A growing body of research examines how and why European political parties appeal to social groups, finding such appeals to be widespread in contemporary politics. Yet the interactive dynamics that shape group appeals remain underexplored. How do parties’ group appeal strategies respond to each other?
This project aims to exploit a hitherto untapped resource of rich texts, to build a novel dataset uniquely suited to exploring these topics: a cross-party corpus of British election addresses from 1892-1931. Starting in the late 19th century, candidates to the British House of Commons disseminated personal manifestos to their constituencies in every general election. These documents, referred to as “election addresses,” complemented the central party platform by introducing each candidate's personal qualities and specific priorities to their local electorate.
British election addresses were written for mass public consumption and disseminated widely; the candidate would give the speech in person at rallies and also distribute it in print form through local newspapers and pamphlets. Systematically archived by the National Liberal Club from 1892 onwards and available on microfilm until 1931, these election addresses serve as a crucial record of the development of political culture and the changing nature of political competition in Great Britain in the early 20th century. Despite their potential value as a resource for scholars, to date the archived British election addresses have not been made into a comprehensive, publicly-available dataset, in part because of the considerable challenges they pose for digitisation.
With a machine-readable corpus of British election addresses, we will be able to compare the use of group appeals by parliamentary candidates in the period, exploring how their strategies varied with and responded to each other. This type of analysis will be enabled by the unique geographic and temporal structure of the data, in which candidates from different parties are observed both across districts and across elections.
One basic interactive hypothesis we can test with this data, for example, is that the use of class-based group appeals by Labour Party candidates in one election might affect the propensity of candidates from other parties in the district to use similar rhetoric in subsequent elections, conditional on the closeness of the electoral result. Alternatively, we might find the successful use of class-based group appeals by Labour candidates induces candidates from other parties to try and de-emphasise the salience of class, for example, by mounting an increased number of nationalistic appeals. These and other hypotheses about the diffusion and dynamic effects of group appeals will be made testable with these data.