General Processing Tree (GPT) models (Hu and Batchelder, 1994), a class of multinomial statistical models, are employed widely in many domains of cognitive psychology, including memory, perception, logical reasoning, and psychometrics (Batchelder and Riefer, 1999). These models represent latent cognitive variables as discrete-state processes (vs. continuous) that occur in serial fashion (vs. parallel). This paper presentation generalizes the GPT class to handle continuous-state processes, through a recursive formulation of the discrete-state GPT model. The recursive GPT model will be illustrated using the Fuzzy Logic Model of Perception (e.g., Massaro, 1987), which, like the Bradley-Terry-Luce choice model (Luce, 1959) and the Rasch model (Rasch, 1960) of Item Response theory, represents cognitive states by a particular rational function of the parameters.
展开▼
