Why You Finish the Series You Stopped Enjoying
The sunk cost effect is real, it is smaller than its reputation, and nobody has ever measured the version of it you are living on a Sunday evening
Key takeaways
- The famous theatre study is genuine and half-quoted. Full-price season-ticket buyers used about four of their first five tickets against about three and a third for buyers who had been given a random discount, and over the last five plays the groups were indistinguishable. The effect ran for roughly six months and then stopped.
- Thirty-six laboratories on three continents, more than six thousand people, found the effect slightly larger than the original study had, which almost never happens. It is still a small effect. Thousands of people have to be counted before it is visible at all, no single viewer could ever demonstrate it, and only half the individual laboratories found it in their own sample.
- The question those laboratories used was Thaler's football ticket, not Arkes and Blumer's theatre experiment, which has never been directly repeated. Saying the sunk cost fallacy replicated and saying the theatre study replicated are two different claims.
- Sunk money effects are the consistent ones. Sunk time effects are markedly more volatile, with several failures on the record, and a television series is a time case that cannot borrow the money literature's confidence.
- Olivola's eight experiments, just over six thousand people, found that people honour costs somebody else paid, unaffected by how close they feel to that person and unaffected by whether anyone is watching. That is the strongest single result in the area and self-justification cannot explain it.
- There is no published research on people finishing series they have stopped enjoying. The completion percentages in circulation are platform analytics, commercial analytics or untraceable marketing, and none of them are used here.
The [sunk cost effect](https://thewhyweproject.com/articles/what-is-the-sunk-cost-fallacy) is real, and the field study everybody cites for it is quoted at roughly half its length. What the term covers and where it fails is in What is the sunk cost fallacy. The missing half changes what it shows.
The first sixty people to buy a season ticket to the Ohio University Theater's 1982 to 1983 season were sold, at random, one of three prices: the ordinary fifteen dollars, two dollars off, or seven dollars off, the discount explained away as a promotion. Everyone went home with an identical booklet and identical access to the same ten plays. The only difference between them was what they had paid, decided by chance and already gone. The stubs were printed in three colours so that attendance could be counted play by play, and the colour was the experiment.
Over the first five plays, the full-price buyers used about four of their tickets. The two discounted groups managed about three and a third. Over the last five plays the three groups were indistinguishable. The effect lasted about six months and then it stopped.
That last sentence is the one popular coverage drops. The line about full-price holders attending a quarter more performances is a rounding of the first half with the second half quietly removed. Keep the stubs in mind, because you are four episodes into something you have stopped enjoying and the next one is already queued.
Sort what you refuse to abandon
One phrase covers four situations that behave differently, and telling them apart is most of the work.
1. You paid money and can still use the thing. A season ticket, a gym membership, a course with twelve modules left in it. The money is gone, the access remains, and what you paid decides how much of the access gets used. This is what the stubs measured.
2. You chose it in public. Barry Staw handed 240 business students a failing division to fund in 1976. Some had picked it themselves, some had it assigned. Same losses, same figures, and the ones who had picked it put in the most. That is escalation of commitment, and the work is done by who made the original call rather than by the money.
3. Somebody else spent it. A manager keeps a bad project running because a predecessor funded it. No authorship, no reputation at risk, and the pull is still there. Escalation cannot be the whole story, and the two ideas are not interchangeable.
4. You have spent only time. No receipt, no witness, nothing recoverable, nothing anybody else can see. You did not commission the series, your judgement is not on trial, and nobody will ever know whether you finished it. Four episodes, and you finish it anyway.
The same error carries a second name from another field. Richard Dawkins and his co-author called it the Concorde fallacy in 1976, after an aeroplane that kept being built because it had already been half built, and they were arguing about birds. The name belongs to the first case, and it describes the error rather than the ego.