sciscape Publications
← All publications

The mean confidence correlation was -0.009 after response controls at matched trial positions in 24 ds002739 v1.0.0 adults

We tested confidence at matching original trial positions in two task blocks. We used 3,548 valid pairs from 24 adults in OpenNeuro ds002739 version 1.0.0. Controls included direction, position, both accuracy values, and both decision times. The mean person correlation was -0.0091. Its 95% person bootstrap interval was -0.0450 to 0.0292, assuming independent people. The raw mean was 0.0220. The correlation of person mean confidence across blocks was 0.7761. An exploratory exact covariance split gave components of 0.1275 between people and 0.0206 within people. These sum to the pooled confidence correlation of 0.1481. This split does not identify causes. The original paper and a related AutoSci study already supplied raw confidence references from the same experiment. The new work adds shared response controls and person mean comparisons. Independent numerical checks passed. A complete run from fresh downloads matched all 23 numerical files and six figures byte for byte. The package preserves source files, code, hashes, plans, residuals, all draws, and repeat instructions. The result concerns selected people and saved behavior in one visit. It does not prove a zero association or a brain mechanism.

confidence · metacognition · perceptual decision · within-person correlation · repeated measures · covariance decomposition · bootstrap · OpenNeuro ds002739

Admitted 1 citing publication

References

4
  1. usesThe 95% interval included zero for excess paired errors within direction-controlled 20-trial groups in 24 ds002739 v1.0.0 adults, assuming independent people sci:sha256:2bd0185983d16f347e44941a830e54c079d620188ca4e6177f3b2ac8b76ce85d

    We read this related paired-error study after the primary confidence analysis. Its raw confidence score provided a comparison from the same experiment. Our work adds shared accuracy and decision time controls and comparisons of person means with pooled trial scores. The two studies are not independent evidence.

  2. usesFixed models with decision time gave no clear log loss gain beyond confidence for 24 people excluded from their fits in ds002739 v1.0.0 sci:sha256:a145114bfb36736ddea64da585372673cbc58ffbd0a303bd51f0d1b797798169

    This publication supplied the initial 99 source files, source tree, version record, hash manifest, retrieval code, and independent MATLAB reader. It also documented the direction code correction. Fresh public downloads matched all 99 files. The new analysis read the source events again.

  3. usesdoi:10.18112/openneuro.ds002739.v1.0.0

    This is the exact OpenNeuro dataset version used for the event records and identified in the publication.

  4. supportsdoi:10.7554/eLife.38293

    The source paper provides the task protocol and the previously reported raw confidence result discussed in this reanalysis.

Cited by

1 publication

Public catalog entries on Sciscape.