Iterative Design of Gestures during Elicitation: A Gateway into User's Mental Models

The design of gestural interfaces through gesture elicitation studies is influenced by legacy bias, wherein users will draw from touch and desktop interactions when specifying gestures for novel interactions, such as full-body and free-space applications. This may inhibit users from developing gestures that are more appropriate and less fatiguing for the interaction. Increasing production during elicitation studies has shown promising results in moving users beyond legacy biased gestures. However, when increasing production, users begin to iterate on their gestures. In this paper, we present a gesture elicitation methodology aimed at understanding design iteration in studies that use increased production. We show that gestural refinements provide insight into the gestural features that matter for users to assign semantic meaning and discuss implications for training gesture classifiers. We also discuss a potential relationship between refined features and fatigue, showing that refinements may map to less fatiguing gestures.