What the data actually shows
For decades, several studies reported a cohabitation effect: people who cohabited before marrying showed somewhat higher rates of marital instability than those who did not. This finding was influential and widely repeated. But researchers increasingly came to see much of it as a selection effect — the kinds of couples who cohabited (for example, those marrying younger, or with fewer economic resources, or with more uncertainty about the relationship) also tended to have higher divorce risk for reasons unrelated to having shared an address.
The picture also depends heavily on cohort. As cohabitation went from uncommon to near-universal, the gap between those who cohabited and those who did not narrowed, because cohabiting before marriage stopped marking out an unusual group. Research by scholars including Wendy Manning and colleagues has pointed to this cohort dependence, with the association weakening or disappearing for more recent marriages. What looked like a stable fact turned out to be a moving one, tied to the era being studied.
More recent analyses add nuance about timing. Work by Michael Rosenfeld and Katharina Roesler (2019) suggested the relationship can differ over the short versus long term — for instance, premarital cohabitation associated with lower divorce risk in the first year of marriage but somewhat higher risk over longer horizons in their data. Other researchers dispute the size and interpretation of such effects. The upshot is a literature that is unsettled, sensitive to how and when the data is measured, and far from a simple yes or no.
Why this feels different from how it actually is
The idea that living together first must affect divorce feels intuitive because it sounds like a controlled test — try before you commit. But couples who cohabit and couples who do not differ in many ways before anyone moves in, so comparing their later divorce rates is not comparing like with like. Much of what early studies attributed to cohabitation was really about pre-existing differences between the two groups.
The belief also lingers because the cohabitation effect was a striking, memorable finding that circulated widely before the picture grew more complicated. Headlines and advice tend to keep repeating the original, cleaner version long after the research has hedged and qualified it. So the cultural story can lag well behind the evidence.
And the topic carries moral weight that distorts how the numbers are read. Because cohabitation has been entangled with values debates, findings on either side get amplified to support a position. That makes it easy to encounter a confident claim — in either direction — that the underlying research does not actually support.
What the research says to do about it
To the extent the research offers practical signal rather than a verdict, several scholars emphasise the difference between sliding into living together and deciding to. Some work suggests that couples who move in gradually, by default or convenience, without a clear shared sense of commitment, may face more difficulty than those who make the decision deliberately — a distinction sometimes framed as deciding versus sliding. This is about how a transition happens, not about cohabitation being inherently risky.
The broader literature on what predicts marital stability points consistently to factors that have little to do with whether you cohabited first: marrying at a very young age, financial strain, and the quality of communication and commitment in the relationship tend to matter far more. Those are the variables the data most reliably connects to outcomes.
Given how contested and cohort-dependent the cohabitation finding is, the honest takeaway is to treat it as weak, debated context rather than a rule to organise your life around. The decision to live together is a personal one shaped by your circumstances and values, and the evidence does not give it the predictive power the older headline implied.
What the research says does not help
Treating the old cohabitation effect as a settled law of relationships does not help, because the research has substantially complicated it. Deciding for or against living together purely to avoid a divorce risk that is contested, modest, and largely explained by selection means acting on a number the evidence no longer firmly supports.
Reading either side's confident claims at face value is also unreliable. Because the topic is value-laden, both the it-causes-divorce and the it-makes-no-difference versions get overstated. The accurate position is genuinely uncertain, and sources that sound certain are usually flattening that uncertainty.
Using these statistics to judge your own or others' choices is not supported by the data and is beside the point. Population-level associations, even where real, say nothing reliable about an individual couple, and the research gives no grounds for moral conclusions about how anyone chooses to build their relationship.
Real numbers in context
There is no single trustworthy figure for how much cohabitation changes divorce risk, because the estimates move with the cohort and the method. Older studies found a measurable gap; more recent analyses find it weakened, absent, or split by time horizon — for example, Rosenfeld and Roesler (2019) reported lower first-year but somewhat higher longer-term divorce risk associated with premarital cohabitation in their data, while other researchers contest those results. Any precise-sounding percentage should be treated with caution.
What is clearer is the backdrop: cohabitation before marriage has become the norm rather than the exception across many countries, so the cohabiting group now looks much like the general population of couples. That alone helps explain why the early effect has faded — the behaviour stopped distinguishing one kind of couple from another.