Simulating the variability of actual outcomes

A method is described for simulating the variability of outcomes encountered in a variety of settings. Outcome values falling within a specified range are randomly generated and presented across subjects and trials, so that, over time, the average values seen by all subjects in a condition are the same. Randomly varying reinforcement dimensions (such as magnitude) over trials for each subject is recommended to enhance both the internal and the external validity of laboratory results. In addition, it can be used to study the effects of the variability or uncertainty of outcome dimensions such as magnitude and duration in investigations of behaviors such as matching, melioration, and behavioral momentum