Networks make us more cautious: using Drift Diffusion Model to measure the learning process in Prisoner’s Dilemma on different network topologies

Montero, E. (Speaker)

Activity: Talk or presentationTalk or presentation at a workshop/seminar


Different types of people learn how to make decisions differently, especially when making complex choices informed by social network interactions. The details about these differences are still largely unknown. In this paper, we investigate how differences in cooperativeness are linked with our deliberation process in different social contexts. We study how long it takes for subjects to learn to play the game and how this depends on the network structure. The Drift Diffusion Model (DDM) (Ratcliff, 2016) gives us a unique perspective and ability to quantify the subject's cautiousness by observing response times. We found that subjects playing Prisoner’s Dilemma in a network act more cautious with respect to a pairwise setting. Moreover, subjects start to learn the game within 10-30 rounds, and from this moment on, they do not change much how they play the game. We then categorise subjects in those who cooperate/defect most of the time regardless of what others do and those who adapt their actions to what the other players do in the previous round. Our DDM analysis predicatively shows that those who react to others’ actions perceive a difference in their cautiousness between playing in a fixed-neighbors setting versus a shuffled-neighbors network setting. We thus hypothesize that these settings enable building a trust relationship which plays a role in this case.
Period12 Nov 2020
Event titleBelgian Network Science Research Meeting
Event typeWorkshop
LocationGhent, Belgium
Degree of RecognitionNational