Spatial variations in economic attitudes and voting behaviour in Britain, 1983-92.

The objective of this thesis is to assess the role of geography in the construction
of economic attitudes and electoral behaviour in Britain during the 1980s and the
early 1990s. Aggregate and individual level data are used to separate two
specific time periods - the economic recovery of the 1980s and the 'new'
recession of the early 1990s. The new recession also coincided with the long campaign
leading to the 1992 General Election, when the Conservatives were
returned for a fourth successive term.
A two stage model of the relationship between social class, geography,
economic attitudes and party support is constructed. Initially the link between
geography and economic attitudes appears enigmatic. However, as the analysis
progresses a clearer picture emerges of the geographic basis of prospective and
retrospective, egocentric and sociotropic economic evaluations.
Analysis of Variance and Multiple Classification Analysis techniques reveal the
extent of the growing geographic divide in party support and certain economic
attitudes during the 1980s. A particularly crucial theme emerges with the
investigation of partisanship during inter-election periods. Groups that tend to
form the core of the Conservative vote in Election years, are identified as
reluctant Conservatives in non-election years.
Important contextual effects are perceived in the analysis of reported vote
intention, geography and economic attitudes in the run-up to the 1992 General
Election. As well as the orthodox personal economic expectations variable,
ascription of economic responsibility and economic approval for the
Government's programme are shown to be critical to levels of Government
support - and are spatially variable. Ordinary Least Squares and Logistic
regression analysis reveal the precise role of geography in economic attitudes
and party support. Here the 'devil is in the detail' as the interactions effects of
the dependent variables reveal that when an individual's economic evaluations
clashes with their geographic context, the contextual effect either dilutes - or
overcomes completely - the economic effect. The analysis of individual level
data represents an advance for electoral geography and for the study of
geographic milieux and local socialisation effects.