Statistics anxiety is widely recognised as a barrier to student learning, and the Statistics Anxiety Rating Scale (STARS) is its most widely used self-report measure. However, it remains unclear whether the STARS captures a construct distinct from mathematics anxiety or reflects a jangle fallacy. Using a large international sample of undergraduate students (N = 6,885) from 83 universities across 33 countries, we examined the empirical distinctiveness of the STARS relative to the Revised Mathematics Anxiety Rating Scale (R-MARS). Across four criteria, the evidence indicated substantial overlap. First, correlations between the STARS and R-MARS were consistently strong. Second, exploratory and confirmatory factor analyses showed that items clustered primarily by type of educational experience (e.g., tests, help-seeking, interpretation) rather than by domain, with limited and unstable domain-specific factors. Third, both scales demonstrated statistically equivalent associations with 11 theoretically related anxiety and education variables. Fourth, the R-MARS explained negligible incremental variance beyond the STARS across these outcomes. Together, the findings suggest that the STARS and R-MARS largely measure the same underlying construct. Their continued separate use risks redundancy, conceptual fragmentation, and statistical artefacts, underscoring the need for clearer construct definition and more precise measurement tools. 

Background: Statistics-related anxiety is a common barrier to learning in higher education, particularly for psychology and other non-specialist students. High prevalence rates are concerning because statistical literacy is integral to academic success, employability, and informed citizenship. Clarifying the nature of statistics anxiety is central to understanding it, yet it remains unclear whether statistics anxiety is genuinely distinct from mathematics anxiety or simply another manifestation of the same construct.

Aims: We examined the degree of overlap (jangle fallacy) between the Statistics Anxiety Rating Scale (STARS) and the Revised Mathematics Anxiety Rating Scale (R-MARS).

Samples: Two cohorts of UK undergraduate psychology students participated: Study 1 (n = 489) and Study 2 (n = 245).

Methods: Each study combined a cross-sectional survey with a between-participants experiment. Surveys included the STARS, R-MARS, parallel forms to control for scale content, and measures of trait and state anxiety. In the experiment, participants were randomly assigned to complete either a statistics or mathematics multiple-choice test, with state anxiety measured before and after.

Results: Results pointed to overlap across all analyses and samples. Scales were highly correlated, items did not separate into distinct domains, individuals rarely reported one without the other, and state anxiety increased similarly across test types.

Conclusions: Findings show that statistics anxiety, as measured by the STARS, is not distinct from mathematics anxiety, as measured by the R-MARS. Recognising statistics anxiety as part of the broader mathematics anxiety construct can streamline measurement, clarify theory, and support more effective interventions for reducing quantitative anxieties in education. 

Self-reported grades are widely used in statistics education research, yet their validity as proxies for official grades is rarely evaluated. This study examined the validity of score interpretations based on self-reported university-level statistics grades, using 649 observations from nine courses across four universities in the UK and Türkiye. Evidence from correlations, intraclass correlations, and equivalence tests supported the interpretation of self-reports as reasonable estimates of official grades. Discrepancies were typically small and did not meaningfully bias associations with nine psychological variables, supporting their use in group-level research. Demographic and psychological predictors explained little variance in discrepancies, which appeared more strongly linked to course-level factors. These findings provide empirical support for the research use of self-reported statistics grades, while highlighting the importance of contextual influences on accuracy. 

Background: The number of autistic students entering higher education (HE) has increased, yet many continue to face systemic barriers that can hinder their academic success. Despite their unique cognitive strengths, such as hyperfocus, attention to detail, and strong analytical skills, many autistic students report challenges with academic learning experiences. This study aimed to develop and validate the Academic Learning Experiences Questionnaire (ALEQ), a tool designed to assess specific learning experiences and inform autism-inclusive educational practices.

Methods: We co-created the ALEQ with autistic and non-autistic students to assess learning experiences across five academic contexts: small and large group teaching, self-directed study, examinations, and coursework. A total of 829 university students (formally-diagnosed autistic: n = 106; self-diagnosed autistic: n = 112; non-autistic: n = 611) completed an online survey comprising the ALEQ and an autism screening measure (SRS-2). To establish the ALEQ’s psychometric properties, we conducted exploratory and confirmatory factor analyses and tested for measurement invariance between the autistic and non-autistic groups.

Results: The ALEQ produced seven theoretically relevant factors with good local and global fit: Microtransitions’, Social Anxiety, Sensory Reactivity, Planning and Prioritising, Monotropic Focus, Group Work and Global Comprehension). Configural and metric invariance were supported for this factor solution, demonstrating equivalence of the factor loadings across groups and warranting examination of group differences in ALEQ scores. Using this final, 34-item version of the ALEQ, autistic students reported significantly more challenges than non-autistic students across all five subscales, with the greatest disparities in Sensory Reactivity and Microtransitions.

Conclusion: The ALEQ provides a structured way to understand the academic challenges that autistic students face in different learning contexts. By identifying key learning experiences, it offers both a practical tool for educators and a measurement instrument for researchers that can identify adjustment needs and, ultimately, enhance accessibility and inclusion in HE.

Despite a negligible aggregate effect size, there is a prevailing narrative that statistics anxiety has a detrimental effect on statistics literacy. Without a cohesive theoretical framework, it is difficult to investigate variables that might explain why the observed relationship between statistics anxiety and statistic literacy is, on average, so small. This systematic review lays a foundation for such a framework by identifying, comparing, and evaluating relevant theories. We searched four databases and, following PRISMA guidelines, identified 56 relevant scientific reports of which 33 minimally evoked a theoretical explanation and were retained for review. The 11 theories mentioned were categorised into three types: Non-linear relationships (there is an optimal degree of statistics anxiety for statistical literacy test performance), motivational (the relationship is mediated and/or moderated by perceived control and/or value of statistics, which in turn influences learning behaviours), and cognitive-interference (statistics anxiety disrupts cognitive processes required for learning and/or recall). Only 18 studies tested the theories and almost all had severe methodological limitations, leaving a critical knowledge gap. An overlapping theoretical nomological network between statistics and mathematics anxieties emerged, suggesting that they could be the same construct. If so, the concept of statistics anxiety should be retired and insights from the more developed mathematics anxiety literature should be applied to statistics contexts just as they are to mathematics. We conclude that examining this potential overlap should be the field’s most urgent priority.