ABSTRACT
Consider a body of information in which knowledge can be assessed by asking solely binary questions—that is, a response to a question or item is only coded as correct or false. In such a case, it makes sense to represent the knowledge state of an individual by the set of all mastered items (i.e., all those items to which the individual is capable of reliably providing correct responses). 1 The collection of all such knowledge states is called the knowledge structure. This approach to knowledge representation was conceived by Doignon and Falmagne (1985) for the design of efficient procedures for the assessment of knowledge, where “assessing knowledge” means uncovering the actual knowledge state of some individual, and “efficient” means that knowledge can be uncovered by testing only a small subset of items. These procedures are not described here (see Falmagne & Doignon, 1988a, 1988b).
