The association of self-efficacy, anxiety sensitivity, and perfectionism with statistics and math anxiety

Abstract

Statistics and math anxiety are pervasive problems for post-secondary students. We hypothesized that self-efficacy would be negatively related to math/statistics anxiety, and that anxiety sensitivity and perfectionism would be positively related to math/statistics anxiety, even when controlling for gender, university program, and education level. Method: Graduate and undergraduate students (N = 447, after exclusions) completed an online self-report questionnaire, including an abbreviated version of the Statistics Anxiety Rating Scale (STARS), math anxiety, self-efficacy, anxiety sensitivity, perfectionism, and demographics. Results: Exploratory factor analysis supported a six-factor structure for statistics anxiety. Self-efficacy was negatively associated with math/statistics anxiety, whereas anxiety sensitivity and perfectionism were positively associated with math/statistics anxiety. Relationships ranged from small-to-moderate, and most relationships persisted after adding covariates. Discussion: Our study suggests the feasibility of a short-form version of the STARS. Moreover, it provides important information on how personality is associated with domain-specific anxiety that can impede statistics education.Statistics education has unique challenges, and many students experience a great deal of anxiety when faced with statistical concepts and mathematical calculations. Though a great deal of research has explored personality risk factors for generalized anxiety in clinical domains, comparatively few studies have examined the relationship between individual differences and domain-specific anxiety in statistics and mathematics classrooms. An interesting finding of the present research was that concern over mistakes and doubts about action (facets of perfectionism) predicted specific classroom-related fears (e.g., test anxiety), yet did not strongly relate to negative attitudes (e.g., attitudes about the value of statistics overall). Our approach applies common rigorous methods in personality science (e.g., basic psychometric soundness, quantitative analysis that accounts for familywise error rates, pre-registration, and open data) to a core topic of importance to educational research (i.e., classroom anxiety). We believe this cross-disciplinary approach is an important strength. Further, we believe that our research can open a dialogue about important personality variables that can impede learning in the classroom. In this way, we hope that our research can help teachers identify students in need of support by revealing personality variables that tend to co-occur with statistics anxiety. A short-form statistics anxiety scale was found to have 6 factors. Perfectionism and anxiety sensitivity positively correlated with math and statistics anxiety. Self-efficacy negatively correlated with math and statistics anxiety. Results held when controlling for gender, university program, and education level.