The Internet, Policy & Politics Conferences

Oxford Internet Institute, University of Oxford

Marit Hinnosaar, Toomas Hinnosaar: Authority Bias

Marit Hinnosaar, Northwestern University
Toomas Hinnosaar, Northwestern University, Collegio Carlo Alberto

Full version of paper (PDF)

Abstract

People at positions with formal authority are often expected to make better decisions and fewer mistakes, and therefore their opinions and contributions are given higher weight. This can be an equilibrium effect: people may be selected to the positions with formal authority because of their knowledge or skills. But respect for authority could also be a behavioral bias. These two explanations have very different implications. Our goal is to measure the authority bias, which we define as the difference between perceived and true quality of contributions by people with formal authority. Identifying the authority bias is complicated by the fact that almost always the observable outcomes include both explanations. We propose a method of identifying the authority bias that allows us to separate it from the equilibrium effect. We estimate the authority bias using a novel dataset from Wikipedia. In Wikipedia, editors at high-rank positions are treated differently, and there is high regional variation. Our preliminary estimation results indicate that the authority bias does not exist in Western Europe, but is large in Eastern Europe. The authority bias more than doubles the time needed for the mistakes made by high-rank editors in Eastern Europe to be corrected.

Key words: media economics, Wikipedia, behavioral economics, authority, social learning.