Can We Prove That Marriage Improves Behavior?

by Jim Manzi

Amanda Marcotte (via Andrew Sullivan) questions whether the evidence often trotted out for the social desirability of marriage — basically that married people score better on various health, economics, and similar indicators — really demonstrates that marriage causes these differences:

[T]hese kinds of studies lump all nonmarried people into one group.  People who are in long term, committed relationships without that piece of paper are put in the same group as people who’ve never held a relationship together. 

Marcotte is not quibbling here, but pointing to a substantial objection. It is a special case of the problem of selection bias — i.e., do married people tend to do better because of the effects of marriage, or because of what kinds of people tend to get married?

Her proposed solution:

I want to see apples to apples comparisons. How do unmarried people who’ve been together for five or 10 years hold up next to people who have been together that long but tied the knot in their first year or two together?

Marcotte’s proposed approach of using long-term cohabitating couples as a control group doesn’t solve the problem. Whether we discover that married couples do better, worse, or the same as this control group on some indicator, how do we disentangle the selection-bias effect from the causal effect of marriage? For example, assume we find that married couples do worse on average than cohabitating couples for some outcome. How do we know that they wouldn’t have done even worse had not the beneficial effects of marriage improved them, and that therefore marriage actually has a positive causal effect on the outcome? We don’t. 

Of course, Marcotte’s selection bias issue has occurred to countless social scientists, and they have therefore created a huge body of analysis that attempts to create exactly such apples-to-apples comparisons, using various case-control matching, regression, and other methods. For reasons that I’ve gone into in detail elsewhere, these are all unreliable in evaluating the direct causal link between marriage and outcomes for the couple under consideration, never mind addressing the broader issue of the social effects of a change in the marriage norm on other people outside this couple. There is no practically available social-scientific method that can provide the kind of proof that Marcotte demands in a non-totalitarian society.

But it doesn’t follow that those who advocate continued legal support for traditional marriage are, as Marcotte puts it, “marriage chauvinists.” As I’ve written about elsewhere at length, I think the general principles for political action that flow from such a lack of scientific knowledge on a topic like this are: (1) a loose status quo preference, and (2) a regime of subsidiarity.