When a new analysis of PISA data on growth mindset appears, the temptation to draw quick conclusions about what schools “should do” is never far away. That temptation only grows when the dataset looks impressive. The PISA 2022 study, for example, covers 74 countries and includes more than half a million students. A recent PLOS One article by Pimmada Charoensilp and colleagues, shared by Carl Hendrick, fits that pattern perfectly. The question I always want to ask, however, is simple: what does this study actually tell us?
The authors examine the association between a self-reported growth mindset and mathematics performance, and explore whether socioeconomic status strengthens or weakens that relationship. Methodologically, the analysis is carefully conducted within the constraints of PISA research. But that is also where its most important limitation lies. To repeat a familiar point, because it really matters: these are correlations, not causal effects. Anyone who has read my earlier post on growth mindset interventions will understand why that distinction is crucial.
The results themselves are anything but straightforward. In most countries, there is a small positive association between a growth mindset and mathematics achievement. At the same time, in many countries, there is no association at all. In a few, the association is even negative. The role of socioeconomic status is equally inconsistent. In some countries, students from lower-SES backgrounds appear to “benefit” more. In others, it is students from higher-SES backgrounds. Moreover, in more than half of the countries, socioeconomic status does not moderate the relationship at all.
Anyone looking for a clear, universal lesson will return empty-handed.
More importantly, the effect sizes are small. Even where associations reach statistical significance, their magnitude is modest. This is especially true when compared with the much stronger relationship between socioeconomic status and achievement. This aligns closely with what earlier meta-analyses have shown. It also supports the conclusion of my previous blog post: growth mindset is not a lever that automatically reduces educational inequality.
What this study clearly highlights is how context-dependent the whole story is. Beliefs do not translate into learning outcomes in a vacuum. A growth mindset may only matter when the environment supports it. Without high-quality education, strong instruction and genuine opportunities to learn, a growth mindset risks becoming a moral demand. It becomes placed on students rather than a responsibility of the education system.
Perhaps that is the real lesson here. Not that students need to think differently, but that we need to think more carefully about what education actually offers them.