Development of an Approach to Measuring Learnability Based on NGOMSL from Perspectives of Extended Learnability

摘要

The present study proposes an approach to measuring learnability, focusing on the learning process from a perspective of extended learnability. In developing a predictive model for time performance without any interruption or error based on NGOMSL (Natural GOMS Language), the concepts of "expertise time," "actual time," "learning deficit," and "learning level" are considered with a few assumptions, and a learning deficit curve for the relationship between learning deficit and learning level is presented. Experimental data on repetitive use in sample users' first use session of a system demonstrate the application of our approach herein. For exploratory tasks on four simulated websites with different levels of usability and aesthetics, responses from 64 users were obtained in terms of time performance, perceived usability, and user satisfaction. Through this application, we were able to discern some information about the dynamic properties of learning in relation to users' subjective responses: (1) learning level according to number of uses, (2) the number of uses required to reach a competency level, (3) effects of usability and aesthetics factors on learning level, and (4) learning level in connection with perceived usability and user satisfaction. The model is expected to have potentials as an effective approach to measuring learnability insofar as it provides useful information in ways that differ from existing methods for learnability.