Hanna Kang

DOI: Vol.25(No.2) 9-43, 2022

The purpose of this study is to investigate the differential effects of two interlocutors, an Artificial Intelligence (AI) avatar and a native speaker (NS) of English, in interaction on ESL learners’ speaking ability and affective aspects. The participants were 61 ESL learners from a college in the United States. They were assigned to the AI-to-learner interaction group (n=31) and the native speaker-to-learner interaction group (n=30). The major findings are as follows. First, interacting with both AI and NS interlocutors had positive effects on improving participants’ speaking ability, but talking with AI avatars brought more significant effects on improving oral performance with respect to accuracy, fluency, coherence, and interaction. Second, the low-level participants showed more speaking improvements when engaging in interaction with AI avatars, whereas the high-level participants benefited more from interactions with NSs. Lastly, all participants’ attitudes towards English learning were positively changed overall, but the interactive activities with NSs did not have a positive effect on increasing high-level participants’ motivation and could not reduce low-level participants’ stress and anxiety levels. Based on the findings, pedagogical suggestions are provided to help learners improve their speaking ability and lower their affective filters in ESL contexts.

ESL speaking ability, artificial intelligence, native speaker and non-native speaker interacti on, human-AI interaction, affective aspects

Na-young Kim

DOI: Vol.25(No.2) 44-63, 2022

This study aims to determine whether MALL directed by artificial intelligence positively affects second language (L2) learning. By comparing the effects of human-directed MALL, this study particularly explores the effectiveness of AI-directed MALL in increasing L2 comprehension skills. The participants were 98 freshmen who were taking a TOEIC course at a university in Korea. For the treatment, they were divided into the AI-directed MALL group (n = 51) and the human-directed MALL group (n = 47) at random. According to their treatment conditions, the participants took the online learning course with their mobile phones directed by either the AI or the human teacher. At the beginning and the end of the study, pre and post-TOEIC tests were administered. Paired-samples t-tests as well as independent samples t-tests were run to compare the mean scores. The findings reveal that all participants in both two groups increased their test scores at a significance level. However, there was a statistically significant difference in mean scores between the two groups. The AI-directed MALL group outperformed the human-directed MALL group. Based on these findings, implications are discussed.

artificial intelligence, mobile-assisted language learning, second language learning, TOEIC, comprehension skills

Jang Ho Lee

DOI: Vol.25(No.2) 64-83, 2022

The monolingual approach has served as a strong impetus for L2 teaching and research since the 20th century. However, post-2000 has seen a growing amount of support for the non-monolingual approach from different strands of research, including sociocultural, psycholinguistics, and SLA studies. Despite such growth, it has been suggested that CALL research has still not yet given much attention to non-monolingual themes, leaving the value of the non-monolingual approach unexamined in the CALL context. In view of this suggestion, the present study was conducted to examine the extent to which the articles published in the current journal (i.e., Multimedia-Assisted Language Learning) over the past two decades have been influenced by the monolingual bias. To this end, a wide range of keywords related to the non-monolingual approach (e.g., bilingual, multilingual, translanguaging, L1, translation) have been employed to identify studies that have addressed non-monolingual themes. The results revealed that non-monolingual themes are rather under-researched in the current journal, and even skewed in terms of the target themes. Based on this finding, we encourage researchers to pay greater attention to the non-monolingual approach, seeking to examine its values in the CALL context.

L1, monolingual bias, monolingualism, multilingualism, non-monolingual approach

Yousun Shin , Gyumi Kim

DOI: Vol.25(No.2) 84-102, 2022

Providing corrective feedback on L2 learners’ writing is considered essential by both learners and instructors. The present study aimed at investigating the effects of the automated written feedback tool, Grammarly, on reducing writing errors of L2 learners in their writing tasks. In this study, seventeen first-year university students majoring in English education completed two types of writing tasks in class and revised their drafts by using Grammarly. A total of 68 writing samples were analyzed by language-related features including the number of words, sentence length, or error types. Overall findings indicated that some significant decrease in the errors was observed between the 1st and 2nd drafts in the two tasks. However, it was also shown that the length of the compositions from the two tasks was positively proportional to the number of errors they made. The most frequent errors identified by Grammarly included ‘determiner use’, ‘incorrect verb forms’, ‘word choice’, ‘punctuation in sentences’, ‘misspelled words’, and ‘wordy sentence.’ The study also presented the positive relationship between their L2 language proficiency and L2 writing ability. The pedagogical implications and future research direction were discussed.

automatic written feedback, Grammarly, English writing, revision process, error types, writing tasks

Sun-young Son

DOI: Vol.25(No.2) 103-131, 2022

The purpose of the present study is to investigate the effects of using machine translation (MT) on Korean middle school students’ second language (L2) writing performance and affective aspects. A total of 30 middle school students participated in the study. For eight weeks, the students performed six writing tasks in three modes: without MT, with MT in the draft stage, and with MT in the revision stage. They wrote two writings in each mode. The students completed pre and post questionnaires to explore their affective aspects of using MT in L2 writing. The findings indicated that students’ writing scores improved when using MT in their L2 writing, and that using MT in the draft stage, where their first language (L1) was relatively dominant, was more helpful. Secondly, while investigating how students accepted the outcomes of MT as feedback and how their writing changed, there were significant changes in surface, stylistic, structural, and contents of writing with MT in the draft stage. Lastly, the students’ questionnaire responses indicated that in L2 writing, they considered writing with MT in the draft stage more positively than when using MT in the revision stage.

English writing, middle school students’ writing, process writing, machine translation

Min-chang Sung

DOI: Vol.25(No.2) 132-151, 2022

This study examines the development of prospective teachers’ adaptive expertise in AI chatbots in 2019 versus 2021. The comparison is based on social and technological changes during the period under study such as the increasing use of AI chatbots and notable improvements in web-based chatbot builders. Two groups of prospective elementary school teachers at a university of education in the Republic of Korea participated in a chatbot workshop in 2019 and 2021, respectively. The two workshops used different chatbot builders at Google Cloud Platform, i.e., Dialogflow ES in 2019 versus Dialogflow CX in 2021. Dialogflow ES uses intents as its building blocks and is suitable for small and simple chatbots, whereas Dialogflow CX uses flows and pages as building blocks and is suitable for larger and more complex chatbots. The participants’ adaptive expertise was measured via pre- and posttest questionnaires about four domains, namely knowledge, experience, attitude, and will. The results indicated that the 2021 group had greater adaptive expertise in AI chatbots than the 2019 group, especially in the domain of attitude. More importantly, the 2021 workshop using Dialogflow CX was found to be more effective in developing prospective teachers’ adaptive expertise than the 2019 workshop using Dialogflow ES. These findings led to discussion of the pedagogical implications for applying new AI technologies to training future teachers in a timely manner.

prospective teachers, adaptive expertise, Dialogflow, teacher training, English education

김낙훈 Kim Rakhun

DOI: Vol.25(No.2) 152-181, 2022

This study investigated the effects of data augmentation on the accuracy of AI chatbot responses for L2 English learning and learners’ satisfaction. Data augmentation is a computational method that can help increase the response accuracy of AI chatbot to enhance the user satisfaction. However, research on L2 learning has not been focused on evaluating the influence of data augmentation on the AI chatbot response accuracy and learner satisfaction. To clarify this influence, 36 Korean high-school English L2 learners were invited to participate in this study. The participants used two hands-on AI chatbots, with and without data augmentation, for English L2 learning. The accuracy of the AI chatbot responses was statistically analyzed, and the participants’ satisfaction in using the AI chatbot was qualitatively and quantitatively investigated. The results demonstrated that the responses of the data-augmented AI chatbot were more accurate than those of the chatbot without data augmentation. Additionally, learners who used the data-augmented AI chatbot were more satisfied with their AI chatbot-based learning experience. This research highlights the necessity of incorporating data augmentation in the design and implementation of AI chatbots for English language learning.

AI chatbot, data augmentation, accuracy of AI chatbot response, corrective recast

이동주 Lee Dongju

DOI: Vol.25(No.2) 182-206, 2022

This study aims to investigate the effects of real-time online classes using Google Forms on TOEIC test scores of Korean EFL undergraduates and to explore their perspectives on this course. Over one winter semester, the online TOEIC course was conducted via Zoom and a total of 40 students participated in this course. During the class, the quizzes and exercise problems, created using Google Forms, were required for the participants to answer and then immediate feedback was provided to them. The data included 1) pre-questionnaires on students’ background, 2) post-questionnaires on their perceptions, and the advantages and disadvantages of the Zoom TOEIC course, 3) pre- and post-TOEIC reading tests. The collected data was analyzed statistically, and open-ended responses were categorized into themes based on the participants’ responses. The results indicated that the students did not achieve significantly higher scores in the post-test compared to the pre-test. The students, however, expressed positive responses about this course. They reported it appears that taking the synchronous TOEIC course improved their overall English abilities and helped to prepare for a TOEIC test. They were also pleased with the comfortable environment, the immediate feedback using Google Forms, and the active communication through online chatting.

online learning, real-time English class, Zoom, Google Forms, TOEIC course

정희정 Jung Hee-jung

DOI: Vol.25(No.2) 207-231, 2022

The purpose of this study is to develop and empirically validate the research model based on value-attitude-behavior (VAB) framework in order to understand English learners’ behavior in the context of augmented reality (AR). To this end, this study proposes four variables (learning connectivity, learning realism, immersive, and learning efficiency) as unique characteristics of AR English learning, which have a positive impact on the perceived value of AR learning. Then the perceived value has a positive impact on the attitude toward AR English learning, positively related to the intention to use AR for English learning. Also, the proposed research model includes the network effect as a moderator within the relationships among VAB variables. For a period of five months, a total of 184 responses is collected using a short survey from 5th and 6th grade students at local elementary schools, children’s libraries, and child care centers participating in the Edutech program by the Ministry of Education. Using AMOS 24.0, the structural equational model was analyzed in order to test hypotheses. The results indicate that all proposed hypotheses were significant, which provides a clear understanding of English learners’ behavior regarding AR. Based on the findings of the study, the implications are discussed with respect to the future direction of AR research in English education.

English learning, augmented reality (AR), value-attitude-behavior (VAB), network effect

진승희 Jin Seunghee

DOI: Vol.25(No.2) 232-254, 2022

This study aims to examine the effects of using an artificial intelligence (AI)-based speaking application (app) for English speaking courses on Korean EFL learners’ speaking skills and their affective attitudes such as motivation, interest, and confidence. Over a semester (16 weeks), a total of 50 students participated in the study. They are assigned to an experimental group who participated in AI app-assisted speaking activities and to a control group who engaged in the speaking activities without the aid of a smartphone app. To collect the data, 1) a pre- and post-test on speaking skills and 2) a pre- and post-survey on affective attitudes were conducted for both groups. The results indicated that an AI app-assisted learning group has shown significant development in speaking skills when compared to the control group. In particular, AI app-assisted learning group’s speaking abilities concerned with vocabulary, grammar, and pronunciation were significantly enhanced. This study also revealed that the AI app-based speaking instruction was effective in improving the students’ affective attitudes such as motivation, interest, and confidence. The results supported that AI app-assisted speaking instruction is an effective way to develop students’ speaking skills and affective attitudes and is more effective than the traditional speaking instruction.

artificial intelligence (AI), smartphone application (app), speaking skills, affective attitudes

하종범 Ha Jongbum

DOI: Vol.25(No.2) 255-284, 2022

With the purpose to explore the optimal online English class model, this study examined students’ perception on video and untact realtime hybrid classes. Participants numbered 146 students from three different English courses at a university in South Korea. Classes were composed of four modules, each consisting of two video and one untact realtime class sessions. During the untact realtime class session, Nearpod was used as the main platform to enhance students’ communication and participation. At the end of the semester, students completed a written survey regarding their perception on each of the video and untact realtime classes as well as overall management. The results showed the students were satisfied with each of the class sessions and overall management, while no significant differences were found between class, year, or sex. The results also showed the deactivation of cameras did not affect their perception during the video sessions, while registering a positive effect in the untact realtime class sessions. Their favorite online class structure was found to be a video and untact realtime hybrid class model as the one used in the current study. The findings suggest that Nearpod can provide pedagogical benefits under systematic organization within and between online sessions.

online hybrid English class, Nearpod, students’ perception, communication, participation