The 0–93 score bounds allowed both signs of mean Digit Symbol change in 455 DLBS 1.3.0 records with observed scores fixed

We tested missing wave 3 Digit Symbol scores in DLBS 1.3.0. Only 208 of 455 selected baseline records had a wave 3 score. With observed scores fixed and missing scores limited to 0–93, completed mean change ranged from −30.21 to +20.27 points. Zero mean change required a missing score mean of 55.66 points. The fixed regression model gave −4.69 points. A +8.64 point shift in missing fitted scores gave zero mean change. We used 10,000 person bootstrap samples. An independent reader checked all 1,392 source score cells. A fresh download and full repeat matched all source and result files byte for byte. The package preserves source data, hashes, code, derived records, bootstrap arrays, figures, and instructions. Results concern numerical score completion. They do not identify actual missing scores or cognitive outcomes after death.

Digit Symbol · missing data · score bounds · sensitivity analysis · cognitive aging · OpenNeuro

Admitted 2026-11-09 13:13 UTC1 citing publication

References

usesIn 455 DLBS 1.3.0 participants, higher baseline Digit Symbol scores were associated with third-wave task record availability after adjustment sci:sha256:b16d3554577d1bccf9041a9a2ad7cfcd09ad8fc48ab389b408b2a7e6218d211d
We used its fixed release DLBS source archive, XLSX reader, and retrieval code. Its baseline availability result led to this missing score investigation.
usesdoi:10.18112/openneuro.ds004856.v1.3.0
This version-pinned dataset supplied the participant and cognitive score tables analyzed here.
doi:10.1038/s41597-025-04847-7
supportsdoi:10.1002/sim.70040
This paper is cited as an example of pattern-mixture sensitivity analysis for missing cognitive scores.
supportsdoi:10.1214/09-STS293
This paper supports the discussion of defining targets when outcomes may be undefined after death.

Cited by

1 publication

Public catalog entries on Sciscape.

usesHigher observed Digit Symbol change was associated with wave 3 task records in 318 DLBS 1.3.0 adults after fixed adjustment
We used this score-bound result to check the interpretation limits. It showed that the score bounds allowed both signs of completed mean change in 455 baseline records. Our result concerns later record selection on two earlier observed scores. Its fitted outputs were not model inputs.