Personality and wellbeing: Using data mining to refine digital behaviour change interventions

We study the relationship between personality and wellbeing using questionnaire data from 14,397 people who participated in a digitally delivered Do Something Different (DSD) behaviour change intervention. Our dataset consists of answers to a pre-intervention questionnaire comprising sections addressing behaviour, wellbeing, anxiety and depression, and habits. For 2,863 of these participants, corresponding post-intervention responses are also available. DSD interventions target various health and wellbeing issues such as stress reduction, weight loss, smoking cessation and diabetes self-management. They are based on the psychological theory of behavioural flexibility, developed in a series of books and papers by Fletcher, Pine and others. This paper describes how we applied regression models to data from DSD interventionsto understand better the role of behaviours and personality in wellbeing,and hence refine the theory of behavioural flexibility. We describe our dataset and present a simple model of how behaviours are related to wellbeing; discover that the 30 behaviours listed in the questionnaire can be classified into 9 “inhibitory” and 21 “facilitatory” behaviours; and identify regressions models that predict wellbeing from behaviours more accurately than the existing behavioural flexibility model.