Assessing Validity of the UCR Standardized English Automatic Oral Test in the MEP Monitoring Test
Guardado en:
| Autores: | , , |
|---|---|
| 格式: | artículo original |
| 狀態: | Versión publicada |
| Fecha de Publicación: | 2026 |
| 實物特徵: | The Program of the Evaluation of Foreign Languages (PELEx) from the University of Costa Rica is at the forefront of the application of automatic speech recognition (ASR) and artificial intelligence (AI) in the assessment of foreign languages within Latin America (LATAM), creating a need for continual research, data collection, and reliability assessment to collect further evidence of validity for these tests. This paper compares the results obtained during the application of Costa Rica’s Ministry of Public Education (MEP) Monitoring Test in 2023 by PELEx of the University of Costa Rica (UCR) with those of 7,454 Costa Rican public high school students through a statistical sampling of the tests, compared with the grading of a certified human tester from PELEx, to produce evidence of the test’s concurrent validity—specifically, the degree of agreement between human and AI scores. It offers an analysis of the ratings as well as recommendations for further validation research and continued collection of evidence of construct validity and scoring reliability for the AI-based oral exam. |
| País: | Portal de Revistas UCR |
| 機構: | Universidad de Costa Rica |
| Repositorio: | Portal de Revistas UCR |
| 語言: | Inglés |
| OAI Identifier: | oai:portal.revistas.ucr.ac.cr:article/8575 |
| 在線閱讀: | https://revistas.ucr.ac.cr/index.php/rrlm/article/view/8575 |
| Palabra clave: | Artificial Intelligence Foreign Language Evaluation Test Validity Speaking Assessment Inteligencia Artificial Evaluación de Lenguas Extranjeras Validez de Pruebas Evaluación del Habla |