The geography of educational voting: Understanding where individuals with similar qualifications vote differently across Britain

This study offers a novel account of the geographical variation in the association of educational attainment with electoral behaviour. It estimates multilevel random-coefficient models using survey data from the British Election Study Internet Panel and data on constituency characteristics to explore the extent to which, and reasons why, the voting behaviours of individuals with identical qualifications varied across different types of parliamentary constituencies in British general elections from 2015 to 2019. We find the geography of this educational cleavage has not been evenly distributed in recent years. While the individual-level association of educational attainment and vote choice indeed varied geographically at the 2015, 2017 and 2019 general elections, there is only limited evidence to suggest that similarly educated individuals voted differently in constituencies with low and high densities of persons with high levels of education. Rather, our findings suggest that constituency left behind-ness (broadly defined) is the most important factor in explaining the spatial heterogeneity of education-based voting in recent British general elections.