What the research actually says
Six entries, in the order the argument unfolded. Critics included on purpose.
Dweck, Mindset (2006). The founding claim: how you think about ability changes how you meet failure. Fixed mindset reads a setback as a verdict. Growth mindset reads it as information.
Blackwell, Trzesniewski and Dweck (2007), Child Development. Among 373 seventh graders tracked across the rocky transition to junior high, those who believed intelligence can grow earned climbing math grades, while believers in fixed ability flatlined. A short workshop teaching the growth view reversed a grade decline that the control group continued. Foundational and small: the workshop arm was 91 students at one school.
Yeager et al. (2019), Nature. The study the field needed: pre-registered, nationally representative, about 12,500 ninth graders in 65 public schools. Two short online sessions, under an hour in total, raised grade point averages among lower-achieving students by 0.10 grade points and lifted enrollment in advanced math by about 3 percentage points. The gains held best in schools where trying hard was not socially costly. Small, real, and cheap.
Sisk, Burgoyne, Sun, Butler and Macnamara (2018), Psychological Science. Two meta-analyses pooling hundreds of studies: across 273 effect sizes, the link between mindset and achievement averaged r = .10, and across 43 intervention studies the average effect was d = 0.08, with 37 of the 43 not distinguishable from zero. The exception the authors themselves flag, from a handful of studies: bigger effects for low-income and academically at-risk students. Which is where Yeager found his.
Macnamara and Burgoyne (2023), Psychological Bulletin. The strongest case against. Rerun the evidence through strict quality criteria and the average effect lands at d = 0.05, no longer significant once publication bias is corrected; in the six highest-quality studies it is d = 0.02. They also report that authors with a financial stake in mindset programs published larger effects than authors without one. This page would be dishonest without this paragraph.
Yeager and Dweck (2020), American Psychologist. Their answer: nobody should expect one hour online to move every student. The effect shows up where it was predicted to: in students who are struggling, in schools whose culture does not fight the message. A 2023 meta-analysis by Burnette et al. lands in between: small overall, d = 0.14 for the students it was aimed at, with wide and honestly reported uncertainty around both numbers.
Where that leaves us. The effect is real and small. It concentrates in students who are struggling. An hour online costs almost nothing, so a small real effect can still be worth having. Believing you can change does not replace the work of changing.
How I have grown
Research is one kind of receipt. This is the other kind. Four items, kept exactly as dictated on a walk, 2026-07-10.
- Cried in front of friends.
- Learned to make friends.
- Learned to deal with anger.
- Learned various counseling techniques.
None of these came installed. Each one was a “not yet” for years.
What a growth mindset is good for
Held honestly, the belief buys you specific things:
- A better relationship with setbacks. In the 2007 studies, students who saw ability as growable set learning goals, blamed failure less on being dumb, and picked strategies over giving up.
- Harder choices, voluntarily. In the national experiment, treated students signed up for harder math at higher rates. The belief changes what you attempt, which changes what you get to practice.
- The most help where there is the most struggle. Every honest reading of the data, including the skeptical one, agrees the effect is largest for people currently failing at something. If everything is already going fine, a mindset session will not add much.
- A cheap thing to try. Under an hour, online, no tutor, no fee. At that price, an effect of 0.10 grade points for struggling students is a bargain, and calling it a bargain is different from calling it a miracle.
- What it is not. Dweck herself warns against the poster version, which she calls false growth mindset: praising effort while changing nothing else. The belief helps when it changes what you attempt and how you meet failing. On its own it is a slogan.
What learning is good for
Zoom out from mindset to the thing mindset is for. The evidence that learning itself pays is older and steadier:
- Memory, in older age. The Synapse Project randomly assigned adults aged 60 to 90 to three months of learning something genuinely hard. The group learning digital photography improved episodic memory clearly; quilting helped less; socializing without learning a new skill did not help at all. The demanding part is what mattered.
- Years of life. A 2024 review in The Lancet Public Health pooled 603 studies from 59 countries and found each additional year of education associated with 1.9 percent lower risk of death from any cause. This is association, and educated lives differ in many ways, but the gradient is steady, dose-shaped, and global.
- Wellbeing. “Keep learning” is one of the Five Ways to Wellbeing distilled from the UK government's Foresight review of the evidence. It is a policy summary rather than a primary study, and the underlying work on adult learning and life satisfaction is largely correlational. Correlates: the honest word.
One card at a time
A little deck of this page's findings, because reading about learning is not the same as keeping any of it. The middle button is the whole site in two words.
Say something back
What are you still learning? Or what should this page do differently? Type it, or press the dot and speak it. Both land in the same store the version modal reads from.