Varieties of learning to use computer tools

This article presents an overview of different approaches to the study of how people learn computers as tools. The varieties of learning most relevant for computer learners and also mostly discussed by human-computer interaction (HCI) researchers are transfer, construction of mental models, and learning by doing. Transfer from prior knowledge is a well-developed theory within HCI. Empirical data show that quantitative predictions of positive transfer from prior knowledge can be made for simple tasks. The occurrence of negative transfer can be predicted, but effects of prior knowledge that conflict with the material to be learned cannot be covered by a simple transfer theory. It is generally acknowledged that users develop mental models by interacting with a system. However, the exact nature of these models are difficult to capture empirically, since mental models have to be inferred from users' behaviour. In normal use, people seem to have difficulty talking about their mental models. People express their mental models mainly when their expectations are violated. It seems relevant to support the forming of adequate mental models by instruction, but the effects of such educational models are not clear-cut. Several factors influence the comprehension and use of the models given. People do not learn very efficiently by doing, because they have difficulty in choosing alternative actions, observing outcomes, and interpreting them. To conclude, users of computer tools mostly learn as a result of performing a task and do not focus on learning the computer system per se. This implies that learning new aspects of a computer tool is more similar to problem solving than to intentional learning.