Growth mindset, twenty years and one honest meta-analysis later

Don’t treat “believing intelligence can grow” as a strategy on its own — the large-scale evidence says that belief shift, by itself, does remarkably little for most people. If a growth mindset approach is going to help, it needs to be paired with a concrete plan for what changes, and it tends to matter most for people who are genuinely struggling, not for those already doing fine.

Why this works, or mostly doesn’t

Growth mindset — the idea that believing your abilities can develop through effort leads to better resilience and achievement — is one of the most culturally successful psychological ideas of the past twenty years, shaping school curricula, corporate training programmes, and a huge share of everyday self-improvement advice. The compelling, memorable framing arrived well before the large-scale evidence did.

When that evidence finally arrived, it was considerably less impressive than the cultural reach suggested. A major 2018 meta-analysis spanning dozens of studies and thousands of participants found the overall effect of growth mindset interventions on academic achievement was d = 0.08 — in plain terms, enough to move an average student from roughly the 50th to the 53rd percentile, a genuinely negligible shift. A later, independent 2022 meta-analysis found an even smaller effect, d = 0.05, which became statistically non-significant once the analysis adjusted for publication bias — the tendency for positive results to get published and studied more than null ones. When researchers restricted analysis to only the highest-quality studies — pre-registered, properly blinded, with checks that the intervention actually changed people’s stated beliefs — the effect on real-world outcomes disappeared entirely. Several attempted replications of Dweck’s original foundational studies on praise and effort also failed to reproduce the original findings.

The story isn’t a total debunking, though, and the nuance matters. A large, well-designed 2019 study published in Nature — specifically addressing the earlier failures by testing at national scale with careful measurement — found that a brief, low-cost mindset intervention did produce a real, if modest, improvement in grades specifically among lower-achieving students, and specifically in school environments where the surrounding culture reinforced the idea (teachers and peers who supported a growth-oriented approach rather than one enthusiastic poster in an otherwise unchanged classroom). Context, in other words, turned out to matter enormously — the same intervention that does almost nothing for an already-thriving student in a status-quo environment can produce a small, real benefit for a struggling student somewhere the surrounding culture actually reinforces it.

How to do it

1. Don’t expect a mindset shift alone — “I believe I can improve” — to meaningfully change outcomes for someone who’s already doing reasonably well. The research specifically doesn’t support that use case.
2. If you’re genuinely struggling with something specific, pairing a growth-oriented belief with a concrete, structured plan (worth combining with an implementation intention, covered elsewhere) is more likely to actually help than the belief shift alone.
3. Environment matters more than the individual belief-change exercise itself — a supportive surrounding culture that reinforces effort and improvement did the heavy lifting in the one large study that found a real effect.
4. Be skeptical of growth mindset framed as a general life philosophy that alone explains success or failure — the evidence doesn’t support that broad a claim, however intuitively appealing it is.
5. If you’re in a position to influence a learning or work environment (as a teacher, manager, or parent), the research suggests structural and cultural support matters more than a one-off talk or poster about believing in yourself.

Common mistakes

• Treating growth mindset as a proven, general-purpose performance booster. The average effect size across large, high-quality studies is close to zero for most people in most contexts.
• Assuming the concept is entirely fake because the average effect is small. The 2019 national study found a real, replicated benefit for a specific group in a specific context — dismissing it entirely overcorrects past what the evidence actually shows.
• Applying a brief intervention (a talk, a poster, a single lesson) and expecting it to work regardless of the surrounding environment. Context was the single biggest factor separating the studies that found real effects from those that didn’t.
• Using “just believe you can improve” as a substitute for an actual plan. The evidence consistently shows belief change without a concrete behavioural plan does little on its own.
• Ignoring the replication failures because the original framing was compelling. Compelling communication and strong evidence are different things, and this is one of the clearer cases in recent psychology where they diverged for years before the evidence caught up.

What should change

Don’t expect measurable change from adopting growth-mindset language alone. If you pair a mindset shift with a concrete plan, particularly around something you’re genuinely struggling with, and you’re in a supportive environment, modest real improvement is plausible based on the best current evidence — but the mindset belief is doing far less of the work than popular framing suggests.

Your next step

If what you actually need is a concrete plan rather than a belief shift, the article on implementation intentions covers the specific, well-evidenced technique for translating an intention into action — likely a more productive place to focus than the mindset question alone.