Gender and active travel: a qualitative data synthesis informed by machine learning
Abstract: Background: Innovative approaches are required to move beyond individual approaches to behaviour change and develop more appropriate insights for the complex challenge of increasing population levels of activity. Recent research has drawn on social practice theory to describe the recursive and relational character of active living but to date most evidence is limited to small-scale qualitative research studies. To ‘upscale’ insights from individual contexts, we pooled data from five qualitative studies and used machine learning software to explore gendered patterns in the context of active travel. Methods: We drew on 280 transcripts from five research projects conducted in the UK, including studies of a range of populations, travel modes and settings, to conduct unsupervised ‘topic modelling analysis’. Text analytics software, Leximancer, was used in the first phase of the analysis to produce inter-topic distance maps to illustrate inter-related ‘concepts’. The outputs from this first phase guided a second researcher-led interpretive analysis of text excerpts to infer meaning from the computer-generated outputs. Results: Guided by social practice theory, we identified ‘interrelated’ and ‘relating’ practices across the pooled datasets. For this study we particularly focused on respondents’ commutes, travelling to and from work, and on differentiated experiences by gender. Women largely described their commute as multifunctional journeys that included the school run or shopping, whereas men described relatively linear journeys from A to B but highlighted ‘relating’ practices resulting from or due to their choice of commute mode or journey such as showering or relaxing. Secondly, we identify a difference in discourses about practices across the included datasets. Women spoke more about ‘subjective’, internal feelings of safety (‘I feel unsafe’), whereas men spoke more about external conditions (‘it is a dangerous road’). Conclusion: This rare application of machine learning to qualitative social science research has helped to identify potentially important differences in co-occurrence of practices and discourses about practice between men’s and women’s accounts of travel across diverse contexts. These findings can inform future research and policy decisions for promoting travel-related social practices associated with increased physical activity that are appropriate across genders.