Growth mindset is probably the most cited psychology idea of the last twenty years outside academia: corporate training, school posters, LinkedIn. The message is simple: believe you can improve and you will.
It is also one of the ideas that held up worst under large-scale testing. Not in the sense that it is false: in the sense that the real effect is much smaller, much more selective and much more conditional than the way it gets told. This guide reconstructs what survives the data, what doesn’t, and which habits have stronger evidence than the ones credited to mindset.
What Dweck actually claimed
The original distinction, formalised by Carol Dweck and popularised in Mindset (2006), concerns implicit theories of intelligence: the belief that your abilities are a fixed trait or something that develops.
The theoretical prediction is not “optimists do better”. It is more precise: someone who treats ability as fixed reads difficulty as evidence of a limit, and therefore avoids tasks where they might visibly fail. Someone who treats it as developable reads the same difficulty as a signal that learning is happening, and stays on the task longer.
That is a testable prediction. It has been tested a great deal.
What happened when it was measured
Here the story gets interesting, and almost no popular article tells it in full.
The two 2018 meta-analyses. Sisk, Burgoyne, Sun, Butler and Macnamara (Psychological Science, 2018) aggregated the literature in two blocks. The first, covering 273 studies and over 365,000 participants, measured the correlation between mindset and academic achievement: r ≈ 0.10. That is a weak relationship, accounting for roughly 1% of the variance. The second, covering 43 intervention studies, measured the effect of mindset interventions: d ≈ 0.08, equally small. The most useful finding, though, isn’t the average: it is that effects clustered among at-risk and low-income students, and largely vanished for everyone else.
The large 2019 experiment. Yeager and colleagues published the National Study of Learning Mindsets in Nature: a very short online intervention (under an hour across two sessions) delivered to 12,490 students in 65 US high schools chosen to be nationally representative. Result: an improvement of about 0.10 grade points among lower-achieving students, and higher enrolment in advanced maths courses.
Two opposite readings of the same number, both defensible: 0.10 grade points is almost nothing for an individual; it is a lot for an intervention that costs under an hour and can be distributed to millions. The more important detail, though, is another one: the effect appeared only where classroom norms supported taking on challenges. Where the environment punished mistakes, mindset alone produced nothing.
2023, with two meta-analyses in disagreement. In the same year and the same journal, Psychological Bulletin, two papers reached diverging conclusions. Macnamara and Burgoyne found that once study quality and publication bias are accounted for, the effect of interventions becomes negligible. Burnette and colleagues, using different inclusion criteria, found small but real effects concentrated in specific subgroups.
This is neither anomalous nor a scandal: it is what a small, context-dependent effect looks like through different lenses. The honest conclusion is that growth mindset is not a general performance multiplier, and it has been sold as if it were.
What has stronger evidence
If the goal is workplace performance rather than school grades, there is a neighbouring construct with a markedly stronger empirical base.
Self-efficacy. This is the belief that you can execute a specific task, formalised by Albert Bandura (Psychological Review, 1977). The meta-analysis by Stajkovic and Luthans (Psychological Bulletin, 1998), covering 114 studies and more than 21,000 workers, found an average correlation of 0.38 with work performance, which the authors translate into roughly a 28% performance gain. That is a different order of magnitude from r = 0.10.
The difference between the two concepts is operational, not philosophical:
| Growth mindset | Self-efficacy | |
|---|---|---|
| Object | Ability in general | One specific task |
| Phrasing | ”I can improve" | "I know I can do this” |
| How it is built | By reframing beliefs | By accumulating verifiable wins |
| Evidence on work performance | Weak | Robust |
The practical point is that self-efficacy is not built by convincing yourself: it is built by doing things and seeing them work. Bandura named direct mastery experience as its primary source, not encouragement, which sat far lower on his list.
The other correction: practice doesn’t explain everything
Growth mindset almost always travels alongside the deliberate-practice idea: ten thousand hours and you’ll be good at anything. That has been measured too.
The meta-analysis by Macnamara, Hambrick and Oswald (Psychological Science, 2014) estimated how much of the variance in performance deliberate practice accounts for: 26% in games, 21% in music, 18% in sports, 4% in education and less than 1% in professions.
That last figure is worth reading twice, because it is the one about us. It does not mean practising is pointless. It means that in real jobs performance depends largely on factors repeated practice doesn’t touch: context, the opportunities that come your way, the people you work with, the kind of problem you get handed.
Put differently: the environment weighs as much as the mindset, which is consistent with the Nature result. If you are somewhere that punishes mistakes, no mental reframing will save you.
What to do with it, concretely
What remains, once the marketing is stripped out, is still useful. But it changes shape.
1. Move beliefs from general to specific. “I can improve” is too vague to produce behaviour. “I can write a SQL query with a JOIN because I did it on Tuesday” is self-efficacy, and that is what holds under pressure. Track what you have completed, not what you intend.
2. Build evidence, not phrases. The most reliable way to believe you can do something is to do a small version of it. That is the logic of a path made of small verifiable steps: it exists to generate evidence about you, not just skill.
3. Ask for feedback in a usable form. Not “how am I doing”, but “what is the one thing you would change here”. How feedback is framed determines whether it helps or hurts performance: that is documented, and we cover it in giving and receiving feedback.
4. Assess the environment before assessing yourself. If every mistake carries a social cost where you are, the explanation for your caution isn’t your mindset: it is the place. There the effective lever is changing context, not repeating a phrase.
5. Separate “not yet” from “not here”. Adding “yet” to “I can’t do this” makes sense when a realistic path exists. If that path doesn’t exist under your current conditions, “yet” becomes a polite way of not deciding.
False growth mindset, according to Dweck herself
This is worth reporting because it is an internal critique, not an external one: in 2015 and then in her work with Yeager (Perspectives on Psychological Science, 2019), Dweck explicitly described the phenomenon of false growth mindset, adopting the label without the content.
The most common forms she flagged:
- claiming “growth mindset” as a personality trait, when the theory describes a state that varies by domain and by moment;
- using the idea as praise for effort as an end in itself, applauding hard work even when the strategy is plainly wrong;
- turning it into an individual responsibility that offloads organisational problems onto people: if a project failed because it was understaffed, that is not a mindset issue.
The last one is how the idea most often enters companies, and it is also why in many workplaces it has a bad reputation it has earned.
My take: the real question is whether it’s worth it
Having read this literature, my conclusion is not that growth mindset is useless: it is that it has been used to answer the wrong question.
The question “do I believe I can improve?” is almost always already settled: hardly anyone genuinely believes they are incapable of learning anything. The question that actually stops people is a different one, “is this worth a year of my life?”, and that is a question of costs, alternatives and context, not of beliefs. No mental reframing helps there. Numbers, available time and an honest look at where you are do.
This hasn’t been measured in any controlled way, so read it as an opinion, not as data.
Frequently asked questions
So is growth mindset nonsense? No. It is a real but small effect, showing up mainly in people who start at a disadvantage and in environments that tolerate mistakes. The problem isn’t the original research, it is the distance between r = 0.10 and how it gets told in training rooms.
Why do two 2023 meta-analyses disagree? Because they apply different criteria for including studies and correcting for publication bias. That is normal when an effect is small: methodological choices weigh as much as the data. Anyone telling you “science has proven” growth mindset is flattening a debate that is still open.
What should I do instead if I want a measurable result? Build self-efficacy on specific tasks: small, verifiable goals with external proof that you hit them. It is the construct with the strongest workplace evidence, and it is built through action, not affirmation.
Does this apply to adults or only students? Nearly all intervention research is set in schools. Extending the results to adults at work is an inference, not a finding, which is another reason to be cautious with corporate courses that cite them.
How do I tell whether a course on this is serious? A good sign is that it cites Sisk 2018 or Macnamara 2023 and explains the limits. If it only cites Dweck 2006 and shows the “two types of people” chart, it is selling you the 2007 version. Our approach to checking sources is described in how we write.
Sources
- C. S. Dweck, Mindset: The New Psychology of Success, 2006. Original formulation of the distinction between implicit theories of intelligence.
- V. F. Sisk, A. P. Burgoyne, J. Sun, J. L. Butler, B. N. Macnamara, To What Extent and Under Which Circumstances Are Growth Mind-Sets Important to Academic Achievement? Two Meta-Analyses, Psychological Science, 2018. Average correlation r ≈ 0.10 across 273 studies; intervention effect d ≈ 0.08.
- D. S. Yeager et al., A national experiment reveals where a growth mindset improves achievement, Nature, 2019. 12,490 students, +0.10 grade points among lower achievers, effect conditional on context norms.
- B. N. Macnamara, A. P. Burgoyne, Do growth mindset interventions impact students’ academic achievement? A systematic review and meta-analysis with recommendations for best practices, Psychological Bulletin, 2023. Negligible effects after correcting for quality and publication bias.
- J. L. Burnette et al., A systematic review and meta-analysis of growth mindset interventions, Psychological Bulletin, 2023. Small but present effects in specific subgroups.
- A. Bandura, Self-efficacy: Toward a Unifying Theory of Behavioral Change, Psychological Review, 1977. Definition of self-efficacy and the hierarchy of its sources.
- A. D. Stajkovic, F. Luthans, Self-Efficacy and Work-Related Performance: A Meta-Analysis, Psychological Bulletin, 1998. Correlation of 0.38 with work performance across 114 studies.
- B. N. Macnamara, D. Z. Hambrick, F. L. Oswald, Deliberate Practice and Performance in Music, Games, Sports, Education, and Professions: A Meta-Analysis, Psychological Science, 2014. Variance explained by deliberate practice, under 1% in professions.
- C. S. Dweck, D. S. Yeager, Mindsets: A View From Two Eras, Perspectives on Psychological Science, 2019. Acknowledgement of “false growth mindset” and recalibration of expectations.



