Academic Anxiety

Learner Variable · All learner variables
Evidence · 135 claims (86 show a difference, 33 mixed, 16 no difference) · 86 studies (36 associational, 17 review, 15 causal, 6 quant-synthesis, 6 theoretical, 5 qualitative, 1 design), q1–q4 · 21 of 86 report an effect size · 128 claims rest on one study

Description

Academic anxiety is fear or tension aroused by a subject or by being evaluated in it. The research mostly studies three forms: mathematics anxiety, test anxiety and foreign-language anxiety, each measured with self-report scales. Anxiety predicts lower performance, partly because worry takes up working memory; it moderates instructional effects (timed or high-stakes formats hurt anxious learners more); and it is an outcome that some designs reduce, such as low-stakes practice testing and expressive writing before an exam. It is a narrower case of affect regulation.

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 anxiety r = −.28 (k = 747), weaker in older samples; test anxiety −.23; general anxiety −.19 and −.09. Claim [+W]

Implications

Context

• High-stakes, timed and public formats raise anxiety; frequent low-stakes practice tends to lower it.
• An anxious learner's score can understate what they know.

Target Learners

• Learners anxious about mathematics, tests or speaking a second language.

Target Learning Objectives

• Performance under assessment, and willingness to take further courses in the subject.

Claims

• A 15-week extensive reading course increased Japanese EFL learners' feelings of comfort with foreign-language reading (r = .54) [+M] — instruction changes it
• Academic emotions—enjoyment, anxiety, and boredom—mediate the association between perceived AI-adaptive environments and SRL, accounting for 54.8% of the total effect [+W] — learners who differ on it differ in outcomes
• AI anxiety attenuates the positive association between AI literacy and trust in AI [~M] — an instructional effect differs with it
• AI anxiety is negatively associated with academic motivation among university students, in both bivariate and regression analyses [+W] — learners who differ on it differ in outcomes
• AI reduces foreign language anxiety and strengthens self-efficacy and self-confidence, supporting the Affective Foundation Pathway [+M] — instruction changes it
• AI simulation induced higher state anxiety than standardized patient simulation in one RCT, while qualitative studies reported reduced anxiety [~M] — instruction changes it
• AI-assisted pair programming significantly reduced programming anxiety compared to individual programming [+M] — instruction changes it
• Anxiety and toxic stress from stereotypes, bullying, and exclusionary practices impede learning by preoccupying the brain with worry and fear [+M] — learners who differ on it differ in outcomes
• Anxiety in foreign language classroom learning experiences is linked with oral performance among Japanese EFL learners [+W] — learners who differ on it differ in outcomes
• Anxiety shows positive concurrent associations with acceptance outcomes for AI-assisted English learning, and should not be read as uniformly negative or beneficial [+M] — learners who differ on it differ in outcomes
• At semester end, intrinsic motivation was positively correlated and programming anxiety negatively correlated with programming performance [+M] — learners who differ on it differ in outcomes
• Brief expressive writing before an exam can raise test-anxious students' scores, but across the wider literature its effect on test performance is negligible [~S] — an instructional effect differs with it
• Cognitive apprenticeship students became more anxious about mathematics but reported gains in self-confidence, rapport, and enjoyment [+W] — instruction changes it
• 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
• Computer-based test advantages can be used to manage the effects of test anxiety as a construct-irrelevant factor [+W] — an instructional effect differs with it
• Dental students in a revised low-stress TBL format reported lower stress and frustration than medical students in the original format, in a non-randomised comparison of two cohorts [+W] — instruction changes it
• Early studies of the relationship between anxiety and achievement produced mixed and confusing results [~W] — learners who differ on it differ in outcomes
• EFL students report that games reduce speaking stress and improve speaking skills [+M] — instruction changes it
• Emotion regulation partially mediates the negative association between AI anxiety and academic motivation [+M] — learners who differ on it differ in outcomes
• Extensive reading reduced Japanese EFL learners' anxiety about foreign-language reading (r = .34) [+M] — instruction changes it
• Guiding analytics for learners can improve mathematics achievement, retention, learning strategies, and reduce math anxiety [+M] — instruction changes it
• Immediate task repetition benefits oral performance, including fluency, self-correction and reduced anxiety [+W] — instruction changes it
• Incidental vocabulary learning anxiety positively predicts L2 incidental vocabulary acquisition through reading [+M] — learners who differ on it differ in outcomes
• Indirect paths through AI learning self-efficacy and perceived usefulness account for about 69.5% of the total association between school support and AI learning anxiety [+M] — instruction changes it
• L2 anxiety is negatively correlated with L2 achievement, with a mean correlation of r = −.36 across 216 effect sizes [+M] — learners who differ on it differ in outcomes
• L2 proficiency, anxiety, and mastery of strategies jointly explain 18% of variance in incidental vocabulary acquisition [+W] — learners who differ on it differ in outcomes
• Lack of practice ranks fourth among speaking anxiety sources, which the authors suggest may be due to teacher-centered Iranian classrooms [+W] — instruction changes it
• Lack of study space and financial issues are associated with higher test anxiety in online exams, and reliable technology with lower [+M] — instruction changes it
• Learners valued safe, repeatable, nonjudgmental AI practice environments that reduced anxiety and bridged the theory-practice gap [+M] — instruction changes it
• Learning GenAI anxiety moderates the intention–procrastination association, which is negative and significant at high anxiety but absent at low anxiety [~M] — an instructional effect differs with it
• Learning strategies relate to test anxiety: persistence positively, environmental structuring negatively, time management not significantly, and women report higher test anxiety [~M] — learners who differ on it differ in outcomes
• Math anxiety has a small-to-moderate negative correlation with math achievement (r = −.28) in a meta-analysis of correlations from 223 studies [+M] — learners who differ on it differ in outcomes
• Mindfulness practices are linked to academic improvements including reduced test anxiety, focus, and reading competence [+W] — instruction changes it
• Most middle and high school students surveyed after classroom retrieval practice programs viewed them positively and said frequent retrieval practice helped them feel less nervous about exams [+W] — instruction changes it
• Students taking online and proctored online exams report higher test anxiety than students taking on-campus exams and non-proctored online exams [+M] — instruction changes it
• Oral assessments can induce more anxiety than written ones, yet many students still prefer them [+W] — instruction changes it
• Perceived school support, AI learning self-efficacy, and AI perceived usefulness each show negative paths to AI learning anxiety [+M] — instruction changes it
• Poorly designed or misused assessment practices can harm students through increased anxiety, lost learning time, and deficit-based messages [+W] — instruction changes it
• Pre-actional metacognitive, motivational, and positive-emotional components are positively associated with regulatory behavior in the actional phase, while anger and anxiety are negatively associated [+W] — learners who differ on it differ in outcomes
• Retrieval practice benefits learners regardless of trait anxiety level, but higher trait or induced anxiety is associated with smaller testing effects [~M] — an instructional effect differs with it
• Self-efficacy does not significantly moderate the relationship between AI anxiety and career decisions [-W] — learners who differ on it differ in outcomes
• Self-reported intrinsic cognitive load correlates positively with anxiety and challenge and negatively with probability of success in an algebra task [+M] — learners who differ on it differ in outcomes
• Social science students report higher job finding anxiety and AI anxiety than natural and applied field students, with small effect sizes [+W] — instruction changes it
• Sociocultural factors, learner motivation, self-efficacy and perceptions of the classroom environment influence FLCA learning outcomes and oral performance [+W] — learners who differ on it differ in outcomes
• Speaking practice opportunities and teacher support reduced classroom anxiety and increased self-efficacy for one learner [+W] — instruction changes it
• Structural paths are stable when continued-use intention and current use are modeled as separate outcomes, and AI anxiety is not directly related to continued use [-W] — learners who differ on it differ in outcomes
• Student anxiety under remote proctoring is mixed: some students report reduced anxiety from home comfort and no travel, while others face anxiety from connectivity issues and fear of wrongful accusations [~M] — instruction changes it
• Student-centered writing environments are argued to lower anxiety, increase confidence, and provide natural language contexts [+W] — instruction changes it
• Students report that tests dominate assessment, carry too much weight, and create stress, while extended performance-based assessments offer better opportunities to learn and showcase learning [+M] — instruction changes it
• Students reported significantly lower anxiety during AIvaluate-mediated viva sessions than during equivalent face-to-face viva assessments [+M] — instruction changes it
• Tablet-PC-supported peer-assisted learning reduced EFL learners' anxiety and promoted motivation and confidence in collaborative reading [+M] — instruction changes it
• Teaching a ChatGPT agent does not significantly change students' test anxiety [-W] — instruction changes it
• Technology anxiety is a positive predictor of AI literacy once other determinants are controlled, despite a negative zero-order correlation [~W] — learners who differ on it differ in outcomes
• Test anxiety is negatively associated with exam performance, and no moderation by proctoring was detected [+M] — learners who differ on it differ in outcomes
• The anxiety–achievement correlation varies across educational levels, target languages, and anxiety types [~M] — learners who differ on it differ in outcomes
• The anxiety–achievement correlation varies across types of language achievement measures [~M] — an instructional effect differs with it
• The authors claim that making teachers' own mistakes, struggles, and fears central to the curriculum lessens students' fear of making mistakes in a new endeavor [+W] — instruction changes it
• The indirect association between AI anxiety and academic motivation through emotion regulation is significant for both male and female students but stronger among male students [+M] — learners who differ on it differ in outcomes
• The mindfulness interventions did not significantly change anticipatory or executive state-MA relative to cognitive support alone [-M] — instruction changes it
• The serial indirect effect of AI literacy on continued use is weaker at higher levels of AI anxiety (moderated mediation) [~M] — an instructional effect differs with it
• Trusting teacher-evaluator relationships lessened evaluation anxiety and encouraged teachers to welcome and use REACH feedback [+M] — instruction changes it
• Two-part collaborative assessment, individual then group answering, transforms summative testing into a learning experience and may reduce test anxiety [~M] — instruction changes it
• Using the ITS reduced executive state-math anxiety and improved algebra learning across conditions [+M] — instruction changes it
• Attentional shifting, but not focusing, significantly moderates the effect of speech anxiety on judged quality [~M] — learners who differ on it differ in outcomes
• Early identification and treatment of math anxiety is important because early anxieties can snowball into avoidance of mathematics [+W] — learners who differ on it differ in outcomes
• End-of-year decoding was not predicted by any mid-year anxiety subscale [-M] — learners who differ on it differ in outcomes
• High achievers show the lowest math anxiety and MLD students the highest across achievement groups [+M] — learners who differ on it differ in outcomes
• Intrinsic math motivation is positively correlated with arithmetic skills and negatively correlated with explicitly measured math anxiety in sixth graders [+M] — learners who differ on it differ in outcomes
• Math anxiety correlates strongly and negatively with mathematics achievement among third graders [+M] — learners who differ on it differ in outcomes
• Math anxiety differs significantly across achievement groups, with MLD students not differing from low achievers [~M] — learners who differ on it differ in outcomes
• Math anxiety has both a direct negative effect and a WM-mediated indirect effect on children's verbal Math Reasoning, but only an indirect effect on Numerical Operations [~M] — learners who differ on it differ in outcomes
• Math anxiety partially mediates the relation between attention problems and both math outcomes [+M] — learners who differ on it differ in outcomes
• Math anxiety predicts mathematics achievement in linear regression (R2=.357) [+M] — learners who differ on it differ in outcomes
• Math anxiety's direct effect on math performance emerges in grade 4 but not grade 3, suggesting a developmental trajectory [~M] — learners who differ on it differ in outcomes
• Mathematical metacognition statistically mediates the association between mathematics anxiety and children's word problem solving after controlling for IQ [+M] — learners who differ on it differ in outcomes
• Modeling mathematics anxiety as the mediator of metacognition's effect on word problem solving yields no mediating effect [-M] — learners who differ on it differ in outcomes
• Public speaking anxiety is associated with poorer judged speech performance only among individuals low in attentional control [~M] — learners who differ on it differ in outcomes
• Self-image and learning mathematics anxiety, but not other metacognition dimensions or evaluation anxiety, predict mathematics achievement after controlling for IQ [+M] — learners who differ on it differ in outcomes
• Speech length mediates the anxiety-to-judged-quality link only for individuals low in attentional control [~M] — learners who differ on it differ in outcomes
• The direct path from mathematics anxiety to word problem solving is non-significant once metacognition is modeled [-M] — learners who differ on it differ in outcomes
• Total anxiety predicts later fluency for girls only, via a significant gender-by-anxiety interaction [~M] — learners who differ on it differ in outcomes
• Working memory fully mediates the effects of math anxiety and attention on children's non-verbal Numerical Operations [+M] — learners who differ on it differ in outcomes
• Working memory mediates the indirect effects of anxiety on math performance [+M] — learners who differ on it differ in outcomes
• Both SEMA subscales — numerical and situational anxiety — are equally negatively related to math achievement [+M] — learners who differ on it differ in outcomes
• Math anxiety arises and relates to math performance already in early primary school (Grades 2-3) [+W] — learners who differ on it differ in outcomes
• Math anxiety in Grade 2 shows an indirect effect on subsequent math performance in Grade 3 [+W] — learners who differ on it differ in outcomes
• Math anxiety is not significantly correlated with reading achievement in young children [-W] — learners who differ on it differ in outcomes
• Math anxiety is not significantly correlated with trait anxiety in young children [-W] — learners who differ on it differ in outcomes
• Math anxiety is significantly negatively correlated with math achievement in second and third graders, independent of trait anxiety and IQ [+W] — learners who differ on it differ in outcomes
• Math anxiety relates more strongly to Mathematical Reasoning than to Numerical Operations in second and third graders [~M] — learners who differ on it differ in outcomes
• Observed anxiety expressions during challenging tasks predict lower teacher-rated work-related and self-regulation skills [+W] — learners who differ on it differ in outcomes
• The association between math anxiety and math performance is significant in Grade 3 after accounting for general anxiety [+W] — learners who differ on it differ in outcomes
• The prevalent direction of the Grade 3 anxiety-performance link runs from math anxiety to performance, not the reverse [+W] — learners who differ on it differ in outcomes
• Self-efficacy positively and test anxiety negatively predicted classroom grades, while test anxiety positively predicted cognitive strategy use but negatively predicted regulatory strategy use [~M] — learners who differ on it differ in outcomes
• Students whose anxiety, surprise or enjoyment decreased over the course achieved more positive learning-outcome change than students with stable trajectories in those emotions. [+M] — learners who differ on it differ in outcomes
• Better metacognitive monitoring and control at the start of the year are correlated with lower boredom and anxiety about English at year's end [+W] — instruction changes it
• Boredom is the only achievement emotion predicting English vocabulary growth in primary school L2 learners; enjoyment and anxiety show no significant effect [-W] — learners who differ on it differ in outcomes
• English vocabulary correlates with confidence and metacognitive control at both time points but with monitoring accuracy only at T0, and with only the negative emotions boredom and anxiety at T1 [~W] — learners who differ on it differ in outcomes
• Over six months, English vocabulary, relative monitoring accuracy and metacognitive control improved, and boredom increased, while enjoyment and anxiety showed no significant change [-W] — instruction changes it
• Brilliance-oriented contexts heighten anxiety among intellectually stigmatized students [+M] — instruction changes it
• Mathematics anxiety does not moderate the effect of mathematics test practice on subsequent performance [-M] — an instructional effect differs with it
• Arousal showed a positive relationship with final high school mathematics grade when controlling for the other test anxiety facets [+M] — learners who differ on it differ in outcomes
• The tribute reports that self-efficacy consistently outperformed anxiety as a predictor of academic performance [+W] — learners who differ on it differ in outcomes
• Worry did not explain unique variance in students' test performance or mathematics grade, challenging the worry-based explanation of the test anxiety-performance link [-M] — learners who differ on it differ in outcomes
• SRL training programmes improve SRL strategies, motivation, academic achievement and reduce anxiety across age groups [+M] — instruction changes it
• Mathematics anxiety shows mixed relations to algebraic thinking across studies [~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
• In non-randomised controlled trials, teacher programs show moderate effects on stress (g=0.50) and small effects on anxiety and wellbeing (g=0.38 each) at post-intervention, with non-significant effects at short-term follow-up [~M] — instruction changes it
• Anxiety, enjoyment and boredom are three critical emotions in language learning that affect outcomes and engagement [+M] — learners who differ on it differ in outcomes
• Trait test anxiety predicts performance-based goals, and in a path model state anxiety and worry, but not emotionality, link performance-avoidance goals to lower exam performance [+M] — learners who differ on it differ in outcomes
• Gender is not a significant moderator of the math anxiety–performance relationship [+M] — learners who differ on it differ in outcomes
• Math anxiety correlates negatively with math performance at a pooled r = −.30 across 169 independent samples [+M] — learners who differ on it differ in outcomes
• Math anxiety correlates weakly with working memory (r = −.18), and WM's mediating role between anxiety and math achievement is negligible [~M] — learners who differ on it differ in outcomes
• Test anxiety correlates negatively with math achievement at a pooled r = −.23, weaker than math anxiety's correlation [+M] — learners who differ on it differ in outcomes
• Test anxiety correlates weakly with working memory (r = −.18), with a much stronger correlation for spatial than verbal material [~W] — learners who differ on it differ in outcomes
• Type of math task moderates the math anxiety–performance link: basic numerical knowledge shows the weakest correlation (r = −.22) and applied math among the strongest (r = −.31) [~M] — learners who differ on it differ in outcomes
• Background variables (need for cognition, test anxiety, mindset, puzzle experience) showed no consistent moderating patterns across the three experiments [~W] — an instructional effect differs with it
• Idiographic diary analyses show daily anxiety relates differently to learning variables in different students, with some students profiting motivationally from anxiety [~M] — learners who differ on it differ in outcomes
• Higher mathematics test anxiety predicts earlier practice-session exit, mediating lower learning gains in an adaptive math program [+M] — learners who differ on it differ in outcomes
• Misconception-based adaptive feedback reduced mathematics anxiety and supported motivation without cognitive outcome differences [+M] — instruction changes it
• Math achievement and math anxiety show reciprocal longitudinal relationships in primary school children: achievement at T1 predicts anxiety at T2 and anxiety at T1 predicts achievement at T2 [+M] — learners who differ on it differ in outcomes
• Mathematics anxiety remains low and does not change significantly during the first three school years [-M] — instruction changes it
• Mathematics anxiety for learning is indirectly related to seventh-grade mathematics achievement through reduced in-class attention [+W] — learners who differ on it differ in outcomes
• Academic anxiety mediates between one-tenth and one-fifth of the social class test gap across social class indicators, anxiety forms, and school domains [+M] — learners who differ on it differ in outcomes
• Whether anxiety explains the income inequality × social class interaction remains inconclusive [~M] — learners who differ on it differ in outcomes
• Academic buoyancy moderates the relation between anxiety and mathematics test performance, with performance highest when anxiety is low and buoyancy high [~M] — an instructional effect differs with it
• Anxiety negatively predicts subsequent mathematics test performance in primary school students when emotions are modeled simultaneously [+M] — learners who differ on it differ in outcomes
• Girls and FARM-enrolled students report higher mathematics anxiety at year's end, controlling for fall anxiety [+M] — instruction changes it
• Retention improves math self-efficacy and reduces math anxiety in the first year, with effects maintained (anxiety showing a positive sleeper effect at Lag 4) [+W] — instruction changes it
• Teachers' beginning-of-year anxiety for teaching mathematics, science, or ELA shows no direct significant association with students' end-of-year anxiety in that content area [-M] — instruction changes it
• Teachers' mathematics anxiety is positively associated with the mathematics anxiety of their low-SES students via a teacher-anxiety-by-FARM interaction [+M] — instruction changes it
• Teachers' science anxiety is positively associated with the science anxiety of their low-SES students via a teacher-anxiety-by-FARM interaction [+M] — instruction changes it
• When enjoyment, boredom, and anxiety are modeled simultaneously, anxiety exerts the strongest predictive power for mathematics achievement [+M] — learners who differ on it differ in outcomes

The evidence by question

• Do learners with higher anxiety about a subject or a test achieve less in it? Math anxiety has a small-to-moderate negative correlation with math achievement (r = −.28) in a meta-analysis of correlations from 223 studies [+M]
• Does the format and setting of an assessment change how anxious learners are about it? Students taking online and proctored online exams report higher test anxiety than students taking on-campus exams and non-proctored online exams [+M]
• Does private, low-pressure and repeated practice lower learners' anxiety about using a second language? AI reduces foreign language anxiety and strengthens self-efficacy and self-confidence, supporting the Affective Foundation Pathway [+M]

Examples

• AEQ-ES adapted to English as L2 for primary school children

Related Learner Variables

• Affect Regulation
• Self-Efficacy
• Working Memory

Page metadata

Field	Value
type	learner-variable
id	academic-anxiety
title	Academic Anxiety
description	Anxiety tied to learning and assessment in a subject, such as mathematics anxiety, test anxiety and foreign-language anxiety.
status	review
canonical	True
generated	by claude/unspecified on 2026-10-15
trust	unverified — no one has confirmed this page's content

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