Most of the analysis before the 2016 referendum on the UK’s continued membership of the European Union based on opinion polling data focused on which groups were more likely to support each of the two options, with less attention to the geography of that support – although some regions, especially London and Scotland, were expected to provide substantial support for Remain. Using a recently developed procedure for detailed exploration of large tables derived from survey data, this paper presents the result of a prediction of the outcome across local authorities in Great Britain using just two variables – age and qualifications. In relative terms, that prediction was reasonably accurate – although, reflecting the polls’ overestimate of support for Remain it underestimated the number of places where Leave gained a majority, as was also the case within local authorities where data were published by ward. The model’s predictive value was enhanced by post hoc incorporation of information on turnout and the number of registered electors, and taking these into account there was little evidence of substantial, additional regional variation in levels of support for Leave. Overall, regions were relatively unimportant as influences on the referendum outcome once the characteristics of the people living there were taken into account.
- local authorities
- predictive modelling