Predicting drop-out in Toastmasters

The use of learning analytics to build predictive models that identify students at risk of not finishing their studies is becoming more common in Higher Education. In professional education, which focuses on skills rather than knowledge-based education, the practice is still relatively unheard of. This thesis focuses on Toastmasters International (TMI); a global public speaking organisation of more than 260,000 Members that provides an educational programme. For the last few years, rates of retention have been around 50% per year with average Member tenure hovering around the three-year mark. The study utilises data from Pathways, the Learning Management System that TMI introduced in 2020, to explore whether predictive models can be applied in the area of professional education; providing data-driven insights into drop-out to support these volunteer-run clubs. The thesis uses a dataset from the UK population of Toastmasters to predict the drop-out rate of 8,122 Current Members. It takes a Mixed Methods approach to the research and comprises three consecutive studies. Study 1, an initial predictive model, a series of interviews in Study 2 to gain insight into the data behind the predictive model and Member experience, before finally bringing this knowledge together to create Study 3 and a final predictive model. Based on these findings the thesis concludes by suggesting some strategies for intervention that TMI should carry out as well as valuable feedback about the use of Pathways itself. By applying predictive modelling techniques usually seen in Higher Education, this thesis demonstrates that such techniques can be applied to a different sphere of professional education and group membership. As such, this study has wider applications than the initial scope of this specific study in Toastmasters.