Using Affective Signals as Implicit Indicators of Information Relevance and Information Processing Strategies

Search engines have become increasingly better at providing information to users. However, they still face major challenges, such as determining how searchers process information, how they make relevance judgments, and how their cognitive or emotional states affect their search progress. We address these challenges by exploring searchers affective dimension. In particular, we investigate how feelings, facial expressions, and electrodermal activity (EDA) could help to understand information relevance, search progress, and information processing strategies (IPS). To meet this goal, we designed an experiment in which 45 participants were exposed to affective stimuli prior to solving a fact-finding search task. Results indicate that initial affective dimensions are linked to IPSs, search progress, and task completion. However, further analyses suggest that affective-related features alone have limited utility in the binary classification of relevance using machine learning techniques. Copyright © 2016 by Association for Information Science and Technology