The Guided Evaluation Method: an Easier Way To Empirically Estimate Trained User Performance for Unfamiliar Keyboard Layouts

Abstract

To determine in a user study whether proposed keyboard layouts, such as OPTI, can surpass QWERTY in performance, extended training through longitudinal studies is crucial. However, addressing the challenge of creating trained users presents a logistical bottleneck. A common alternative involves having participants type the same word or phrase repeatedly. We conducted two separate studies to investigate this alternative. The findings reveal that both approaches, repeatedly typing words or phrases, have limitations in accurately estimating trained user performance. Thus, we propose the Guided Evaluation Method (GEM), a novel approach to quickly estimate trained user performance with novices. Our results reveal that in a matter of minutes, participants exhibited performance similar to an existing longitudinal study - OPTI outperforms QWERTY. As it eliminates the need for resource-intensive longitudinal studies, our new GEM thus enables much faster estimation of trained user performance. This outcome will potentially reignite research on better text entry methods.