Statistical measures for the double factorial paradigm.

摘要

Much of scientific psychology and cognitive science can be viewed as a search to understand the mechanisms and dynamics of perception, thought and action. Two processing attributes of particular interest to psychologists are the architecture, or temporal relationships between sub-processes of the system, and the stopping rule, which dictates how many of the sub-processes must be completed for the system to finish. The Survivor Interaction Contrast (SIC) is a powerful tool for assessing the architecture and stopping rule of a mental process model. Despite its demonstrated utility, the measure has lacked a method for statistical testing. This paper demonstrates basic statistical properties of the measure and develop two statistical tests for use with the SIC, one null-hypothesis-significance test and one non-parametric Bayesian test. A third processing attribute that commonly arises in the study of cognition is the effect of varying the system workload. The capacity coefficient is a measure on response times for quantifying changes in performance due to workload. This paper demonstrates the basic statistical properties of the estimated cumulative hazard function of an unlimited-capacity, independent, parallel model, which is the basis of the capacity coefficient. Using these properties, I develop a null-hypothesis-significance test for the capacity coefficient. Each of the four methods for determining model characteristics are first developed theoretically and then explored using simulation. At the end of each section, the use of the test is demonstrated with data from a simple detection task.