Correlations between political party size and voter memory: A statistical analysis of opinion polls

Abstract

This paper describes the application of statistical methods to political polling data in order to look for correlations and memory effects. We propose measures for quantifying the political memory using the correlation function and scaling analysis. These methods reveal time correlations and self-affine scaling properties respectively, and they have been applied to polling data from Norway. Power-law dependencies have been found between correlation measures and party size, and different scaling behaviour has been found for large and small parties.