Integrating Natural Language Processing and Multimedia Databases in CALL Software: Development and Evaluation of an ICALL Application for EFL listening Comprehension

Abstract

This paper presents an original computer-assisted language learning (CALL) app for EFL listening comprehension. The software, Listening Hacked, utilized a multimedia database and natural language processing (NLP) to create a personalized, autonomous learning environment for EFL learners. The paper is organized into two parts. In the first part, the paper describes the development of the software, including its theoretical underpinnings and developed functionalities. The second part, reports on the evaluation of the software which involved an experiment with 53 English-major Vietnamese students. The students were randomly assigned to the Experimental Group (EG) and Control Group (CG). The EG learned EFL listening by watching English-speaking movies, doing paused transcription tasks, and using various help options available on the platform to complete the tasks. The CG learned by doing traditional listening exercises with comprehension questions on Google classroom. The t-test and repeated measures of ANOVA results indicate that after 12 weeks of study, the students in the EG showed improvement in their EFL listening performance and that they performed better on the posttest than those who learned with traditional methods. The paper also discusses some implications of the findings in the context of researching, developing, and implementing CALL software for L2 listening.