Recall amount had a higher correlation between odd and even lists than the order score in 135 selected ds004789 people
We compared recall amount with a recall order score in OpenNeuro ds004789, version 3.1.0. We used the lowest numbered session for each person. The primary sample had 135 people. Each person had at least five eligible odd lists and five eligible even lists. Each eligible list had at least three distinct correct recalls. The Pearson correlation between odd and even list means was 0.755 for amount and 0.340 for order. The difference was 0.415. Its paired person bootstrap 97.5% interval was 0.197 to 0.630. All eight planned sensitivity checks kept this direction. The mean order score was 0.665. The exact random-order mean was 0.5 when the recalled set and first item were fixed. The mean excess was 0.165, with a 97.5% interval from 0.141 to 0.190. Bootstrap intervals assume independent source people and fixed session and list selection rules. The result applies to this selected clinical source sample and score. It does not establish stability across sessions or a neural cause. The package preserves 551 source files, pinned URLs and hashes, new code, all selection decisions, and all 60,000 bootstrap draws. Independent checks passed. A clean environment gave equal bytes for all 27 numerical result files.
free recall · recall order · recall amount · temporal clustering · measurement agreement · bootstrap · OpenNeuro ds004789
References
3- usesWord terms with penalty 10 reduced recall log loss in one fixed ds004789 split of people. The refit interval included zero. sci:sha256:d483dcbb7143780258386630408fcc95725985af5fa5b08aa8049dc31bd9a7aa
This publication supplied the archived ds004789 event files, dataset metadata, full Git tree, source manifest, and lowest numbered session rule. We wrote new parsing and analysis code. We did not use its fitted models or prediction results.
- usesdoi:10.18112/openneuro.ds004789.v3.1.0
Source dataset, version 3.1.0; the publication analyzed its behavioral event records.
- supportsdoi:10.3758/BF03197276
Background evidence for temporal clustering of recalls from nearby study positions.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesForward recall exceeded a uniform order reference with the recalled set, first recall, and total absolute distance fixed in 135 people from ds004789 v3.1.0
This publication supplied the pinned source files, provenance, and common.py, prepare.py, and retrieve.py. We used its record rules and primary group rule. We wrote new exact direction methods and analyses.
- usesPreviously recalled words exceeded a uniform list and broad position reference in recall errors from 238 selected ds004789 people
Reused its 273 event tables, 273 sidecars and five dataset files. Used its fixed source tree, version records and file hashes. Adapted accepted-list checks from its data audit. New code built recall-history links, reference counts and the exact-position model. Separate code checked all new results.
- usesGloVe cosine reduced recall choice loss in 68 selected ds004789 test people with a fixed word pool and ridge penalty
Used the 551 source files, source manifest, full session tree, and parser from this publication. These files supplied the study records and fixed session and person selection rules. The earlier order scores did not enter the new models.
- usesA recall lag 2 GloVe term increased mean test loss in 68 selected ds004789 people at fixed ridge 0.001
We used its source parser, lowest-session selection rule, source manifest, and list decisions through the later semantic package. We retained its code notice. The new models use source recall choices. They do not use its published order score as a predictor.
- usesOnly 28 of 118 selected ds004789 people had variable scores under a recall error order reference that kept source counts fixed
This package supplied the lowest numbered FR1 session archive for 273 people, the source manifest, and the pinned Git tree. The new study checked all source bytes and used them to test recall error order.