Letter Tracing: a Serious Game to Teach Handwriting and Assess Proficiency through Machine Learning.

Area of Science:

• Pediatric occupational therapy
• Educational technology
• Developmental psychology

Background:

• Handwriting is a crucial skill for academic success.
• Assessing handwriting development and identifying delays can be challenging.
• Digital tools offer potential for objective assessment and intervention.

Purpose of the Study:

• To introduce and evaluate the "Letter Tracing" serious game for teaching letter formation.
• To assess the game's ability to capture the evolution of handwriting abilities and proficiency in children.
• To utilize machine learning for predicting grade level and handwriting delay risk.

Main Methods:

• Co-design of the "Letter Tracing" game with clinicians and technicians.
• Testing with 35 first- and second-grade children (9 first-graders, 26 second-graders).
• Children traced letters in a gamified environment; indicators from raw data trained machine learning models.

Main Results:

• Machine learning models achieved 71% accuracy in predicting grade level.
• Models also achieved 71% accuracy in predicting the risk of handwriting delay.
• The game effectively modeled the maturation of children's handwriting skills.

Conclusions:

• The "Letter Tracing" game shows promise as an effective training and assessment tool.
• It can objectively measure handwriting development and identify potential delays.
• This serious game offers a novel approach to supporting children's handwriting proficiency.