Predictions as a better way to elicit preference structure

Sonja Radas, Institute of Economics Zagreb

In this paper we propose a new method of eliciting market research information. Instead of asking respondents for their personal choices and preferences (which is the standard market research approach), we propose to ask respondents to predict the choices of their peers. In this way people are requested to engage their peer-knowledge, a rich source of information that accrues from individuals’ immersion in social networks.

We illustrate the benefit of peer-knowledge with conjoint analysis as our methodological test bed. We develop a theoretical argument that shows that relying on peer-knowledge yields improvements: namely, larger social circles allow for more accurate utility estimates and for better predictions of preference share. We show that this is true regardless of the underlying segmentation. In addition, we show that our results remain valid even when the information from social circles is biased. These theoretical findings were confirmed in three online experiments.



About the speaker

Sonja Radas is a Senior Research Fellow at the Institute of Economics Zagreb, and a Research Affiliate at the Massachusetts Institute of Technology, US. She received a PhD in Mathematics and a PhD in Business, both at the University of Florida, Gainesville, US. Previously she was on the faculty of Washington University in St. Louis. She held a number of visiting positions, most recently a Visiting Researcher at the Massachusetts Institute of Technology, Sloan School of Business. From 2015-2018 she has been a Marie Curie Fellow. 



Venue: Conference room (0.16 & 0.17)

Date: 08 March 2018

Time: 12:00 - 13:00