A large proportion of users of sophisticated software will never use a discretionary help system such as the help button or help menu. This number includes both inexperienced users and supposedly expert users. We believe there is a pressing need to find a more effective way of engaging people with the available help, especially for less-experienced users.. Since these users may not understand the purpose of help or even realize it exists, the use of embedded help is an obvious solution to improving user task performance. In order to test the effectiveness of embedded help, we created a Web-based timesheet simulation where users were assigned one of two available interfaces — help menu alone or help menu plus embedded help. Seventeen users were then asked to complete timesheets using one of the interfaces with all their actions recorded. Users completed a final survey at the end of the session and were asked follow-up questions a week later. Results show that the embedded help was accessed five times more often than menu help alone; and that embedded help users were more confident and relaxed with their interface than menu help-only users. The usefulness of embedded help for all users is discussed.
展开▼