The learning opportunities hiding in our failures


Informative failures

Eskreis-Winkler and Fishbach believe a key factor is that many of us simply don’t realise how informative failures can be. To test this experimentally, they created a stripped-down task designed to model real-life situations in which the key to success is avoiding mistakes.

They wanted to see if volunteers would avoid sharing their failures even though they were more informative than their successes.

For the task, dozens of online volunteers opened two mystery boxes from an array of three, for the chance to win money. One box contained 20 cents, another 80 cents, while the last was a dud and would cost them a cent. Next, they had the opportunity to share information about one of the boxes they opened to help the next participant in the game. As an incentive, they were told this other player would soon have the chance to reciprocate by sharing information with them.

One study showed how volunteers opened virtual boxes of money and would only share failures (losing money) than successes (winning money) with others (Credit: Alamy)

Crucially, the researchers contrived things so that each volunteer always opened a losing box and the 20-cent box. This meant, objectively, that it was always more useful if volunteers shared their failure – that is, the location of the money-losing box – than their relative success, the 20-cent box. Sharing the failure would allow the next player to dodge it, while sharing their success would still risk the other player opening the losing box. Yet, Eskreis-Winkler and Fishbach found that, across several studies, between one third to half of the volunteers chose to share success over failure – even though sharing failure would have been more beneficial to the other player.

Not only is failure feedback more readily embraced, but it’s also integrated in [the person’s] plans to reach the wish and to actually fulfil the wish – Gabriele Oettingen

The researchers uncovered more evidence for the way we overlook the value of failure in a follow-up quiz-style experiment, but this time they also found it was quite easy to remedy the bias. Online volunteers guessed the meaning of ancient symbols, choosing from two possible answers for each one. For one set, the researchers told the participants there wasn’t time to give them their results. For the other, the researchers told them they’d answered everything incorrectly. What’s particularly revealing is that when the researchers asked the volunteers which set they knew more about and could help other people with, 70% of them opted for the set for which they’d received no feedback, rather than the set for which they knew they’d failed so badly but which, due to the binary forced-choice format, they now effectively knew all the correct answers.

As with the money-box task, the problem again seemed to be the volunteers’ ignorance of how informative failures can be. Then, when Eskreis-Winkler and Fishbach nudged another group of volunteers into appreciating that learning they’d got all the answers wrong meant that they now knew the correct answers, this increased their willingness to share their knowledge about the symbol set they’d failed on.

Pg 3/5

Contact Us