ABSTRACT
Tests of the neighbourhood effect hypothesis at general elections in the United Kingdom have largely been precluded by the absence of suitable xivsmall-scale data. This article uses specially generated bespoke neighbourhoods for each respondent to the 1997 British Election Study, which give contextual data on areas around each respondent’s home, ranging from the nearest 500 to the nearest 10,000. Analyses of voting by socio-economic class and area characteristics show that within each class, the lower the status of the local neighbourhood (at all scales) the lower the probability that they voted Conservative in 1997. This relationship holds when a range of other variables are held constant, though with different forms in urban and rural areas. Investigations of the various scale effects show that in making their voting decisions respondents were sensitive to conditions in both the wider environment (represented by their constituency) and their local context (the bespoke neighbourhood for the 5,000 residents nearest to their homes).
