Investigating the Impacts of Expectation Disconfirmation on Web Search

Expectation disconfirmation refers to the situation where a user's perceived performance of a system disconfirms her original expectation. Previous information systems studies have demonstrated that expectation disconfirmation can significantly affect a system user's behavior and experience. Inspired by this finding, we go beyond the traditional approach that focuses on the final post-search perception and study the expectation disconfirmation problem in Web search. Our study investigates task difficulty expectation disconfirmation and demonstrates that: (1) unexpectedly difficult task can significantly decrease a user's perceived level of search success and increase the perceived time pressure; (2) the size and direction of task difficulty expectation disconfirmation are significantly associated with Web search behavior; (3) it is possible to predict the state of expectation disconfirmation (especially the negative, unexpectedly difficult cases) based on search behavioral features. This study demonstrates the value of integrating expectation disconfirmation approach with interactive IR research and thus may encourage future researchers to further explore the effects of other aspects of users' expectations and post-search perceptions.