Growth Mindset Research: Limitations, Criticism and Open Questions
Growth mindset is one of the most influential — and most contested — ideas in modern education. This page summarises what the strongest available evidence actually shows, where the criticism has landed, and what those debates mean for how anyone (including you) should read a mindset test result. It aims to be neutral and evidence-based rather than either promotional or dismissive.
What large meta-analyses show about average effect sizes
Two systematic reviews are the usual starting point. Sisk, Burgoyne, Sun, Butler and Macnamara (2018, Psychological Science) synthesised 273 effect sizes across 129 studies. They reported a weak overall correlation between measured mindset and academic achievement (roughly r ≈ 0.10), and small average effects of mindset interventions on academic outcomes (roughly d ≈ 0.08). Both numbers are far below the popular perception that mindset is a large lever for grades.
Macnamara and Burgoyne's more recent meta-analysis in Psychological Bulletin (2022) applied stricter inclusion criteria and risk-of-bias adjustments to the intervention literature. After correction, the average intervention effect on academic achievement was close to zero, with the residual signal concentrated in a small number of studies with methodological features that tend to inflate estimates. Burnette and colleagues (2023), also in Psychological Bulletin, reached broadly similar overall averages but argued that moderator analyses — for whom, and under what conditions — are more informative than a single headline number.
The honest summary is: on average, brief growth mindset interventions produce small effects on academic outcomes. That is neither "mindset works" in the popular sense nor "mindset is a myth" in the polemical sense. It is a modest average with important heterogeneity underneath.
Why effects appear larger for some groups than others
Underneath the small average is a consistent pattern of moderators. The National Study of Learning Mindsets, a pre-registered randomised trial in 65 US high schools with about 12,000 ninth-grade students (Yeager et al., 2019, Nature), found that a short online mindset module produced no reliable gains on average but a small, reliable improvement in grades for lower-achieving students. A follow-up analysis (Yeager et al., 2022, Psychological Science) showed that even that subgroup effect appeared only in schools where teachers themselves held incremental beliefs and where the peer environment supported effortful learning.
Sisk et al. (2018) reported the same broad pattern across the wider literature: effects are largest, on average, for academically at-risk students and students from lower-SES backgrounds. Two mechanisms are usually proposed. First, students who are already succeeding academically have less room for a belief-level intervention to change outcomes. Second, students who are struggling are more likely to be running on strongly fixed assumptions ("I'm just not a maths person"), so a plausible alternative framing can meaningfully change how they respond to the next setback.
Neither mechanism justifies claims that mindset alone is a large lever for everyone. It does mean that the case for well-designed mindset support is strongest exactly where students face the largest obstacles — provided the surrounding teaching, feedback and resources actually make different behaviour possible.
Replication concerns
Several high-profile early mindset findings have not replicated cleanly at larger sample sizes. Li and Bates (2019, Journal of Experimental Psychology: General) ran three well-powered studies in UK schools and failed to replicate central claims about mindset and grade trajectories and about the effect of process versus person praise. Bahník and Vranka (2017) found essentially no association between mindset and scholastic aptitude in more than 5,000 Czech university applicants.
This is a common pattern in social and educational psychology and is not unique to mindset research. Effects estimated from small, exploratory studies tend to shrink when the same question is asked with pre-registered designs and larger samples. It does not mean the underlying construct is fake; it means the headline effect sizes from the earliest studies were, in retrospect, likely overestimates.
Measurement problems with self-report mindset scales
Most published mindset findings rely on short self-report items with wording close to "You have a certain amount of intelligence, and you can't really do much to change it." These scales are brief, easy to run and translate well across contexts, but they carry known measurement weaknesses.
Responses shift with wording changes, ordering effects, and the immediate framing (a lesson on brain plasticity right before the scale reliably nudges answers toward the growth end). Self-report also confounds beliefs with self-presentation: in classroom or research contexts where a growth answer feels expected, social desirability inflates growth scores. And there is a well-documented gap between what people say they believe about ability and how they actually behave when facing challenge — the behavioural measure is often the more informative one, but it is harder to collect.
Cross-cultural and cross-language use adds another layer. Measurement-invariance analyses — which check whether a scale means the same thing to speakers of different languages — are relatively rare for mindset items, so most cross-country comparisons should be read with caution.
Publication bias
Mindset research has been affected by the same publication-bias pressures as much of social psychology. Studies with statistically significant positive results are more likely to be written up, submitted, and accepted than null results, which biases the visible literature upward. Both Sisk et al. (2018) and Macnamara and Burgoyne (2022) reported statistical signatures consistent with small-study effects and publication bias, and both estimated smaller "corrected" effects than the raw meta-analytic averages.
The shift toward pre-registration and registered reports over the past decade has improved the situation. Pre-registered mindset trials — including the Yeager et al. (2019) national study — tend to report smaller, more clearly bounded effects than the earlier open-ended literature. Any responsible reading of the mindset evidence should now weight pre-registered work more heavily than legacy findings.
"False growth mindset": Dweck's own critique of shallow implementations
Dweck herself has been among the sharpest critics of how her work has been implemented in schools. In a widely cited 2015 Education Week piece, and in later interviews, she coined the term "false growth mindset" for practices that use the language of growth mindset without changing anything underneath: praising effort per se without pointing at strategy, telling struggling students to "just try harder", posting "yet" posters on the wall while continuing to sort students by ability, or using the word "mindset" as a synonym for grit or generic motivation.
The concern is not that these implementations are ineffective (though they generally are); it is that they can be actively unhelpful. A student who is told that effort alone is what matters, but who is not given the strategies, feedback, or curriculum access needed to succeed, may reasonably conclude that persistent failure is now their own fault. That is the opposite of the intended message.
Why mindset interventions cannot replace good teaching, resources and support
Even the most favourable readings of the evidence describe mindset as a small ingredient in a much larger system. High-quality curriculum, skilled teaching, timely and specific feedback, safe and predictable classrooms, adequate materials, and equitable opportunity outside school explain far more variance in learning than any measured belief about ability.
This has two practical implications. First, mindset work is most useful as a complement to good teaching, not a substitute. A short online module or an assembly on brain plasticity is not going to close attainment gaps on its own, and treating it as if it could tends to displace attention from the structural changes that actually would. Second, mindset framing works best when the surrounding system makes new behaviour genuinely possible: teachers who reward strategy and iteration, feedback that points at process, and access to the resources needed to act on that feedback.
This is not a defeatist conclusion. It is a specific claim about the size and shape of the effect: small, real, and conditional on the environment.
What this means for how to read your test result
The Mindset Test on this site is a research-informed self-reflection tool, not a validated psychometric instrument. Given everything above, a few reading rules follow directly from the evidence.
• Treat your overall 1–6 score as a snapshot of what you reported today, not a permanent trait. Retaking the test after a few weeks in a different mood or context will typically move the number.
• The intelligence and talent subscores are usually more informative than the overall average. A large gap between the two is a common and interesting pattern to reflect on.
• Do not treat the descriptive band label (fixed-leaning, mixed, growth-leaning) as a personality type. It is a label for a region of the scale, not a category of person.
• Do not draw conclusions from a small difference between two attempts, or from a single item. Short self-report scales are noisy at the item and single-administration level.
• Never use the result — yours or anyone else's — for hiring, admissions, grading, streaming, or clinical judgement. The evidence does not support that use, and the test has not been validated for it.
• If you are a teacher looking at aggregated Classroom Mode results, discuss patterns and strategies, not individuals. The most useful move after any mindset activity is usually a concrete change in feedback, task design, or study strategy — not a change in vocabulary alone.
Open questions worth watching
• For whom, and under what specific school and family conditions, do growth mindset interventions produce meaningful, durable gains?
• How stable are measured mindsets across weeks and months, and how much of the variance is domain-general versus domain-specific (intelligence vs. specific talents)?
• How much of the observed effect of mindset messaging is really about beliefs, and how much is about the strategies and feedback that get bundled in alongside?
• Can behavioural or physiological measures of implicit theories replace or supplement self-report scales at scale?
• How should schools and platforms evaluate the trade-off between simple, scalable mindset messaging and deeper pedagogical redesign that the evidence suggests matters more?
Selected sources
• Sisk, V. F., Burgoyne, A. P., Sun, J., Butler, J. L., & Macnamara, B. N. (2018). To what extent and under which circumstances are growth mind-sets important to academic achievement? Two meta-analyses. Psychological Science, 29(4), 549–571.
• Macnamara, B. N., & Burgoyne, A. P. (2022). Do growth mindset interventions impact students' academic achievement? A systematic review and meta-analysis. Psychological Bulletin, 149(3–4), 133–173.
• Burnette, J. L., Billingsley, J., Banks, G. C., Knouse, L. E., Hoyt, C. L., Pollack, J. M., & Simon, S. (2023). A systematic review and meta-analysis of growth mindset interventions: For whom, how, and why. Psychological Bulletin, 149(3–4), 174–205.
• Yeager, D. S., et al. (2019). A national experiment reveals where a growth mindset improves achievement. Nature, 573, 364–369.
• Yeager, D. S., Carroll, J. M., Buontempo, J., et al. (2022). Teacher mindsets help explain where a growth-mindset intervention does and doesn't work. Psychological Science, 33(1), 18–32.
• Li, Y., & Bates, T. C. (2019). Testing the role of growth mindset on response to setbacks, educational attainment, and cognitive ability. Journal of Experimental Psychology: General, 148(9), 1640–1655.
• Bahník, Š., & Vranka, M. A. (2017). Growth mindset is not associated with scholastic aptitude in a large sample of university applicants. Personality and Individual Differences, 117, 139–143.
• Dweck, C. S. (2015). Carol Dweck revisits the growth mindset. Education Week.