Confidence and reasoning: Cognitive roots and societal implications (CORCORS)
This fundamental research project is supported by an ANR Access ERC grant (ANR-23-AERC-0006).
The CORCORS project studies how confidence shapes the way people think and make decisions. In particular, it examines whether feeling less confident in a response leads people to slow down and think more carefully.
The project has three main goals:
• understand different forms of confidence and how they affect reasoning,
• test whether confidence directly influences deliberate thinking,
• explore the longer-term impacts of this relationship, including in education and career choices.
By combining experiments with computational work, CORCORS explores the mechanisms linking confidence and reasoning. This work also looks at how confidence contributes to persistent gender differences in participation in science, technology, engineering, and mathematics (STEM) via the confidence-deliberation mechanism.
Scientific outputs include:
Purcell, Z. A., Desender, K. & De Neys (2025; https://doi.org/10.31234/osf.io/qxzae_v1)
Purcell, Z. A., Charbit, L., Borst, G., & Nussberger, A.-M. (2025; https://doi.org/10.1016/j.cognition.2025.106198)
Why do people disbelieve in climate science?
This research project was supported by an Institute of Advanced Studies Toulouse: Multidisciplinary Prize. Principal investigators include B. Bago, S. Thieme, and Z. Purcell.
Collective climate action requires shared climate beliefs in line with the scientific consensus. However, climate discourse is flooded with fallacious information and people often fail to evaluate it correctly, accepting weak arguments and rejecting strong ones. Understanding these failures is critical for developing evidence-based interventions, yet it remains unclear whether they are driven by partisan identity protective tendencies, entrenched prior beliefs, or failures in logical reasoning. This project applies cognitive science principals and techniques to unravel these effects and begin to build effective, scalable interventions.
Scientific outputs include:
Bago, B. & Purcell, Z. A. (2025; https://osf.io/preprints/psyarxiv/fxc9m_v1)