Evaluating Intelligibility Usage and Usefulness in a Context-Aware Application

Intelligibility has been proposed to help end-users understand context-aware applications with their complex inference and implicit sensing. Usable explanations can be generated and designed to improve user understanding. However, will users want to use these intelligibility features? How much intelligibility will they use, and will this be sufficient to improve their understanding? We present a quasi-field experiment of how participants used the intelligibility features of a context-aware application. We investigated how many explanations they viewed, how that affected their understanding of the application's behavior, and suggestions they had for improving its behavior. We discuss what constitutes successful intelligibility usage, and provide recommendations for designing intelligibility to promote its effective use.