Distinct Classes of Statistical Anxiety: Latent Profile and Network Psychometrics Analysis of University Students.

Purpose: Many university students will experience statistical anxiety. Consequentially, the relationship between such anxiety and learning performance has been of concern to various educational researchers. To date, however, there has been no consistent resolution to this problem. Because previous studies have mainly used the perspective of variant-centered analysis rather than taking into account individual differences, this study argues that the different classes of statistical anxiety among university students may be an important influencing factor. Participants and Methods: In this study, 1607 Chinese university students who had just completed a statistics course were assessed using the Statistical Anxiety Scale, Statistics Learning Self-Efficacy Scale, and Learning Engagement Scale, and an exploratory study was conducted to determine whether university students' statistical anxiety could be divided into different classes. Latent profile and network psychometrics analyses were then used to analyze the data. Results: (1) The latent profile analysis found that university students' statistical anxiety could be divided into three different latent classes: mild test anxiety, moderate text anxiety, and severe statistical anxiety. (2) The correlation analysis showed that the relationships among the three latent classes of statistical anxiety and learning performance were not entirely consistent, indicating that there was heterogeneity in the statistical anxiety of these university students. (3) Further network psychometrics analysis showed that the statistical anxiety network structure of the three latent classes has different core nodes that reflected the most important symptoms of statistical anxiety. Conclusion: There is heterogeneity in university students' statistical anxiety that can be divided into three latent classes. These core nodes in the statistical anxiety networks of the three latent classes were different, helping statistics instructors to better understand the nature of these latent classes, take different intervention measures for different latent classes of university students.