Three Things That Affect the Decision to Leave a Relationship
The reasons people stay are rarely the ones they would name. What the research on commitment, conflict, and grief says about leaving a relationship.
We’ve all heard it from friends or relatives when we admit to being exhausted. The advice is free, immediate, and always the same. “Leave.” It tends to come from people who have never had to.
What’s missed is that the person hearing it already agrees. They have been agreeing for months. Agreeing has gotten them nowhere, and now they have two problems instead of one: the relationship and the shame of still being in it.
Staying is not a failure of nerve. The part of you that agrees with the advice isn’t the part that makes the decision.
Satisfaction is one number out of three
Social psychologist Caryl Rusbult spent her career demonstrating that commitment and happiness are not the same. Three factors predict whether someone will stay: their level of satisfaction, how the alternatives look, and how much they have already invested—years, money, a mortgage, a dog whose vet knows both their names, a network of friends, and a version of themselves that has been built around this person. Unhappiness pulls one way. The other two factors outweigh unhappiness more often than expected.
That is why the arithmetic looks impossible from inside and also obvious from outside. Your friend or relative with the wise advice is running one number. You are running three.
Rusbult and Martz tested this where it matters most, with women who had left violent partners for a shelter. What predicted who went back was commitment, built on what they had invested and how thin their alternatives looked; not on how bad the domestic violence had been. Wanting to leave and leaving are separate acts, and only one of them shows up in the advice.
The couples in trouble are not the ones fighting
Fighting is a poor diagnostic sign of a dysfunctional relationship.
Psychologist John Gottman’s observational work found that durable couples were not the ones who avoided conflict but the ones who could interrupt it and repair. A joke, a hand on the arm, a change of subject that both parties agree to accept. He also found couples who rarely argued and stayed together, and couples who shouted their way through four decades and stayed together. There is no correct pattern. His prediction percentages have taken fair criticism for being fitted and tested on the same couples, and they probably deserve less confidence than they get, but the observation underneath has held: What separates couples is not whether they fight but whether repair actually lands.
Which means fighting is a weak reason to leave, and never fighting is a poor reason to stay.
You learned to read the room before you could read
Long before anyone asked for your opinion on love, you were a child solving a problem: how to get what you needed from adults. A parent with unpredictable moods teaches a child to read faces quickly and react fast. A parent who falls silent when asked for something teaches a different lesson; ask for less, and you’ll lose less. Both children are right about the house they live in.
Neither child gives up those adaptive behaviors at eighteen. The first child grows into an anxious adult who can interpret four hours of a partner’s silence and explain its meaning. This person keeps the relationship alive in their mind for months after it has long since ended. The second child becomes an avoidant adult who leaves a fifteen-year marriage dry-eyed and is admired for handling it so well. Eight months later, they find themselves crying in the car outside a supermarket. Mikulincer and Shaver explain this delay. Shutting down requires continuous effort. When a person is exhausted, or frightened, this effort fails, and the feeling arrives all at once.
Your forecast is wrong in one direction
People often stay in relationships for reasons other than their partner. They’re afraid of what comes next: a bare apartment, awkward questions at Christmas when you’re alone, and the challenges of starting over in your forties.
This picture is flawed. Eastwick and his colleagues studied students going through breakups and compared their predictions of how they would feel with their actual feelings. People overestimated how much it would hurt and how long the pain would last. Those who were still in love overestimated the most.
This does not make the fear stupid or irrelevant. Forecasting pain is what keeps our hands off hot stoves. But one factor influencing the decision is consistently inaccurate, and it happens to be the loudest voice in the room. It’s worth knowing this before treating it as fact.
Leaving without saying why makes a loss without closure
When it does end, the cruelest exit is the silent one. Pauline Boss gave a name to loss without confirmation: ambiguous loss. The mind cannot file what has not been declared over, so it audits instead—rereading messages, running the last conversation again, searching for the sentence that explains it. That search can carry on in a loop for a long time.
A short, unpleasant conversation explaining why it’s over is not just a courtesy you extend to someone you no longer love. It lets them stop to try to figure it out.
Then, on both sides, there is grief to process. Grief for what you wished for and what was never there and never will be. It’s not just grieving for someone who is still alive and posting holiday photos. It’s also grieving for a comfortable routine and a fantasized future. Kenneth Doka calls this disenfranchised grief: a loss that no one allows you to grieve. There’s no funeral, no cards, and no time off work. By day nine, a friend asks if you’re over it yet. People try to skip this stage, but it lurks around the corner, waiting for them with a vengeance.
The goal of couples therapy isn’t always to keep the couple together. Some people who come to my office are not going to stay together, and they both kind of know it when they arrive. The goal is sometimes to end things better than they would manage alone. It’s about saying the truth instead of something hurtful, dividing a decade without making it seem like criminal evidence, and allowing each person to mourn something real instead of rewriting it as wasted time.
What the struggle actually indicates
The question that arises in therapy is rarely about the partner. It’s usually some version of “What’s wrong with me?” But nothing is actually wrong. The pull you are fighting is the same mechanism that allows you to form attachments. It is operating as designed, but in a situation it was not built for. You can’t love deeply and detach cleanly. You can’t have one without the other. Deal with it.
What you can get is better arithmetic: Three numbers instead of one. A forecast that is what it is: You can get an honest look at which reasons for staying are facts and which are memories pretending to be facts. Some relationships end because they have run their course, and acknowledging that is not the same as calling them a mistake.
But nobody makes that call from inside a fog. Clear the fog first. Then, whichever way it goes, the decision will belong to someone who can see.
Research references
Caryl E. Rusbult, John M. Martz, and Christopher R. Agnew, “The Investment Model Scale: Measuring Commitment Level, Satisfaction Level, Quality of Alternatives, and Investment Size,” Personal Relationships 5, no. 4 (1998): 357–87, https://doi.org/10.1111/j.1475-6811.1998.tb00177.x.
Caryl E. Rusbult and John M. Martz, “Remaining in an Abusive Relationship: An Investment Model Analysis of Nonvoluntary Dependence,” Personality and Social Psychology Bulletin 21, no. 6 (1995): 558–71, https://doi.org/10.1177/0146167295216002.
John M. Gottman, “The Roles of Conflict Engagement, Escalation, and Avoidance in Marital Interaction: A Longitudinal View of Five Types of Couples,” Journal of Consulting and Clinical Psychology 61, no. 1 (1993): 6–15.
John M. Gottman et al., “Predicting Marital Happiness and Stability from Newlywed Interactions,” Journal of Marriage and the Family 60, no. 1 (1998): 5–22.
Richard E. Heyman and Amy M. Smith Slep, “The Hazards of Predicting Divorce without Crossvalidation,” Journal of Marriage and Family 63, no. 2 (2001): 473–79, https://doi.org/10.1111/j.1741-3737.2001.00473.x.
Mario Mikulincer and Phillip R. Shaver, Attachment in Adulthood: Structure, Dynamics, and Change, 2nd ed. (New York: Guilford Press, 2016).
Paul W. Eastwick et al., “Mispredicting Distress Following Romantic Breakup: Revealing the Time Course of the Affective Forecasting Error,” Journal of Experimental Social Psychology 44, no. 3 (2008): 800–807.
Pauline Boss, Ambiguous Loss: Learning to Live with Unresolved Grief (Cambridge, MA: Harvard University Press, 1999).
Kenneth J. Doka, ed., Disenfranchised Grief: Recognizing Hidden Sorrow (Lexington, MA: Lexington Books, 1989).