A Meta-Analysis on Teachers' Growth Mindset

Mindsets (implicit theories) reflect beliefs about the malleability of attributes like intelligence. A growth mindset implies these attributes can be changed, whereas a fixed mindset views them as stable. While student-focused research has proliferated, evidence on teachers’ growth mindset and its implications for teacher motivation, instructional practices, and student outcomes remains fragmented and mixed. The study addresses whether teachers’ growth mindset relates to (a) teacher outcomes (self-efficacy, achievement goals, teachers’ achievement test performance), (b) instructional practices (classroom goal structures), and (c) student outcomes (achievement). It also examines measurement and study moderators (dimensionality, referent, mindset domain; world region, educational level, publication year/status, study quality) to clarify when and how teacher mindset associations vary. The meta-analysis aims to provide the first quantitative synthesis focused specifically on teachers’ growth mindset and educationally relevant outcomes.

Prior mindset research with students suggests growth mindset may modestly benefit motivation and achievement, though average effects in meta-analyses are small and heterogeneous. Theoretically, teachers’ growth mindset could shape their own motivation (more mastery goals, higher self-efficacy), inform instructional practices (mastery goal structures), and influence student achievement through expectations and classroom climate. Empirical findings in teacher samples have been mixed: some studies report positive links with self-efficacy and mastery goals; links to performance goals, teachers’ test performance, and student achievement are inconsistent or trivial. Meta-analyses outside teacher samples report small correlations between mindset and achievement goals (e.g., Payne et al., 2007; Burnette et al., 2013). Measurement debates persist regarding whether mindset is bipartite (fixed and growth scales) or unipolar and the impact of item referent (self, you, students, teachers) and domain (intelligence vs. ability) on observed effects. Cultural and contextual differences (e.g., US/Canada vs. Europe) may also shape associations.

Literature search was conducted March 11–15, 2024 in PsycINFO, Web of Science, and ProQuest Dissertations using comprehensive keywords combining teacher roles, mindset terminology (implicit theories, entity/incremental, fixed/growth mindset, beliefs about intelligence/ability), and outcomes (self-efficacy, achievement goals, goal structures, achievement/performance). Initial retrieval N=1,759 (850 Web of Science; 715 PsycINFO; 194 ProQuest), with N=1,495 after deduplication. Screening included quantitative studies involving teachers, preservice teachers, or university instructors assessing Dweck-consistent mindset and targeted outcomes, with full texts available. Interrater agreement for inclusion was 96.04%. After exclusions (e.g., missing zero-order correlations, ambiguous mindset measures, unsuitable outcomes, data overlap, unmatched teacher–student levels, unavailable full texts), an expanded search via references, citation tracking, and expert recommendations yielded 16 additional studies. Final sample: 50 studies (52 independent samples), 81 effect sizes, N=19,555 teachers/preservice teachers. Regional distribution: US/Canada 67.3% (k=35), Europe 25% (k=13), Asia 5.8% (k=3), Australia 1.9% (k=1). Educational levels: primary (k=10), secondary (k=16), tertiary (k=4), mixed (k=10), preservice (k=9). Publication status: 27 published, 25 unpublished. Coding: Two coders independently extracted growth mindset–outcome correlations (93.7% agreement). Fixed mindset scales were sign-reversed to represent growth mindset. For multiple self-efficacy domains within sample, Fisher z-transformed correlations were averaged and back-transformed. When both teacher-focused and self-focused self-efficacy were present, self-focused was used. Moderators coded included dimensionality (recoded vs. not recoded), referent (teachers in general, students, self, you, mixed/other), content (intelligence, ability, mixed/other), educational level, world region, publication year/status, and study quality (0–5 index per adapted Newcastle Ottawa Scale). Statistical analyses: Three-level random-effects meta-analysis (maximum likelihood) synthesized all outcomes, modeling heterogeneity at participant, within-study, and between-study levels, allowing inclusion of dependent effect sizes. Two-level random-effects meta-analyses were conducted within outcome subgroups (independent effects). Heterogeneity was summarized via I² thresholds. Moderation was tested with single three-level mixed-effects meta-regressions across all correlations and two-level regressions within outcome subsets (dummy coding of nominal predictors). Dissemination bias was assessed using methods suitable for multilevel meta-analysis: Egger’s regression with multilevel sandwich estimator (Egger MLM), Egger sandwich, and three-parameter selection model (3PSM), interpreting p<.10 as indicative of bias. Analyses were conducted in R. Transparency: coding file and scripts available on OSF (https://osf.io/zyt4b/).

Overall association across outcomes: small positive effect r=.111, p<.001 (three-level model; k=80 effects nested in 52 studies). Outcome-specific meta-analytic correlations (two-level models): - Self-efficacy: r=.180, p<.001 (k=29), significant small-to-typical positive association. - Mastery-approach goals: r=.148, p<.001 (k=7), significant small positive association. - Performance-approach goals: r=.033, p=.084 (k=7), trivial, non-significant. - Performance-avoidance goals: r=.077, p=.074 (k=6), trivial, non-significant. - Mastery goal structures: r=.137, p=.001 (k=8), significant small positive association. - Performance goal structures: r=.014, p=.717 (k=6), trivial, non-significant. - Teachers’ achievement test performance: r=.013, p=.852 (k=8), trivial, non-significant. - Students’ achievement: r=-.024, p=.309 (k=10), trivial, non-significant. Moderator analyses (selected significant effects): - Across all outcomes: publication year b=-0.008, p=.008 (decline effect); Europe b=-0.085, p=.005 and Australia b=-0.673, p<.001 vs US/Canada (smaller effects); publication status (unpublished larger) b=0.130, p<.001. - Self-efficacy: Europe smaller b=-0.142, p=.005; unpublished larger b=0.116, p=.006. - Performance-avoidance goals: referent “you” vs mixed/other b=0.180, p=.003 (larger); unpublished larger b=0.199, p=.001; educational level smaller in tertiary b=-0.195, p=.002 and mixed levels b=-0.243, p=.006 vs primary. - Teachers’ test performance: unpublished larger b=0.262, p=.022; Europe smaller b=-0.335, p=.016 vs US/Canada; higher study quality associated with smaller effects b=-0.130, p=.024. - Mastery goal structures: nonrecoded mindset measures yielded smaller effects b=-0.247, p<.001 vs recoded; higher study quality associated with larger effects b=0.115, p=.011. - Performance goal structures: referent “you” yielded smaller effects b=-0.358, p=.012 vs mixed/other; content mixed/other larger b=0.333, p=.017 vs intelligence; secondary education smaller b=-0.358, p=.012 vs primary. - Student achievement: nonrecoded mindset measures yielded larger effects b=0.276, p=.021 vs recoded (still overall non-significant average association). Additional subgroup analysis: preservice teachers showed larger associations with teachers’ test performance (r=.115) than mixed educational levels (r=-.259), Q=15.287, df=1, p<.001. Dissemination bias tests: no evidence of bias (Egger MLM p=.316; Egger sandwich p=.259; 3PSM p=.930).

Teachers’ growth mindset shows conceptually aligned, small-to-typical positive links with mastery-oriented motivation (self-efficacy, mastery-approach goals) and with mastery goal structures, indicating relevance for teachers’ adaptive motivational beliefs and classroom climates emphasizing learning and improvement. Associations with performance-oriented goals, performance goal structures, teachers’ own achievement test performance, and student achievement were trivial and non-significant, suggesting limited practical salience of teacher mindset for these outcomes on average. Moderator findings highlight context and measurement sensitivities: effects have declined over time, are generally stronger in US/Canada than in Europe (and Australia, cautiously), and are larger in unpublished work (reverse publication bias, though formal bias tests did not indicate confounding). Measurement nuances matter: mindset item referent and content shape some associations, and dimensionality (recoded vs. nonrecoded) differentially relates to mastery goal structures and student achievement, suggesting fixed-mindset items recoded to growth may discriminate better in some contexts. Educational level moderated performance-oriented associations (smaller/more negative at higher levels), aligning with claims that performance emphasis increases in advanced settings. Study quality influenced estimates (smaller for teacher test performance; larger for mastery goal structures), underscoring methodological impacts. Overall, findings refine theory by situating teacher growth mindset primarily within mastery-related motivational and instructional domains, with limited direct ties to performance outcomes and student achievement.

This first meta-analysis focused on teachers’ growth mindset indicates small, meaningful associations with teachers’ self-efficacy, mastery-approach goals, and mastery goal structures, but trivial, non-significant links with performance goals, performance goal structures, teachers’ achievement test performance, and student achievement. Results caution against overemphasizing teacher growth mindset as a broad lever for performance outcomes but support its role in fostering mastery-oriented motivation and classroom climates. Future research should: develop and evaluate teacher-focused mindset interventions; examine heterogeneous effects (e.g., for at-risk students); employ diversified, less bias-prone assessments (observations, student reports, implicit measures, vignettes); compare mindset measurement approaches (recoded vs. nonrecoded; referent; domain specificity); and expand nuanced analyses of goal structures (e.g., approach vs. avoidance; appearance vs. normative) to detect potentially stronger, subtype-specific associations.

The synthesis relied on zero-order correlations and cannot establish causality; intervention or experimental work targeting teacher mindset is needed. Some outcome subsets had few studies, limiting power for moderator analyses and precision of estimates. Studies primarily used self-report measures for teacher variables and goal structures, which may be susceptible to social desirability and response biases; triangulation with observations, student ratings, implicit measures, and situational judgment tests is recommended. Goal structure measures often used a mastery–performance dichotomy, not reflecting newer, more nuanced subtypes (approach vs. avoidance; appearance vs. normative), potentially masking stronger subtype-specific links. Heterogeneous effects (e.g., benefits for specific student subgroups) were not tested. Decline effects and regional differences suggest contextual and temporal influences; measurement characteristics (dimensionality, referent, domain content) may alter effect sizes, complicating generalization.