Growth Mindset Training Alone May Not Boost Academic Outcomes, Researchers Say

Growth mindset training has become one of the most widely adopted interventions in American education, marketed to schools at every level as a route to higher achievement. Yet a new review published in Educational Psychology Review argues that the evidence behind this popularity is far weaker than most educators realize, and that the real promise of mindset training may lie in combining it with a scientifically robust but far less appealing strategy: retrieval practice. The review, written by Anne M. Cleary of Colorado State University and Daniel H. Robinson of the University of Texas at Arlington, contrasts the mixed evidence base for growth mindset training with the overwhelming empirical support for retrieval practice, and proposes that future research test whether the two interventions work better together than either does alone.

A growth mindset is the belief that learning abilities are malleable and can be improved through effort, in contrast to a fixed mindset, which holds that some people are simply better learners than others. Growth mindset training aims to shift students toward the former belief, on the theory that this shift will translate into persistence and better academic outcomes. The appeal is obvious: the intervention can be delivered in a single brief online session of as little as ten minutes, it is intuitive, easy to grasp, and it feels good to learners. According to the ERIC database, 181 journal article titles published between 2020 and 2025 contained the phrase “growth mindset,” a measure of the concept’s grip on educational research. But the authors caution that these same qualities—simplicity, speed, and feel-good appeal—are precisely the properties that pseudoscientific concepts like learning styles share, and they should prompt scientific scrutiny rather than reassure educators.

That scrutiny, the review shows, has not been kind to growth mindset training as a standalone treatment. In a systematic review and meta-analysis published in Psychological Bulletin, Macnamara and Burgoyne found a mean effect size of just d = 0.05 for growth mindset interventions on academic achievement. When they controlled for publication bias, the estimate fell to between 0.01 and 0.03, effectively no effect at all. When they restricted the analysis to studies in which the intervention actually changed students’ measured mindsets, or to the few studies meeting best-practices criteria, the effect on academic performance disappeared entirely. The authors concluded that most studies in the growth mindset literature suffered from serious design and interpretation flaws, including manipulations that co-varied with confounds such as instructions to work harder or personalized study plans, and cluster-randomized designs that assigned whole schools to conditions without adequately accounting for the resulting dependency in the data.

The meta-analysis also flagged a troubling pattern of potential conflicts of interest. Authors with a financial stake in growth mindset products were 2.5 times as likely to report significant effects as authors without such interests. Macnamara and Burgoyne pointed to the for-profit company Mindset Works, which markets a growth mindset software product called Brainology at a listed price of $4,499 per school per academic year, even though the Institute of Education Sciences awarded the company a $3.5 million grant in 2015 to study whether the product actually works. The reviewers compared this to obtaining federal funding for a randomized controlled trial of a medication after it had already been marketed and sold to consumers. Defenders of the intervention have pushed back: a competing meta-analysis by Burnette and colleagues reported an average effect of 0.09, and Tipton and colleagues argued that multilevel models accounting for heterogeneity yield somewhat larger estimates. But Cleary and Robinson note that even these differences are small in absolute terms, and that the debate over analytical choices does not overturn the core finding of weak average effects.

Field evidence from outside the United States points in the same direction. A randomized controlled trial in England involving more than 100 schools and over 5,000 students found no academic advantage for students who received growth mindset training. In the Czech Republic, a study of more than 5,000 university applicants found no relationship between growth mindset and admissions test scores. Meanwhile, high-profile correlational studies continue to appear in top journals—for example, an Educational Researcher article titled “Students with a Growth Mindset Learn More in School”—but as Cleary and Robinson emphasize, such studies cannot establish causation. Students who report a growth mindset may simply have higher academic ability to begin with, or a greater tolerance for the discomfort that effective learning strategies inherently involve.

That discomfort is central to the second half of the review. Retrieval practice, also known as the testing effect, refers to the finding that actively retrieving information from memory produces better long-term retention than spending the same amount of time restudying it. The phenomenon was documented more than a century ago, and it is now supported by hundreds of studies, including meta-analytic reviews. Rowland’s meta-analysis of 159 studies found that people were 2.5 times as likely to remember information after testing as after restudying, a figure rising to 4.35 times when the testing format involved free recall. Benefits have been demonstrated in simulated and real classrooms across medical education, undergraduate science, statistics, and biology, and a recent systematic review by Agarwal and colleagues confirmed medium to large effects of testing over restudying in authentic classroom contexts. Testing even reduces mind-wandering during online lectures and improves retention when delivered through intermittent smartwatch prompts.

Yet despite this mountain of evidence, retrieval practice remains underused in schools, and students rarely choose it on their own. The reason, the review argues, lies in metacognitive illusion. Retrieval is a “desirable difficulty”: it feels harder in the moment than re-reading, and learners systematically misread that difficulty as poor learning. Students consistently report that restudying feels more effective than self-testing, even though the opposite is true, and this misinterpretation of effort drives them toward the less effective strategy. The disconnect between impressions of learning and actual learning outcomes is one of the most robust findings in the science of learning, and it helps explain why interventions that feel easy and pleasant—like growth mindset training—spread readily, while interventions that demand sustained effort and tolerance of difficulty—like retrieval practice and spaced study—languish. Cleary and Robinson illustrate the asymmetry with the research methods typically used to study each: a scan of 2024 ERIC ProQuest records suggested that popular methods such as growth mindset are rarely examined with intervention studies, whereas less popular methods like retrieval practice are overwhelmingly tested experimentally.

Here is where the authors’ proposal becomes concrete. Rather than discarding growth mindset training altogether, they suggest it may operate indirectly—by building persistence and a tolerance for the unpleasantness of difficult learning—rather than directly improving achievement on its own. Supporting this interpretation, studies have found that growth mindset did not directly affect math achievement but worked through boosting self-efficacy and reasoning ability, and that mindset interventions selectively helped students from disadvantaged groups while leaving advantaged students unaffected. The authors hypothesize that at-risk students benefit because they disproportionately lack prior experience with effective study strategies and the resilience to endure the difficulty those strategies produce. Growth mindset training, in this framing, could serve as a stepping-stone: it instills the belief that learning can be improved, which then needs to be paired with explicit instruction in what students should actually do with their effort.

The authors propose bundling growth mindset training with training in retrieval practice and spacing, along with instruction on desirable difficulties and the metacognitive miscalibration that makes effective strategies feel ineffective. They point to existing frameworks such as McDaniel and colleagues’ knowledge, belief, commitment, and planning (KBCP) model, in which students must first know a strategy works, believe it will work for them, commit to using it, and then plan to use it. Within that structure, growth mindset training could supply the belief component while retrieval practice training supplies the behavioral component. Yan and Schuetze have similarly suggested broadening interventions to combine science-of-learning approaches with socio-motivational ones, and research shows that explicit training can change students’ incorrect beliefs about learning and increase their self-regulated use of retrieval practice in authentic educational settings.

The reviewers acknowledge that scaling a bundled intervention across classrooms and school systems would be logistically demanding, and that stronger evidence is still needed before broad implementation claims can be made—particularly regarding claims about benefits for at-risk students. But they argue that because growth mindset training is already deeply embedded in American education, attaching retrieval practice training and resilience training to existing mindset programs could be a viable way to deploy an evidence-based learning strategy at scale. The alternative, they suggest, is a continuation of a familiar pattern in education: quick, intuitive, feel-good interventions flourish while scientifically supported methods that demand effort remain marginal. Bundling, if the research bears it out, could finally give the science of learning the popularity it deserves—and give the growth mindset movement the evidence base it has long lacked.