Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0889
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-08-26T06:04:52.646727Z
status: published
lens: g_space
sector: crime
headline: Forensic analysis of dominant versus non-dominant handwriting: Statistical and machine learning insights
dek: Systematic, paired quantitative evidence on how handwriting changes when the non-dominant hand is used remains limited in forensic document examination, despite its frequent relevance in cases involving disguise and authorship concealment. This study presents the first large-scale within-writer analysis examining general and individual handwriting characteristics using statistical testing and predictive modeling. Handwriting samples were collected from 94 right-handed participants, each of whom produced the same…
gain_title: Supervised logistic regression distinguished dominant versus non-dominant handwriting with 92.98% test accuracy using execution-quality features, providing quantitative benchmarks to support forensic evaluation of suspected off-hand disguise.
problem_title: (none)
trace_subject: (none)
gain_reading: Supervised logistic regression distinguished dominant versus non-dominant handwriting with 92.98% test accuracy using execution-quality features, providing quantitative benchmarks to support forensic evaluation of suspected off-hand disguise.
gain_evidence: A model using 10 handwriting characteristics achieved a test accuracy of 92.98%
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers collected paired samples from 94 right-handed participants who each wrote the same standardized text with dominant and non-dominant hands, then scored 13 general and 19 individual characteristics. They found statistically significant differences in 61.5% of general and 57.9% of individual characteristics and trained a supervised logistic regression model to distinguish hand use.

The work matters for forensic document examination because it provides paired quantitative benchmarks for which features degrade with non-dominant writing versus which remain stable, potentially aiding assessment of disguise. Uncertainty remains because the sample was restricted to right-handed writers and controlled text, leaving open how models perform on left-handed writers, naturalistic disguise, or varied writing conditions.
limitation: Findings are limited to right-handed writers producing the same standardized text in a controlled collection, not varied real-world disguise scenarios or left-handed writers.
tag: Evidence-backed gain
key_points: Within-writer paired design with 94 right-handed participants writing identical standardized text with both hands. | 13 general and 19 individual characteristics evaluated; 61.5% of general and 57.9% of individual characteristics showed statistically significant differences. | General features most affected included writing speed, graphic maturity, slant stability, letter connectivity, neatness, and legibility. | Reduced model retained 91.23% accuracy, indicating execution-quality features carry strongest diagnostic value.
rundown: The study evaluated handwriting using established forensic criteria, finding structural letter forms showed high concordance across hands while fine motor features like loop formation and diacritic construction differed, supporting persistence of individuality despite altered hand use.

Predictive modeling used cross-validation and feature selection; the 10-feature model reached 92.98% test accuracy and a reduced model reached 91.23%, pointing to execution-quality features as most diagnostic for hand-use classification.
sources:
- peer_reviewed | Journal of Forensic Sciences | https://doi.org/10.1111/1556-4029.70456 | 2026-08-25
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

4f5c688200f09cc133903d9e7871b666bfe354155413b8ee6a4bdca29e225b9e

previous

0000000000000000000000000000000000000000000000000000000000000000