Predicting Trust in Wikipedia Articles

Predicting Trust in Wikipedia Articles - scientific work related to Wikipedia quality published in 2011, written by K.L. Cheung.

Overview

Wikipedia is a very popular source of encyclopedic information and studies have shown that its information quality is high. Due to the open source character, evaluation of the source is not possible. Therefore users need to use other features to assess the credibility of the articles. This study focused on measurable features, which are in this case ‘surface’ features. Studies suggest some important features to predict article quality and that are deemed important by the users themselves. These are ‘references’, ‘internal links’, ‘pictures’, and ‘length’. An online experiment was conducted to test whether these features could positively predict the perceived credibility of articles. Authors did not find any effects of manipulations. The manipulated features reflected that, based on these features, participants did not trust ‘featured articles’ more than ‘random articles’. However, many motivations of the participants were about these features, especially the number of references. This means that the participants did think the features were important. One explanation for this result is that the interpretation may be the same for both conditions of manipulations. Models for evaluating computer credibility may explain this phenomenon.