The 95% interval for feedback gain included zero in 84 ds005123 adults with run 1 to run 2 logistic prediction and penalty 1

We tested whether past displayed partner outcomes improved later trust choice prediction in OpenNeuro ds005123 version 1.1.3. Fixed logistic fits used run 1 and tested run 2 in 84 selected adults. The penalty was 1. The baseline used offers, personal choice tendencies, previous choices, button side, and trial progress. The gain from adding feedback was 0.000177 nats per choice. Its central 95% interval was -0.003856 to 0.002349. The interval used 5,000 paired person draws. Both models were fitted again for each draw. All five planned secondary intervals also included zero. The test did not establish lower loss from the specified feedback variable. This does not show absence of learning. Both runs occurred in one visit. The package contains fixed source files, a checked trial history, all fits and draws, and repeat instructions.

trust decisions · social feedback · behavioral prediction · logistic regression · OpenNeuro · bootstrap · fMRI dataset