Strategies for remediating the impact of math anxiety on high school math performance

The study addresses how math anxiety impairs high school students’ performance and evaluates two feasible, classroom-based interventions aimed at mitigating its effects. Math anxiety can hinder performance via intrusive thoughts that consume working memory and via maladaptive study habits that reduce effective learning and promote avoidance. The research compares an Emotion Regulation (ER) intervention teaching cognitive reappraisal strategies to reframe anxious responses during math, and a Study Skills (SS) intervention promoting evidence-based learning techniques (spaced practice and retrieval practice/self-testing). The primary questions are whether either intervention improves grades relative to pre-intervention performance and whether one is more effective at reducing the negative association between math anxiety and grades. Secondary questions examine the durability of effects across the semester and whether other forms of anxiety (trait or test anxiety) account for outcomes.

Prior work links math anxiety to decrements in performance through intrusive worries that tax working memory resources. Evidence-based study strategies—especially retrieval practice and spaced studying—enhance learning and test performance in lab and classroom contexts and can reduce test anxiety. Intensive math tutoring has been shown to remediate math anxiety and alter amygdala activity, though such approaches are costly and difficult to scale. Emotion-focused approaches, including therapy and cognitive reappraisal, can reduce negative affect and sometimes improve performance, but are often individualized and evaluated with lab tasks. Cognitive reappraisal decreases physiological arousal and negative emotion, improves control, and for highly math-anxious individuals can improve math task performance and modulate neural activity in arithmetic-related regions (e.g., intraparietal sulcus). Classroom-adjacent strategies like expressive writing before tests can reduce rumination and improve performance by facilitating reframing. This literature motivated testing, in real classrooms, whether a brief ER (reappraisal) or SS (retrieval practice and spacing) intervention can mitigate math-anxiety-related performance deficits.

Design and participants: Two high schools in distinct U.S. regions participated. School 1 (rural New England) contributed N=91 analyzable students (60% female; ages 13–18) from Algebra I/II and Geometry across six classes (two instructors). School 2 (mid-Atlantic) contributed N=156 analyzable students (53% female; ages 14–19) from 13 year-long classes (six instructors) in Algebra I, Algebra II, and Algebra II Honors. Combined dataset N=224. Procedures were approved by Dartmouth CPHS and school administrators; School 2 was opt-in with parental consent; School 1 used opt-out with no opt-outs. Assignment: Within each class, students were pseudo-randomly split in half and assigned to either the Emotion Regulation (ER) or Study Skills (SS) intervention. Assignment was counterbalanced on gender and prior performance (GPA/standardized test or previous math grades) to balance groups within classes. No separate no-contact control was included due to class splitting; first-semester grades served as within-subject controls. Interventions: At the start of the third quarter, each group received a 20–30 minute small-group, researcher-led discussion with a worksheet tailored to the assigned strategy. ER taught cognitive reappraisal via emotional distancing and stress-reframing (e.g., viewing arousal as helpful; seeing challenges vs obstacles), aiming to reduce negative affect and encourage approach. SS taught spaced studying and retrieval practice, emphasizing self-testing techniques (practice problems, quizzes, flashcards) and planning regular study sessions. Follow-up activities: Before major tests during Q3 (and reminders into Q4), students completed brief, targeted writing tasks. ER students wrote about anticipated thoughts and feelings and described reappraisal strategies to use during the test. SS students wrote about anticipated problem types and strategies to solve them (a content-focused planning/self-testing prompt). Students also completed brief recurring surveys in Q3 (less frequent in Q4) that assessed anxiety, understanding, confidence, and usage of assigned strategies; in School 2, these surveys included additional reminders of the assigned technique. Measures: Baseline surveys included standardized instruments: Test Anxiety Inventory (TAI), Math Anxiety Rating Scale (MARS), State-Trait Anxiety Inventory (STAI), Emotion Regulation Questionnaire (ERQ), and the Academic Anxiety Inventory (AAI) with subscales for math, tests, trait, science, and writing. Primary outcome measures were course grades collected from school records. For School 1, Q3 and Q4 quarter grades (composites including tests, quizzes, homework, etc.) were analyzed; prior performance included earlier same-course grades or previous course grades. For School 2, analyses focused on grades from classwork, homework, and quizzes during Q3–Q4 to isolate current-term performance; prior performance used earlier quarter grades. Analysis plan: Linear mixed models (LMMs) were used with grades as outcomes, fixed effects for intervention group and anxiety measures, and random effects for participant, course subject, teacher, and previous math performance. Preliminary models showed significant grade differences by subject (χ²(3)=37.17, p<.001) and teacher (χ²(6)=53.87, p<.001), but not by school (χ²(1)=0.30, p=.58), motivating inclusion of subject and teacher as random effects. Quadratic anxiety terms were tested in supplementary analyses and did not improve fit over linear models. Additional linear models compared pre/post differences using a Pre/Post Grade Difference score: Q3 minus the average of Q1 and Q2. Further LMMs tested durability of effects across Q3 vs Q4 and evaluated other anxiety sources (trait and test anxiety) and gender effects. All grade analyses controlled for previous performance.

Pre-intervention equivalence: In first semester (pre-intervention), higher math anxiety predicted lower grades, F(1,192)=55.69, p<.001. There were no group differences prior to intervention, F(1,192)=0.16, p=.69, and no anxiety-by-group interaction, F(1,192)=1.27, p=.26. Intervention-period grades (Q3–Q4 combined): Higher math anxiety predicted lower grades overall, χ²(1)=8.42, p=.004 (β=−3.87, t(160.89)=−2.90). No main effect of group, χ²(1)=0.003, p=.95. Critically, math anxiety interacted with intervention group, χ²(1)=6.73, p=.010. In the ER group, increasing anxiety was associated with lower grades; in the SS group, the negative anxiety–grade relationship was ameliorated (reported β=−3.17, t(135.59)=−2.59 for slope comparison). Among students highest in math anxiety (top quartile), SS yielded grades about 6.29 points higher (~half a letter grade) than ER during the intervention period. Within-student pre/post change (Q3 – mean(Q1,Q2)): Groups did not differ in pre-intervention average grades (ER=80.88, SS=81.40; t(220.6)=−0.29, p=.77). There was a significant anxiety-by-group interaction on Pre/Post Grade Difference, F(1,174)=4.72, p=.03. In ER, higher anxiety predicted larger grade decreases relative to prior performance; in SS, this negative relationship was attenuated (β=−2.19, t(174)=−2.17), with more anxious students more likely to maintain or improve relative to their own prior grades. Durability across Q3 vs Q4: Main effects included anxiety, χ²(1)=8.44, p=.004, and quarter, χ²(1)=24.64, p<.001 (overall grades decreased from Q3 to Q4). No main effect of group, χ²(1)=0.004, p=.95. The anxiety-by-group interaction persisted across both quarters, χ²(1)=6.74, p=.010; there was no anxiety-by-quarter interaction, χ²(1)=1.53, p=.22; no group-by-quarter interaction, χ²(1)=3.75, p=.053; and no three-way interaction, χ²(1)=0.03, p=.86. Thus, SS benefits for highly anxious students extended into Q4. Other anxieties and gender: Analyses substituting trait or test anxiety showed no significant interactions with group (all p>.05), and no significant two-way or three-way interactions with quarter (all p>.05), suggesting specificity to math anxiety. Gender showed a main effect on grades, χ²(1)=10.73, p=.001, with female students higher (M=80.8, SE=5.41) than male students (M=72.2, SE=5.47), but no interactions with anxiety or group (all p>.05).

Findings indicate that while both interventions were easy to implement, only the Study Skills (SS) intervention effectively mitigated the negative association between math anxiety and math performance among highly anxious students. SS promoted spaced practice and retrieval practice/self-testing, likely counteracting avoidance and strengthening knowledge, making it more resilient to anxiety during evaluations. The benefits were evident during the intervention quarter and sustained into the following quarter, implying potential durability with intermittent reminders. For students low in math anxiety, ER yielded relatively better grades than SS, possibly because cognitive reappraisal is resource-intensive and easier to apply when anxiety is low and working memory resources are more available. The discrepancy with lab-based findings that reappraisal benefits highly anxious individuals may reflect context differences: lab tasks are low-stakes and often rely on mastered skills, whereas real classrooms impose higher stakes and concurrent learning demands that reduce capacity for reappraisal among highly anxious students. The SS group’s pre-test writing (problem/strategy anticipation) may also have reduced arousal and supported retrieval, providing a “double dose” (better study techniques plus anxiety relief), and improved study may reciprocally reduce anxiety over time, shifting from a vicious to a virtuous cycle. Overall, the results suggest that empowering students—especially those high in math anxiety—to engage in effective study behaviors can ameliorate anxiety-related performance deficits in authentic classroom settings.

This study demonstrates that a brief, scalable Study Skills intervention emphasizing spaced studying and retrieval practice can reverse math-anxiety-related performance deficits and improve high school math grades, with effects persisting into the post-intervention quarter. Emotion Regulation via cognitive reappraisal benefited students low in math anxiety but did not offset anxiety-related declines among highly anxious students in this real-world context. The work contributes evidence that targeting study behaviors can reduce avoidance, bolster learning, and indirectly mitigate the impact of anxiety on performance. Future research should incorporate a no-contact/waitlist control group, test combined SS and anxiety-support interventions (e.g., reappraisal plus retrieval practice and structured writing), examine mechanisms (e.g., working memory demands, desensitization), and evaluate longer-term, longitudinal dynamics of the reciprocal relations between achievement, study behavior, and anxiety, potentially including moderators such as mindset and self-efficacy.

Key limitations include: absence of a no-contact control group within the same cohorts (necessitated by splitting each class into two active conditions), relying on first-semester grades as within-subject controls; variability across schools (course structures, grading compositions, timing/frequency of assessments) and minor implementation differences; potential attenuation of effects due to real-world heterogeneity; modest class-level sample sizes limiting power for additional conditions; and context-specificity to high school algebra/geometry classes that may affect generalizability. The study was not preregistered. Data sharing is restricted due to FERPA and privacy considerations.