Re:Business Advice Plagued by Survivor Bias
by jerf on Monday August 17, 2009 @09:45PM
(#769565)
Are you sure about that? Journals don't (in general) publish negative results. The result of the pressure to publish is not that negative results get published; it's that the scientist has an incentive to create positive results. Not necessarily through data falsification, but through choosing experiments very likely to generate "results", which, unfortunately, also implies that the information content of such experiments is necessarily lower than it could be if the scientist had more freedom to perform an experiment that had a higher likelihood of failure.
[ Parent
| on HN ]