Background. As they guide people in performing a task, procedures are the heart of most manuals. It is, therefore, somewhat surprising that the theoretical and empirical knowledge of their nature has remained somewhat elusive. This paper describes a theoretical framework for procedures, summarized as the four components model, which is grounded in systems theory and rhetoric. Aim. The study addresses two research questions: (1) What are procedures made of? and (2) Which design guidelines for procedures can be abstracted from theory and research? Results. The model distinguishes between: goals, prerequisite states, unwanted states (warnings and problem-solving information), and actions and reactions. For each component pertinent research findings are summarized and lead to the formulation of design guidelines. Occasionally these guidelines are compared with existing procedures from a sample of 104 manuals to see how well theory and practice agree. Conclusion. The model offers a manageable and expandable framework for creating user support that is based on scientific research. It can be used for a systematic analysis of procedures and for their (re)design.

The study set out to validate the supportive role of screen captures for switching attention. Forty-two participants learned how to work with Microsoft Excel with a paper manual. There were three types of manuals: a textual manual, a visual manual with full-screen captures, and a visual manual with a mixture of partial- and full-screen captures. The findings show that participants in all conditions looked up from the manual to the screen on about 97% of the cases in which such a switch was called for rank order analyses showed that users of the visual manuals switched attention significantly more often than did users of the textual manual. No differences were found between conditions on learning effects and training time.

The study examines the use of screen captures in manuals. Three designs of manuals were compared, one textual and two visual manuals. The two visual manuals differed in the type of screen capture that was used. One had screen captures that showed only the relevant part of the screen, whereas the other consisted of captures of the full screen. All manuals contained exactly the same textual information. We examined the time used on carrying out procedures (manual used as a job aid) and the results on retention tests (manual used for learning). We expected to find a trade-off between gain in time and learning effects. That is, we expected that higher scores on the retention tests involved an increase in time used and, vice versa, that gains in time would lead to lower retention test scores. We also explored the influence of manual design on user motivation. For job-aid purposes, there were no differences between manuals. For learning, the full-screen captures manual and the textual manual were significantly better than the partial-screen captures manual. There was no proof for the expected trade-off. More learning was not caused by an increase in time used. We found no effects on user motivation. This study does not yield convincing evidence to support the presence of screen captures in manuals. However, if one wants to include screen captures, this study gives clarity for the type of screen capture to choose. The use of full-screen captures is preferable to partial ones. Finally, we conclude that documentation designed to expedite the execution of tasks does not necessarily hamper the learning that may result.