Effect of dependency and tail behavior on a probability inequality occurring in modeling cognitive processes

A central idea in modeling performance in cognitive tasks is dynamic competition among processes in separate channels, known as "race model". This model implies a certain inequality between associated probability distributions under rather general conditions. The inequality represents an important empirical test of the race model, but its usefulness is limited since it requires the assumption of stochastic independence between the channels. Using the stop signal paradigm as reference, we investigate more general forms of stochastic dependence that still imply the inequality using the concepts of copula and heavy-tailed marginal distributions.