Marriage doesn't make relationships more stable, a group of social scientists in the UK found:
Parents who are cohabiting when their child is born are three times more likely to split up by the time their child is five than married parents (27% compared to 9%). However they are also typically younger, less well off, less likely to own their own homes, have fewer educational qualifications and are less likely to plan their pregnancies than married people. Once these differences between the two groups are accounted for, the difference in the likelihood of separation almost disappears (falling to 2 percentage points).
The IFS analysis shows that relationship stability is mainly determined not by marriage but by other factors such as age, education, occupation and income, and delaying and planning pregnancy. These factors are also influential in whether people choose to marry or not. So while married couples have more stable relationships than couples who cohabit, this is not because they are married, but because of the other characteristics they have that lead to marriage.
Prepare for all sorts of whining from the moral majority crowd about how this study was biased, doesn't apply to the US, yada yada yada, mainly because it doesn't fit neatly in with their values. Reality has a terrible leftwing bias, which is a great reason to just ignore it.