Designing ecologically valid studies that accurately gauge the learnability, usability and overall efficiency is often difficult. Previous research on sketch based interaction techniques are often evaluated using simplified tasks. In this paper, we propose evaluating new pen/stylus interfaces using experiments with higher ecological validity. As an example, we present a study of a novel interaction technique – the inferred mode protocol. The study looks at learnability in the real world by considering multiple scenarios: users with varying level of training, and different ways of presenting computational intelligence. The study design helped extract previously undiscovered important information on the learnability and usability of the inferred mode protocol.

Matei Negulescu, Jaime Ruiz, and Edward Lank, Designing Studies with High Ecological Validity, In CHI '10 Workshop: Designing Sketch Recognition Interfaces, 2010