This research used recorded preferences to estimate the average preferences for government activity (left-right positions) of British voters between 1950 and 2005. It explained variations in these positions using economic data together with data on public spending and taxation. It also estimated voters average positions on Europe between 1961 and 2005. The estimated positions were then combined with data on party positions and partisanship to estimate the distance between the parties and the median voter. This evidence is then used to explain election outcomes between. There are three time series data sets: (1) The original set of policy preferences based on responses to survey questions asked between 1950 and 2005 by market research organisations (Gallup, MORI, ICM, NOP, YouGov) and social scientific studies (BES, BSA, Eurobarometer). (2) The estimated preferences for government activity (or left-right position) series 1950-2005, together with parallel estimates of preferences towards Europe and voters’ self-reported positions and the economic and government expenditure data that is used to explain variations in preferences for government activity. (3) The macro polity data base consisting of vote shares at each election, partisanship, preferences for government activity, party positions based on MRG estimates. This research examines the relationship between aggregate election outcomes and the party positions between 1945 and 2005. It will use all available public opinion data to estimate the left-right position of the average voter at each election. It will then use separate analyses of the contents of election manifestos to establish the parties' positions on this dimension at each election. It will then locate both parties and voters on the same scale in order to establish the 'distance' between the two. Finally, it will estimate the relationship between these distances and the parties' vote share at each election taking into account voters' long-term loyalties. In order to provide a fuller picture of the relationship between party activity and election outcomes this research will also assess the relationship between government activity (as reflected in the state of the economy and public spending) and public opinion. In particular, it will assess the general proposition that demands for government activity increase under Conservative governments and decrease under Labour governments. This part of the research will, therefore, provide a fuller understanding of the trends and cycles in public opinion that have been noted by many scholars, journalists and politicians.

The data is chiefly drawn from published sources: (e.g., Gallup Political Index) and social scientific studies (e.g., BES). The data sets are large and so it is not possible to list all sources now (though it will in the data sets and codebooks). No specific permissions were sought. The use polling data has never been challenged by market research organisations (they actively encourage political scientists to use data). Usage is covered by general permissions in case of social scientific studies. These are aggregate political opinions and behaviour. No individuals can possibly be identified.