Information Aggregation in Polls
Author(s) -
John Morgan,
Phillip C. Stocken
Publication year - 2008
Publication title -
american economic review
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 16.936
H-Index - 297
eISSN - 1944-7981
pISSN - 0002-8282
DOI - 10.1257/aer.98.3.864
Subject(s) - polling , economics , estimator , construct (python library) , sort , yield (engineering) , econometrics , information transmission , microeconomics , statistics , computer science , mathematics , computer network , metallurgy , information retrieval , materials science , programming language , operating system
We study information transmission via polling. A policymaker polls constituents, who differ in their information and ideology, to determine policy. Full revelation is an equilibrium in a poll with a small sample, but not with a large one. In large polls, full information aggregation can arise in an equilibrium where constituents endogenously sort themselves into centrists, who respond truthfully, and extremists, who do not. We find polling statistics that ignore strategic behavior yield biased estimators and mischaracterize the poll's margin of error. We construct estimators that account for strategic behavior. Finally, we compare polls and elections.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom