We’re excited to kick off our inaugural blog post on statistics education research with insights from Paddy and his colleagues! They’ll explore students’ perspectives on strategies to reduce statistics anxiety and how to make learning statistics more enjoyable. 

Patrick (Paddy), a Senior Lecturer (Education) in the School of Psychology, Queen’s University Belfast, also leads the Research and Scholarship Special Interest Group (SIG) at RoSE. Learn more about his work and contributions here.

The British Psychological Society (BPS) emphasizes the importance of practical classes in research methods and statistics for psychology students, as they are crucial for completing final-year theses and preparing for research careers (BPS, 2019; University of California Press, 2023). However, these classes often cause significant anxiety, necessitating effective interventions (Allen et al., 2016; Murtonen et al., 2008; O’Connor & Lee, 2023). A previous mixed-methods study (Thompson, Wylie & Hanna, 2016) used focus groups to identify interventions to reduce anxiety, yielding mixed results. However, the study noted the potential benefits of using real-life examples in teaching quantitative methods to reduce anxiety. The use of focus groups may have been problematic due to groupthink, unbalanced dynamics, or hesitation to discuss the topic (Janis, 1972; Krueger, 2014). Additionally, involving both undergraduate and postgraduate students may have skewed the results, as postgraduate students typically have more experience with statistics while less experienced students may be more prone to anxiety (Murtonen & Lehtinen, 2003). This study aimed to find an alternative method to generate anxiety-reducing ideas for statistics classes. In Study 1, first-year students anonymously suggested interventions via Qualtrics. In Study 2, a new cohort of first-year students rated the effectiveness of these interventions. 

Eighty-five Level 1 undergraduate students (61 females, 1 non-binary; Mage = 20.20 years, SD = 4.11 years) from Queen’s University Belfast participated in Study 1 during the 2022-2023 academic year. Participants generated ideas to reduce anxiety in research methods and statistics lab classes via an open-ended Qualtrics survey. Responses were coded inductively by a primary rater (Chandra & Shang, 2019), who generated an initial set of codes. The rater refined these codes through multiple iterations to minimize the number of distinct codes while preserving their meaningfulness. Inter-rater reliability was calculated, showing 83% agreement with Rater 1 and 86% with Rater 2. Disagreements were resolved through discussion, resulting in a total of 164 codes. Table 1 presents the most commonly reported codes, which accounted for 47% of all suggestions (77 codes in total). 

Eighty-five Level 1 undergraduate students (93 females, 2 non-binary) from the same undergraduate program participated in Study 2. Most participants (90.8%) were aged between 18-21. Participants completed a Qualtrics survey where they rated the effectiveness of the top eight interventions identified in Study 1 on a 0-100 scale. Figure 1 displays the mean ratings, highlighting three interventions rated higher than others: Anonymity, Worked Examples, and Breaking Things Down. Kolmogorov-Smirnov tests indicated non-normality for all variables, so Wilcoxon signed-ranks tests were used to assess differences in ratings. Ratings for Anonymity, were significantly higher, compared to the other interventions (all p values < .05) except Break Things Down (p = .139). Ratings for Worked Examples were significantly higher, compared to the other interventions (all p values < .05), but did not significantly differ from Slower Teaching (p = .061), One-on-one Support (p = .160), and Break Things Down (p = .724). Ratings for Break Things Down were significantly higher compared to other interventions (all p values < .05), except for the previously mentioned comparisons.

The current study aimed to identify and evaluate interventions to reduce anxiety in practical statistics classes for psychology students. High ratings for Anonymity indicate that students feel more comfortable engaging with material without fear of judgment. Implementing anonymous response systems in classes could reduce anxiety and encourage participation. The high rating for Worked Examples shows that concrete examples and step-by-step problem-solving may help make abstract concepts more relatable and less intimidating. This aligns with Thompson et al. (2016), who also found real-life examples helpful in reducing anxiety. Breaking Down the Material received consistently high ratings, indicating that segmenting complex information into manageable parts significantly could potentially reduce anxiety. Perhaps instructors should introduce concepts incrementally, ensuring students grasp foundational elements before progressing.

These findings have significant implications for teaching practices, However, there are limitations to this research. Different cohorts generated and rated the codes, so results should be interpreted with caution. Tailored interventions based on specific strengths and weaknesses of different cohorts may be necessary. The sample was limited to first-year psychology students from a single institution, affecting the generalizability of the results. Future research should explore these interventions across different cohorts and university courses to validate their effectiveness and identify variations in their impact. Additionally, investigating the long-term effects of these interventions on students' performance and attitudes towards statistics would provide valuable insights into their overall efficacy.

In conclusion, this study highlights effective strategies for reducing anxiety in statistics education for psychology students. The findings suggest testable interventions for reducing anxiety among first-year students. Incorporating these strategies into teaching practices may help create a more supportive learning environment, enhancing student engagement and performance.