Chosen and Arranged, and What the Evidence Shows

Which produces better outcomes, a marriage arranged by families or a marriage chosen freely by the couple?

The question is asked constantly and answered with enormous confidence in both directions. The literature does not support the confidence.

Findings conflict. Several studies across different countries report more positive feeling in freely chosen marriages. An often-cited older study found love marriages declining in reported love with duration while arranged marriages rose. A 2012 study comparing arranged and love marriages among Indian adults in the United States found no significant difference on marital satisfaction, commitment, companionate love or passionate love.

That last result is a finding rather than a failure to find one, and it is the kind of result that does not travel, because nobody wants to hear that their strong opinion describes nothing.

Before going further, the deeper problem: the categories are not stable.

Arranged covers an enormous range. At one end, a marriage settled between families with the couple barely consulted. At the other, and far more common today, families make introductions and the individuals hold a full veto and often an active choice. Studies sampling different points on that range are not measuring the same thing, and then get compared as though they were. A good deal of the apparent conflict in this literature is definitional rather than empirical.

And free choice is less free than it presents itself.

Who you might marry is determined, long before any choosing happens, by where you live, where you studied, where you work, where you worship and who your friends already are. The filtering is severe, it happens upstream, and it is performed by circumstance rather than by relatives. The difference between the systems is not filtered against unfiltered. It is who does the filtering and whether they announce themselves.

The most useful way to read the comparison is to ask what each system is optimised for.

Arrangement optimises for compatibility as judged by people who are not in love. That is a genuine advantage, since the assessment is made with a clear head about things that matter over decades: temperament, family, expectations, money. It is also a genuine risk, because those assessors have interests of their own, and the couple's happiness is not always the first of them.

Free choice optimises for the quality of the initial feeling. That is the best available predictor of whether two people will tolerate their first difficult year, and it is close to the worst available predictor of whether they will still be together in twenty.

Neither is obviously better, and both are extremely good at doing the thing they were built to do.

What the evidence does suggest, and this is the point the last book of this course develops, is that the system matters far less than what happens afterwards. By the third year neither the family's judgement nor the initial feeling is doing much work. What predicts the outcome from there is a set of small repeated behaviours that no method of selection installs and that both systems leave entirely to chance.