Absract:

While individuals hold, exchange, and update opinions over multiple issues, opinions are often correlated and a unidimensional spectrum is enough to summarize them. But when should one expect opinions to be unidimensional? And how important is the underlying structure of communication? Our experimental results: i) validate the crisp predictions by De Marzo et al. (2003) when individuals update their opinions on a fixed network always trusting the same neighbors, ii) jointly with new theoretical results and simulations indicate the prevalence of unidimensionality as an expected outcome even when communication is less structured with individuals' network possibly varying over time, and iii) highlight the importance of the communication structure in predicting whether individuals hold relatively moderate or extreme opinions.

Joint work with Philippos Louis and Orestis Troumpounis

Location:

Room 3E34, CEPS ENS Paris-Saclay
4 avenue des Sciences, 91190, Gif-sur-Yvette