The sentence was seven words long. “You must be smart at these problems.” Claudia Mueller and Carol Dweck said it to roughly 200 fifth graders after they solved a set of easy puzzles, and then they sat back and watched what those seven words did to the children over the course of six experiments.
The other 200 kids heard something different. “You must have worked hard at these problems.” Same puzzles. Same success. One word swapped: smart for hard-working. That was the entire intervention.
What followed is one of the most replicated findings in motivational psychology, and one of the most ignored in management. The “smart” kids fell apart.
What seven words did to 400 kids
After the praise, both groups were offered a choice: try a harder set of puzzles where they might learn something, or try another easy set where they’d look good. The effort-praised kids chose the harder puzzles by a wide margin. They wanted the challenge. The intelligence-praised kids mostly chose the easy set. They wanted to protect the label they’d just been given.
Then Dweck did something cruel and necessary. She gave every child a set of problems that were two grade levels too hard. Everyone failed. But the two groups failed differently.
The “hard-working” kids assumed they hadn’t worked hard enough. They leaned in, tried new strategies, reported enjoying the challenge. The “smart” kids assumed they weren’t actually smart. Their enjoyment cratered. Several lied about their scores when asked to report them to another student. They weren’t just discouraged. They were ashamed.
Here is the part that should keep managers up at night. After the impossible problems, Dweck gave both groups a final test at the same difficulty as the original easy set. The effort-praised children improved their scores. The intelligence-praised children scored 20% lower than their first attempt. The praise had not just failed to help. It had made them measurably worse at something they already knew how to do.
The sentence you probably said last Tuesday
“You’re a natural.” “You’re one of the sharpest people on this team.” “We hired you because you’re brilliant.” These are the workplace versions of “you must be smart at these problems,” and I’d bet money you said something like them in the last week.
Dweck’s research, which she later built into the terms “fixed mindset” and “growth mindset,” explains a pattern I’ve seen in every organization I’ve observed. The highest-credentialed people are often the most risk-averse. They were selected for being smart. They were promoted for being smart. They have been told they are smart for so long that “smart” has become their identity, and they will do almost anything to avoid a situation that might suggest otherwise.
This is the person who won’t volunteer for the new product launch because it might fail publicly. The VP who avoids learning a new tool because being a beginner feels threatening. The senior engineer who only takes on projects where the outcome is already certain. They are not lazy or unambitious. They are protecting a label that was welded onto them by thousands of well-intentioned compliments.
The way you deliver feedback either reinforces this pattern or breaks it. Every time you praise a person’s traits (“you’re so talented”) instead of their process (“the way you restructured that analysis after the client call changed the whole argument”), you are running Dweck’s experiment on your team. You just can’t see the data because nobody is measuring it.
Where this gets complicated
I should be honest about the limits. Dweck’s 1998 study has been cited over 4,000 times and has held up in multiple replications, but the effect sizes in later studies have been smaller than the original. A 2018 meta-analysis found that growth mindset interventions work, but modestly, and they work best for students from lower socioeconomic backgrounds. Some psychologists argue the whole thing has been oversimplified into a slogan.
Dweck agrees with that critique, by the way. She’s written about the problem of “false growth mindset,” where teachers (and I’d add managers) adopt the language without changing anything real. Telling someone “just try harder” while leaving every other system unchanged is not growth mindset. It’s a pep talk.
But I think the workplace application is actually sturdier than the classroom version. You’re not trying to reshape a child’s entire identity. You’re trying to give an adult professional feedback that makes them want to take on harder work and learn from failure rather than hide from it. That’s a narrower problem, and Dweck’s data addresses it directly.
The practical shift is small enough to start tomorrow. Stop praising who people are. Start praising what they did and how they did it. Pay attention to the specific language in your performance reviews, your Slack messages, your one-on-ones. Most managers find, when they actually track it, that 80% of their praise is trait-based. Flip that ratio.
And watch what happens when your team encounters something hard. The ones who’ve been told they’re “brilliant” will flinch. The ones who’ve been told they “worked the problem well” will lean in. It’s a seven-word difference, and Dweck proved it matters more than most of us want to admit.