Improving system knowledge and performance with proactive intelligent on-line help

Abstract

Computer software is becoming increasingly powerful and complex, but most people only utilize a small fraction of the capabilities of the software that they use to accomplish their daily tasks (Antsey, 1988; Fischer, 1987, 1993; Sutcliffe & Old, 1987). Carroll and Rosson (1987) described this problem as the production paradox. Typical methods of training and help, such as tutorials, manuals, and on-line help, fail to help intermediate-level users overcome this problem (Elkerton & Williges, 1989). The use of proactive, intelligent help systems (e.g., Coaches and Critics) may provide a more effective approach (Selker, 1994). Coach, Critic, and tutorial training were compared to determine which intelligent help design would provide the most effective instruction. Sixty-two users were trained, tested, and evaluated on advanced word processing functions during three sessions to assess the acquisition, retention, and application of efficient methods. The Coach and tutorial groups both improved their performance times and efficiency by the second test. However, the Critic group did not perform as well. These results suggest that providing users with frequent, proactive, context-specific feedback and help may encourage them to learn and utilize more efficient software methods.