Multimodal versus unimodal instruction in a complex learning context

Multimodal instruction with text and pictures was compared with unimodal, text-only instruction. More specifically, 44 students used a visual or a textual manual to learn a complex software application. During 2 103-116-min training sessions, cognitive load, and time and ability to recover from errors were measured. After training, the authors tested students' learning on trained and untrained tasks. The results for cognitive load, training time, and learning effects initially supported dual coding theory. The results show that even in this complex situation, multimodal instruction led to a better performance than unimodal instruction. That is, the multimodal manual led to a stronger mental model of the computer program, improved identification of window elements and objects, and speeded up the location of window elements and objects.