Psychological Bulletin, Vol 152(2), Feb 2026, 207-253; doi:10.1037/bul0000514
Math anxiety and math achievement are reciprocally related, which likely impacts individuals’ agency; their educational and career trajectories in science, technology, engineering, and mathematics fields; and countries’ economic growth in these areas. Previous meta-analyses on this relationship have faced limitations from studies by using small convenience samples and have encountered methodological issues, such as variance restriction, low statistical power, and ecological bias. To address these challenges, the present research synthesis is the first to use a comprehensive collection of representative individual participant data from international large-scale assessments. This analysis incorporated cumulative evidence from 1980 to 2022, including 452 probability samples from 90 countries, and covered student and adult populations. The meta-analytic average correlation between math anxiety and math achievement was r = −.26. Moderator analyses revealed novel evidence that this relationship is sensitive to both construct characteristics and individual and contextual factors. Specifically, the negative relationship was weaker for the worry facet of math anxiety compared to its affective and cognitive interference facets, and it was weaker following the onset of the COVID-19 pandemic in 2020. Additionally, the negative relationship was stronger for individuals with average and high levels of math achievement and in countries with higher gross domestic product per capita. Gender differences in the relationship were largely negligible after 2010, although female individuals exhibited a stronger negative relationship in the 1980s and 1990s than male individuals. In summary, synthesizing individual participant data from international large-scale assessments provided novel, nuanced, and robust evidence for research, practice, and policy. (PsycInfo Database Record (c) 2026 APA, all rights reserved)