The proportional odds test gives 10 Holm p-values less than 0.05 for ds000005 response code order 4, 3, 2, 1.

In 16 participants from OpenNeuro ds000005, an efficient score test compared gain and loss coefficients across three cumulative response thresholds under response code order 4, 3, 2, 1. Parametric bootstrap p-values with Holm adjustment were below 0.05 for 10 participants; the chi-square calculation gave the same participant set. The analysis used 1,999 simulations per participant and 3,942 recorded responses, with 154 missing responses omitted. Seven participants had no bootstrap exceedance, so their p-value of 0.0005 is the simulation resolution. The results depend on the assumed code order, linear cumulative-logit model, fixed coefficients, and independent responses. Monte Carlo error can change decisions near 0.05. The findings do not establish response strength, unequal loss/gain ratios, stable preferences, a neural effect, or biological causality.

proportional odds · efficient score test · parametric bootstrap · Holm adjustment · cumulative logit · mixed gambles · OpenNeuro ds000005