This study analyzes the results of the UK's 2016 referendumBrexitby means of territorial data regarding votes and characteristics of the British Local Government Districts. The main variable that the present paper intends to explain is the share of votes for Leave in each LGD. The explanatory variables are represented by a set of political, demographic and socio-economic variables. The purpose is that of checking the impact on the referendum outcome of aspects belonging to various domains: in particular, political partisanship, the economic setting, the occupations, the educational level, and foreign immigration. Since the main dependent variable was the share of votes for Leave, i.e. a fractional response variable assuming continuous values but within the 0, 1 range, the statistical analysis has used a fractional logit regression model. The results show that the impact of political partisanship was significant, provided there was no internal split, and that Leave was territorially associated with working class occupations and intermediate educational levels, but neither with the presence of those in the most unfavourable socio-economic conditions, nor with the size of the immigrant population.