A stochastic version of a knowledge space is developed in which knowledge states are considered as possible epochs in a subject's learning history. It specifies how a subject is channeled through and progresses along a "gradation." The model's application to artificial data is described, based on maximum likelihood methods. (TJH)

A model is presented for the growth of knowledge reflected by 24 progress tests completed by approximately 600 students at the University of Limburg (Netherlands) Medical School. Based on the Rasch model, this model treats both the person's ability and the difficulty of the question as random variables. (SLD)