Likelihood of split
Conflicting studies on the effect of cohabitation on subsequent marriage have been published. In countries where the majority of people disapprove of unmarried individuals living together, or a minority of the population cohabits before marriage then marriages resulting from cohabitation are more prone to divorce. But in a study on European countries, those where around half of the population cohabits before marriage, cohabitation is not selective of divorce-prone individuals, and no difference in couples that have cohabited before and after marriage is observed. In countries such as Italy, the increased risk of marital disruption for people who experienced premarital cohabitation can be entirely attributed to the selection of the most divorce-prone into cohabitation.
In 2002 the CDC found that for married couples the percentage of the relationship ending after 5 years is 20%, for unmarried cohabitators the percentage is 49%. After 10 years the percentage for the relationship to end is 33% for married couples and 62% for unmarried cohabitators.  
A 2004 study of 136 couples (272 individuals) from researchers at the University of Denver found differences among couples that cohabited before engagement, after engagement, or not until marriage. The longitudinal study collected survey data collected before marriage and 10 months into marriage, with findings suggesting those who cohabit before engagement are at greater risk for poor marital outcomes than those who cohabit only after engagement or at marriage. A followup survey by the researches of over 1,000 married men and women married in the past 10 years found those who moved in with a lover before engagement or marriage reported significantly lower quality marriages and a greater possibility of a separation than other couples.  About 20 percent of those who cohabited before getting engaged had since suggested splitting - compared with only 12 percent of those who only moved in together after getting engaged and 10 percent who did not cohabit prior to marriage.
The researchers from Denver suggest that relationships with pre-engagement cohabitation "may wind up sliding into marriage", whereas those that only cohabit post engagement or marriage make a more clear decision. This could explain their 2006 study of 197 heterosexual couples finding that men who cohabited with their spouse before engagement were less dedicated than men who cohabited only after engagement or not at all before marriage. In some heterosexual couples, women are more likely to understand cohabitation as an intermediary step preceding marriage, and men more likely to perceive it without an explicit connection to marriage.  
An analysis of data from the CDC's National Survey of Family Growth data from 1988, 1995, and 2002 suggests that the positive relationship between premarital cohabitation and marital instability has weakened for more recent birth and marriage cohorts, as the total number of couples cohabitating before marriage has increased.
Later CDC work found that between 2002 and 2006-2010, the number of couples in opposite-sex cohabiting relationships increased from 9.0% to 11.2% for women, and from 9.2% to 12.2% for women. Drawing on the 2006-2008 data, Princeton university researchers examined whether and to what extent variation in premarital cohabitation experiences influence marital stability. They found that the relationship between cohabitation and marital instability is complex and depends in part on marriage cohort, race/ethnicity, and marriage plans. Their analyses reveal that a ‘cohabitation effect’ exists only for women married prior to 1996, and that, until marriage plans are considered, there is no cohabitation effect among women married since 1996. 
Recent research from 2011 by the Pew Research Center has found that the number of couples that...