Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
INTRODUCTION Discovering what the users of a computer software system do know and should know are important goals in current research on human-computer interaction. Research on the kinds of knowledge people have when they use computers, including the concept of a mental mode} of the system, is one of the major topics that is bringing the field of human-computer interaction from the tra- dition of human factors closer to that of experimental/cognitive psychology. Traditional human factors work has focused principal attention on behavior and performance itself, and has avoided the problem of describing the conceptual causes and effects of that behavior. On the other hand, while academic cognitive psychol- ogy does concern itself with theoretical interpretations of mental processes, it has focused on narrowly restricted mental processes, such as particular aspects of learning, memory, problem solving, or planning, and has studied them in the context of highly controlled and contrived laboratory tasks. The study of knowledge represen- tations of users of computer-based systems affords an opportunity to explore both the theoretical base of behavior as well as specific behaviors in tasks that involve many different cognitive processes in concert. Because a number of researchers are concerned with mental representations, and because this topic has an impact on cognitive psychology and software human factors, there is an emerging need to clarify the concepts underlying knowledge representation and mental models as they apply to human-computer interaction. We intend to fill this need by reviewing relevant current research and presenting a preliminary framework of the kinds of mental representations of procedures people might have. MODELS OF WHAT, HEED BY WHOM? Several key distinctions need to be recognized in discussing mental representations and mental models in human-computer in- teraction. For example, various individuals are concerned with using or designing a piece of software, and they hold different conceptions of it. These individuals include the user, the software 3
engineer, the human factors analyst, and the cognitive psycholo- gist. Furthermore, there are different aspects of the system to be known: the task, knowing what the goal is and in general what subtasks need to be accomplished to achieve the goal; the system interface, knowing how to accomplish the sequence of subtasks in this system, given the data presentation and interaction languages of this system; and the system architecture, knowing the way the data are stored, the internal processes the interactions invoke, and in general how the system works. Confusion has surrounded the term mental mode! because different authors have referred to different owners of the models (the user, the software engineer, etc.) and are not clear as to what the mode] actually represents (the task, the architecture, etc.~. For example, some researchers and human factors analysts acknowledge that it is important to know the way users themselves are built and work, what their memory limits are, their common strategies in problem solving, their individual differences, and so on, in order to build useful, usable software. A system that requires the user to remember a list of 100 codes that represent areas of the country or the types of transactions that are required (as in some airline or automobile reservation systems) is predictably difficult because our mode! of the user includes a long-term memory that is confused by similar meaningless items. These researchers have sometimes used the term mental mode' to refer to the model that they, as researchers, have of the user's mental architecture. Similarly, software engineers have ideas about what the user wants to do and how the system itself is structured that dictate how they will program the system and how it will operate to serve the users' needs. Engineers have mental models of their design. This highlights another distinction, that between descriptive and prescriptive representations. Researchers want to be able to analyze what the user currently knows so they can explain why he or she is having difficulty, which aspects are learned and which are confused, and so on. In this case, they are using a descriptive model, one that tells us what the user knows. Designers, however, want to construct a mode! of what the user should know. This representation could be used to analyze, for example, whether a proposed system will be too difficult to learn or where the errors might be. And, in designing commands and screen presentations, designers would like to invoke a mode} in the user that fits the dialog; they would like to get the user to build a mental model 4