APPLICATION OF AI IN GRADING NEW QUESTION TYPES IN THE 2025 VIETNAMESE HIGH SCHOOL GRADUATION EXAM

Tran Quang Hieu, Nguyen Thị Phuong, Tran Thi Thu

Abstract


The 2025 Vietnamese National High School Graduation Exam introduces a new three-format structure (multiple choice, True/False, short answer), while still relying on traditional OMR answer sheets that often lead to mechanical errors. This study proposes an AI-assisted grading system combining a direct-answer sheet format with Mathpix OCR handwriting recognition to address these limitations. Using a five-stage simplified Borg & Gall model—regulatory analysis, system development, expert validation, dual-phase trials, and pilot testing on 93 student papers—the system achieved 98.5% character recognition accuracy and high user satisfaction (Likert ≥ 4.0/5) in reducing errors and supporting all question types. Findings show that AI-enhanced direct-answer formats maintain OMR-level reliability while improving test validity. A proposed implementation roadmap includes expanding handwriting datasets, standardizing 300 dpi devices, province-wide teacher training, and integration into the official grading process.

 

Article visualizations:

Hit counter


Keywords


AI-based grading, handwriting recognition, high school examination, direct-answer format, assessment validity

Full Text:

PDF

References


Afifi, M., & Hussain, K. F. (2019). The achievement of higher flexibility in multiple-choice-based tests using image classification techniques. International Journal on Document Analysis and Recognition (IJDAR), 22(2), 127–142. https://doi.org/10.1007/s10032-019-00322-3

Agarwal, S., Varun, M., & Prabakeran, S. (2023). OMR reader info scanner. 2023 9th International Conference on Advanced Computing and Communication Systems (ICACCS), 1, 205–209.

Aloisi, C. (2023). The future of standardised assessment: Validity and trust in algorithm s for assessment and scoring. European Journal of Education, 58(1), 98–110. https://doi.org/10.1111/ejed.12542

Ascencio, H. E., Pena, C. F., Vasquez, K. R., Cardona, M., & Gutierrez, S. (n.d.). Automatic Multiple Choice Test Grader using Computer Vision. https://doi.org/10.1109/mhtc52069.2021.9419920

Borg, W. R., & Gall, M. D. (1984). Educational research: An introduction. British Journal of Educational Studies, 32(3).

Caraeni, A., Scarlatos, A., & Lan, A. (2024). Evaluating GPT-4 at Grading Handwritten Solutions in Math Exams. ArXiv.Org. https://doi.org/10.48550/ARXIV.2411.05231

de Elias, E. M., Tasinaffo, P. M., & Hirata Jr, R. (2021). Optical mark recognition: Advances, difficulties, and limitations. SN Computer Science, 2(5), 367.

Gardner, J., O’Leary, M., & Yuan, L. (2021). Artificial intelligence in educational assessment: ‘Breakthrough? Or buncombe and ballyhoo?’ Journal of Computer Assisted Learning, 37(5), 1207–1216. https://doi.org/10.1111/jcal.12577

Hien, N. T., & Toai, D. B. (2024). The Study of the Development and Changes in Vietnam’s University Entrance Examination System Over Forty-Five Years. World Journal of Educational Research, 11(6), p.80. https://doi.org/10.22158/wjer.v11n6p80

Jain, V., Malik, S., & Bhatia, V. (2022). Robust image processing-based real-time optical mark recognition system. 2022 IEEE 6th Conference on Information and Communication Technology (CICT), 1–5.

Mahmud, S., Biswas, K., Alam, A., Al Mamun Rudro, R., Anannya, N. J., Mouri, I. J., & Nur, K. (2024). Automatic Multiple Choice Question Evaluation Using Tesseract OCR and YOLOv8. 246–252. https://doi.org/10.1109/cai59869.2024.00054

Ministry of Education and Training. (2024). Quyết định số 764/QĐ BGDĐT: Quy định về cấu trúc định dạng đề thi Kỳ thi tốt nghiệp THPT từ năm 2025.

Ministry of Education and Training. (2025). Công văn số 1239/BGDĐT QLCL: Hướng dẫn một số nội dung tổ chức Kỳ thi tốt nghiệp THPT năm 2025.

Muangprathub, J., Shichim, O., Jaroensuk, Y., & Kajornkasirat, S. (2018). Automatic Grading of Scanned Multiple Choice Answer Sheets. International Journal of Engineering & Technology, 7(2.23), 175. https://doi.org/10.14419/ijet.v7i2.23.11910

Muller, D., Calhoun, E., & Orling, R. (1972). Test Reliability as A Function of Answer Sheet Mode. Journal of Educational Measurement, 9(4), 321–324. https://doi.org/10.1111/j.1745-3984.1972.tb00964.x

Ragolane, M., Patel, S., & Salikram, P. (2024). AI Versus Human Graders: Assessing the Role of Large Language Models i n Higher Education. Asian Journal of Education and Social Studies, 50(10), 244–263. https://doi.org/10.9734/ajess/2024/v50i101616

Rodrigo, T. C., F., Z., Rogério, N., & J., A. Q.-G. (2016). An application for automatic multiple-choice test grading on Android.

Scarfe, P., Watcham, K., Clarke, A., & Roesch, E. (2024). A real-world test of artificial intelligence infiltration of a university's examination system: A “Turing Test” case study. PLOS ONE, 19(6), e0305354. https://doi.org/10.1371/journal.pone.0305354

Tabassum, K., & Rahman, Z. (2024). Optical Mark Recognition with Object Detection and Clustering. 2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT), 352–357.

Tinh, P. D., & Minh, T. Q. (2024). Automated Paper-based Multiple Choice Scoring Framework using Fast Object Detection Algorithm. International Journal of Advanced Computer Science & Applications, 15(1).




DOI: http://dx.doi.org/10.46827/ejes.v12i4.6007

Refbacks

  • There are currently no refbacks.


Copyright (c) 2025 Tran Quang Hieu, Nguyen Thị Phuong, Tran Thi Thu

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2015-2026. European Journal of Education Studies (ISSN 2501 - 1111) is a registered trademark of Open Access Publishing Group. All rights reserved.

This journal is a serial publication uniquely identified by an International Standard Serial Number (ISSN) serial number certificate issued by Romanian National Library (Biblioteca Nationala a Romaniei). All the research works are uniquely identified by a CrossRef DOI digital object identifier supplied by indexing and repository platforms. All authors who send their manuscripts to this journal and whose articles are published on this journal retain full copyright of their articles. All the research works published on this journal are meeting the Open Access Publishing requirements and can be freely accessed, shared, modified, distributed and used in educational, commercial and non-commercial purposes under a Creative Commons Attribution 4.0 International License (CC BY 4.0).