant completed a 12-item questionnaire assessing their anxious and avoidant attachment styles on a 1-to-7 scale and then wrote about a recent, unresolved conflict in which their partner had wronged them. Even though both members of each couple took part, each person recalled a conflict that stood out to them personally rather than both partners reporting on the same event.
After writing about the conflict, participants rated the extent to which their partner had responded with a nonapology, such as shifting blame onto them. They also rated their own destructive reactions on 1-to-7 scales across three areas: holding a grudge characterized by disdain and contempt, avoiding or withdrawing from their partner, and withholding forgiveness. The researchers combined these three ratings into a single overall score for destructive responses and statistically adjusted for avoidant attachment in all primary analyses to isolate the specific role of attachment anxiety.
In Study 1, looking at simple correlations alone suggested that people with higher attachment anxiety reported more destructive reactions, with an overall sample average of 1.79 out of 7 on the destructive response scale. Once the researchers accounted for avoidant attachment, however, the direct link between attachment anxiety and destructive responses was not statistically significant.
Instead, the relationship depended on how the partner had reacted after the wrongdoing: when participants perceived low levels of nonapologetic behavior from their partner, attachment anxiety had no statistically significant link to destructive responses, but when participants perceived high levels of nonapology, higher attachment anxiety was associated with more destructive behavior.
To verify these results in a larger, independent group and improve the precision of their attachment measure, the researchers conducted a second study with 538 adults in exclusive romantic relationships lasting at least six months. Participants in Study 2 had an average age of 25 and an average relationship length of 5.5 years, and unlike the first study, only one person from any given couple was allowed to participate. The researchers also replaced the brief attachment survey from Study 1 with a longer, 36-item attachment questionnaire before asking participants to recall an unresolved partner transgression and rate their partner’s nonapology and their own destructive reactions.
After accounting for avoidant attachment in Study 2, higher attachment anxiety was actually linked to slightly fewer destructive responses overall, the opposite of what the researchers predicted. But the pattern depended on the partner’s response, replicating the interaction from the first study. When participants perceived their partner as showing low levels of nonapology, the association between attachment anxiety and destructive responses was not statistically significant. When partners were seen as highly nonapologetic, however, victims with higher attachment anxiety reported more destructive responses overall, and the link between perceiving a nonapology and reacting destructively was more than twice as strong among high-anxiety victims as among low-anxiety victims.
“When an individual is wronged by their romantic partner, how their partner responds afterward can be just as important as the wrongdoing itself,” Davison explained. “For individuals with an anxious attachment style, a partner who refuses to apologize, shifts blame, denies their actions, or makes excuses may contribute to more destructive responses, such as an individual holding grudges, avoiding their partner, and struggling to forgive. Put simply, an individual’s attachment style may set the stage, but their romantic partner’s response following the wrongdoing can shape how the story unfolds.”
Because the first two studies were correlational, they could not establish whether a partner’s behavior actually causes destructive reactions. To test this experimentally, Study 3 recruited 334 undergraduate students in romantic relationships lasting an average of 2.5 years, with an average age of 22. Participants wrote about a real unresolved conflict in which their partner had hurt them and were then randomly assigned to imagine one of three scenarios immediately afterward: receiving an apologetic text message from their partner, receiving a nonapologetic text message in which their partner blamed and lashed out at them, or a control condition in which their partner did not respond and they instead received a routine university notification on their phone.
After imagining the scenario, participants answered questions checking their comprehension of the message and completed the attachment questionnaire, along with 1-to-7 scales rating how emotionally available and responsive their partner seemed and how destructively they would react. Comparing the apology condition against the other two conditions combined did not produce a statistically significant interaction with attachment anxiety.
Comparing the nonapology and no-response conditions, however, produced a surprising result: both high-anxiety and low-anxiety participants reported more destructive reactions when imagining a defensive nonapology than when imagining no response, with low-anxiety participants showing an especially sharp jump, in comparisons the authors describe as exploratory. As a result, within the nonapology group itself, the difference between high- and low-anxiety participants was not statistically significant.
In the control condition where the partner sent no response at all, however, participants with higher attachment anxiety reported more destructive reactions than those with lower attachment anxiety. The researchers also ran a statistical mediation analysis to see whether perceptions of partner availability and responsiveness connected attachment anxiety to destructive behavior, though they noted that because these perceptions were measured rather than experimentally manipulated, the analysis cannot establish a definitive causal chain.
Contrary to their hypothesis, partner responsiveness was not a statistically significant pathway in the nonapology condition; instead, in the no-response condition, higher attachment anxiety was associated with viewing the silent partner as less available and responsive, which in turn was related to more destructive conflict responses.
In their discussion of the findings, the authors offer two possible explanations for why getting no response in the control condition was so troubling for anxiously attached participants. First, the researchers suggest that even though a nonapologetic text message is hostile, it still acknowledges the victim and leaves an opening for further conversation and eventual relationship repair, whereas being completely ignored or “ghosted” after a fight may signal total disengagement. Second, the authors propose that a partner’s silence creates ambiguity, and because anxiously attached individuals tend to interpret uncertain situations pessimistically, receiving no text back may lead them to assume their partner is unavailable and uncaring.
These results align with an earlier study on responses to hypothetical infidelity, which similarly linked higher attachment anxiety to unforgiving reactions such as revenge and avoidance motivations, although that study examined imagined sexual unfaithfulness rather than a partner’s refusal to apologize after everyday relationship conflicts. Study 3’s finding that a partner’s silence was linked to more destructive responses among anxiously attached participants also echoes a daily diary study of romantic couples, which focused on everyday phone snubbing rather than post-conflict nonapologies but likewise indicated that anxiously attached partners are more prone to retaliate when they feel ignored or slighted by their partner.
There are a few things to keep in mind when interpreting these findings. Across the studies, overall destructive response scores were relatively low, averaging 1.79 out of 7 in Study 1 and 2.05 out of 7 in Study 2, and the way attachment anxiety and a partner’s nonapology combined to predict those responses was modest in size. In addition, the researchers relied on one partner’s personal recollection of a conflict rather than independent observations or reports from both partners about the same argument.
Because people high in attachment anxiety tend to notice and remember negative relationship details more readily than positive ones, the authors note that highly anxious participants in the first two studies may have overlooked softer gestures from their partner and blended their distress over the argument itself with their rating of their partner’s nonapology.
The experimental setup in the third study also carries some caveats. Because imagining no response from a partner after a fight turned out to be distressing in its own right for anxiously attached individuals, the control condition did not function as a truly neutral baseline, and running extra exploratory comparisons across the experimental groups increased the statistical chance of a false-positive finding.
Participants in Study 3 also filled out the attachment questionnaire after imagining their partner’s text message, which could have colored how anxious they rated themselves. In addition, they answered comprehension questions about the message right before rating their destructive responses, which the authors note may have drawn attention to the manipulation’s purpose and inflated the observed effects.
Finally, the conflicts participants wrote about across all three studies were generally moderate in severity. Future research will be needed to see whether these dynamics look different for minor daily annoyances that build up over time or for major betrayals.
The study, “The Association Between Victims’ Anxious Attachment and Transgressors’ Nonapology on Conflict Resolution in Romantic Relationships,” was authored by Brianna Davison, C. Ward Struthers, Jingyuan Sophie Li, and Amy Muise.
