If you are staying in a marriage after finding out about infidelity, you may have heard some version of "forgiveness just takes time." A new piece of betrayal trauma research suggests that idea leaves out the most important part of the picture. Barlık and Kocabas (2026) built machine learning models to predict marital forgiveness after infidelity. The models used data on trust, trauma, attachment, distress, and relationship quality from married individuals who stayed together after disclosure.
What the study asked
The researchers wanted to know which factors best predict whether a betrayed spouse moves toward forgiveness after an affair. They looked at relational trauma, attachment anxiety, attachment avoidance, psychological distress, dyadic trust, and relationship quality, and asked which of these predicted forgiveness most strongly.
To answer this, they collected data from 548 married individuals in Turkey. All had experienced emotional, sexual, or combined infidelity and remained in their relationships after disclosure. Participants were recruited from counseling centers, family therapy clinics, and online support networks. The researchers then used several machine learning models, including XGBoost, LightGBM, and Random Forest, to see which one could most accurately predict forgiveness scores. They also used a technique called SHAP to find out which factors each model relied on most.
What it found
The study found that forgiveness was closely tied to specific relational and psychological factors.
- Forgiveness was negatively associated with relational trauma, attachment anxiety, attachment avoidance, and psychological distress. In other words, higher trauma and higher distress went along with lower forgiveness.
- Forgiveness was positively associated with dyadic trust and relationship quality. Higher trust and higher relationship quality went along with higher forgiveness.
- The best performing model, XGBoost, explained 84% of the variance in forgiveness scores. LightGBM and Random Forest performed slightly less well but were still strong.
- When the researchers ranked which factors the model relied on most, dyadic trust came out as the single strongest predictor. Next came relational trauma, relationship quality, attachment anxiety, psychological distress, attachment avoidance, and time since disclosure.
Time since disclosure was on the list, but it ranked last. Trust, trauma, and relationship quality carried far more weight in predicting forgiveness than how much time had passed.
What this means for you
If you are a betrayed partner, this betrayal trauma research offers something concrete. Forgiveness appears to depend on three things. These include whether trust is being rebuilt, whether the trauma is being addressed, and whether the relationship feels solid again. If you feel far from forgiveness even after a long time, it may mean trust and trauma still need direct attention.
If you are the partner who broke trust, this study points to where your effort has the most impact. Rebuilding trust and helping your partner process the trauma of betrayal appear to have more impact on the outcome than time alone. Attachment patterns and psychological distress also played a role, which suggests that both partners' internal states shape how the relationship heals.
For clinicians, this study suggests that trauma-informed and attachment-focused work with couples fits what predicts forgiveness. This approach may work better than asking couples to reach forgiveness on a set timeline. Trust restoration and processing of relational trauma stood out as the strongest predictors in this sample. This may help clinicians decide where to focus early sessions.
What it does not prove
This was a cross-sectional study, meaning everyone was measured at one point in time. That means the researchers can say trust and trauma were strongly associated with forgiveness. They cannot say for certain that building trust causes forgiveness. It is possible that couples who already feel closer to forgiveness also report higher trust, rather than trust coming first.
All the data came from self-report questionnaires. The findings reflect participants' own descriptions of their trust, trauma, and distress. The sample was drawn from married individuals in Turkey who were recruited through counseling centers, clinics, and online support networks. All had chosen to stay in their relationships after disclosure. This means the findings may not apply the same way to couples who separated, to unmarried couples, or to people in other countries and cultures. The machine learning models were highly accurate at predicting scores in this dataset. Strong prediction is not the same as proof of what causes forgiveness in real life.
Read the original
This study is open access, and the full citation is below if you want to read it in full.
About this study
Meral Barlık, Sevde Kocabas (2026). Predicting Marital Forgiveness after Infidelity Using Explainable Artificial Intelligence and Relational Trauma Indicators. Research and Practice in Couple Therapy. https://doi.org/10.61838/rpct.4.2.7
Open access: the full paper is free to read at the link above.
Cite this summary: Sanatio Research (2026). New betrayal trauma research on trust and forgiveness. https://sanatioresearch.org/blog/betrayal-trauma-trust-forgiveness
Reviewed by Spencer Posey, Licensed Marriage and Family Therapist (LMFT #141641), CSAT. Editor, Sanatio Research.
Educational content, not medical or therapeutic advice. Reading this creates no therapist-client relationship. If you are in crisis, call or text 988.