In this test, the criterion drift method gives two score differences more than their recorded ds000005 values.

This fixed-design simulation tested whether choice-independent criterion drift can produce two recorded response-history score differences from OpenNeuro ds000005. In 500 simulated cohorts at each of nine settings, 24 cohorts at stationary criterion SD 2 logit units and persistence 0.99 exceeded both recorded differences (4.8%; exact 95% Monte Carlo interval 3.099316879350537% to 7.058079625732981%). Removing participant sub-05 in a post hoc check reduced this count to 7/500. The result shows compatibility under this conditional simulation model; it does not establish a human mechanism. The analysis used 16 participants, 4,096 trials, and fixed missing-response masks. Parameter recovery, a direct response-update comparison, and empirical assessment of the drift settings were not performed.

criterion drift · response history · decision making · mixed gambles · logistic regression · simulation · OpenNeuro ds000005 · Monte Carlo