The Effects of Interaction Sequencing on User Experience and Preference
Interactive computing tasks are composed of sequences of sub-interactions (or “moments”), each of which yields a slightly different user experience. Prior work, predominantly from the psychology literature, suggests that the order of these moments can affect people‟s retrospective evaluation of experiences. Several kinds of sequencing effects have been examined, including primacy, recency, and peak-end effects. We review previous research on sequencing effects and their potential application in Human-Computer Interaction, of which prior work has found mixed results regarding the influence of interaction sequence on preference – possibly because the magnitude of experiential changes caused by interactive tasks are weaker than those studied in psychological experiments. However, sequencing effects are still of great importance to interface design, because when they occur, they have the potential to substantially change user preferences for common interactions. To explore the subtlety of sequencing effects in HCI, we describe two experiments that examined user preferences for series of interactions with different orderings that created positive and negative recency and primacy effects. Positive and negative experiences were created with simulated system assistance that either worked well (aiding the user in drag-and-drop tasks) or worked poorly (hindering the user). In both experiments, the series differed only in the order of positive and negative momentary experiences. Results of Experiment 1 were mixed: the study provided some support for recency effects, but without strong evidence. Experiment 2 modified the experimental method to better accentuate the positive and negative experiences, and produced results showing strong effects of recency, but not of primacy. We discuss reasons for these results, consider overall explanations for the subtle nature of sequencing effects on HCI tasks, and provide an agenda for further research and design lessons regarding recency effects. Overall, we contribute new understanding of a phenomenon that can have a substantial impact on user experience, but that is currently underexplored in HCI.