Copy the page URI to the clipboard
McConway, Kevin
(1995).
DOI: https://doi.org/10.1007/BF02562628
URL: http://www.springer.com/statistics/journal/11749
Abstract
Abstract of main article: An expert (for You) is here defined as someone who shares Your world-view, but knows more than You do, so that were She to reveal Her current opinion to You, You would adopt it as Your own. When You have access to different experts, with differing information, You require a combination formula to aggregate their various opinions. A number of formulae have been suggested, but here we explore the fundamental requirement of coherence to relate such a formula to Your joint distribution for the experts' opinions. In particular, in the context of opinions about an uncertain event A, we investigate coherence properties of the linear, harmonic and logarithmic opinion pools. Some general results on coherence of the joint forecast distribution are also developed.