Exploring the Relationship between Interactions and Learning Performance in Robot-Assisted Language Learning

Ao, Yutong; Yu, Zhonggen

doi:https://doi.org/10.1155/2022/1958317

Education Research International

On this page

Abstract Introduction Literature Review Results and Discussion Conclusion Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Review Article | Open Access

Volume 2022 | Article ID 1958317 | https://doi.org/10.1155/2022/1958317

Exploring the Relationship between Interactions and Learning Performance in Robot-Assisted Language Learning

Yutong Ao¹and Zhonggen Yu¹

Academic Editor: Bilal Khalid

Received04 Feb 2022

Accepted05 Apr 2022

Published20 Apr 2022

Abstract

Robot-assisted language learning is a pattern that uses a social robot in a class to enhance learning performance through human-robot interaction. While this pattern receives increasing attention, few review articles generalize whether each interactive behavior improves learning performance. This study delves into the answers to three research questions. The results show that the synthesized and prerecorded human voices are appropriate for different teaching activities. Verbal communication except for egocentric small talk and L1 translation between the robot and learners are indispensable to improve students’ confidence and learning gains. Moderate employment of nonverbal interactions helps students increase concentration, motivation, and retention of vocabulary, while undue interactions give rise to counterproductive effects. Based on the findings, future research on robot-assisted language learning is suggested to pay attention to the effectiveness of an independent interaction. Another valuable focus is the proper combination of interactive behaviors that suit different teaching tasks. A suitable level of robotic sociability is also worthy of exploration. Educators teaching with a robot should make full use of verbal and nonverbal interactive behaviors to boost students’ confidence, motivation, engagement, and learning gains. Meanwhile, they need to be cautious about the excessive employment of interactions.

1. Introduction

Technology-assisted language learning (TALL) makes up numerous affordances that face-to-face environments cannot achieve, such as interactive feedback and tailored instruction. As a branch of TALL, robot-assisted language learning (RALL) is regarded as a prospective subfield [1]. Robots, qualifying properties of inference, learning, object seeking, problem-solving, and adaptiveness, have undertaken tasks in education.

RALL is defined as the employment of a robot or robots to enhance language skills, from vocabulary recognition to text comprehension [1]. Scholars acknowledge the efficacy of robots in foreign language learning [2], as evidenced by past studies that learners remembered words faster [3], used grammar more accurately [2], and communicated more fluently [4].

Previous scholars adopted the term “social robot” to discriminate robots in language learning from those used in engineering or surgery. A social robot features the behavioral norms of human beings and provides interactions with the audience [5]. Unlike virtual images of robots online, social robots are physical entities and thus allow learners to touch them. Although some are animal-shaped robots, most are anthropomorphized versions with movable heads, arms, legs, and even waists.

Social robots are regarded as distinctive assistants from other technologies (e.g., gamification and virtual reality) because of their interactive behaviors. The two levels of interaction are verbal and nonverbal interactions. On the one hand, robots being available to perform oral communications, students can get direct knowledge. On the other hand, robots use eye-gazes or other signals to transfer information that verbal interactions cannot.

The impetus of this review is to explore the efficacy of each interactive behavior robots perform, both verbally and nonverbally. Based on those findings, the study then provides suggestions for future RALL researchers and educators. Because social robots transmit knowledge through interactions, the efficacy of each interaction is crucial to research.

2. Literature Review

To disperse the mass gathering of people in one location, universities gradually shift from the in-person pattern of education to distant learning [6]. In the wake of COVID-19, TALL has become the trend in pedagogical activities. It is beneficial not only on the lifestyles but also on language learning activities [7, 8]. Tech-based education can be effectively adopted in smart classrooms as a supplement of the traditional approach in order to advance the learning procedure [7, 9]. As TALL is increasingly used in popularity, instructors and researchers develop many technology-supported tools as educational auxiliaries.

The widely adopted technologies can be classified into three patterns: online recorded courses, APPs, and web conferences. Taking the advent of Language Massive Open Online Courses (LMOOCs) as an example, it is a practical approach to offer lifelong learning opportunities. LMOOCs enable people to learn in formal and informal settings, such as studying at home and work [10, 11]. Besides, it attracts students on the ground that it offers freedom and enables students to study at their own pace [6, 12, 13].

A majority of studies in LMOOCs are carried out from the learner’s perspective. Researchers examined five elements that affect learners’ motivation: relatedness, utility value, pressure, effort, and intrinsic value [14]. In another study, researchers identified the percentage of highly self-determined learners among all LMOOC participants and found a surprisingly low proportion, accounting for 4.31% [15]. Zeng et al. [16] explored how LMOOC learners allocated their attention. The results showed that high-performing learners paid most attention, not to the start of a new course or specific course units but the maintenance and circulation of knowledge to other learning units. Besides, high-performing learners always followed the predesigned course outline. Chong and Reinders [17] analyzed 100 LMOOC courses and generalized their three limitations: the scarcity of interactions, the neglection of a student-centered approach, and the ignoring of individual differences. In terms of future trends of LMOOC studies, researchers pointed out four respects, comprising of the distinctive nature of LMOOCs, the intrinsic motivation and experience of participants, the reflection of teachers’ roles in LMOOCs, and the combination of sociability in LMOOCs [18].

With regard to APPs, social networking sets comprising of Facebook and WeChat have been witnessed a wide use in education because they positively impact language learning [19–23]. The applications facilitate information literacy learning [24], incidental vocabulary acquisition, L2 pragmatic awareness [25–27], and instructor-student communications [28]. Barrot [19] scrutinized 396 documents and concluded three reasons why social media applications have an increasing status among language learners. They are flexible in communication, have a wide range of geographical distributions, and contain a considerable number of active users. Other merits are attributed to their enhancement of intercultural and sociopragmatic awareness, the facilitation of language user identities, and improving literacies, as Reinhardt [29] discussed.

Web-based computer conference, for instance, ZOOM and Tencent Classroom, provides authentic and direct interactions between teachers and students [30, 31]. Researchers have found out the affirmative effects of online learning platforms in advancing learner’s self-awareness, self-regulation, and affective skills [32]. Almusharraf and Bailey [33] investigated 329 students using computer conferences and concluded that the collaborative language learning orientation, mediated by agent engagement, significantly predicts academic learning expectations. Furthermore, active students probably have more positive class expectations than those inactive. Another study conducted a four-month experiment and made a comparative analysis between traditional and online conferencing classes. It has been found that the latter approach elicited numerous nonverbal responses and could expand verbal results. An emerging issue was also pointed out that reticent students had difficulties fitting into the class [34].

Apart from educational technologies above, RALL, the first of which appeared in 2004, is a new tech-based language learning approach. In its initial study, researchers intended to test whether children learned from robots as they learned from peers [35]. Robots in RALL have the attribute of socialness. They can identify, speak to, and respond to the audience through implicit and explicit expressions. Although the definitions of social robots vary in detail, all highlight four points: they (1) have physical entities, (2) resemble creatures, (3) adapt themselves to a social community, and (4) interact with humans in an effective manner.

The four features of robots in RALL separate them from other electronic approaches, especially the simulation of an authentic interactive environment. Effective interaction should be considered when designing the robots. Arkin et al. [36] proposed three demands of interaction for social robots: high-level dialogue communication, perception and presentation of emotions, and the ability to use natural cues. Former studies discussed the efficacy of interactions in terms of subjects and learners’ age [35, 37–39]. Studies also generalized RALL as a whole [5] but failed to focus on interactive skills specifically. It leads to the first and the second question of this study: what verbal interactions in RALL could improve learning performance, and what nonverbal interactions in RALL could improve learning performance?

Overall, previous studies lacked an in-depth analysis of the interactions between the robot and learners. It is possible that researchers and educators neglect several useful functions when setting robotic interactive behaviors. Therefore, this study proposed the third question: what are the suggestions for RALL researchers and educators?

3. Methodology

The literature review followed the documentary analysis method proposed by Spolaôr and Benitti [40]. Starting from searching for keywords, this study depicted the strategy of identifying exclusion criteria, quality criteria, and operations, as well as ways of retrieving and synthesizing information.

3.1. Keyword Searching Strategy

This review collected empirical studies through the Web of Science database. The author first identified keywords relevant to robot-assisted language learning. The article followed the opinion of van den Berghe et al. [5]. It did not restrict the quality of journals in case of missing information, in that RsALL is a new area, and there is the possibility that the number of articles with high quality is not adequate to review. The searching string was “(robot) AND (“language learning” OR “English learning”). To be more accurate, the author set the search scope in “Topic.” The author then constrained “Subject” in education to narrow the number of articles.

3.2. Identification of Exclusion Criteria

This section set the exclusion criteria (EC), which were partly drawn from the previous two papers [1, 5]. Studies colliding with one or more exclusion criteria were not considered. Twelve EC are shown in Table 1.

3.3. Establishment of Quality Criteria and Strategy

The study framed quality criteria (QC) to measure if each remaining publications are of high quality (Table 2). Answers to QC were yes or no. By answering QCs, 1 or 0 was recorded, respectively, providing quantitative evidence of the publications’ quality. Final scores were the sum of all criteria scores [40–42].

3.4. Information Retrieval

In light of a previous publication reviewing robotics in education, there were 26 Items (IE), highlighting the essential information of each study and horizontally comparing information under the same items. The items were packeted into four groups, as presented in Table 3. Seven out of 26 items were extracted to collect helpful information for this article (Table 4).

3.5. Information Synthesis

The following discussions were centered on empirical studies and review articles that met the demands of this review. The article answered three research questions by integrating information from the newly collected studies. Figure 1 is the filtering process of prior studies.

4. Results and Discussion

Due to the newness of robot-assisted language learning, studies on this subject are sparse. Twenty empirical studies were included after passing the filter criteria. This section answered three research questions. Robots possessing both verbal and nonverbal functions interact with learners in multiple ways. For example, robots speak and perform to learners, who respond to them in turn. The following paragraphs explored interactions from verbal and nonverbal perspective separately.

4.1. RQ1: What Verbal Interactions in RALL Could Improve Learning Performance?

Social robots perform verbal interactions through communication, in which the type of voice is set beforehand.

4.1.1. Voice

Designers program the robotic voice to be mechanically synthetic or prerecorded by humans. The mechanical voice, although not natural, is human-like and can qualify features of human beings’ voice. A large proportion of empirical articles this paper reviewed did not mention the voice type; others either reported it explicitly or were inferred between the lines.

Firstly, the synthesized voice is of use due to its flexibility to present different characteristics such as age and gender [43]. In an experiment, researchers designed English storytelling practices for 22 young children assisted by a robot with a synthesized and varied voice. The results showed that children in this group were more interactive and concentrated than the control group without a robot [44]. Researchers inferred that the male child-like voice narrowed the psychological distance and made the robot more acceptable to children. Another study set the robot to perform a story with either a male or female voice. After five weeks of learning, teachers in the class positively evaluated the voice variation as it could make the class more vivid and engaging [45].

Although it is welcomed by children and teachers, the mechanical voice has a distinctive flaw. In the study described above, teachers also pointed out the robotic difficulty expressing emotions indispensable to language learning [45]. In a study, 33 first-year students of the English teaching program learned English vocabulary assisted by the NAO robot with a mechanical voice, while 32 students learned through human teachers. The results showed no significant difference between both groups in terms of learning gains. Some students commented that the voice was emotionless and robotic while they listened [46]. Another experiment doing field trails confirmed had the same results [47], but currently, no experiment supported its adverse effects on children’s engagement and performance.

The prerecorded voice can perfectly express emotion as an alternative to the synthesized voice [48], but it created fewer identities in the storytelling practice than the synthesized voice. In two experiments, researchers changed the pitch of prerecorded adult voice to make it sound child-like [37]. Researchers in a Japanese experiment used a recorded voice of a native English speaker to communicate with students. After the interactions, they found that the utterances were too fast for the young Japanese students to comprehend [37], but this negative impact could be moderated by the adjusted rate of the recorded utterances.

4.1.2. Type of Verbal Communication

Verbal communication is essential for social robots to interact with the audience. Researchers adopted different communication frames to adapt to language abilities, namely, two-way communication (dialogue and quiz) and one-way communication (egocentric small talk, feedback, and L1 translation).

The two-way communication directly imbues learners with knowledge and fosters comprehensibility. It is a conversation that consists of questions, answers, and comments. Not only does it provide higher exposure, but it also drives learners to focus. At the first stage of RALL, the robotic talk assists learners to have a vague impression of the new knowledge. Next, robots tell stories [39] and ask questions with specific answers [4] or open questions [49] to either grab attention or facilitate understandings [1]. Another form of two-way communication is a dialogue that guides students to form discussions themselves. A study found that students’ communicative competence increased, and the new knowledge was internalized through two-way communication [40]. Researchers also input multiple expressions inside the robot. For example, the robot could say, “Can you show me the word X?” or “I don’t know the word X, can you show it to me” to increase the reliability of the robot [50].

Some forms of one-way communication benefit learners, while others are controversial. One-way communication is an essential component of robotic verbal ability, including small egocentric talk, robot-to-learner feedback, and L1 translation support. They are discussed in the following paragraphs.

Egocentric small talk can be either positive or negative owing to students’ individuality. In the experiment, researchers adopted the Furhat robot [49] and programmed it to dominate a discussion, only asking trivial questions and telling learners facts it wanted to tell [49]. As some learners were intrigued by the story, others performed passively, causing a decline in motivation. This difference happened presumably due to students’ differences (e.g., character).

As for the feedback, robots providing positive feedback may result in higher learning performance than other feedback, but students receiving explicit feedback require less assistance from human teachers while learning. In a study testing three types of feedbacks, namely, explicitly negative, explicitly positive and implicitly negative, and no feedback, researchers found no difference in children’s engagement, but those receiving explicitly negative feedback showed a more independent performance [51]. Nevertheless, in both groups containing feedback, the robotic roles were not unified, which were peer (explicitly negative feedback) and adult (explicitly positive and implicitly negative), making it hard to decide to what extent roles influenced the research outcomes. Another study found that positive feedback from the robot resulted in a significant improvement in concentration and learning gains. Neutral and negative comments also benefited learners, although learning gains were moderate [52]. Some other articles also implemented the robotic feedback when designing the experiment but did not discuss its impact.

Only one experiment explored the effect of robotic translation, according to which the bilingual robot attracted children’s attention, but the one speaking the second language stimulated more favorable learning performance [3]. In this study, [28] Turkish kindergartners learned Dutch vocabularies, half taught by the bilingual robot with Turkish translation, half by a monolingual Dutch robot. Contrary to the hypothesis, researchers found the children remembered fewer vocabularies from the bilingual robot than the Dutch-only ones.

4.2. RQ2: What Nonverbal Interactions in RALL Could Improve Learning Performance?

Many studies have provided evidence for the efficacy of nonverbal behaviors in enhancing learners’ understanding of the language knowledge [53]. Nonverbal behaviors function better when the robot is programmed in a personalized condition [54]. There are two typical nonverbal interactions throughout the gathered information: facial expressions and gestures [1].

4.2.1. Facial Expression

It is observed that facial expressions improve learners’ concentration and motivation. Among all robots researchers got access to, only those countenances in the form of screen accomplished various emotions, including shyness, laughs, frustration, yawning, thinking, and eye-gazing [50, 55]. The most widely utilized NAO robot can only animate countenance through gaze and facial light as the robot Comm U [4].

4.2.2. Gaze

Gazes gather learners’ attention, but undue gazes increase students’ learning anxiety. When the robot and the learner have eye contact, the latter is prone to focus on the learning activity [56]. A study supported the usefulness of eye-gazing. In this experiment, the robotic eyes moved casually to the children’s direction to impress them of being relaxed, accompanied by other nonverbal behaviors [39]. However, in this study, the effectiveness of gaze lacked empirical proof.

In contrast, a study using an antisocial robot (a robot that avoided eye contacts) in the class resulted in better learning performance of the students than those learned from a normal robot [38]. Since this was an experiment assisted by a tablet, students having eye contact with the sociable robot looked less at the knowledge presented on the tablet. Another empirical study provided evidence for this conclusion [57]. Some students were watched by the robot while doing independent tasks. In this situation, the nature of eye contacts changes from attention-grabbing to watchfulness, causing anxiety.

4.2.3. Facial Lights

Researchers adopt facial light as auxiliaries to strengthen the expressiveness of the robot. In two experiments [4, 55], facial lights were embedded in the cheeks of the robot with the aim of enhancing expressiveness. In a study teaching English to young children, the robot assistant switched its facial lamps to provide positive or negative responses [58]. In another study aiming to enhance 9 Japanese adults’ oral speaking, researchers added red LEDs to the two cheeks for strengthening the robotic expressiveness [4].

4.2.4. Gestures

Gestures can improve engagement and motivation more or less. In this part, gestures are discussed in the narrow sense: limbs and body movements. Gestures are generally classified into three types: body orientation, deictic gestures, and iconic gestures. Two encouraging gestures, thumbs-up and applauding, are also involved. Iconic gestures move with limbs and the torso (e.g., bend the arms to make a circle to indicate the noun circle), and deictic gestures are index pointing gestures that implicate an entity or a place [59].

Body orientation is crucial for the clarity of interactive behaviors. By changing the body direction, the robot and learners directly see each other. There are three studies that employed one-on-one interactions, where researchers placed the robot in front of the learners beforehand, so there was no need to move [51, 60, 61]. However, institutions cannot adopt the robot for each student because of the cost. In this case, most are three or above students simultaneously learned from the robot, and thus, students failed to face the robot straightforward, which might affect language learning. In another study, researchers placed the robot at a 90-degree angle to the children so that the robot and the children had a similar frame of reference, but it might reduce the clarity of gestures and hinder comprehensibility [62].

Iconic gestures have instant benefits towards learning performance. A study found that a robot conducting iconic gestures was conducive to children [63]. In this study, Dutch children from two groups learned six animal words from the robot, who introduced them verbally and performed iconic behaviors in one group but did not in another. Although research results showed no difference in the immediate posttest, children who saw iconic gestures retained more words than other children in a delayed posttest a week later.

Another two studies provided evidence for the distractive effect of iconic gestures. Researchers in the first study generalized three reasons. Above all, the robot faced the children in a 90-angle position, which might affect the clarity of gestures. Another reason was that the same iconic gestures were conducted many times when a new word appeared, causing distraction. The degree of children enacting the robotic gestures was also affected [62].

Researchers in the second study believed that some iconic gestures are similar to deictic gestures. In this study, children answered a questionnaire to test their perception of anthropomorphism towards the robot [61]. Both groups, one seeing deictic gestures only and the other perceiving iconic and deictic gestures, had the same perception of anthropomorphism. Because the robot showed iconic gestures repetitively when introducing the same vocabulary, and because the robot failed to perform exactly iconic when teaching abstract words, such as “more,” its human-likeness did not differ from the deictic gesture-only robot, as the researchers concluded.

Robots with deictic motions are more human-like than those without in the eyes of children, and a human-like robot provides a sense of confidence and cosiness [54]. They are designed to point out a particular object the new word refers to. Empirical studies in the scope of RALL all used deictic gestures except the one with robot Furhat (it only has a head), but no study explored in detail. Therefore, the efficacy of deictic gestures to learning performance is still awaiting exploration.

Other body gestures can absorb learners’ attention and thus increase learning engagement and motivation. Researchers add them as an auxiliary to amplify the effect of intended interactions. During learning activities, the robot introduced itself accompanied by dancing to create closeness [63]. Stooping [4], nodding [56], applauding [58], and thumbs-up [47] were also designed to adapt the learning scenario and to give a positive response. A benefit for the robot to pose while teaching is instructing children to use appropriate gestures while speaking under a specific social context [45].

4.3. RQ3: What Are the Suggestions for RALL Researchers and Educators?

4.3.1. Suggestions for RALL Researchers

Future studies could attach importance to the effectiveness of independent interactions, rather than the pooled impact of a series of interactive behaviors. Specifically, researchers should delve into differences of two types of voice, the mechanical and prerecorded human voice, while remaining other settings the same. Some scholars recommended robots speak in the most natural voice [64], but present studies did not have inadequate proof to support the superiority of one voice type to another in terms of learning performance. Future studies are expected to conduct more experiments of two types of voice to determine what are their suitable learning activities. Other potential functions to research are diverse communicative methods (dialogue, quiz, egocentric small talk, feedback, and translation), facial expressions (gaze and facial light), and different gestures (body orientation, iconic gestures, dancing, and thumbing-up). Since some interactions like iconic gestures may decline learning performance, researchers should explore the proper combination of interactions. Robots in language learning do not function solely based on a single interactive behavior but a series of behaviors to achieve learning gains.

While some studies have been committed to the added benefits of social interactions, few have discussed a suitable level of sociability for those behaviors. Proper interactions could enhance learning motivation, engagement, and gains, but inadequate or undue behaviors could not function the most effectively, sometimes even leading to counterproductive effects. For instance, excessive gazes caused distraction and learning anxiety [56]. Researchers were expected to find the linear relation between different interactions and learning performance [65].

4.3.2. Suggestions for RALL Educators

Educators should make use of interactions to optimize learning performance. If storytelling practices are necessary to foreign language learning, then the synthesized voice may be the priority, as it creates more characters and makes the story more attractive. In other learning activities, a robot equipped with a human voice is easier to familiarize students with the robot. Teachers should first clarify their class activities and then set certain interactions of the educational robot.

Educators can combine robotic feedback with learning activities. Positive feedback boosts students’ confidence, concentration, and learning gains [52], while explicitly negative feedback presumably lessens students’ interest and confidence in learning, although it makes students more independent. Positive and implicitly negative feedback can be used when students are demotivated by factors, such as previous mistakes [51]. When teachers set the verbal output of robots, they should input commands for the robot to either praise students or point out where students can improve themselves.

Excessive gazes from the robot negatively affect students’ performance [56]. Therefore, teachers need to employ the gaze interaction properly, although there is no approved standard to maximize the performance. Iconic gestures are effective for educators to adopt. The robot performing iconic gestures helps students retain new vocabularies longer. But teachers also need to be careful of the distractive effect of iconic gestures [47]. How to improve the expressiveness of an iconic gesture is also worthy of consideration. Abstract vocabularies like “less” are hard to be presented by image. Gestures like dance, thumbing-up, or applauding could enhance learning motivation and shorten the distance to the robot [58]. Researchers could consider excluding iconic gestures when designing the robot-learner interaction if the words taught are abstract.

5. Conclusion

This study presented a literature review on the relationship between interactions and learning performance in robot-assisted language learning.

5.1. Major Findings

RQ1 discussed verbal interactions in RALL that could improve learning performance. The synthesized voice is widely used. It shows various identities to make the storytelling process more engaging and concentrated, but some teachers and students regard it as cold and emotionless, reducing the effectiveness of language learning. The prerecorded human voice is natural and can express emotions powerfully, but not flexible to change the present multiple identities. Quiz and dialogues are indispensable in verbal interactions to transmit knowledge and foster comprehensibility. Feedbacks also function in RALL. However, egocentric small talk declines motivation, and students having a bilingual robot perform worse than those with a monolingual robot.

RQ2 discussed nonverbal interactions in RALL that could improve learning performance. Facial expressions of a robot increase concentration and motivation. To mention one point, gaze from the robot should be moderate in case of distraction and learning anxiety. The position of the robot and learners influences the clarity of gestures. Iconic gestures help learners retain vocabulary. Many other gestures, including thumbing up, applauding, and dancing, enhance language motivation and confidence. Still, these behaviors may distract learners’ attention and reduce their performance.

RQ3 proposed four suggestions for future RALL research and educators. As for the researchers, they are expected to determine the degree of effectiveness of each independent interactive behavior. How to combine different behaviors to adapt to different learning activities and how to maximize learning performance are also waiting to explore. Another suggested direction is to determine a suitable level of sociability for the robot in order to avoid excessive interactions. It is suggested that in the future, scholars conduct empirical studies, in which they set a solely interactive behavior each time to compare their effectiveness. Through the permutation and combination of each interaction, researchers can test some more practical groups of interactions than others. In terms of the proper behavior to particular teaching activity, researchers should select potential behaviors and compare their effectiveness by empirical studies.

Suggestions for educators are specific. Educators should choose a suitable type of voice to accord to teaching activities. If the activities require a variation of characteristics, the synthesized voice probably functions better than the recorded human voice. Positive feedback and implicitly negative feedback are functional, as they boost confidence and stimulate learners. Gazes and gestures are indispensable to interactions, but to what extent the robot performs needs careful design beforehand. The match of iconic gestures to the new knowledge is another concern to enhance comprehension.

5.2. Limitations of This Study

The main limitation is that discussions on effects of the aspects except interactions are out of the scope of this study. These factors are different robots, different roles a robot plays, different participants, different degrees of human teacher interference, different skills being taught, and different learning environments.

Another limitation lies in the limited studies this literature explored. It is inevitable to miss some studies because of the limited number of pathways or the loss of some relevant keywords while searching. Additionally, some studies are inaccessible to acquire full texts and thus are excluded in this study.

It would be grateful if scholars or other readers put forward precious suggestions for the improvement of this study. Future researchers are expected to determine whether each interactive behavior contributes to learning performance and the degree of its effectiveness when interrupted by other factors.

Data Availability

The data used to support the findings of this study are available from the first author upon request.

Conflicts of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

The authors would like to express their acknowledgement to their supervisor Professor Zhonggen Yu, who inspired them on the topic of the paper and helped them revise the organization and language of the paper. This work is supported by the 2019 MOOC of Beijing Language and Culture University (MOOC201902) (Important) “Introduction to Linguistics”; “Introduction to Linguistics” of online and offline mixed courses in Beijing Language and Culture University in 2020; and Special fund of Beijing Co-construction Project-Research and reform of the “Undergraduate Teaching Reform and Innovation Project” of Beijing Higher Education in 2020-innovative “multilingual +” excellent talent training system (202010032003).

References

M. Hein and D. Nathan-Roberts, “Socially interactive robots can teach young students language skills: a systematic review,” Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 62, no. 1, pp. 1083–1087, 2018.
View at: Publisher Site | Google Scholar
A. B. Khalifa, T. Kato, and S. Yamamoto, “Learning effect of implicit learning in joining-in-type robot-assisted language learning system,” International Journal of Emerging Technologies in Learning (iJET), vol. 14, no. 2, 2019.
View at: Publisher Site | Google Scholar
H. Leeuwestein, M. Barking, H. Sodacı et al., “TeachingTurkish‐Dutchkindergartners Dutch vocabulary with a social robot: does the robot's use of Turkish translations benefit children's Dutch vocabulary learning,” Computer Assisted Learning, vol. 37, no. 3, pp. 603–620, 2021.
View at: Publisher Site | Google Scholar
T. Iio, R. Maeda, K. Ogawa et al., “Improvement of Japanese adults' English speaking skills via experiences speaking to a robot,” Journal of Computer Assisted Learning, vol. 35, no. 2, pp. 228–245, 2019.
View at: Publisher Site | Google Scholar
R. van den Berghe, J. Verhagen, O. Oudgenoeg-Paz, and P. Leseman, “Social robots for language learning: a review,” Review of Educational Research, vol. 89, no. 2, pp. 259–295, 2019.
View at: Publisher Site | Google Scholar
C. Muangmee, S. Kot, N. Meekaewkunchorn, N. Kassakorn, S. Tiranawatananun, and B. Khalid, “Students’ use behavior towards e-learning tools during COVID-19 pandemics: case study of higher educational institutions of Thailand,” International Journal of Evaluation and Research in Education (IJERE), vol. 10, no. 4, p. 1166, 2021.
View at: Publisher Site | Google Scholar
A. Hariharasudan and S. Kot, “A scoping review on digital English and education 4.0 for industry 4.0,” Social Sciences, vol. 7, no. 11, p. 227, 2018.
View at: Publisher Site | Google Scholar
M. Kondratiuk-Nierodzińska, “New knowledge generation capabilities and economic performance of Polish regions,” Equilibrium. Quarterly Journal of Economics and Economic Policy, vol. 11, pp. 451–471, 2016.
View at: Publisher Site | Google Scholar
J. Jo and H. Lim, “A study on effectiveness of smart classrooms through interaction analysis,” Advanced Science Letters, vol. 21, no. 3, pp. 557–561, 2015.
View at: Publisher Site | Google Scholar
A. Tuijnman and A. K. Boström, “Changing notions of lifelong education and lifelong learning,” International Review of Education, vol. 48, no. 1, pp. 93–110, 2002.
View at: Publisher Site | Google Scholar
K. Steffens, “Competences, learning theories and MOOCs: recent developments in lifelong learning,” European Journal of Education, vol. 50, no. 1, pp. 41–59, 2015.
View at: Publisher Site | Google Scholar
R. Stefko, R. Fedorko, and R. Bacik, “Website content quality in terms of perceived image of higher education institution,” Polish Journal of Management Studies, vol. 13, pp. 153–163, 2016.
View at: Publisher Site | Google Scholar
M. Stuss, K. Szczepańska-Woszczyna, and Z. Makie&lstrok;a, “Competences of graduates of higher education business studies in labor market I,” Sustainability, vol. 11, no. 18, p. 4988, 2019.
View at: Publisher Site | Google Scholar
T. Beaven, T. Codreanu, and A. Creuzé, “Motivation in a language MOOC: issues for course designers,” Providing Learning, Transcending Boundaries, pp. 48–64, 2014.
View at: Publisher Site | Google Scholar
N. Agonács, J. F. Matos, D. Bartalesi-Graf, and D. N. O’Steen, “Are you ready? Self-determined learning readiness of language MOOC learners,” Education and Information Technologies, vol. 25, no. 2, pp. 1161–1179, 2020.
View at: Publisher Site | Google Scholar
S. Zeng, J. Zhang, M. Gao, K. M. Xu, and J. Zhang, “Using learning analytics to understand collective attention in language MOOCs,” Computer Assisted Language Learning, pp. 1–26, 2020.
View at: Publisher Site | Google Scholar
S. W. Chong and H. Reinders, “Autonomy of English language learners: a scoping review of research and practice,” Language Teaching Research, p. 136216882210758, 2022.
View at: Publisher Site | Google Scholar
M. Sallam, E. Martin-Monje, and Y. Li, “Research trends in language MOOC studies: a systematic review of the published literature (2012-2018),” Computer Assisted Language Learning, pp. 1–28, 2020.
View at: Publisher Site | Google Scholar
J. S. Barrot, “Social media as a language learning environment: a systematic review of the literature (2008-2019),” Computer Assisted Language Learning, pp. 1–29, 2021.
View at: Publisher Site | Google Scholar
J. S. Barrot, “Effects of Facebook-based e-portfolio on ESL learners’ writing performance,” Language, Culture, and Curriculum, vol. 34, no. 1, pp. 95–111, 2021.
View at: Publisher Site | Google Scholar
C. Jia and K. F. T. Hew, “Supporting lower-level processes in EFL listening: the effect on learners’ listening proficiency of a dictation program supported by a mobile instant messaging app,” Computer Assisted Language Learning, pp. 1–28, 2019.
View at: Publisher Site | Google Scholar
S. Manca, “Snapping, pinning, liking or texting: investigating social media in higher education beyond Facebook,” The Internet and Higher Education, vol. 44, p. 100707, 2020.
View at: Publisher Site | Google Scholar
J. Z. Paul and E. Friginal, “The effects of symmetric and asymmetric social networks on second language communication,” Computer Assisted Language Learning, vol. 32, no. 5–6, pp. 587–618, 2019.
View at: Publisher Site | Google Scholar
J. C. Guo and J. Huang, “Information literacy education in WeChat environment at academic libraries in China,” Journal of Academic Librarianship, vol. 46, no. 1, p. 102073, 2020.
View at: Publisher Site | Google Scholar
C. Prichard, “Training L2 learners to use Facebook appropriately and effectively,” CALICO Journal, vol. 30, no. 2, pp. 204–225, 2013.
View at: Publisher Site | Google Scholar
B. Smith, “Computer-mediated negotiated interaction and lexical acquisition,” Studies in Second Language Acquisition, vol. 26, no. 3, pp. 365–398, 2004.
View at: Publisher Site | Google Scholar
S. L. Thorne, “Artifacts and cultures-of-use in international communication,” Language Learning and Technology, vol. 7, no. 2, pp. 246–259, 2003.
View at: Publisher Site | Google Scholar
Z. G. Yu, “Extending the learning technology acceptance model of WeChat by adding new psychological constructs,” Journal of Educational Computing Research, vol. 58, no. 6, pp. 1121–1143, 2020.
View at: Publisher Site | Google Scholar
J. Reinhardt, “Social media in second and foreign language teaching and learning: blogs, wikis, and social networking,” Language Teaching, vol. 52, no. 1, pp. 1–39, 2019.
View at: Publisher Site | Google Scholar
H. K. Evans, “An experimental investigation of videotaped lectures in online courses,” TechTrends, vol. 58, no. 3, pp. 63–70, 2014.
View at: Publisher Site | Google Scholar
L. Kohnke and B. L. Moorhouse, “Facilitating synchronous online language learning through Zoom,” RELC Journal, p. 003368822093723, 2020.
View at: Publisher Site | Google Scholar
K. Jeong, “Integrating a web-based platform to promote creativity and authenticity in language classrooms,” International Journal of Knowledge and Learning., vol. 11, no. 2–3, p. 127, 2016.
View at: Publisher Site | Google Scholar
N. Almusharraf and D. Bailey, “Online engagement during COVID-19: role of agency on collaborative learning orientation and learning expectations,” Journal of Computer Assisted Learning, vol. 37, no. 5, pp. 1285–1295, 2021.
View at: Publisher Site | Google Scholar
A. Cheung, “Synchronous online teaching, a blessing or a curse? Insights from EFL primary students’ interaction during online English lessons,” System, vol. 100, p. 102566, 2021.
View at: Publisher Site | Google Scholar
C. Bartneck and J. Forlizzi, “A design-centred framework for social human-robot interaction,” in 13th IEEE International Workshop on Robot and Human Interactive Communication., pp. 591–594, Kurashiki, Japan, Sept. 2004.
View at: Publisher Site | Google Scholar
R. Arkin, M. Fujita, T. Takagi, and R. Hasegawa, “An ethological and emotional basis for human-robot interaction,” Robotics and Autonomous Systems, vol. 42, no. 3-4, pp. 191–201, 2003.
View at: Publisher Site | Google Scholar
T. Kanda, R. Sato, N. Saiwaki, and H. Ishiguro, “A two-month field trial in an elementary school for long-term human–robot interaction,” IEEE Transactions on Robotics, vol. 23, no. 5, pp. 962–971, 2007.
View at: Publisher Site | Google Scholar
J. Kennedy, P. Baxter, E. Senft, and T. Belpaeme, “Higher nonverbal immediacy leads to greater learning gains in child-robot tutoring interactions,” International Conference on Social Robotics, Springer, Cham, pp. 327–336, 2015.
View at: Publisher Site | Google Scholar
J. Kennedy, P. Baxter, E. Senft, and T. Belpaeme, “Social robot tutoring for child second language learning,” in 2016 11th ACM/IEEE International Conference on Human-Robot Interaction, Christchurch, New Zealand, March 2016.
View at: Publisher Site | Google Scholar
N. Spolaôr and F. Benitti, “Robotics applications grounded in learning theories on tertiary education: a systematic review,” Computers and Education, vol. 112, pp. 97–107, 2017.
View at: Publisher Site | Google Scholar
I. K. Crombie, Pocket Guide to Critical Appraisal, Bmj Publishing, 1996.
B. Kitchenham, E. Mendes, and G. H. Travassos, “A systematic review of cross- vs. within-company cost estimation studies,” in The 10th International conference on Evaluation and Assessment in Software Engineer, pp. 81–90, Rio de Janeiro, Brazil, April 2006.
View at: Publisher Site | Google Scholar
W.-C. V. Wu, R.-J. Wang, and N.-S. Chen, “Instructional design using an in-house built teaching assistant robot to enhance elementary school English-as-a-foreign-language learning,” Interactive Learning Environments, vol. 23, no. 6, pp. 696–714, 2015.
View at: Publisher Site | Google Scholar
S. Meiirbekov, K. Balkibekov, Z. Jalankuzov, and A. Sandygulova, “You win, I lose: towards adapting robot’s teaching strategy,” in 2016 11th ACM/IEEE International Conference on Human-Robot Interaction, pp. 475-476, Christchurch, New Zealand, March 2016.
View at: Publisher Site | Google Scholar
C.-W. Chang, J.-H. Lee, P.-Y. Chao, C.-Y. Wang, and G.-D. Chen, “Exploring the possibility of using humanoid robots as instructional tools for teaching a second language in primary school,” Educational Technology & Society, vol. 13, no. 2, pp. 13–24, 2010.
View at: Google Scholar
H. Banaeian and I. Gilanlioglu, “Influence of the NAO robot as a teaching assistant on university students’ vocabulary learning and attitudes,” Australasian Journal of Educational Technology, vol. 37, no. 3, pp. 71–87, 2021.
View at: Publisher Site | Google Scholar
Z.-J. You, C.-Y. Shen, C.-W. Chang, B.-J. Liu, and G.-D. Chen, “A robot as a teaching assistant in an English class,” in The 6th IEEE International Conference on Advanced Learning Technologies, pp. 87–91, Kerkrade, July 2006.
View at: Publisher Site | Google Scholar
J. K. Westlund, G. Gordon, S. Spaulding et al., “Learning a second language with a socially assistive robot,” in In Proceedings of New Friends: The 1st International Conference on Social Robots in Therapy and Education, Almere, The Netherlands, 2015.
View at: Google Scholar
O. Engwall and J. Lopes, “Interaction and collaboration in robot-assisted language learning for adults,” Computer Assisted Language Learning, pp. 1–37, 2020.
View at: Publisher Site | Google Scholar
G. Gordon, C. Breazeal, and S. Engel, “Can children catch curiosity from a social robot,” in ACM/IEEE International Conference on Human-Robot Interaction, pp. 91–98, Portland, Oregon, USA, 2015.
View at: Publisher Site | Google Scholar
J. de Wit, T. Schodde, B. Willemsen et al., “The effect of a robot’s gestures and adaptive tutoring on children’s acquisition of second language vocabularies,” in Proceedings of the 2018 ACM/IEEE International Conference on Human-Robot Interaction, pp. 50–58, Chicago, IL, USA, March 2018.
View at: Publisher Site | Google Scholar
D. P. Newton and L. D. Newton, “Humanoid robots as teachers and a proposed code of practice,” Frontiers in Education, vol. 4, p. 125, 2019.
View at: Publisher Site | Google Scholar
T. Ishino, M. Goto, and A. Kashihara, “Robot lecture for enhancing presentation in lecture,” Research and Practice in Technology Enhanced Learning, vol. 17, no. 1, 2022.
View at: Publisher Site | Google Scholar
N. Randall, “A survey of robot-assisted language learning (RALL),” ACM Transactions on Human-Robot Interaction, vol. 9, no. 1, pp. 1–36, 2020.
View at: Publisher Site | Google Scholar
E.-j. Hyun, S.-y. Kim, S. Jang, and S. Park, “Comparative study of effects of language instruction program using intelligence robot and multimedia on linguistic ability of young children,” in IEEE International Symposium on Robot & Human Interactive Communication, Munich, Aug. 2008.
View at: Publisher Site | Google Scholar
J. Shin and D.-H. Shin, “Robot as a facilitator in language conversation class,” in Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts, pp. 11-12, Portland, Oregon, USA, March 2015.
View at: Publisher Site | Google Scholar
J. S. Herberg, S. Feller, I. Yengin, and M. Saerbeck, “Robot watchfulness hinders learning performance,” in 2015 24th IEEE International Symposium on Robot and Human Interactive Communication, pp. 153–160, Kobe, Japan, Sept. 2015.
View at: Publisher Site | Google Scholar
Z. Hong, Y. Huang, M. Hsu, and W. Shen, “Authoring robot-assisted instructional materials for improving learning performance and motivation in EFL classrooms,” Educational Technology and Society, vol. 19, no. 1, pp. 337–349, 2016.
View at: Google Scholar
Ö. E. Demir-Lira, J. Kanero, C. Oranç et al., “L2 vocabulary teaching by social robots: the role of gestures and on-screen cues as scaffolds,” Frontiers in Education, vol. 5, 2020.
View at: Publisher Site | Google Scholar
R. van den Berghe, M. de Haas, O. Oudgenoeg‐Paz et al., “A toy or a friend? Children's anthropomorphic beliefs about robots and how these relate to second-language word learning,” Computer Assisted Learning, vol. 37, no. 2, pp. 396–410, 2021.
View at: Publisher Site | Google Scholar
R. van den Berghe, M. de Haas, O. Oudgenoeg‐Paz et al., “Individual differences in children’s (language) learning skills moderate effects of robot-assisted second language learning,” Frontiers in Robotics and AI, vol. 8, 2021.
View at: Publisher Site | Google Scholar
J. K. Westlund and C. Breazeal, “The interplay of robot language level with children's language learning during storytelling,” in The Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction, pp. 65-66, Portland, Oregon, USA, 2015.
View at: Publisher Site | Google Scholar
M. de Haas, P. Baxter, C. de Jong, E. Krahmer, and P. Vogt, “Exploring different types of feedback in preschooler and robot interaction,” in Proceedings of the Companion of the 2017 ACM/IEEE International Conference on Human-Robot Interaction, pp. 127-128, Vienna, Austria, 2017.
View at: Publisher Site | Google Scholar
M. Saerbeck, T. Schut, C. Bartneck, and M. D. Janse, “Expressive robots in education varying the degree of social supportive behavior of a robotic tutor,” in International Conference on Human Factors in Computing Systems, pp. 1613–1622, Atlanta, Georgia, USA, April 2010.
View at: Publisher Site | Google Scholar
K. M. Lee, N. Park, and H. Song, “Can a robot be perceived as a developing creature?: effects of a robot's long-term cognitive developments on its social presence and people's social responses toward it,” Human Communication Research, vol. 31, pp. 538–563, 2005.
View at: Publisher Site | Google Scholar
A. Muneeb, O. Mubin, S. Shahid, and J. Orlando, “Robot’s adaptive emotional feedback sustains children’s social engagement and promotes their vocabulary learning: a long-term child–robot interaction study,” Adaptive Behavior, vol. 27, no. 4, pp. 243–266, 2019.
View at: Publisher Site | Google Scholar
P. Vogt, R. van den Berghe, M. de Haas et al., “Second language tutoring using social robots: a large-scale study,” in 2019 14th ACM/IEEE International Conference on Human-Robot Interaction, pp. 497–505, Daegu, Korea (South), March 2019.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Yutong Ao and Zhonggen Yu. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

813

Downloads

885

Citations