Playing the game and dodging the blame: an experimental study of trust perceptions, personality traits, and politicians’ blame avoidance behaviour.

Popularised by Kent Weaver and later Christopher Hood, blame avoidance behaviour (BAB) refers to those strategies employed by political elites to minimise blame and manage risk. However, studies of politicians’ BAB have largely relied on secondary data recorded in the media or parliamentary records (e.g. Hansard), and none have gone far enough to consider the individual differences of those who engage in BAB or the ways in which these characteristics might encourage or mitigate BAB. This paper addresses each of these gaps in the existing research base. Firstly, I seek to replicate two existing insights from observational studies of BAB using experimental data gathered directly from politicians. Principally, I test a reactivity hypothesis whereby officeholders are more likely to engage in BAB when levels of blame are high (H1) and a sequencing hypothesis whereby officeholders will ‘retreat’ through types of BAB in predictable patterns (H2). Secondly, I analyse the importance of two salient sets of psychological characteristics: namely politicians’ personality traits (i.e. the Big Five) and their ‘felt’ trust and distrust. On one hand, I propose that politicians with high scores for extraversion and openness will be less likely to exhibit BAB regardless of the level of blame surrounding them, whilst politicians with high scores for neuroticism, agreeableness and conscientiousness will be more likely (H3). On the other hand, I propose that politicians who feel or perceive high levels of public trust/distrust will be less/more likely to use and then double down on BAB to compensate for, or anticipate, the mediated reception of political choices amongst their electors (H4). To test H1-4, I present the findings of a survey experiment fielded to elected councillors (local politicians) in the United Kingdom (N=440) as part of a larger funded study of trust and governance. Focusing specifically on presentational strategies following a hypothetical scandal, this experiment placed participants in three treatment conditions that varied the level of blame between low, moderate and high. Participants could then choose between three responses indicative of PD, PA+RD, and PA+RA. If participants opted for either of the first two strategies, they were presented with more information about the development of the scandal. In this follow-up scenario, which takes places 24 hours later, the level of blame increases further and politicians are once again presented with three possible response options indicative of PD, PA+RD, and PA+RA. Politicians also completed the Ten-Item Personality Inventory (TIPI) and a 24-item battery of felt trust and distrust (PTB-24). The significance of this paper is two-fold. Firstly, I contribute methodologically to our understanding of how BAB might be studied by replicating observational findings using experimental data gathered directly from elected representatives. Secondly, I contribute substantively to our understanding of why BAB occurs in politics by analysing the impact of politicians’ individual differences (both trait-based and cognitive) on BAB.