School-based dataset for children's low back pain prevention, global trunk mobility and trunk function

This school-based dataset contains anthropometric measurements, global trunk (thorax-segment) range of motion, peak angular velocities, trunk extensor endurance, isometric deadlift force and task-related pain ratings from 226 children and adolescents aged 9–16 years.

Version 3 preserves the 34 measurement and descriptor fields of version 2 and adds a variable-specific QC_Flags column. It provides updated metadata, a flag dictionary, the historical quality-control log, a reconstruction provenance ledger, a sensitivity dataset, reconstructed processing code and reproducible validation results.

Mobility measurements describe global thorax-segment motion and may include hip and pelvic contributions; they do not isolate spinal motion. Speed variables represent peak angular velocity. Additional screening includes ROM/peak-velocity duration lower bounds, two-sided statistical fences and constant-increment or repeated-value sequences. Review flags indicate observations requiring caution and do not establish measurement error.

Original device/session records and the original processing scripts are no longer available. The reconstructed pipeline reproduces the documented historical transformations and revision analyses but does not independently validate the original measurements. Six historical replacements for participant 1 remain unverified and are explicitly flagged; a separate sensitivity dataset treats those six cells as missing. Participant-1 BMI is independently recalculable from body mass and stature. The 24 measurement cells previously set to missing remain missing.

The primary dataset is scientific_data_dataset_v3.csv. File contents, processing instructions and limitations are described in README_v3.txt. Earlier versions remain available for provenance; version 3 is recommended for new reuse, with explicit consideration of the quality flags and sensitivity dataset.