Learn by failing? Not so easy

Alice G. Walton | Dec 16, 2019

Failure is often framed as a great teacher—the successful sometimes crow about the number of times they failed before hitting it big. But failure may not always teach us as much as it could, often because ego gets in the way, suggests research by Chicago Booth postdoctoral fellow Lauren Eskreis-Winkler and Booth’s Ayelet Fishbach.  

In an experiment, the researchers had telemarketers take a 10-question quiz about customer service. Each question offered two potential answers, only one of which was correct. For example, one question asked, “How much money, annually, do US companies lose due to poor customer service?” The participants had to choose between approximately $90 billion and approximately $60 billion. The correct answer was $60 billion.

Half of the participants received “success” feedback for questions they answered correctly, and the other half received “failure” feedback for wrong answers. 

When tested later on what they’d learned, the telemarketers who had received success feedback fared better, answering 62 percent of the follow-up questions correctly, compared with 48 percent for those in the other group. 

The researchers then performed a similar experiment that involved symbols, with participants from Amazon Mechanical Turk. As before, success feedback produced much better performance in a follow-up test: participants in the success condition got 80 percent of the answers right, versus 59 percent in the failure condition. 

In each successive variation of the experiment, a question only had two possible answers, thus the two types of feedback provided the same amount of information. In that case, why was success a better teacher than failure?

Eskreis-Winkler and Fishbach write that failure can be a big hit to one’s ego, which may reduce motivation. When study participants reported their self-esteem levels following a task, those in the failure condition registered lower self-esteem. And when the researchers removed ego from the equation by having some people learn from others’ wrong answers, not their own, participants learned equally from failures and successes. “Because people find failure ego-threatening, they will disengage from the experience, which means they stop paying attention, or, tune out,” the researchers write.

The results suggest that feedback about failure should be given with caution. And when it is absolutely necessary to give failure feedback—perhaps after a fumbled presentation at work—the researchers suggest that ego be removed from the equation as much as possible, as “reducing the degree to which failure involves the ego will promote learning.”

Eskreis-Winkler and Fishbach further find that although it’s hard to learn from failure, people should still try—there may be benefits for society. Their follow-up work finds that because people don’t realize failures contain valuable information, they don’t share them with others. People’s erroneous belief that failures don’t contain information produces an asymmetrical world of information where failures are common in private but hidden in public.

Ego makes people reluctant to both learn from failure and share much about their failures. But “information on failures is a public good,” says Fishbach. “When it is shared, society wins.”