Mathematics anxiety (MA) is a global phenomenon which induces underperformance or failure in university courses that entail knowledge of high-level mathematics (Ashcraft & Krause, 2007; Chang & Beilock, 2016; Foley et al., 2017; Mammarella et al., 2019; OECD , PISA, 2014, 2016). Research shows that MA is potentially caused by students’ previous unpleasant learning experiences and their interpretations of them (Ramirez et al., 2018a; Szücs et al., 2019). Once acquired, MA causes disruption to attention and working memory, inhibits students’ ability to ask questions and seek help, and they end up avoiding classes and coursework (Ashcraft & Krause, 2007; Beilock & Carr, 2005; Carey et al., 2016; Mammarella et al., 2019; Ramirez et al., 2018a; Szűcs et al., 2019). There is a plethora of research on the potential causes and consequences of MA, but less is known about efficacy of interventions to protect students from it and how interventions work. This study aimed to investigate the impact of a small-scale intervention designed to tackle MA and understand, from students’ perspectives, the main reasons for the effectiveness of this intervention. This doctoral research and its proposed intervention are grounded in an overarching theoretical framework combining Social Cognitive Theory (SCT) (Bandura, 1986) with Symbolic Interactionism (SI) (Blumer, 1969), which is its unique feature. SCT highlights the central role of self-efficacy in learners’ choices, persistence, perseverance, and anxiety levels during the learning process. SI attributes key importance to meanings individuals assign to situations they encounter, such as learning experiences in mathematics.
The intervention was implemented in statistics courses for non-specialist undergraduate students in a higher education (HE) institution in the United Kingdom (UK), during the academic year commencing in 2018. In total, 176 students from three courses participated in the study.
A multi-mixed methods research (Onwuegbuzie & Hitchcock, 2019) was conducted to assess the intervention’s impact on reducing reported MA and to understand the main reasons for its effectiveness. Two research designs (RDs) with different student cohorts were used namely RD1 and RD2 i.e. a mixed methods sequential explanatory design (Creswell, 2003; Creswell et al., 2003; Ivankova et al., 2006) and an experimental design respectively. Methods of data collection involved measures of students’ reported MA using the AMAS instrument (Hopko et al., 2003) at the beginning and at the end of the intervention in RD1. AMAS was also used to measure MA reported by the experimental and control groups in RD2. Analysis of paired data from 53 participants in RD1 who received the intervention showed a significant reduction in the post-intervention reported MA (d= 0.175). Results from the RD2 with 123 participants indicated a statistically significant reduction in the reported MA for students in the experimental group relative to the control group (d= 0.559). According to interviews with 19 participants in RD1 all of them reported feeling included. Students generally reported favourable class perceptions and enhanced ability to focus attention and ask questions in class. Blended learning was favourably viewed by some students, who reported positive changes in their ascriptions of encounters with mathematical applications in statistics. Students generally believed that MA could be managed and acknowledged the practical value of statistics for future progression. Self-efficacy and self-concept were enhanced according to some students’ perceptions, and a few students reported benefitting from collaboration and enjoying it. Smaller class sizes, and interactive learning environments were recommended by some students. This research could be a valuable contribution to MA research, which commenced more than six decades ago and has continued to evolve globally. It offers an intervention that, on average, helps to reduce MA, and also provides principles underpinned by theory that could be adopted for teaching other quantitative courses. It puts forward an innovative overarching theoretical framework for future MA research. Its feedback model shows the potential of being developed into an exemplar that could address the international dissatisfaction with feedback in HE, particularly in mathematics education. Although the study was restricted to one HE institution, which limits the generalisability of results, the inclusion of participants according to gender and nationalities allowed for some diversity in perspectives.