Growth Mindset, Between Hype and Truth — From the Praise Experiment to a Nation-Scale Test

Introduction — the Two Words That Swept Through Education

"Children who believe ability is not fixed grow more." Carol Dweck's growth mindset is among the most famous psychological concepts of the past 20 years. It went up on classroom walls, entered corporate training, and changed how parents praise their children.

Fame brought a backlash of matching size. Criticism poured in — "the effect does not replicate," "an overhyped fad" — and today the internet serves up "growth mindset is science" and "growth mindset is pseudoscience" side by side. In this second installment of Psychology, Straight from the Papers, we walk through the controversy in the order of the original papers. To give away the ending: the truth sits in a place that makes both camps a little uncomfortable.

The Starting Point — 1998, an Experiment About One Sentence of Praise

Contrary to the popular image, the landmark paper of this field is not a study saying "believe and you will succeed" — it is an experiment about praise. The 1998 paper by Claudia Mueller and Dweck is a bundle of six experiments with some 400 fifth graders.

The design is elegantly simple. The children solved moderately difficult puzzles (Raven's matrices), were told they had done well regardless of their actual score, and then — depending on the condition — received just one different sentence of praise.

• Intelligence praise condition: "You must be really smart."
• Process praise condition: "You must have worked really hard." (or praise for the method)
• Control condition: no additional praise

The children were then given choices, deliberately confronted with hard problems, and given easy problems again. The pattern of results was consistent.

The children praised for intelligence picked "easy problems I can definitely do well on" as their next task, lost their enjoyment faster in front of hard problems, and saw their scores drop more sharply after failure. Even more striking, in a situation where they reported their score to a child at another school, the share who inflated their score was significantly higher (about 40% versus the 10% range in one experiment). The children praised for process, by contrast, chose challenging tasks at higher rates, and after failure their scores held steady or improved.

The logic of the paper goes like this. Praising "smart" turns performance into a verdict on ability. In a game where a verdict is at stake, failure is dangerous, so the child switches from learning to defending the verdict (easy problems, inflated scores). One sentence of praise changed the rules of the game.

The Backlash — the Meta-analysis That Found No Effect

The trouble came next. Intervention studies claiming "change mindset beliefs and grades go up" poured out, and the early ones had small samples and large effects — the classic trajectory of a fad.

In 2018, the meta-analysis by the team of Brooke Macnamara threw cold water on all of it. Aggregating hundreds of effects, it found that the correlation between mindset and achievement was weak, and the average effect size of the interventions was very small (below 0.1 in standardized terms). The interpretation that "money spent on growth mindset education is wasted" made the headlines.

This is the moment to bring out one of the core habits of this series: a small average effect and no effect are two different sentences. Look inside the meta-analysis and the heterogeneity of the effects was large. In other words, the picture could well be one where some groups benefit and others do not, and averaging flattened it into a small number. Testing that hypothesis required a single large, well-designed experiment.

The Definitive Test — 2019 Nature, an Experiment at National Scale

In 2019, a large research team including David Yeager and Dweck published a "national experiment" in Nature. It was a preregistered experiment that sampled US high schools with national representativeness and randomly assigned two 50-minute online sessions (learning about brain plasticity, then writing advice for younger students) to more than 12,000 incoming students.

The results were honestly nuanced.

A gain of 0.10 grade points is not a life-changing size. But look at the cost of this intervention: no teacher training, no curriculum reform — less than one hour of online sessions per student. Converted to cost-effectiveness, it ranks near the top of educational interventions, and above all, where the effect lives (lower-achieving students, a supportive peer culture) was confirmed in a preregistered design. Then, in 2023, the Yeager and Dweck side and the critical Macnamara side each published follow-up meta-analyses, and the axis of the debate shifted to "under which conditions does it work." A fight of fad versus pseudoscience turned into normal science that specifies conditions.

What We Take Home — a Three-Line Summary and One Warning

1. Praise the process. The lesson of the 1998 experiment is still practical 25 years later. With children or with colleagues, pointing at strategy and process ("starting the check from the edge cases was a good approach") instead of "you are smart" is what keeps people taking on challenges. It transplants directly into code review and feedback culture.

2. The effect of belief is real but small — and cheap. Growth mindset is not a cure-all; it is a small intervention that costs almost nothing. The arithmetic of small effect × low cost × large scale also means this: do not dismiss cheap levers like the wording of onboarding materials and the language of evaluation systems.

3. Belief alone is not enough. Mindset interventions worked in places where challenge was actually supported. Belief translates into performance when "effort makes you better" comes together with the conditions of deliberate practice — feedback, appropriate difficulty. Belief is the ignition, not the engine.

There is one thing to guard against. The moment growth mindset is used as language for shifting blame — "your failure is your mindset" — the concept becomes a tool for the exact opposite of what the original research showed. Dweck herself warned against this, calling it "false growth mindset."

Reading Guide

• Praise experiment: Mueller, C. M., & Dweck, C. S. (1998). Praise for intelligence can undermine children's motivation and performance. Journal of Personality and Social Psychology, 75(1), 33-52.
• Critical meta-analysis: 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? Psychological Science, 29(4), 549-571.
• National experiment: Yeager, D. S., et al. (2019). A national experiment reveals where a growth mindset improves achievement. Nature, 573, 364-369.

Reading tip: in the 1998 paper, Experiment 3 (score inflation) alone is worth the price of admission. In the 2019 Nature paper, the core is not the main text but Figure 1 (the effect for lower-achieving students) and the school-context analysis in the extended data. The next installment is the original paper behind the 10,000-hour debate.