An Evaluation of Text-Only and Text-3D Model Instruction Formats in Augmented Reality Training

Augmented reality (AR) exhibits potential for delivering training instructions. Trainee learning and transfer can be influenced by the different extraneous cognitive loads implicit in text and graphical instruction formats. Our research evaluated this influence using a GoCube™ training procedure delivered in AR. Learning, transfer and user experience were evaluated using mental rotations, physiological ratings, eye gaze, facial expressions, memory recall and questionnaire responses. Results showed that trainees using text-only instruction were significantly faster in training and recall than their counterparts using a combined text and 3D-model instruction format. This correlated to mental rotation baselines, gaze shifts and cognitive load.