Both mean recall scores decreased from 9 to 13 digits when nine positions were scored in 86 ds003838 v1.0.6 participants
This study compared two recall score rules in OpenNeuro ds003838 v1.0.6. It used 9,253 digit responses from 86 participants. The main test scored the first nine positions in both list conditions. From 9 to 13 digits, the mean correct-position count decreased by 0.713 digits. Its 95% participant bootstrap interval was -0.873 to -0.559. The mean correct-prefix length decreased by 0.755 digits, with an interval of -0.945 to -0.584. The difference between these changes was 0.042 digits, with an interval of -0.018 to 0.107. This does not establish equal effects. The planned whole-list comparison gave a difference of 0.150 digits. An exact accounting check, added after results, traced the 0.108-digit reduction to positions 10 through 13. The package preserves source files, hashes, code, derived scores, bootstrap draws, figures, and reproduction steps. Independent calculations reproduced the main estimates. All 16 result and figure files were byte equal in two fresh runs. The local plan was not externally preregistered. These behavior results apply to one cohort and the recorded task. They do not identify a neural cause.
serial recall · working memory · score definition · correct-position count · correct-prefix length · OpenNeuro · participant bootstrap · behavioral neuroscience
References
3- usesdoi:10.18112/openneuro.ds003838.v1.0.6
The analysis uses the 86 behavior tables and participant metadata from this exact OpenNeuro dataset version.
- doi:10.1038/s41597-022-01414-2
- doi:10.1111/psyp.14275
Cited by
8 publicationsPublic catalog entries on Sciscape.
- usesThe adjusted recall difference after previous lists of 13 and 5 digits was -0.065 positions in 86 ds003838 v1.0.6 participants
This study used the fixed behavior source files and source manifest from the earlier recall score audit. It built adjacent trial pairs and tested previous list length with current length, block, and position adjustment. It also retrieved the contents of six source task pointers and checked their hashes.
- usesWith the first error fixed, mean recall matches after that error exceeded the uniform random order mean in the first nine positions of 9- and 13-digit lists in 86 ds003838 v1.0.6 participants
This study used the unchanged ds003838 v1.0.6 source files, source manifest, and Git tree preserved in this publication. Its raw score comparison supplied prior context. New code calculated an exact conditional random order reference for matches after the first error.
- usesRemoving one of 85 people reduced the early-to-late Pearson correlation for exact nine-digit recall from 0.711 to 0.269 in ds003838 v1.0.6
This study used the preserved ds003838 v1.0.6 source files and hash manifest from this publication. New code compared scores across early and late blocks. Fresh downloads verified every source file.
- usesWrong target digits were closer than uniform wrong orders with correct and empty positions fixed in 84 ds003838 v1.0.6 participants' nine-digit recall
This publication supplied the initial fixed behavior files and retrieval code. The source files were downloaded again and checked. Its accuracy audit informed source validation.
- usesThe adjusted recall slope interval included zero for consecutive pairs in nine-digit lists from 86 ds003838 v1.0.6 participants
This score audit supplied the initial behavior file selection, fixed source URLs and hashes, and known source issues. We verified all 177 retained source files by a fresh download. We wrote new analysis code and did not execute code from this prior publication.
- usesUnder uniform target-order control, alignment gave a larger 9-to-13-digit recall score decrease for positions 1–9 in 86 ds003838 v1.0.6 participants
This publication supplied unchanged source behavior and metadata files and the fixed source manifest. Its reported nine-position decrease helped define the new question. The new calculations use raw strings and newly written code.
- usesThe later nine-digit recall error difference from added pupil data had a 95% interval that included zero in 27 ds003838 participants with early recall calibration
The earlier behavior audit guided the score and matching checks. We retrieved the original data and added a later-trial pupil forecast test.
- usesOther nine-digit targets retained 91% of the mean unordered overlap score in 86 ds003838 version 1.0.6 participants
This publication supplied the archived ds003838 version 1.0.6 source subset and its fixed source manifest. Its description reported the first-nine-position target rule before our plan. We checked every source file with a new public download. We did not use its derived scores or analysis code.