Browse Theses

Animated demonstrations show people procedural tasks through graphical animation. They seem to be a natural way of helping people learn procedures for computer-based tasks. Yet, we know little about their effectiveness. Two studies were conducted which compared animated demonstrations to procedural text to measure initial acquisition, retention, and transfer. In both studies, users learned and performed HyperCard$\sp{\rm {TM}}$ authoring tasks on the Macintosh$\sp{\rm {TM}}$. In a training session, users either watched a demonstration or read procedures for a task. Later, users performed identical and similar tasks without the aid of any instructions to measure retention and transfer of interface skills. Both studies suggested that demonstration users were mimicking the training demonstrations and had not processed them sufficiently for later retention and transfer. Demonstration users were faster and more accurate than text users when initially acquiring procedural tasks. After a delay, the performance of the demonstration users deteriorated. Unlike text users, demonstration users had greater difficulty transferring to similar procedural tasks. Thus, demonstrations are an effective way to show users how to perform procedural tasks immediately, but limit later retention and transfer of procedural skills. The results were surprisingly similar even when spoken text was added to the demonstration, indicating that the simplicity of watching demonstrations may encourage imitation and disregard for the text. From these experimental results, a model is proposed for acquisition of procedures from animated demonstrations. This model asserts that users simply watch and play back demonstrated procedures, similar to a computer macro. To direct further research, a set of testable hypotheses is derived from the mimicry model. Testing these hypotheses will help strengthen the model and provide design guidelines for the effective use of animated demonstrations.