The 95% interval included zero for excess paired errors within direction-controlled 20-trial groups in 24 ds002739 v1.0.0 adults, assuming independent people

We tested errors at matching trial positions in two task blocks. We used 24 adults from OpenNeuro ds002739 version 1.0.0 and 3,548 valid pairs. The primary score controlled motion direction and fixed groups of 20 original trials. Mean excess paired errors were 0.137 percentage points. The 95% person bootstrap interval was -0.241 to 0.530 percentage points, assuming independent people. The interval included zero. This result does not establish zero agreement. Without these controls, excess paired errors were 1.166 percentage points. An exploratory exact identity divided that score into direction, local order, and residual terms. Their means were 0.814, 0.215, and 0.137 percentage points. These terms do not show causes. A fixed 40-trial check had a positive, unadjusted interval. The result thus depended on the local group rule. The package contains a checked pair dataset, source files, code, provenance, all random scores, figures, and instructions. Independent readers and calculations passed. A complete run from fresh public downloads reproduced all 15 numerical result files byte for byte. The results concern source-selected adults and saved paired behavior during one visit. They do not identify a brain mechanism.

perceptual decision-making · repeat errors · confidence · motion direction · trial order · bootstrap interval · OpenNeuro