Arranged and chosen marriages

Why do marriages arranged by families often report satisfaction climbing over the years while chosen ones drift down?

The claim is repeated so often it has stopped sounding like a claim: a marriage that begins between near-strangers grows warmer, while one that begins in love cools. It is offered as evidence that romance is a poor foundation and that families choose better than hearts do.

Before asking why it happens, notice the verb hiding in the question. Arranged marriages report satisfaction climbing. A report is not a measurement of a marriage; it is a person comparing their marriage against something. What are they comparing it against, and does that something start in the same place in both cases?

Ask someone to rate their marriage from one to seven. They do not read a quantity off a dial. They set what they have beside what they expected and answer with the difference. So a satisfaction score behaves roughly like reality minus expectation — and the trend in that score is the change in reality minus the change in expectation. That second term is usually ignored, and it is not small.

Now ask where each kind of marriage starts on that scale. A chosen marriage is entered at, or very near, the peak of felt intensity — you married when it felt best, which is close to a definition. Any later reading is taken from a high-water mark, and the expectation formed at that mark stays as the yardstick. A marriage arranged between people who barely know each other starts with almost no accumulated feeling and a modest expectation. Ordinary familiarity — learning someone's habits, surviving a first crisis together — then registers as a gain, because the yardstick was set low. Two marriages could be identical in every observable way at year ten and still produce opposite trend lines, purely from where the comparison began.

Second, ask who is still in the sample to be asked. Where divorce is available and carries little stigma, the unhappiest marriages leave the survey early. Where exit is costly, stigmatised, or economically impossible, they stay in — and are still asked how satisfied they are. And a person who cannot leave has a cheaper adjustment available than a person who can: revaluing what they have. That is not cynicism about anyone's honesty; it is the ordinary way people reduce the discomfort of an unchangeable situation.

Third, something that is not an artefact at all. A chosen marriage is typically asked to supply companionship, passion, practical partnership and personal growth from one relationship, while an arranged marriage usually sits inside a kin network supplying much of that. More demand loaded onto fewer people means more ways to fall short.

Now the evidence, softest first, because it matters which parts of this rest on what. The specific crossover picture traces back to a small study of about fifty couples in Jaipur published in the early 1980s — I am recalling that figure, and it is exactly the finding not to lean on: one city, one moment, tiny sample. A later comparison of arranged marriages in India with chosen marriages in the United States found no significant difference in marital satisfaction at all; again recalled, and again small. The honest position is that the dramatic crossover is not well established, while the reference-point argument holds regardless of whether it is.

Can we test the account in both directions? If the reference point sets the sign, then chosen marriages that began without a peak — long friendships that turned slowly — should decline least, and the arranged rise should be weaker in behavioural markers such as time spent together than in self-report. Pushing the other way: if the crossover were a real relational gain, it should survive in a sample where exit is equally available to both forms, and it should not evaporate when divorce becomes cheap. In populations where arranged marriage remains normal but divorce has become socially possible, divorce rates rise rather than staying flat — which is what you expect if some reported contentment was the absence of an exit. What would break my account is a large longitudinal study under one legal regime, following both forms with behavioural measures, and finding the crossover intact.

This is not an argument that arranged marriages are secretly unhappy or that chosen ones are secretly fine. It is narrower: the sign of a reported trend is set largely by where the expectation started and by who remains available to answer, so the trend is weak evidence about the marriage itself. Note too that "arranged" spans everything from a veto-free match to a parental introduction the couple can decline, and those are not one thing.

Two neighbours in this collection sit close by. fading-attraction explains why the traits that first drew someone stop being noticed — that is about the level of feeling and its perceptual habituation. My fixed point is the instrument rather than the feeling: the same underlying marriage yields opposite trend lines depending on the entry price. hedonic-adaptation explains the return to baseline after any improvement; here the comparative claim is that the two forms differ in their starting baselines, not in their capacity to adapt. And widowhood-and-health, written alongside this one, takes the marriage as a regulator of a spouse's daily behaviour — a function that runs identically whichever way the marriage began, and which no satisfaction score measures.

• How divorce liberalisation changes reported satisfaction in the marriages that survive it.
• What behavioural markers — shared time, conflict recovery, illness care — track marital quality better than a self-rating does.