neur.ro

Illusion of control

type: effect · origin: study · evidence: contested · themes: judgements and probabilities

Claim assessed: People act as if they could influence outcomes that depend only on chance.

Published · how we work

In short

The illusion of control means behaving as if we could influence things that depend only on chance. The idea comes from Ellen Langer’s 1975 studies and became a basis for explaining problem gambling. Its best-known part, that simply choosing your ticket makes you think your chances are higher, wasn’t confirmed in large replications. Other forms of the illusion, such as the effect of an early run of wins, replicated only partly.

What it says

The lottery tickets. Langer (1975) offered office workers a 1-dollar lottery ticket showing a football player. Some chose their ticket, others were given one. Before the draw, they were offered the chance to sell it back, at whatever price they wanted. Those who had chosen asked on average 8.67 dollars, the others 1.96. In another study, those who learned their ticket’s numbers gradually, on different days, more often kept their ticket rather than swapping it for one in a lottery with visibly better odds: 64% against 32%. Langer proposed four factors that create the illusion: choice, involvement, competition and familiarity. We haven’t read the original study; we describe it from Clark and Wohl (2022) and Klusowski et al. (2021).

Top: in Langer's 1975 study, those who chose their 1-dollar lottery ticket asked on average 8.67 dollars to sell it back, and those who were given one, 1.96 dollars. Bottom: in Study 1 by Klusowski and colleagues, with 794 participants, the felt chance of winning was 5.49 for those who chose their numbers and 5.45 for those who were assigned them, on a scale from 1 to 9. Choosing the ticket: then and now Langer, 1975; Klusowski et al., 2021 Asking price for a 1-dollar ticket Langer (1975), 53 office workers chose their ticket $8.67 were given one $1.96 How likely they felt they were to win Klusowski et al. (2021), Study 1, 794 people chose their numbers 5.49 were assigned them 5.45 Choice rarely changed the felt chance of winning.
Top: Langer (1975), Study 2, as reported by Clark and Wohl (2022); the mean price asked to sell the ticket back. Bottom: Klusowski et al. (2021), Table 1, the 1-to-9 scale. The measures differ; compare the direction, not the size.

Choice, tested again. Klusowski, Small and Simmons (2021) ran 17 preregistered experiments with 10,825 participants. In the first nine, with dice lotteries, roulette and chocolates, those who chose didn’t feel closer to a good outcome than those who were assigned an option. Choice mattered only when it actually raised their chances. In the few experiments where choosers seemed more confident, the cause was something else: some participants believed beforehand, despite being told otherwise, that one option was better, and that was the one they chose. The authors also note that Langer’s result has other possible explanations: participants may simply have liked the ticket with the player they picked more, or feared the regret of selling it and seeing it win.

The run of wins. Langer and Roth (1975) had participants guess the outcome of 30 coin tosses. Without knowing it, everyone got exactly half right, but some won mostly at the start and others mostly at the end. Those with early wins rated themselves better at guessing. Eben et al. (2023) repeated the study online, preregistered, with 289 participants. The direction held, but the effect was small to medium, weaker than in the original. Only one of the five questions passed the threshold set in advance, and the total score, computed after the data were collected, showed a medium difference.

What the meta-analysis says. Ștefan and David (2013) found, on average, a moderate effect in illusion of control experiments. They note, however, that the studies measure very different things, so the same name may cover different phenomena.

In gamblers. Orgaz, Estévez and Matute (2013) gave a group of 49 people diagnosed with pathological gambling and a group of 51 people without a diagnosis a task unrelated to gambling: they decided whether to give a fictitious medicine to fictitious patients. Recovery didn’t depend on the medicine at all. Both groups still believed the medicine helped, and the gamblers more so: 70.6 against 57.2, on a scale from 0 to 100.

Example

Someone throws the dice harder when they need a high number and more gently when they need a low one. The outcome doesn’t depend on the force of the throw, but the feeling of control is strong. The example is an illustration, not a study cited here.

How to apply it

The steps below are a practical way to apply it that we propose, based on the studies cited.

  1. Ask whether your action changes the probability, not just how you feel. A reel-stopping button on a gaming machine or the numbers you pick in a lottery don’t change the odds.
  2. Don’t be fooled by a lucky start. A few early wins in a row don’t show that you have a talent for a game of chance.
  3. Watch out for products that give you a feeling of control. Clark and Wohl describe how modern gaming machines add buttons and options that don’t change the average payout but strengthen the impression of control.
  4. Don’t assume that letting people choose makes them more confident. Large replications show that choosing between equal options doesn’t change perceived odds.

Limits and nuances

Sources

See also: Endowment effect and status quo bias, Gambler's fallacy, Sunk cost