Semantic Networks Analysis of Political Party Platforms: Coalition Prediction Based on Semantic Distances in Scottish Elections 1999–2011

Abstract

This paper conducts a semantic network analysis of Scottish political party manifestos from the 1999, 2003, 2007 and 2011 elections to predict coalition formation. A manifesto is a tool of political communication that represents a party’s ideology and discusses issues considered crucial for each election. Semantic analysis of party manifestos represents an innovative method for examining the relations among political parties. Parties that are closer in the semantic network have a greater potential to form political coalitions. Semantic distances and two-mode semantic network analysis also proved to be a viable method for describing a country’s political climate and power structure, without requiring prior knowledge of country’s political makeup, and without pre-conceived notions. The results indicate that a semantic analysis of political party manifestos can be used to predict coalitions, but factors other than semantic similarity must be considered. Semantic network analysis provides an accurate picture of the distribution of power and centrality of issues for any given election.