She and Lee went to the MIT PlayLab at the Boston Children’s Museum and recruited 103 visiting infants, aged 13 to 18 months. As the babies watched, Leonard tried to retrieve a toy from a container, and detach some keys from a carabiner, narrating her efforts along the way. In front of some babies, she succeeded at each task immediately, performing each three times in the span of 30 seconds. In front of others, she spent the same period struggling, and only retrieved the toy and keys just before the time ran out.
“Now it’s your turn to play with a toy,” she said to the infants. She then handed them a music box that she had already activated. The box came with a large, conspicuous, and completely useless button. Pressing it did nothing, but it was the act of pressing that mattered. Leonard found that babies who had seen her struggling with her own objects prodded the button more often than those who had seen her succeed effortlessly. They had inferred that hard work pays.
Leonard learned the same lesson herself. Her results came in just as psychologists were starting to grapple with their reproducibility crisis—a deep concern that many of the results in published papers might be unreliable due to poorly-designed studies and sloppy practices. To weed out such results, many psychologists have said that their field should put more emphasis on replication—repeating studies to check if their findings hold up. Others believe that more experiments should be preregistered—that is, scientists should specify their research plans ahead of time, and stick to them, obviating the temptation to fiddle with their studies on the fly in ways that guarantee exciting but ultimately misleading results.
So after Leonard had spent a year studying the value of persistence, her advisor Laura Schulz told her to do the experiment again. “It was a very meta moment,” she says. She recruited another 120 infants, and she preregistered her plans. And to her delight, she got exactly the same results.
Critically, the babies were learning a mindset from the adults, rather than a specific set of actions. “It’s a nice case study of infants making rich abstract inferences from sparse data,” says Schulz. They saw Leonard fiddling with a container and a carabiner, but then generalized her behavior to a completely different toy that they had never seen before—a music box. They weren’t just learning about containers and carabiners; they were drawing broader inferences about the value of hard work. “It fits with a lot of prior research showing that infants are good at imitating adults’ goals,” says Liz Gunderson from Temple University, “but it goes one step further in showing that infants can imitate adults’ persistence toward a goal.”
This isn’t quite the same as the much-hyped concept of “grit,” which is more of a stable, long-term character trait. “A lot of the work on grit and persistence focuses on school-age children,” says Leonard. “We wanted to go younger and look at the foundations of that behavior.”