The Growth Mindset, Sized Honestly

Fixed vs growth mindset became psychology's biggest export, then a replication-crisis casualty, then something more interesting: a small, cheap, real effect for the students who need it most. Here's what actually survived, plus a self-check to see where you lean.

July 13, 20269 min read

Two students get the same exam back with the same bad grade. One reads it as a verdict: she's not a math person, and the rational move is to stop investing in something she'll never be good at. The other reads it as a status report: this is where I am right now, and the interesting question is what to change before the next one. Carol Dweck spent four decades arguing that this difference in interpretation, which she eventually named fixed versus growth mindset, quietly shapes what people attempt, how long they persist, and what they end up able to do.

The idea became one of the most successful exports psychology has ever produced, and one of the most mangled. It got printed on classroom posters, stuffed into corporate values decks, dragged through the replication crisis, and declared dead by roughly the same internet that had declared it a law of nature ten years earlier. The honest picture in 2026 looks different from both the TED-talk version and the "it's all debunked" version, and the way it differs is instructive.

Where the idea comes from

In the early 1970s, Dweck sat children down with puzzles that were rigged to become unsolvable. Some kids fell apart fast: they blamed their intelligence, their problem-solving degraded, and they'd refuse puzzles they had already solved minutes earlier. A few did something strange instead. They perked up, treated the impossible puzzles as the fun part, and kept trying new approaches. Ability didn't predict who ended up in which group. Beliefs did. The collapsing kids thought the failure revealed something permanent about them; the resilient ones thought it revealed something about the puzzle.

In the academic literature this became "entity theory" versus "incremental theory" of intelligence. The 2006 book Mindset rebranded the pair as fixed and growth, and the rebrand is what conquered the world.

The most famous evidence came earlier, in Mueller and Dweck's 1998 paper: six studies with fifth graders. Kids solved a set of moderately hard puzzles, then got exactly one sentence of feedback. Half heard they must be smart at this. Half heard they must have worked really hard. Then everyone hit a wall of much harder problems.

One sentence of praise

After the setback, the children praised for intelligence preferred easier tasks, enjoyed the work less, blamed their ability, performed worse on a final set of problems, and misreported their scores to peers more often than the children praised for effort. The entire manipulation was a single sentence.

That result reframed a whole era of parenting. The self-esteem movement had spent the 90s telling adults to praise kids' talent. Dweck's data said that kind of praise teaches a child that outcomes measure who they are, which makes the next failure a threat to be managed rather than a problem to be solved.

A clay child figure assembling an oversized frosted glass puzzle, one teal piece glowing as it clicks into place — The original puzzle studies: ability didn't predict who collapsed after failure. Beliefs did.

What a mindset actually changes

A mindset doesn't change your IQ, your working memory, or how fast you learn. It changes what events mean, and meaning drives behavior. The same situation reads completely differently through the two lenses:

Situation	Fixed reading	Growth reading
A hard problem	A test of whether you're smart	A workout
Effort	Evidence you lack talent	The mechanism of improvement
A setback	A verdict	A status report
Someone better than you	A threat	A source of information

The effort row is the sneaky one. In a fixed frame, needing to work hard is humiliating, because naturals shouldn't need to. So people protect the "talented" identity by not trying, which guarantees the outcome they feared. Dweck called this the logic behind a lot of quiet quitting on hard subjects: it's safer to fail without trying than to try and confirm the verdict.

One thing Dweck has insisted on, loudly and repeatedly, is that a growth mindset was never just about effort. Her version includes trying new strategies and seeking input when you're stuck. Effort is a means to learning, and effort that isn't producing learning is a signal to change approach, not a virtue to be praised.

A mindset doesn't change what you can do today. It changes what you do next.

A clay path splitting at a fallen boulder: one branch loops back, the other climbs over on floating glass steps — Same boulder, two paths. The mindset decides which one looks reasonable.

Curious where you actually land? Rate the eight statements below honestly, not aspirationally.

The replication fight

Then psychology's replication crisis arrived, and mindset research was a prime target. In 2018, Sisk, Burgoyne, Sun, Taylor and Macnamara published two meta-analyses in Psychological Science finding weak average relationships between mindset and achievement, and weak average intervention effects. In 2019, Li and Bates ran close replications of core studies, including the famous praise experiment, and found the interventions changed what students said about intelligence without changing how they performed. A later Macnamara and Burgoyne meta-analysis went further, arguing that apparent effects largely reflect design flaws and bias, and noting that authors with a financial stake in mindset programs reported bigger effects than authors without one.

If the story ended there, "debunked" would be fair. It doesn't end there.

In 2019, Nature published the National Study of Learning Mindsets: a preregistered randomized trial across tens of US high schools and more than ten thousand ninth graders, administered by an independent research firm, with the analysis corroborated by a blinded Bayesian team. The intervention was two online sessions totaling under an hour. The result: about a 0.1 grade-point improvement among lower-achieving students, plus increased enrollment in advanced math courses. Students who were already doing well got no grade boost. And the effect wasn't uniform across schools; it stuck where peer norms supported challenge-seeking, and fizzled where the surrounding culture pushed against it. A separate team replicated the design in Norway with over 6,500 students.

A clay balance scale weighing one large frosted glass sphere against a few small solid clay pebbles, pans nearly level — The verdict of the replication fight: the claims deflated, but something small and solid remained on the scale.

So how big is 0.1 grade points? Small. Critics like Timothy Bates argue an effect that size can't be a serious part of fixing education, and that's a fair reading. The other fair reading is that the intervention costs almost nothing, takes under an hour, and concentrates its benefit precisely on the students most at risk of sliding off track, where a tenth of a grade point can be the difference between passing and repeating a course. Both camps agree on the numbers more than the rhetoric suggests; they disagree about whether a small, cheap, targeted effect deserves the word "works."

What genuinely died in the fight was the silver-bullet version: the idea that mindset is a master variable, that posters and pep talks move grades, and that the original effect sizes were typical. What survived is narrower and more useful: a small, real effect for specific students in supportive contexts, and a body of praise research whose central finding (person praise teaches kids that outcomes measure their worth) has held up well as a parenting insight even where the intervention literature wobbled.

The false growth mindset

The strangest chapter is that the sharpest critique of pop growth mindset came from Dweck herself, in 2015. Watching schools adopt her work, she saw it curdle into what she called a false growth mindset: praising effort as a consolation prize ("great effort!") for students who weren't learning anything, telling kids everyone is smart, and teachers simply declaring themselves growth-minded, as if it were a team to join rather than a practice.

The tell

Effort praise that isn't attached to learning is the self-esteem movement wearing a new shirt. Dweck's version means telling the truth about where a student currently is, then working on strategies to change it. The praise goes to the process that produced progress, not to trying as such.

She also walked back the binary. People aren't fixed-mindset people or growth-mindset people; everyone is a mixture, and the fixed side gets triggered by specific things: harsh criticism, being outshone by a peer, a setback in the one domain your identity leans on. In her framing, the mature skill is learning to notice when your fixed mindset fires, not performing the growth identity. Which means the classroom posters got the psychology backwards: they treated mindset as a trait to announce, and the whole point was that traits-you-announce is the fixed way of thinking.

What's worth keeping

Strip away the hype and the counter-hype, and a few things remain solid enough to act on. Praise process, and be specific: the strategy that worked, the progress made, the decision to ask for help. Skip both "you're so smart" and hollow "great effort." When effort stalls, treat that as a signal to change strategy rather than a virtue in itself; grinding the same failed approach harder is not a growth mindset, it's a ritual. And treat setbacks as data about the approach, not verdicts about the person, which is easier when you remember the mindsets are domain-specific: plenty of people are growth-minded about code and rigidly fixed about sales calls, or the reverse.

The last thing worth keeping is calibration. A belief doesn't do any work by itself, and nobody effort-praised their way past a genuine ceiling. What the belief changes is whether you're still in the game when compounding gets a chance to operate: whether the second attempt happens, whether the harder math class gets picked, whether the stinging feedback gets sought out or dodged. That's why the measured effects are small on average and concentrated at the margin, in the students hovering between quitting and continuing. Most of life's interesting outcomes are decided at exactly that margin.

A small clay seedling growing from a crack in a clay slab, supported by a delicate frosted glass scaffold — Not a transformation. A small, real edge at the point where people decide whether to continue.