Before going further, one thing is worth saying plainly: this research applies to relationships marked by ordinary, painful conflict and distance, not to relationships involving abuse or coercive control. If that’s what someone is facing, the research on recovery below simply doesn’t apply, and safety, not repair, is the priority. With that said, for the far more common experience of a relationship that has genuinely deteriorated through disconnection, resentment, or years of unresolved conflict, two studies suggest real recovery is more possible than people often assume.
The clearest evidence comes from a 2004 study by Andrew Christensen and colleagues, published in the Journal of Consulting and Clinical Psychology. The researchers ran a randomized controlled trial with 134 married couples who were, by design, significantly and chronically distressed, not couples with mild, everyday friction but ones who had been struggling seriously for a long time. Couples were randomly assigned to one of two established therapy approaches and stratified by how severe their distress was at the start. By the end of treatment, 71 percent of couples in one approach and 59 percent in the other showed clinically significant improvement on a standard measure of marital adjustment, and this recovery showed up even among the most severely distressed couples in the study, not only those who started out relatively better off. Genuine, measurable improvement wasn’t limited to relationships that were merely struggling, it extended to marriages that had been in real trouble for years.
A second study helps explain something related but different: why people stay in a difficult relationship in the first place, and what that decision is actually built on. In a 2003 meta-analysis by Benjamin Le and Christopher Agnew, published in Personal Relationships, the researchers statistically pooled 52 studies covering roughly 11,000 participants across five countries, testing the investment model of commitment. Satisfaction with the relationship, the quality of available alternatives, and how much a person had already invested in the relationship, time, shared history, shared resources, together accounted for a large, consistent share of how committed people felt to staying. Notably, commitment itself, not satisfaction alone, was the strongest predictor of whether a relationship actually persisted, meaning people sometimes stay in relationships they’re not currently satisfied with because of investment and a lack of appealing alternatives, not necessarily because they still believe things can improve.
Together, these two studies offer a genuinely useful, two-part answer to whether staying in a difficult relationship ever makes sense. Christensen and colleagues’ research shows that real, clinically meaningful recovery is possible even in marriages that have been seriously and chronically distressed, provided both people engage in structured, evidence-based repair work. Le and Agnew’s research offers an important caution alongside that hope: people often stay for reasons that have little to do with genuine belief in recovery, investment and a lack of alternatives can keep someone in a relationship independent of whether real repair is actually happening or even being attempted. Read together, the research suggests recovery from serious distress is realistic, but only when it’s an active, mutual process, not simply staying because leaving feels harder than staying does.
It’s worth being honest about what these two studies don’t establish. Christensen and colleagues’ research measured outcomes at the end of a structured therapy program with trained clinicians, so it speaks to what’s possible with real, sustained repair effort from both partners, not to relationships where only one person is trying or where no active change is happening at all. Le and Agnew’s meta-analysis describes what predicts staying, not what should predict it, and a relationship persisting because of high investment or poor alternatives is a very different thing from a relationship persisting because it’s genuinely improving.
Within those honest limits, and setting aside anything involving abuse or coercive control, where the research above simply doesn’t apply, the answer this research points toward is a conditional one. Staying can make sense when real, mutual repair work is actually happening and showing signs of working, the way it did for most couples in Christensen’s trial. It makes far less sense when the reasons for staying are really just sunk investment and a shortage of visible alternatives, since neither of those, on their own, predicts anything getting better.