Do the Facts Matter? Effects Statistics and Exemplars in Competitive Contexts

• In Event: What to Believe? Facts in a Post-truth World

Abstract

Democratic accountability mechanisms require that citizens and politicians have information about societal problems and the impact of public policy. Factual evidence may shape policy-making directly by influencing politicians’ policy preferences or indirectly by shaping citizens’ policy preferences which politicians’ in turn anticipate in their policy-making. Policy information infrastructures have been established to generate factual evidence for elite decision makers and the mass public about welfare issues and the impact of public policy. Yet, does factual evidence shape politicians' and citizens' welfare policy preferences?
In this article we advance understanding of this question. We investigate the effect of statistical evidence on politicians’ and citizens’ welfare policy preferences in competitive information contexts where the statistics are challenged by just a single exemplar portraying a person that provides competing perspective. In the hierarchy of evidence, statistical evidence is typically considered at the top whereas single exemplars range further down due to limitations to external validity, reliability and representativeness of the overall target group. Yet Exemplars are widely used in political communication and news reporting. However, to the extent that single exemplars are not representative of the underlying target group they may foster misperceptions and foster misinformed opinions among elites and in the mass public. Therefore it is important to gain a better scientific understanding of the resilience of the effect of statistical evidence on policy opinions in the face of competition from a single exemplar.
Extending recent research on psychological biases for vivid social information we argue that competitive exposure to just a single exemplar can crowd out the effect of statistical evidence on policy preferences. Drawing on research on the deservingness heuristic, cheater detection biases, and loss aversion we further argue that this may be particularly the case when statistical evidence that describes the sufferings of a welfare target group is countered by a single exemplar of a morally fraudulent welfare claimant. We test the predictions in two studies including pre-registered online survey experiments fielded to a population sample of politicians running for national office and to approximately representative samples of citizens. The results show that statistical evidence about experiences of psychological sufferings among welfare benefit claimants significantly reduces support for restrictive welfare policy requirements among citizens but not politicians. Importantly, the findings also show that competitive exposure to just a single exemplar with a fraudulent welfare recipient overwhelms the effect of the factual statistical evidence leading to a significant increase in support for welfare burdens among citizens. Among politicians, there is no statistically significant effect of competitive exposure to just a single exemplar with a fraudulent welfare recipient. In a context with rising concerns about the rise of misinformation the findings highlight how just a single competing exemplars can crowd out the effect of general statistical evidence on policy preferences in the mass public. In terms of policy-implications, this suggests that in competitive information environments, exposure to single exemplars may especially force out the effect of statistical evidence on policy-making through the indirect pathway where politicians tune policy-making according to citizens’ preferences.

Authors

• Lene Aarøe, Aarhus University

• Miceal Canavan

• Julian Christensen, Aarhus University