It is a fact widely accepted by psychologists that people’s judgments are influenced by their beliefs and unconscious biases. Although companies that provide Internet-related services excel in tracking user behavior online, little or no thought is given to the beliefs and unconscious biases that may influence the decisions of users. In this paper, White studies search-related biases through various probes: (1) a log analysis of user search behavior in a web search engine, (2) human labeling of captions and results returned by that search engine, and (3) a survey that reflects the retrospective view of the search. White also analyzes the users’ health domain searches with yes or no questions in which he observes that the searchers themselves have their own biases. The paper also shows that the search engines, irrespective of the truth, always favor positive results. These findings can have profound implications on information retrieval system design, ranking, and recommendation systems.
Findings of this novel research, which thoroughly investigates the biases of users, can have significant implications on how search engines interpret user search behavior and beliefs. The research sheds light on the cognitive biases of users, and on how a search engine can misinterpret these biases and, in turn, produce wrong search results. This research can indeed inspire new research collaborations between cognitive and computer sciences to further investigate user beliefs and unconscious biases that can influence online behavior.