Article,
Do Experts or Collective Intelligence Write with More Bias? Evidence from Encyclopædia Britannica and Wikipedia
Harvard Business School Working Paper, (March 2016)
Abstract
Organizations today have the option of using crowds or experts for knowledge production. While prior work focuses on comparing the factual accuracy of knowledge from crowd-based and expert-based production models, we compare bias from these two models when knowledge is contested. Contested knowledge is endemic to topics involving subjective, unverifiable, or controversial information, and addressing it successfully lay behind the promise of the production model based on collective intelligence. Using data from Encyclopædia Britannica, an encyclopedia authored by experts, and Wikipedia, an encyclopedia produced by an online community, we compare the slant and bias of pairs of articles on identical political topics. Our slant measure is less (more) than zero when an article leans towards Democratic (Republican) viewpoints, while bias is the absolute value of the slant. We find that Wikipedia articles are more slanted towards Democratic views than are Britannica articles, as well as more biased. The difference for a pair of articles decreases with more revisions of Wikipedia articles. The bias on a per word basis hardly differs between the sources, pointing towards the key role of article length in online communities. We stress a mechanism for resolving disputes in online communities: contributors tend to add text instead of reducing it, taking advantage of the lower costs to acquiring, storing, and revising information, as well as the absence of organizational discipline to restrain additions. These results highlight the pros and cons of each knowledge production model, and have implications for how organizations manage crowd-based knowledge production.
