sažetak izlaganja sa skupa
Learning Objectives:
After attending this presentation, participants will gain knowledge about the differences in handwriting characteristics when writing with the dominant and non-dominant hands. They will also gain insight into the potential application of statistical models in identifying text written by a non-dominant hand.
Impact Statement:
This presentation will contribute to forensic handwriting examination by expanding knowledge on identifying disguised handwriting and emphasizing the importance of specific class and individual characteristics when analyzing handwriting samples written with the non-dominant hand. By implementing the statistical model developed in this study, forensic experts can improve their performance in identifying non-dominant handwriting, leading to more evidence-based conclusions.
Abstract Text:
Handwriting is a relatively stable and individualized trait, developed through years of practice and automatization. Because of this, individuals attempting to disguise their identity sometimes write with their non-dominant hand when committing criminal activities, such as fraud, threats, or anonymous letters. Although non-dominant handwriting is typically less controlled and skillful, it retains identifiable individual features, allowing forensic experts to link it to the writer.¹,²
In forensic handwriting examination, it is crucial to recognize such attempts and apply an appropriate analytical approach. This study aimed to explore within-individual differences in class and individual handwriting characteristics when using the dominant and non-dominant hand. Additionally, a statistical model was developed to distinguish between handwriting samples produced by each hand.
The study included 94 adult right-handed participants from Croatia and Bosnia and Herzegovina, who were asked to copy the sentence: “Ništa veliko neće biti postignuto bez velikih ljudi, a čovjek može biti veliki samo ukoliko je odlučan da to i postane.” Each participant wrote the sentence twice—once with their dominant hand and once with their non-dominant hand. The handwriting samples were analyzed based on 13 general (class) characteristics and 18 individual characteristics.³-⁵
Categorical handwriting features were compared using McNemar’s test, while features measured in clock positions were converted to angles and compared using paired samples t-tests. To develop a classification model, both handwriting samples from each participant were treated independently, forming a dataset of 188 texts. The dataset was split into training (70%) and testing (30%) subsets, with 5-fold cross-validation to ensure robustness. A classification model was built using logistic regression combined with recursive feature elimination to identify the optimal number and combination of predictive features.
Results showed that 7 out of 13 class characteristics exhibited significant differences (p < 0.05) between dominant and non-dominant writing, including overall handwriting quality, legibility, neatness, spacing, and fluency. Among 18 individual characteristics, only 5 features were significantly different (p < 0.05): T higher cross stroke, E loop, N arched, S loop, and O angle. The classification model achieved 89% accuracy in cross-validation and 91% accuracy in the test set. It demonstrated a Positive Predictive Value (PPV) of 0.85, meaning that when the model identified handwriting as non-dominant, there was an 85% chance it was correct. The Negative Predictive Value (NPV) of 1.00 indicated that all handwriting classified as dominant was correctly identified.
The findings suggest that class characteristics may be sufficient to distinguish between dominant and non-dominant handwriting while maintaining the writer’s individuality. This means that after identifying a text as potentially written with the non-dominant hand, further analysis of class and individual characteristics can still provide valuable insights. However, specific caution should be exercised when interpreting characteristics that showed significant variation. Since this study only included right-handed individuals, future research should expand to left-handed participants to validate the model across a broader population and strengthen its generalizability.
Handwriting; Questioned Documents; Regression Analysis; Forensic Document Examination; Disguised Writing
Razvojem virtualizacijskih tehnologija raste kompleksnost IT sustava, dok korisnici istovremeno zahtijevaju brži pristup resursima, napredne integracije i jednostavnost korištenja. Jedno od ključnih rješenja za ove izazove primjena je…
Ovaj rad istražuje i uspoređuje mogućnosti prosljeđivanja grafičkih kartica na Hyper-V i KVM hipervizorima prema virtualnim strojevima, detaljno opisujući metode, postupke i razlike u arhitekturi koje utječu na…
The ever-growing usage and adaptation of HTTP/3 in web traffic leverages Quick UDP Internet Connections (QUIC) as a transport protocol. QUIC is widely recognized for reduced latency, enhanced…
