Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session B: Wednesday, August 13, 1:00 – 4:00 pm, de Brug & E‑Hall

Integrating explicit reliability for optimal choices: effect of trustworthiness on decisions and meta-decisions

Keiji Ota¹, Anthony Ciston, Patrick Haggard, Thibault Gajdos Preuss, Lucie Charles; ¹Queen Mary, University of London

Presenter: Lucie Charles

A key challenge in today’s fast-paced digital world is to integrate information from diverse sources with different reliability. Beyond estimating the reliability of information based on prior knowledge, it is also fundamental to understand whether people can use explicit information about the reliability of the source. In particular, a question that remains underexplored is how people use probabilistic information about the likelihood of a source to give correct information. Here, we investigated how such explicit probabilistic estimates of reliability are encoded and integrated into decision processes. To do so, we developed a novel paradigm that required participants to combine evidence from sources with different explicit levels of reliability to estimate among two responses which one was more likely to be correct. Additionally, participants had to rate after each choice the extent to which they felt they were influenced by a given source of information. Through computational modelling, we found that participants misrepresented the reliability of sources, distorting the probability a source to give correct information. As a results, they gave too much weight to unreliable source and too little weight to sources that were reliably wrong sources. However, we found that subjective report of influence correctly predicted the effective influence a source had on the decision. These findings suggest that participants were at least partially aware of what bias their choices.

Topic Area: Predictive Processing & Cognitive Control

Extended Abstract: Full Text PDF