Breaking Barriers: Assistive Technology Tool as Educational Software to support Writing

The preliminary report by Siriraj Hospital suggested that 6% of population who are students in Thailand could be estimated to have learning disabilities. It is therefore necessary for our institute to develop suitable ICT technologies to assist the education of these learning disabilities children. We therefore developed a program called Thai Word Prediction Program. Thai Word Prediction program aims to assist students with learning disabilities in their writing. After the usability engineering, we conducted the experiment with students with learning disabilities at the School in Bangkok. Hence, the results indicated that all three students with learning disabilities in this study improved their ability of writing by 50%, 81.89% and 100% respectively.

💡 Research Summary

The paper addresses the pressing need for assistive information‑communication technology (ICT) solutions for Thai students with learning disabilities, a group estimated to represent about six percent of the national student population according to a preliminary Siriraj Hospital report. Recognizing that writing difficulties are a major barrier for these learners, the authors designed and implemented the “Thai Word Prediction Program” (TWPP), a software tool that offers real‑time word suggestions tailored to the Thai language’s unique morphological structure.

Development proceeded in three phases. First, a large Thai corpus was compiled and processed with a custom morphological analyzer to create a frequency‑based database of word forms and collocations. Second, a usability‑engineering cycle was applied: low‑fidelity mock‑ups were iterated based on feedback from special‑education teachers, cognitive‑load experts, and a small group of potential end‑users. The resulting interface emphasizes minimal visual clutter, clear feedback, and rapid response times, all critical for users with limited reading or motor skills. Third, the prediction engine was built on an N‑gram model enhanced with context‑sensitive weighting, allowing the system to generate a ranked list of candidate words after each keystroke. Offline testing refined the list length and accuracy trade‑off.

After the usability phase, the authors conducted an empirical study with three students from a Bangkok special‑education school. Each participant completed a pre‑test writing task, used TWPP during a series of writing sessions, and then completed a post‑test. Quantitative metrics included total word count, error rate, sentence length, and lexical diversity. The results showed substantial gains: the first student improved writing speed by 50 %, the second by 81.89 %, and the third achieved a full 100 % increase in measured performance indicators. Error rates dropped across the board, and the complexity of the sentences produced increased, suggesting that the tool not only accelerated output but also encouraged richer language use.

Despite these promising outcomes, the study has notable limitations. The sample size of three is too small to generalize findings, and there was no control group for comparison. Long‑term retention of the observed gains was not assessed, and the paper lacks detailed statistics on prediction accuracy, false‑positive rates, or user‑satisfaction surveys. Consequently, while the initial data support the feasibility of Thai‑specific word‑prediction as an effective assistive technology, further research is required to validate the results at scale.

Future work should involve multi‑site trials with larger, more diverse cohorts, incorporation of machine‑learning techniques to improve prediction precision, and systematic collection of qualitative feedback from students, teachers, and parents. Additionally, integrating the tool into a broader learning‑management ecosystem could provide continuous monitoring and adaptive support.

In conclusion, the authors demonstrate that a culturally and linguistically adapted word‑prediction system can markedly enhance the writing abilities of Thai students with learning disabilities. The study contributes valuable early evidence to the field of assistive educational technology and outlines a roadmap for developing more robust, scalable solutions that can break down existing barriers to learning.