Exploratory study of cross-device search tasks

Cross-device search is an emerging subject in the study of information retrieval. This paper explores cross-device search behavior through the characteristics of cross-device search tasks. Unlike previous research on transaction log analysis, this paper extracted cross-device search tasks from descriptions of real-situation cross-device search experiences collected by a crowdsourcing survey targeting global users. A total of 343 valid responses were used for the content analysis, and the coding scheme was grounded in the Multiple Information Seeking Episodes (MISE) model, which was proposed for explaining successive multiple-episode search. Characteristics of cross-device search tasks were uncovered by coded categories of Topic, Type, Complexity of Knowledge Dimension, Complexity of Cognitive Dimension, Environment; Device Switch, and Switching Demand. The results show the most frequently searched topics are Arts, Shopping, Reference, and Computers. Task types focus on factual tasks, indicating a clear information need. Task complexity depends heavily on the user's cognition. Eight reasons for switching device are identified in understanding device switch demand. Finally, implications for designing cross-device search tasks are proposed based on the correlation among task attributes. Limitations on the degree to which respondents answered recall-based questions accurately have been acknowledged.