We review experimental research on a basic question for document designers: what information should a text contain and to what extent should that information be elaborated? Obviously, there can be no absolute answers to questions of content. The answers must depend on several factors, such as the writer's purpose, the readers' intentions and abilities (i.e., their reasons for reading and their prior familiarity with the subject matter), general human capacities for acquiring information or skills, in addition to conventional constraints on the form of the text.
The general strategy for our research has been to produce several versions of a computer manual that differ in systematic ways. We ask participants in the studies (generally college students) to read the manuals. After removing the manuals, we ask the participants to demonstrate what they have learned by performing a set of tasks on the computer. By comparing the performance of participants who read the different versions of the manual, we can draw inferences about the characteristics of the manuals that led to better performance. We have employed experienced computer users as well as computer novices, readers who open a manual with a particular task in mind as well as readers who have no particular agenda. We found that students learned most from manuals in which only certain classes of information were elaborated. Participants did not benefit from elaborations of general concepts (e.g., what is a disk drive) or from elaborations offering advice on when to apply specific procedures. However, they did benefit from elaborations on how to apply procedures, in particular, from well-chosen situational examples. We found that problem solving exercises that forced the learner to apply the instruction offered were more helpful than tutorials or "discovery" tasks. We also found that while some task-oriented readers can get by with very little elaboration, readers with less specific goals benefit from greater elaboration, particularly concerning the correct execution of procedures.