Adrien Aufort
Graduate Student

Ecole doctorale : Economies, Espaces, Sociétés, Civilisations (ED 382)

UFR Géographie, Histoire, Sciences de la Société

Directeur de thèse : Gérald Bronner

1ère année de doctorat


Collective decision making: an experimental and empirical approach to the factors and frames of choice, between experts and laypersons

The present thesis is focused on expert and layperson (non-expert) decision making.

The empirical side, focuses on panels of experts, like the CCNE (Comité consultatif national d'éthique) in which members collectively emit recommendations regarding matters of bio ethics (abortions rights, biodiversity and health, euthanasia, stem cell use...).

The experimental side focuses on replicating a decision structure in which key variables could be controlled: number of participants, diversity, amount of information on the topic, decision modality (open vote, secret vote, consensus...), physical local (round table, bench...), in order to test a set of hypotheses:

  • a single expert decision is likely to be more efficient than a single layperson's
  • group decisions are likely to be more efficient than a single individual decision (regardless of expertise)
  • a group of laypersons is likely to produce better decisions than a single expert
  • a group of experts is likely to produce better decisions than a group of laypersons


  • inclusive heterogeneous groups (where minority opinions and dissent can form) are likely to produce more efficient decisions than homogeneous ones (where conformity overpowers individual initiative)
  • groups that deliberate are likely to produce better decisions than those that don't

On one hand, aggregative models posit that group decisions result from the process of adding individual preferences (as well as structuring environmental factors), on the other, the deliberative models posit that group decisions emerge as personal views intertwine and evolve based on internal group dynamics (as well as structuring environmental factors). Bearing these models in mind, how (in terms of group size and conditions) can heterogeneous groups of non-experts outperform experts?

Social Media Auto Publish Powered By :