Fig. 2.

Overview of method: A between-groups experiment is simulated whereby there simulated participants have decision parameters (drift and boundary separation) sampled from defined distributions. From these defined distributions, which contain either a true or null difference between groups, simulated participants are sampled. From these sampled participant-level parameters, simulated decisions are generated, using the DDM, which generates behavioral outcome variables (RT and accuracy). For each participant, these outcome variables are modeled using the DDM to recover an estimated drift parameter. A test of group differences is then performed on the generated accuracy and RTs and on the estimated drifts. This is compared to the known difference in drift to categorize the test as correctly detecting a true difference between groups in participant discrimination (a hit), or incorrectly detecting a difference when there is none (a false alarm). Over many simulated experiments, and a range of parameter values for simulated sample size and size of true group differences in drift, the average probability of a hit and a false alarm, and the sensitivity (d’) are calculated