A queue-series model for reaction time, with discrete-stage and continuous-flow models as special cases

This article presents a new reaction time model that includes both sequential-stage (discrete) and overlapping-stage (continuous-flow) models as special cases. In the new model, task performance is carried out by a series of distinct processing stages, each of which functions as a queue. A stimulus conveys 1 or more distinct components of information (e.g., features), and each stage can begin processing as soon as it receives 1 component from its predecessor. If a stimulus activates only 1 component, successive stages operate in strict sequence; if it activates multiple components, successive stages operate with temporal overlap. Within this class of models, experimental factors affecting different processing stages always have additive effects on reaction time with sequential stages but rarely do so with overlapping stages. Within this class of models, then, observations of factor additivity support discrete-stage models.