We checked for a potential selection bias within the sample from banks’ customers: we asked CEOs of banks to collect
socio-demographic and economic information of sampled individuals, extracted from the Customer Relationship Management (CRM) database, and to indicate whether they agreed or declined to take part in the task. In total, 332 individuals were invited and 222 (66.87%) agreed to participate in the experiment. Comparisons between those who accepted and refused to take part to the experiment allow us to rule out a selection bias.2 This control is omitted for samples of employees obtained from international asset managers and financial advisory companies (precisely, 84 asset managers and 150 financial advisors), because this recruitment is naturally biased by the sampling criterion itself.3