Because of the sheer volume of consumer reviews posted to the Internet, a manual approach for the detection and analysis of fake reviews is not practical. However, automated detection of fake reviews is a very challenging research problem given the fact that fake reviews could just look like legitimate reviews. Guided by the design science research methodology, one of the main contributions of our research work is the development of a novel methodology and an instantiation which can effectively detect untruthful consumer reviews. The results of our experiment confirm that the proposed methodology outperforms other well-known baseline methods for detecting untruthful reviews collected from amazon.com. Above all, the designed artifacts enable us to conduct an econometric analysis to examine the impact of fake reviews on product sales. To the best of our knowledge, this is the first empirical study conducted to analyze the economic impact of fake consumer reviews.
Lau, Raymond Y.K.; Liao, Stephen S.Y.; and Xu, Kaiquan, "An Empirical Study of Online Consumer Review Spam: A Design Science Approach" (2010). ICIS 2010 Proceedings. 103.