Tracking Personalisation in British Maiden Speeches from 1945ꟷ2017

Several studies have pointed to important changes in the nature of political representation. Politics is thus increasingly described as ‘personalized’ in the sense that MPs see their representational task as more individualized (McAllister, 2007). This is a crucial development as it has important implications for party democracy. Still, we know little about if the behavior of politicians also have become more personalized (Rahat & Sheafer, 2007; Karvonen, 2010). The aim of this paper is to contribute to close this gap. More specifically, this paper focus on personalization in parliamentary speeches. A source, which Rahat and Kenig (2018) has proposed for identifying behavioral personalization. However, they do not pursue the task empirically. Still, Balmas et al. (2014) are the only study, who have investigated personalization in speeches. They propose self-references - the use of first person singular - as a proxy to measure personalization. The question is if counting how many times politicians say ‘I’ really captures an abstract concept like personalization. Therefore, this paper contributes with a clarification of the concept of behavioral personalization and discuss possible ways to operationalize the concept. Even though, Balmas et al.’s innovative study was a good first step in this area of personalization, one could argue that we need more evidence than one study of speeches from 30 politicians. Therefore, this paper provides evidence from a large N study (N=2842) over a long period. More specifically, this paper investigates British MPs’ first speech in parliament - their maiden speech - from 1945-2017. The British case allows us to investigate personalization in a setting with strong parties and an electoral system that provides an incentive for MPs to personalize their behavior. Furthermore, maiden speeches provide an unique opportunity to track personalization over a long time span, in a speech where floor time is not restricted by the party and where factors such as seniority and leadership position is kept constant. To analyze the large corpus of speeches, this paper utilize quantitative text analysis and develops a general dictionary for analyzing political personalization in political texts.