Comparing memorability of gesture sets in an extended reality application

Abstract

In free-form gesture sets, memorability is an important yet often under-explored metric, despite evidence that the usability of interfaces improves when designed with more memorable input gestures. This study examines the memorability of three free-form gesture sets in the HoloLens 2: user-defined, elicitation-defined, and expert-defined. In addition, we examine gestures selected by the participants using common techniques from previous elicitation studies. We found that the user-defined gesture set was the most memorable, with an 88.57% recall rate. And was significantly more unforgettable than the expert-defined (72.73% recall) and the elicitation-defined (59.87% recall). This study also analyzed the user-defined gestures from this experiment. Although this was not an elicitation study, many of the methods commonly used in elicitation studies were used here. This analysis found a higher agreement rate when users were primed with a single gesture set before creating their own and a decrease in agreement when showing them two gesture sets beforehand. Given these results, we propose that designing systems with user-defined gestures will result in the most memorable sets; however, expert-defined gesture sets are also highly memorable and may better suit application design constraints.