Academic Self-Concept

Learner Variable · All learner variables
Evidence · 106 claims (76 show a difference, 22 mixed, 8 no difference) · 53 studies (27 associational, 10 review, 9 causal, 2 quant-synthesis, 2 qualitative, 2 design, 1 theoretical), q1–q4 · 13 of 53 report an effect size · 103 claims rest on one study

Description

Academic self-concept is a learner's self-perception of ability in a domain ("I am good at maths"). It is domain-specific: mathematics self-concept and verbal self-concept are only weakly related, because learners compare their own subjects with one another as well as with other learners (the internal/external frame of reference model). It is formed partly by comparison with classmates, so the same learner may report a lower self-concept in a high-achieving class (the big-fish-little-pond effect). Self-concept and achievement affect each other over time. It is related to but distinct from self-efficacy, which is confidence about succeeding at specific tasks.

How strongly it relates to achievement

Mathematics. These are correlations across meta-analyses (Breit et al. 2025, a systematic review of 30 meta-analyses; overview [~W]): how strongly learners who differ on this characteristic differ in mathematics achievement, not how much changing it would change achievement. Many of these characteristics are correlated with one another, and the review is about mathematics only.

• Mathematics self-concept r = .43, the largest of the 18 personality-category correlates; verbal self-concept .12. Claim [+W]

Implications

Context

• Grouping, feedback and comparison practices shape self-concept.

Target Learners

• Learners with a low view of their ability in a subject.

Target Learning Objectives

• Achievement, course choice and persistence in a subject.

Claims

• Cognitive-only coaching increased deep and achieving learning approaches, academic self-concept, and reduced test anxiety and anxiety/depression, but academic performance declined relative to control; behavioural-only coaching decreased test anxiety and increased academic performance with no other effects. [+M] — instruction changes it
• Early math exposure in young children is linked to development, math confidence, and later math use (authors' assertion) [+W] — instruction changes it
• Female students show stronger academic self-confidence, more positive attitudes, higher achievement, and fewer absences, but this advantage diminishes almost completely under mastery learning [~M] — learners who differ on it differ in outcomes
• Girls show less confidence and interest in math and science from early adolescence, and a strong math/science self-concept predicts course choice and performance for both genders [+M] — learners who differ on it differ in outcomes
• Interpersonal audio discussions with feedback and self-assessment elicit a statistically significant increase in introductory L2 learners' perceived pronunciation ability [+W] — instruction changes it
• Lack of confidence and inappropriate language use were observed obstacles to students' mathematical communication during the experiment [+M] — learners who differ on it differ in outcomes
• Learners' task-specific self-efficacy beliefs predict whether they initiate, persist through, and recover from difficulty on academic tasks. [+M] — learners who differ on it differ in outcomes
• Pre-service primary teachers' science teaching self-efficacy beliefs correlate weakly but positively with their ability to determine suitable taxonomy, strategy, and method-techniques [+W] — learners who differ on it differ in outcomes
• Project-based and experiential learning approaches are reported as most effective for AI literacy education, and early exposure enhances learners' self-efficacy [+W] — instruction changes it
• Prolonged reading failure leads students to attribute failure to fixed low ability and define reading as a failure situation [+M] — instruction changes it
• Students confident about their mathematical ability are mostly also convinced of the relevance of mathematics, forming a motivational basis; low-confidence students are hard to motivate [+M] — learners who differ on it differ in outcomes
• Students reported significantly less difficulty with comprehensibility and accuracy of pronunciation between the first and final audio discussion activity [+W] — instruction changes it
• Math self-concept is a stronger predictor of math achievement, whereas value beliefs are stronger predictors of self-reported effort [+M] — learners who differ on it differ in outcomes
• Only the self-concept-by-low-cost interaction adds predictive power for achievement beyond the self-concept-by-global-value interaction [+W] — learners who differ on it differ in outcomes
• Self-concept and global value interact synergistically and positively in predicting achievement, effort, and teacher-rated engagement [+M] — learners who differ on it differ in outcomes
• Self-concept and value beliefs emerge as equally important predictors of teacher-rated behavioral engagement [+M] — learners who differ on it differ in outcomes
• Adolescents in the Stable High motivation trajectory earned higher science grades, took more AP science courses, and reported higher STEM career aspirations than the other trajectory groups [+M] — learners who differ on it differ in outcomes
• Community-based mathematical investigations gave students confidence to use mathematics for social change [+M] — instruction changes it
• Math circle participants reported positive mathematics attitudes, with most respondents indicating mathematics was important, interesting, liked, and that they felt confident solving problems [+M] — instruction changes it
• Ninth grade marked a turning point where transitory motivation trajectories reversed direction, and these inconsistent trajectories were associated with lower AP enrollment and science performance [+M] — learners who differ on it differ in outcomes
• Students in steadily or sharply declining motivation trajectories showed lower STEM career aspirations and lower science grades than stable-trajectory peers [+M] — learners who differ on it differ in outcomes
• Academic achievement is moderately correlated with academic self-concept in African American middle school youth [+M] — learners who differ on it differ in outcomes
• A 15-minute exposure to higher-performing virtual classmates did not significantly affect sixth-graders' dispositional domain-specific self-concept [-M] — instruction changes it
• A high proportion of high-performing classmates lowers sixth-graders' situational academic self-concept, an experimental BFLPE at the extreme contrast [+M] — instruction changes it
• Content knowledge does not significantly predict the number of experiments conducted; general cognitive ability shows a quadratic association and self-concept a negative linear association [+M] — learners who differ on it differ in outcomes
• Dispositional factors add little to predicting inquiry performance except physics self-concept, and extraneous load has a significant negative effect [+M] — learners who differ on it differ in outcomes
• Latent profile analysis identifies three learner profiles differing in cognitive factors, self-concept, and extraneous load, with significant performance differences [~M] — learners who differ on it differ in outcomes
• Students who perceived the class performance level as higher reported lower situational self-concept in the IVR classroom [+M] — instruction changes it
• Students' perceived class performance level fully mediates the effect of classmates' hand-raising behavior on situational self-concept [+M] — instruction changes it
• Changes in mathematics self-concept and interest are independent of prior achievement and do not predict later achievement [-W] — learners who differ on it differ in outcomes
• Co-teaching does not differ from solo teaching in the development of students' mathematics self-concept and interest [-W] — instruction changes it
• Sixth-graders' mathematics self-concept and individual interest both decline over one school year [+W] — instruction changes it
• Domain-specific self-concept consistently predicts metacognitive judgments and accuracy: higher self-concept means higher judgments, higher sensitivity, lower specificity, and more overconfidence [+M] — learners who differ on it differ in outcomes
• Global predictions rely on domain-specific self-concept but not on test performance, unlike postdictions [+M] — learners who differ on it differ in outcomes
• Domain-specific self-concept relates more strongly to global than to local postdicted judgments, while most other motivational and personality variables do not differ [~M] — learners who differ on it differ in outcomes
• A meta-analysis of 68 longitudinal studies found prior achievement predicted subsequent self-concept and prior self-concept predicted subsequent achievement [+M] — learners who differ on it differ in outcomes
• Across basic and extended RI-CLPMs and CLPMs of five-wave secondary-school data, math self-concept and achievement show consistent reciprocal (bidirectional) directional ordering [+M] — learners who differ on it differ in outcomes
• Changes in math self-concept correlate positively with changes in both comparison-related beliefs after the COMPASS intervention [+M] — instruction changes it
• Dimensional comparisons: achievement–self-concept paths are positive in similar subjects but negative in dissimilar subjects [~M] — learners who differ on it differ in outcomes
• Huang's meta-analysis of 39 independent samples reported average cross-lagged correlations between self-concept and achievement of roughly .19 to .27 [+M] — learners who differ on it differ in outcomes
• Math self-concept gains from the COMPASS intervention are sustained at three- and six-month follow-ups in Group A, but the three-month gain in waiting-list Group B was not statistically significant [~W] — instruction changes it
• Meta-analyses show prior academic self-concept predicts subsequent achievement, controlling prior achievement [+M] — learners who differ on it differ in outcomes
• Reanalyses of simulated and PALMA data provide converging support for the extended reciprocal effects model, including an effect of academic self-concept on achievement [+M] — learners who differ on it differ in outcomes
• School grades tend to be more highly correlated with academic self-concept than standardized test scores [+M] — learners who differ on it differ in outcomes
• Students' domain-specific self-concept moderates non-interactive teaching: low self-concept students profit, high self-concept students do not [+M] — an instructional effect differs with it
• The COMPASS intervention significantly increases students' math self-concept immediately after participation in both an intervention group and a waiting-list control group [+M] — instruction changes it
• Student characteristics beyond prior knowledge bear on inquiry learning: inhibitory control predicts inquiry outcomes, and in one study a direct-instruction advantage over inquiry held only for students with low domain self-concept [+M] — an instructional effect differs with it
• Adding a quadratic achievement term leaves the negative contextual-mean effect essentially unchanged [+M] — instruction changes it
• Class-average achievement negatively predicts mathematics self-concept after controlling individual achievement (BFLPE replication in TIMSS 2019) [+M] — instruction changes it
• Contextual-mean effects show little moderation by individual achievement; apparent positive moderation is illusory when the quadratic term is omitted [+M] — instruction changes it
• Integrative models indicate much of the apparent ordinal-rank effect reflects distributionally scaled relative achievement, while contextual means remain the most robust comparative indicator [+W] — instruction changes it
• Nonlinear epistemic belief profiles show a diverging pattern for science self-concept: multiplistic profiles score similarly or lower than absolutistic, while evidence-based profiles score similarly or higher than sophisticated [~M] — instruction changes it
• The 2I/E model is supported: social comparison effects are moderate, dimensional comparison effects small to moderate, and temporal comparison effects small on math and verbal self-concepts [+M] — instruction changes it
• The contextual-mean effect remains negative across every achievement decile and across MSC quantiles, with no systematic attenuation at the top [+M] — instruction changes it
• The negative contextual-mean effect persists for the top-ranked student in each class, constraining rank-sufficiency accounts [+M] — instruction changes it
• In German secondary-school mathematics, effects of self-concept on achievement are lagged (lag1) whereas effects of achievement on self-concept are more contemporaneous (lag0) [+M] — learners who differ on it differ in outcomes
• Abolishing grades in elementary schools showed no statistically significant effects on achievement, self-concepts, anxiety, joy of learning, well-being, or classroom climate by end of Grade 3, except math interest [-M] — instruction changes it
• Additional modified models suggest a positive effect of change in achievement on contemporaneous change in self-concept [+M] — instruction changes it
• Dimensional comparison effects extend to subjective task values, with contrast effects strongest for intrinsic value and costs and lowest for utility value, often mediated by self-concept [+M] — learners who differ on it differ in outcomes
• Dimensional comparisons carried out by teachers and classmates show no contrast effects on students' self-concepts, unlike students' own comparisons [~M] — instruction changes it
• Dimensional comparisons predict cross-domain academic course and major choice, with negative effects of verbal-domain beliefs on math-related choices [+M] — learners who differ on it differ in outcomes
• Meta-analytic path analyses show math and verbal achievements correlate highly positively while the corresponding self-concepts correlate near zero, with significant negative cross-domain paths [+M] — instruction changes it
• STARTS model shows a positive effect of prior achievement on subsequent self-concept but no effect of prior self-concept on subsequent achievement [+M] — instruction changes it
• Subject similarity moderates dimensional comparison effects: far comparisons (e.g., math vs. verbal) yield strong contrast effects, near comparisons yield weaker contrast or assimilation effects [~M] — instruction changes it
• Academic self-concept predicts lagged (across-semester) changes in academic skills in both mathematics and French [+M] — learners who differ on it differ in outcomes
• Academic skills predict contemporaneous (within-semester) changes in academic self-concept in both mathematics and French [+M] — instruction changes it
• Reciprocal effects are robust to time-invariant confounders, but the French lagged self-concept effect did not survive the lag2 carryover stress test [~M] — learners who differ on it differ in outcomes
• The temporal asymmetry holds across gender and initial ability levels, with a cautious gender difference in the French lagged effect [~M] — learners who differ on it differ in outcomes
• The temporal asymmetry in effect sizes: contemporaneous skill effects on self-concept are moderate (d = 0.33 in French) while alternative cross-effects are null or negligible [~M] — instruction changes it
• Adolescents' self-understandings are informed by their perceptions of teacher expectations, which may differ across student groups [~M] — instruction changes it
• Class-average achievement shows a small negative indirect effect on subsequent mathematics achievement mediated through mathematics self-concept [+M] — learners who differ on it differ in outcomes
• The Big-Fish-Little-Pond effect on mathematics self-concept is negative and statistically significant in all three countries under a fully specified doubly latent model [+M] — instruction changes it
• When compositional effects are constrained equal across countries, the overall peer spillover effect is negative and non-significant while the BFLPE remains negative and significant [~W] — instruction changes it
• The parental aspiration-expectation gap negatively predicts student expectations, academic self-concept, GPA, and age-26 outcomes [+W] — instruction changes it
• Year 10 teacher and parent educational expectations have consistently positive effects on student expectations, academic self-concept, GPA, and age-26 outcomes [+W] — instruction changes it
• Contrast and assimilation are competing compositional processes that can operate simultaneously [~M] — instruction changes it
• School-average achievement negatively predicts academic self-concept (BFLPE) and GPA in the comprehensive doubly-latent model [+M] — instruction changes it
• Starting contemporaneous effects on wave 2 in the STARTS model yields positive contemporaneous effects of achievement on academic self-concept and positive lagged effects of self-concept on achievement [+M] — learners who differ on it differ in outcomes
• Receiver characteristics such as academic self-concept, self-efficacy, and achievement are proposed to moderate learners' engagement with feedback, but the achievement relation is conflicting [~M] — an instructional effect differs with it
• Classroom-average mathematics achievement shows no statistically significant association with ASC development when clique achievement is modeled simultaneously [-M] — instruction changes it
• Clique-average mathematics achievement negatively predicts the development of adolescents' academic self-concept (a BFLPE at the clique level) [+M] — instruction changes it
• Girls exhibited a stronger decrease in mathematics ASC than boys over the school year, while school type did not predict ASC development [~M] — instruction changes it
• Individual mathematics achievement is positively associated with students' ASC development over the school year [+M] — instruction changes it
• No statistically significant assimilation effect (BIRGE) of belonging to a higher-achieving clique within the classroom on ASC change [-M] — instruction changes it
• Variance in ASC change is statistically significant at individual, clique, and classroom levels, with a slight average decline in ASC across the school year [+M] — instruction changes it
• Adaptive task difficulty compensates low interest but not low self-concept among vocational-school students during repeated practice [-W] — an instructional effect differs with it
• Negative feedback lowered math self-concept indirectly via enjoyment and hope; positive feedback raised it via hope [+W] — instruction changes it
• Prior math self-concept moderated positive feedback on self-concept: positive feedback strengthened pre-existing self-concept [~W] — an instructional effect differs with it
• Equally able students report lower academic self-concepts in high-achieving classrooms than in low-achieving classrooms (the Big-Fish-Little-Pond Effect) [+M] — instruction changes it
• The BFLPE exists in both elementary-school and secondary-school contexts [+M] — instruction changes it
• The BFLPE from elementary school diminishes after transition to secondary school and is no longer discernible after one year [~M] — instruction changes it
• In more competitive school environments, students from lower SES backgrounds report more positive self-views than their more advantaged peers, narrowing SES gaps in self-views [~W] — instruction changes it
• Academic self-concept relates most strongly to achievement, with achievement goals, achievement motivation, autonomous motivation, and effort beliefs each adding unique variance [+W] — learners who differ on it differ in outcomes
• Educational certificates are the most important tracking feature shaping students' self-beliefs and school disengagement at the end of schooling [+W] — instruction changes it
• Schoolmates' achievement (big-fish-little-pond effect) relates only to academic self-concepts, not to labor market beliefs or disengagement [+W] — instruction changes it
• Students with the intermediate school-leaving certificate report higher academic self-concepts, better perceived labor market chances, and less disengagement than students with the low certificate [+W] — instruction changes it
• Track level did not predict students' self-beliefs or school disengagement at the end of schooling [-W] — instruction changes it
• Grade retention has positive effects on math self-concept in the first year following retention, maintained over subsequent years [+M] — instruction changes it
• A one-SD increase in mathematics self-concept raises predicted probability of advanced mathematics course choice by 16%; higher German self-concept lowers it [+M] — learners who differ on it differ in outcomes
• Differences in self-concept and value partly, but not fully, explain gender differences in STEM study intentions [~M] — learners who differ on it differ in outcomes
• Domain-specific academic self-concept predicts advanced course choice, whereas value predicts STEM study intentions [+M] — learners who differ on it differ in outcomes
• Grades show mediated dimensional comparison effects on course choices via self-concept and value [+M] — learners who differ on it differ in outcomes
• BFLPE is stronger on the regular class level and strongest on the domain-specific math class level [+M] — instruction changes it
• The school-level BFLPE on math self-concept is weak once class-level achievement is controlled [~M] — instruction changes it

Related Learner Variables

Examples

Key Sources

• Breit, M., Schneider, M., & Preckel, F. (2025). Mathematics achievement and learner characteristics: A systematic review of meta-analyses. Learning and Individual Differences, 118, 102621. doi:10.1016/j.lindif.2024.102621

Page metadata

Field	Value
type	learner-variable
id	academic-self-concept
title	Academic Self-Concept
description	A learner's perception of their own ability in an academic domain, such as how good they think they are at mathematics or at reading.
status	review
canonical	True
generated	by claude/unspecified on 2026-10-16
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.