Growth mindset and academic outcomes: a comparison of US and Chinese students

The study examines cross-cultural differences in intelligence mindsets and their relation to academic outcomes, focusing on Chinese and US students. Prior work suggests East Asian students attribute success to effort more than innate ability, while US students emphasize ability. Contrary to expectations that effort emphasis would align with growth mindsets, some studies show Chinese and Japanese students endorse more fixed intelligence mindsets than Western peers, yet still succeed academically. The authors hypothesize that (1) cultural differences in mindset reflect differences in how intelligence is defined (fluid vs. crystallized) and (2) the association between intelligence mindsets and academic outcomes (effort attribution, performance) may be weaker in China than in the US. The research aims to test these propositions using large-scale assessment data (PISA 2018) and college samples, linking mindset, intelligence definitions, and achievement.

Background research indicates early-emerging cultural differences: Chinese and Japanese contexts emphasize diligence and effort, including parental attributions favoring effort. In WEIRD contexts, endorsing malleable intelligence correlates with effort attributions and academic success. However, cross-cultural evidence shows Chinese and Japanese students often hold more fixed intelligence beliefs, and even ethnic Chinese in the US endorse more growth mindsets than peers in China. Lay conceptions of intelligence differ: Chinese samples often highlight reasoning, creativity, and memory; US samples emphasize practical problem-solving, verbal ability, and social competence. Mindset endorsements can shift depending on how intelligence is defined (fluid vs. crystallized). Additionally, Chinese students view school performance as more malleable than intelligence and may not link intelligence malleability tightly to effort or achievement, aligning with cultural values that emphasize hard work irrespective of innate ability.

Two studies were conducted. Study 1 analyzed PISA 2018 data to compare mindset endorsements and their relation to mathematics performance in China (mainland provinces Beijing, Jiangsu, Shanghai, Guangdong; N=11,979 after excluding missing data) and the US (N=4,663 after exclusions). Mindset was assessed with the item: “Your intelligence is something about you that you can't change very much” (1=Strongly disagree to 4=Strongly agree). Mathematics performance used 10 plausible values (PV1MATH–PV10MATH). Analyses followed PISA procedures (intsvy in R), accounting for demographics, school-level effects, and weights. Regressions tested country differences in mindset and country × mindset interactions predicting math scores; emphasis was placed on effect sizes, CIs, and visualizations, given the large sample sizes. Pairwise comparisons of math scores across mindset response categories were computed per country with effect sizes. Study 2 recruited college students from the University of Michigan (US; N≈189) and Beijing Normal University (China; N=171). Measures included: (a) an open-ended spontaneous definition of intelligence coded into four categories: fluid-only, crystallized-only, both, or neither (interrater reliability substantial in both languages; discrepancies resolved by discussion); (b) Implicit Theories of Intelligence (ITIQ; general intelligence mindset; Likert scales as per instruments used; descriptive reporting used a 6-point Likert scale summary in Table 3); (c) effort-ability attributions for mathematics achievements at school level (e.g., getting an A in algebra) and expert level (e.g., becoming a math professor; Fields Medal), rated 1=Pure ability to 7=Pure effort; (d) mindsets for fluid and crystallized intelligence using modified ITIQ after providing definitions. Procedure: Block 1 collected spontaneous definitions and general mindset; Block 2 presented definitions and measured mindsets for fluid/crystallized intelligence; Block 3 assessed effort-ability attributions; Block 4 captured demographics. Analyses included t-tests, correlations, mixed ANOVAs for country × intelligence-definition (self-defined, fluid, crystallized) on mindset, post hoc contrasts with adjusted p-values, chi-square tests comparing distribution of definition categories across countries, and ANOVAs linking definition category to mindset ratings.

Study 1: • Mindset distribution: 68.39% of US vs. 55.61% of Chinese students endorsed malleability (responses 1–2). Regression showed Chinese students held more fixed mindsets than US students (βcountry=0.24, SE=0.02, t=11.59, 95% CI [0.20, 0.28]). • Math performance associations: Country and mindset predicted math scores (βcountry=37.89, SE=6.30, t=6.01; βmindset=-59.40, SE=3.77, t=-15.77). The interaction was significant (βcountry*mindset=34.16, SE=2.28, t=14.99), indicating different associations across countries. • Pairwise comparisons (math score differences, effect sizes): In China, more fixed mindsets associated with slightly higher scores (Mdiff=10.51–24.86; d=0.13–0.31). In the US, more fixed mindsets associated with lower scores (Mdiff=-22.64 to -72.78; d=-0.25 to -0.86). Study 2: • Mindset levels: Chinese students reported more fixed mindsets than US students for general intelligence (China M=3.08, SD=0.83; US M=4.12, SD=1.09; t(358)=-10.08, p<0.001, d=1.07). • Effort-ability attributions: For school-level achievement, Chinese students were more effort-oriented (China M=5.53, SD=1.09; US M=4.85, SD=1.14; t(358)=5.83, p<0.001, d=0.61). For expert-level achievement, Chinese students were more ability-oriented than US (China M=2.55, SD=1.09; US M=3.44, SD=1.38; t(358)=-6.72, p<0.001, d=0.71). • Correlations between mindset and attributions: School-level achievement: US r=0.22, p=0.003; China r=0.10, p=0.196 (Z=1.16, p=0.123). Expert-level achievement: US r=0.33, p<0.001; China r=0.15, p=0.048 (Z=1.80, p=0.036). Overall attribution composite: US r=0.41, p<0.001; China r=0.18, p=0.016 (Z=2.38, p=0.009). • Mindsets by definition type (mixed ANOVA): Significant main effects of country (F(1,358)=55.88, p<0.001, η²=0.14), intelligence definition (self-defined, fluid, crystallized; F(2,716)=205.58, p<0.001, η²=0.37), and their interaction (F(2,716)=23.08, p<0.001, η²=0.06). Across self-defined, fluid, and crystallized definitions, Chinese students endorsed more fixed mindsets than US students. • Post hoc patterns: In the US, fluid intelligence was rated more fixed than self-defined and crystallized; self-defined more fixed than crystallized. In China, self-defined was most fixed, followed by fluid, then crystallized. • Spontaneous definitions and mindsets: Definition category affected mindset (F(2,292)=5.11, p=0.007). Crystallized-only definitions associated with more growth mindsets than fluid-only (t(204)=-6.54, p<0.001, d=0.92) and combined definitions (t(183)=-4.64, p<0.001, d=0.68). • Cross-cultural distribution of definitions: Chinese students more often gave fluid-only and less often crystallized-only definitions than US students (χ²(2)=39.44, p<0.001), aligning with their more fixed mindsets.

Findings show robust cross-cultural differences in intelligence mindsets and their connections to academic outcomes. Despite cultural emphasis on effort, Chinese students endorsed more fixed intelligence mindsets than US students. In the US, growth mindsets were moderately to strongly associated with higher mathematics performance and more effort-oriented attributions. In China, fixed mindsets were slightly associated with higher mathematics scores, and mindset-attitude links were weaker. These patterns are explained by cultural differences in how intelligence is conceptualized: US students tended to define intelligence more in crystallized terms (perceived as more malleable), whereas Chinese students emphasized fluid components (perceived as less malleable). Consequently, the meaning and implications of “growth mindset” vary by culture. The results suggest that focusing on intelligence malleability may be less relevant to Chinese students’ motivation and achievement, where effort is valued independently of beliefs about intelligence. The work cautions against universalizing WEIRD-based mindset models and underscores the importance of cultural context when interpreting mindset–achievement relations and designing interventions.

Across a large PISA sample and college student samples, the study demonstrates that Chinese students hold more fixed intelligence mindsets than US students while simultaneously valuing effort and achieving strongly in school. Two key explanations are supported: (1) cross-cultural differences in definitions of intelligence (fluid vs. crystallized) align with differences in mindset endorsements, and (2) the link between intelligence mindsets and academic outcomes is markedly weaker in China than in the US. The findings challenge the universality of growth mindset effects and suggest interventions might be more effective if they emphasize the malleability of mastering school subjects and effort attributions rather than intelligence per se, especially in contexts where effort is culturally central. Future research should broaden cultural coverage beyond the US and China and elaborate on culturally specific conceptions of intelligence beyond the fluid–crystallized framework.

• Cultural scope: Analyses focus on US and mainland Chinese samples; results may not generalize to other cultures. • Construct coding: Study 2’s coding of intelligence definitions within a fluid–crystallized framework may not capture all culturally specific conceptions (notably, 23% of Chinese definitions fell into neither category). • Measurement: Study 1 used a single-item mindset measure; more nuanced scales might capture additional variance. • Interpretation of large-sample significance: With very large N in Study 1, many effects are statistically significant; emphasis on effect sizes and CIs mitigates but does not eliminate concerns. • Generalizability across educational levels: Study 2 used highly selective universities; findings may differ in other postsecondary contexts or age groups.