The Belief That Conquered the Classroom
In 2006, Carol Dweck published Mindset: The New Psychology of Success, and within a decade the idea had conquered education. Students who believe their abilities can grow with effort, the theory held, show resilience, work harder, and earn higher grades, while those who believe ability is fixed avoid challenges and decline. School districts bought programs, and teachers threaded the message through their lessons. Parents learned to praise effort instead of talent, and federal agencies funded trials at scale.
The commercial version promised more than the evidence ever did. Vendors sold growth-mindset curricula as a lever that would lift achievement for all students, a cheap psychological fix for expensive educational problems. By 2018, with these programs running in classrooms around the world, five researchers decided to add up everything the field had actually found.
Two Meta-Analyses, 423,000 Students
Victoria Sisk, Alexander Burgoyne, Jingze Sun, Jennifer Butler, and Brooke Macnamara ran two meta-analyses, both published in Psychological Science. The first pooled 273 studies and 365,915 students on whether students with growth mindsets get better grades, while the second pooled 43 intervention trials and 57,155 students on whether teaching a growth mindset raises achievement, bringing the combined total past 423,000 students.
The answers were weak on both counts: the link between holding a growth mindset and earning higher grades was small, and the average effect of interventions was d = 0.08, with a 95 percent confidence interval from 0.02 to 0.14. Even the top of that interval is a small effect by any convention, and the bottom is nearly zero.
The details were worse than the headline number. More than a third of intervention studies never checked whether the program changed anyone's mindset, the one thing it was supposed to do, and of the studies that ran the check, nearly half failed it. The authors concluded that claims of "striking effects on educational achievement" were unwarranted, quoting the very phrase mindset advocates had used.
What 0.08 Standard Deviations Buys You
Effect sizes are hard to feel, so here is a translation the paper never ran: a d of 0.08 moves the average treated student from the 50th to about the 53rd percentile of achievement, and converted to a common-language effect size, the probability that a randomly picked treated student outscores a randomly picked control student, the figure is 52.3 percent against a coin flip's 50. That gap, 2.3 percentage points above chance, is the entire average payoff. Barely a nudge.
Matthew Kraft's benchmarks for large-scale education programs classify effects below 0.05 standard deviations as small and 0.05 to 0.20 as medium, which lands 0.08 in the "medium" bin on the field's generous curve. Absolute scale tells a plainer story: high-dosage tutoring, the best-documented lever for raising achievement, produces effects around 0.37 standard deviations, roughly four and a half times larger. On that scale, growth-mindset interventions are among the weakest tools education research has measured.
The Strongest Counterargument
The fiercest defense of mindset theory does not dispute the numbers; it disputes the question. Dweck's camp argues the theory never promised effects for every student: a growth mindset can only help where students actually hold fixed beliefs and where better strategies are available to deploy, which makes averaging that over all students the wrong statistic, a single number summarizing a distribution it was never meant to describe.
And the defenders have a study to point to. In 2019, David Yeager and colleagues ran a pre-registered national randomized trial of a short online mindset intervention across 65 U.S. high schools, 12,490 ninth graders in total, published in Nature. Lower-achieving students gained 0.10 grade points in core courses, a d of 0.11, and the share falling below a 2.0 GPA dropped from 46 to 41 percent, an 11 percent relative reduction. Effects ran larger, 0.15 to 0.17 grade points, in schools whose peer norms supported taking on challenge, while the intervention raised Algebra II-or-higher enrollment the next year by 3 percentage points. Real, but small.
A 2023 meta-analysis by Jeni Burnette and colleagues sharpened the same point: among targeted subsamples with high implementation fidelity, the achievement effect was d = 0.14, though the 95 percent prediction interval ran from -0.08 to 0.35, meaning a future study could plausibly find a small negative effect or a solid positive one. The effect looks real in the right soil and vanishes elsewhere, which is why Yeager's metaphor, a seed that grows to fruition only in fertile ground, keeps getting quoted in this debate.
There is also a measurement worry running the other direction. As growth mindset became the officially correct belief, admitting a fixed mindset grew socially costly, which may have flattened the measured relationship between stated beliefs and outcomes. If the survey instrument is compromised by social desirability, every estimate built on it, including Sisk's, is noisier than it appears.
What We Didn't Prove
This meta-analysis did not prove that believing in malleability is useless, let alone harmful; it measured the average effect of brief interventions on grades, mostly in American schools, and found it small, though beliefs could still matter enormously for individuals or work through channels this literature does not capture.
The subgroup finding deserves its own caution, because the paper reports that low-income and academically at-risk students might benefit more, yet only a handful of studies fed that analysis, and the authors flagged it as tentative. Averages hide distributions: with a prediction interval crossing zero, some implementations plausibly backfire.
Finally, the evidence base is thinnest where the marketing is loudest. Study quality varied widely, later reviewers flagged expectancy effects and selective reporting, and elementary-age children, the group receiving the heaviest commercial pitch, have almost no rigorous trials behind the claims.
The Bottom Line
Growth mindset is not a lie, but it was sold as a lever and measured as a nudge. Across more than 423,000 students, teaching kids that ability grows with effort moves achievement by less than a tenth of a standard deviation, lifting the median student about three percentile points. It helps most where students are already struggling and the school around them supports the message, while everywhere else the effect dissolves into noise. Not a lie. Not a lever.
What You Can Do
Before buying another schoolwide program, ask for evidence that it changed beliefs among students like yours, not testimonials and not the original bestseller's sales figures, since more than a third of published trials never even ran that check.
Pair the message with the means, because Yeager's national trial worked where schools offered challenging coursework and supportive norms. Believing effort matters only pays when effort has somewhere to go. If your school cannot offer the advanced class, fix that before buying the poster.
Spend the tutoring budget on tutoring, since at roughly four and a half times the effect size, high-dosage tutoring remains the best-documented way to move grades, while mindset work, cheap and harmless as it is, is not a substitute.
Praise the process, not the poster: telling a struggling student their brain can grow is fine, but sitting with them through the hard problem and then making the next one slightly harder is the actual intervention.