Two-Dimensional Analysis of Indigenous Language Learning Effectiveness: Proficiency Advancement and Within-Level Mastery

Abstract

Taiwan is an inclusive maritime nation characterized by rich linguistic and cultural diversity. Its Indigenous languages constitute the Austronesian language family. This study evaluates the learning effectiveness of an AI-based learning platform for Indigenous learners in Taiwan. Longitudinal learning data were extracted from Indigenous users from the Indigenous Language Audio Challenge platform. Utilizing quantitative methods, this research analyzed Indigenous learners’ assessment records stored within the system database. Logistic regression and Ordinary Least Squares (OLS) regression were employed to examine learning effectiveness through two lenses: proficiency advancement and within-level mastery. The findings suggest potential avenues for language education, cultural preservation, and technological innovation. The primary results indicate: (1) prior ability exhibits a stable effect on sustained Indigenous language learning; (2) advancement thresholds and performance disparities manifest across ethnic groups; and (3) school environments exert structural interference on learners’ mastery of Indigenous languages.