How explicit expected value information affects tax compliance decisions and information acquisition

In a MouselabWEB experiment with 345 participants, we investigated whether different presentations of expected value information in tax compliance decisions increase conformity with classical deterrence models' assumptions. Recording both choice and process data, we compare conditions of verbal explanation only, verbal explanation plus numerical cue, verbal explanation plus visual cue, and a control condition without expected value information. Only when the expected value was presented as a visual cue the option with the higher expected value (i.e., evasion) was chosen more often than the control condition (58.3% vs. 38.4%). Nevertheless, individuals were more compliant than predicted by the deterrence model. While we identified differences between the experimental conditions in information acquisition patterns and decision times, they do not suggest that one way of presenting expected value information was easier to process than the others or that the main behavioral effect can be explained by higher saliency of the visual cue. These results indicate that individuals' decisions are not predominantly driven by outcome maximization, even when explicit expected value information is provided.