Exploring Factors Influencing Students’ Attitudes and Engagement in Duolingo-Based Mobile-Assisted Language Learning Using Decision Tree Analysis

Abstract

This study investigates the factors influencing students’ attitudes toward and engagement with using Duolingo as a Mobile-Assisted Language Learning (MALL) platform. While previous research has primarily applied linear models such as the Technology Acceptance Model (TAM), limited studies have explored non-linear relationships among learner perception variables using rule-based analytical approaches. To address this gap, this study employs Decision Tree analysis to examine how perceived usefulness, perceived ease of use, and learning motivation influence students’ attitudes toward using Duolingo. Data were collected in 2024 from 102 undergraduate Business Administration students through an online Likert-scale questionnaire. The J48 decision tree algorithm was applied to identify hierarchical decision rules and key predictor combinations explaining students’ attitudes toward application use. The results indicate that perceived usefulness is the most influential factor affecting learner attitudes, followed by ease of use and learning motivation. The model achieved an overall classification accuracy of 80.2%, with a Kappa coefficient of 0.6485, indicating substantial agreement. The findings provide insights for educators and application developers in designing more effective mobile language-learning environments that enhance student engagement and motivation.