This is a classic example of correlation versus causation. Most likely, those couples who didn't live together prior to marriage are more likely to be affiliated with religious groups that frown upon divorce. The cohabitation isn't the issue, the underlying cultural ethos are.
I'd be interested to see this cross-tabbed against various religious groups to see if this statistic holds true for conservative evangelicals, atheists, Muslims, etc.
It’s “propter” hoc, not “prompter.”