Abstract
Subjectivity in the decision environment allows for distorting beliefs to enable desired behaviors (e.g., Konow 2000; Haisley and Weber 2010; Shalvi et al. 2011; Exley 2015; Gneezy, Saccardo, and van Veldhuizen 2018; Gneezy et al. 2020; Falk, Neuber, and Szech 2020). In this vein, the strategic use of uncertainty to take more selfish choices has been show experimentally both under risk (Exley, 2016) and ambiguity (Garcia et al., 2020): uncertainty is used as an excuse not to give. We rely on the experimental setting of Exley (2016) and Garcia et al. (2020) to detect excuse-driven behaviour under risk in order to explore belief updating and its possible distortion with more or less informative signals: 4 scenarios are implemented through the use of multiple choice lists where subjects choose between a safe option and a risky option, and these 4 scenarios differ in whether the beneficiary of the safe option and of the risky option is the same subject who is choosing or a charity. We add to this setting a signal on the outcome of the lottery, that can be a positive or a null outcome. This signal could be more or less informative. Recent literature has shown overreaction to weakly informative signals informative signals (Augenblick, Lazarus and Thaler, 2021) and even non-informative signals are incorporated (Kieren and Martin Weber, 2022). Our setting contributes to this strand of literature enabling to answering the following questions: Are both informative and weakly informative signals incorporated? If yes, do they affect choices in a self-serving manner? Preliminary evidence shows that both informative and weakly informative signals are incorporated, although the size and significance of the effect of the weakly informative signal is lower. Additionally, the behavioural reaction to the signal is stronger when this is aligned with self-interest.