Daring Role-play and Speaking Anxiety : A Case Study in an Online English Speaking Class

Though studies about role-plays may be frequent, studies specifically investigating the possible relation between role-plays and speaking anxiety may still be very rare, particularly in the Indonesian second/foreign (L2) context, despite anxiety being a quite widespread phenomenon among language learners. Hence, the present study was conducted to investigate Indonesian learners' extent of participation in role-play activities in an online English speaking class and learners’ and teachers' views about role-plays concerning speaking anxiety. The study employed a qualitative case study using online observations and interviews. The participants were 25 English education department learners and two teachers implementing role-plays. Through observations, the study found that role-plays were unable to make most of the learner participants voluntarily participate in class. Through interviews analysed using Thematic Analysis, the study found three themes. Learners' nervousness and fear of making mistakes negatively affected their role-play performances. Authentic situations in role-plays allowed them to feel more comfortable. Next, preparing role-play for performances in groups lessened learners’ anxiety. It is suggested that teachers offered learners choices whether to do live role-plays or show pre-recorded one. The latter may reduce anxiety and make it possible for learners to redo some parts allowing them to showcase their best possible performance.


INTRODUCTION
In a context where English is used as a foreign language, the use of the language is largely limited to the classroom context (Subekti, 2018b). Learners in such situations tend to be selfconscious about how well they can communicate (Baker & MacIntyre, 2003) due to the limited exposure to the language, and thus are very susceptible to anxious feelings when using the language (Subekti, 2018b). Furthermore, Katemba (2013) stated that anxiety seems to be a crucial issue among foreign language learners. In their canon publication on the field of speaking anxiety, Horwitz et al. (1986) defined foreign language anxiety, focusing on speaking anxiety, as "a distinct complex of self-perceptions, beliefs, feelings, and behaviours related to classroom language learning arising from the uniqueness of the language learning process" (p. 128), thus specifying this kind of anxiety exclusive to foreign language environments.
After the seminal work of Horwitz et al. (1986), countless studies have been conducted in the field of speaking anxiety in various second language (L2) learning contexts (Gürler, 2015;Mak, 2011;Subekti, 2018b), suggesting the growing importance of the topic in the field. Several studies suggested that speaking anxiety negatively correlated with language achievements (Hewitt & Stephenson, 2012;Rachmawati & Jurianto, 2020;Shao et al., 2013;Subekti, 2018b), perhaps suggesting the debilitating effects of anxiety towards L2 learning.
Studies have identified several factors attributed to speaking anxiety. One of the most common ones is fear of negative evaluations (Akkakoson, 2016;Tzoannopoulou, 2016). A study in a Thai context by Akkakoson (2016) and a recent study in Indonesia by Subekti (2020b) found that learners tended to be even more apprehensive if they perceived themselves to be lagging behind their peers. Khusnia (2017) noted that at times learners underestimate their speaking ability and this leads to lack of confidence and speaking anxiety. Probably due to this fear of negative evaluations, learner participants in various studies reported that they needed time to prepare for their language performance, otherwise, they would feel apprehensive (e.g.: Mak, 2011;Subekti, 2019). Though not specifically in L2 literature, a study by Cotterill (2014) also found that learners used preparation to help them to cope with their anxious feeling and to lessen mistakes they might create when performing role-plays. Learners with good preparation tended to have more organised performance than those with less preparation (Cotterill, 2014).
Another possible factor is teacher-learner interaction (Al-Saraj, 2014). In an Indonesian university context, Subekti (2018a) found that teachers' supportive attitudes shown through their supportive facial expressions could help lessen learners' anxiety and make them feel more comfortable in L2 speaking. A learner participant in the study by Subekti (2018a) stated "If they [the teachers] give "nice" expression, we will feel confident. But when they show disappointed and upset face, we will become anxious like we have just made mistakes (p. 231). Furthermore, several authors also suggested that teachers need to build a friendly and lesspressure environment in class to help learners lessen their anxiety (Akkakoson, 2016;Subekti, 2020a). Once learners feel they are in a familiar term with their teachers, the class atmosphere tends to become more relaxing and this condition allows more risk-taking behaviours among learners (Akkakoson, 2016).
Furthermore, though several learners may need more than just words of encouragement from their teachers (Wong et al., 2019), in general, teachers' encouragement was found to be able to boost learners' participation in class despite their limitations. This could be seen from the findings of several studies in the L2 context and beyond (Ducca-Díaz, 2014;Guéguen et al., 2015;Mustapha et al., 2010;Russell & Shepherd, 2010). An experimental study in a French context by Guéguen et al. (2015), for instance, reported that more learners succeeded in the verbal encouragement group (82%) than in the control group (47%), indicating the positive effect of encouragement towards learners' success. A qualitative study in Malaysia by Mustapha et al. (2010) also reported that teachers' verbal encouragement and understanding of learners' limitations made learners feel more at ease and comfortable enough to share their thoughts. In line with that, Subekti (2018a) also reported that when learners perceived their teachers to be very attentive to learners' progress and make effort for their sake, learners also tended to be more confident and less anxious in language class.
One prominent class activity often thought to boost learners' participation is role-play (Aliakbari & Jamalvandi, 2010;Eppendi, 2016). Fadilah (2016) argued that role-play activities allow learners to practice communicating in various social contexts and different roles. The activities also prepare learners to understand more about social situations since the activities require learners to act or pretend as someone else in various situations often found in real contexts (Umam, 2011). In other words, role-plays could help create real-life situations (Duong, 2014;Hartatik & Rahmah, 2016) and these real-life situations help learners to feel comfortable when they participate in the class activities (Russell & Shepherd, 2010). In an experimental study in an Indonesian Junior High School, Eppendi (2016) found that the learner participants taught using role-plays had significantly better speaking achievement than those who were not taught using role-plays, suggesting the positive impact of role-plays on speaking performance. In line with that, in an Iraqi college context, an experimental study by Krebt (2017) also reported learners who participated and practiced together in group role-plays experienced a significant improvement in speaking skills, indicating the positive role of role-plays in positively influencing learners' speaking skills.
Learners from English education majors may also experience doing role-plays during their studies, in their speaking classes, for example (Clarita et al., 2020). Several studies in Indonesia reported that learners from English education majors still feel anxious when speaking English (Fitriah & Muna, 2019;Mitha et al., 2019;Putri & Marlina, 2019;Tridinanti, 2018). Hence, investigating such learners' speaking anxiety concerning their uses of role-plays, repeatedly reported to be effective in making learners feel more comfortable and in improving speaking achievement, could be strategic and important. Involving teachers in implementing role-plays could also provide another important angle on seeing the phenomenon since teachers have a big role in determining class activities (Xu & Huang, 2010). Teachers might also have a wide understanding of learners' speaking ability and speaking anxiety. The plethora of literature, mainly in the forms of quantitative studies as previously mentioned, on role-plays and speaking anxiety suggested that these phenomena are widely researched, but studies specifically investigating the possible relation between role-plays and speaking anxiety may still be very rare, particularly in the Indonesian L2 context. Furthermore, conducting a qualitative study involving both Indonesian L2 learners and teachers of English in this field could offer in-depth and multi-angled perspectives enriching the literature.
Considering the rationales of the study, the following research questions are addressed in the present study: a.
To what extent do learners participate in role-plays in an online English Speaking class? b.
What are learners' and teachers' views towards role-play in relation to speaking anxiety?

RESEARCH METHODS Research design and participants
The present study used a qualitative case study as the design. It was a study aimed at investigating a phenomenon in-depth and provide rich descriptions of the phenomenon focusing on its uniqueness (Basit, 2010). With this purpose in mind, the study employed qualitative methods of conducting three online observations and individual online semi-structured interviews (Gray, 2014). Observations were conducted to know learners' level of participation in role-play activities, answering the first research question, and select several learners for interviews based on their visible anxiety behaviours. Furthermore, online interviews were conducted to investigate learners' and teachers' subjective and in-depth views on the use of role-playing in relation to speaking anxiety. Interview checklist consisting of fifteen questions for learners and seven questions for teachers was developed to ensure the interviews were relevant to the research questions. An example of questions asked to learners is "What is your opinion about role-play in speaking class? Are you relaxed or nervous when role-playing?" The participants of the present study were 25 learners in an online speaking class conducted once a week at an English Language Education Department (ELED) at a university in Yogyakarta, Indonesia, and two English teachers implementing role-plays at the same institution. Class instruction was delivered through Zoom teleconferences. From the three observations, six learners were chosen based on their anxiety level as observed: two learners with high speaking anxiety, two learners with medium anxiety, and two learners with low anxiety. Learners' level of anxiety was seen from how they performed their role-plays. The categorisation was loosely based on several authors' reiterations and research findings on the behaviours of anxious learners such as silence (Al-Saraj, 2014;Maher & King, 2020), stammering when talking (Suleimenova, 2013), and making mistakes on things that could had been easy for learners (Subekti, 2018a). Suleimenova (2013) mentioned some other anxiety behaviours such as squirming, playing with hair or clothing, and nervously touching objects. However, as the observations were conducted online, there were limits on what could be observed. Hence, learners who spoke in a stammering voice and made numerous mistakes whilst performing were considered to have high anxiety. Learners whose voice sounded a little breathy and who made fewer mistakes whilst performing were considered to have medium anxiety. Finally, learners who had a relaxed voice and demonstrated fluency in performing were considered to have low anxiety.

Data collection and analysis
After permission to conduct the study was granted from the Head of the ELED as the gatekeeper (Ramrathan et al., 2016), we contacted the teachers of the class to observe to see the feasible schedules for three online observations. Then, online observations were conducted three times in a row within October 2020, the first being on the 8 th , the second on the 15 th , and the third on the 22 nd . The results of the three observations were used to answer the first research question on learners' degree of participation in role-play activities and to select six learner participants based on their anxiety level as previously mentioned. The next step was to conduct individual interviews in the Indonesian language with these six selected learners and two teachers implementing role-play. These interviews, each of which ranging from twenty to thirty minutes in duration, were conducted within November 2020. The interview data were transcribed, translated into English, and analysed using Thematic Analysis to answer the second research question. Thematic Analysis was conducted by getting familiar with the interview data, formulating initial codes, for example "Anxiety---Worse RP", "Group---Better RP" and "Authenticity---Less Anxiety", generating themes from these codes, reviewing these themes, formulating the final themes for the report (Braun & Clarke, 2006). This process was conducted through reading and rereading the interview data, making annotations in the documents, and making separate notes as necessary.

Ethical considerations
The present study employed several principles of research ethics. First, the participants' voluntary participation was guaranteed through the use of a consent form (Gray, 2014) distributed to the participants before each interview. It was also to ensure that they knew the objective of the study, their rights and responsibilities as participants before they decided to participate. Pseudonyms were used throughout this report to protect the confidentiality of participants and to avoid possible harm (Hall, 2010).

First observation
This class had 25 learners but the number of learners who participated was 22. From these 22 learners, there were just around nineteen learners who actively participated in roleplay activities through teleconference. The duration of the role-play for each group was seven to ten minutes. Seven learners were participating actively and taking voluntary actions in the role-plays. Four of them always offered themselves to read the script in the role-plays by saying "Miss, I will/want to do it". One of them always raised a hand and said "Choose me Miss" in a seemingly excited tone. The other three learners sometimes offered themselves to do the roleplays especially the simple ones. Furthermore, the other twelve learners participated in the roleplays when the teacher picked or asked them to do that. The teacher called their names one by one in each of the role-play activities. Among the learners chosen by the teacher, four learners sounded stuttered whilst reading scripts and three learners were able to speak easily and comfortably whilst role-playing.

Second observation
The number of learners who participated was 23. Not all learners could actively participate in role-playing. A few of them had some technical problems. Three learners were willing to present their group's role plays first and they seemed to be confident in performing the role-play seen from the way they delivered their role-play in which they looked comfortable, and they did not make mistakes. Additionally, they also volunteered themselves by saying "Miss, we want to show our performance first because we are ready". These learners' performance was well-delivered, and they showed their mastery of the script. The teacher also gave some compliments such as "Good job" and "You guys have a nice preparation". Other than these three learners, the other 20 learners were picked by the teacher to perform their roleplay because they were not volunteering themselves. Among these 20 learners, there were also several who were confident to do group role-plays although some of their friends in groups still spoke in a stuttering voice, indicating nervousness.

Third observation
Learners were asked to present the role-plays that they had prepared from the previous meeting. Several learners were willing to present their role-plays. They said "Our group wants to perform first, Ma'am" and the teacher let them perform. After the first group, there was no other group volunteering themselves. In the role-play activity, there was one learner who did improvisation. He added some words that were not mentioned in the script like "Oh, thank you for your service". He showed good improvisation seen from the tone of his voice and in the way he delivered the role-play. In comparison, there were several learners making mistakes whilst performing their role-plays and speaking with a stuttering voice. The teacher helped give corrections and encouraged them to continue, "It's okay, just continue, everyone makes mistakes. We all still learn". The encouragement words from the teacher seemed helpful because the learners continued to finish their lines rather than stopping.
Based on the findings of the first and second observations, not all learners participated voluntarily in the role-plays. Only eight up to twelve of them volunteered themselves. The other learners were waiting until the teacher called them by their names. Moreover, there was a small number of learners who spoke with a stuttering voice and made unnecessary mistakes whilst performing. This observation finding indicated that several learners still experienced anxiety when performing role-plays. This finding was in line with the reiterations of several authors on anxiety behaviours such as silence (Al-Saraj, 2014;Maher & King, 2020), stammering when talking (Suleimenova, 2013), and making mistakes on things they could have mastered (Subekti, 2018a). Furthermore, in relation with role-plays, this observation finding was in contrast with findings of several previous studies (Duong, 2014;Puspitorini, 2018;Rahman & Maarof, 2018;Ruslan, 2020) in which they found that by utilising role-play activities, learners could be braver and more confident in doing speaking performance. The present study's findings may suggest that role-play may not be a one-size-fits-all strategy to alleviate learners' anxiety. That was because role-plays may not suit all types of learners as several learners might find it hard to deal with their nervousness whilst role-playing. The online role-play environment in the present study could also play a part in which those who were performing may feel that they were in the spotlight of the teleconference as the only ones with microphones on. Furthermore, in response to several learners' observable anxiety behaviours, as seen in the third observation, the teacher encouraged learners to make them continue their performance despite the mistakes. Regarding this, several authors reiterated the role of teachers' encouragement to boost learners' participation (Guéguen et al., 2015;Mustapha et al., 2010;Russell & Shepherd, 2010). Furthermore, the present study's finding confirmed the bigger role teachers played to engage learners in role-plays in online settings (see Russell & Shepherd, 2010).

Learners' and teachers' views towards role-plays in relation with speaking anxiety
To facilitate reference-tracing of each interview excerpt, the following codes were used: "F" for Female, "M" for Male, "HA" for High Anxiety, "MA" for Medium Anxiety, "LA" for Low Anxiety, and "Tc" for Teacher. Thus, "[Wiwi/F/HA]", for example, means that the excerpt was from Wiwi, a female learner with high anxiety. All names were pseudonyms.
Through Thematic Analysis, the present study found three themes on learners' and teachers' views towards role-plays in relation to speaking anxiety. These could be seen in Table  1. Learners' anxiety (feeling nervousness, fear of failure) negatively affected their roleplay performance. Theme 2.
Role-plays created authentic situations making learners feel less anxious in performing role-play. Theme 3.
Role-play preparation in groups reduced learners' fear of failure whilst performing.

Theme 1: Learners' anxiety (feeling nervousness, fear of failure) negatively affected their role-play performance.
Learners' feelings of nervousness and their fear of failure were reported to have affected their performance negatively. Regarding this, Wiwi reported: "When I feel nervous, I make many mistakes. I can speak easily when I'm alone but when I perform in front of my friends, suddenly… I forget everything that I memorised… I'm afraid I will make mistakes." [Wiwi/F/HA] Didin and Reinald also shared almost similar experiences. They stated: "I feel nervous, just a little. Maybe when I make mistakes and make wrong pronunciation, I feel like everyone will laugh at me." [Didin/M/HA] "I still feel nervous and stutter when I speak. I often speak with wrong pronunciation and I'm afraid that my friend will laugh at me...." [Reinald/M/LA] As seen in the excerpts, all learners from high, medium, and low anxiety seemed to have a convergent view that their fear of failure negatively affected their role-play performances. Regarding this, Aliyah, one of the teachers, also confirmed what these learners stated. She reported: "They are afraid that their friends might laugh at them when they make mistakes." [Aliyah/F/Tc] The present study's finding was in line with the findings of several previous studies (Subekti, 2018a(Subekti, , 2018b. The present study's finding could give some kind of qualitative confirmation of a quantitative finding in the study by Subekti (2018b) that learners' anxiety was negatively associated with speaking performance.
Regarding teachers' efforts to help learners deal with their anxiety, learners mentioned that their teachers encouraged them to participate in role-play activities to develop their speaking skills. Didin, for example, reported: "The teacher also encouraged me to play a role in role-play even though I was nervous and afraid of making mistakes." [Didin/M/HA/St] Didin's statement corresponded with Aliyah's statement. She stated: "I encourage them by saying 'don't be shy, this is just practice, everyone makes mistakes.' Also, I make sure they participate in the role-plays. Usually, I will call them by their names and ask them to perform. I believe that if other learners can then they also can." [Aliyah/F/Tc] Regarding this finding, several authors reported the role of teachers' encouragement to engage learners to actively participate in class activities (Ducca-Díaz, 2014;Guéguen et al., 2015). The present study's finding may also serve as a qualitative confirmation of a finding in a quantitative study by Guéguen et al. (2015) in a French context in which learners who received encouragement more often succeeded in reaching their goals than the learners who received no encouragement. However, the present study's finding was in contrast with a finding of study by Wong et al. (2019) reporting that though several learners felt engaged when receiving words of encouragement, some others still needed 'more than just words' to feel engaged. The different findings may suggest that to engage learners, teachers, aside from encouraging, needed to facilitate other contributing factors such as interesting, low-stress activities, and the right amount of challenge.

Theme 2: Role-plays created authentic situations making learners feel less anxious in performing role-play.
Both teacher participants argued that role-plays provided authentic situations through which learners could connect materials in class with real-life situations. Maia, for example, stated: "Roleplay allows us to act like someone else and see something from their perspective. It is effective because it is the bridge to connect the materials in class with the real-life situation." [Maia/F/Tc] About speaking anxiety, Hillary, a learner with medium anxiety, reported that because role-plays mirrored real-life situations, she feels more comfortable performing them. She stated: Regarding the finding that authentic situations made learners feel more comfortable, a previous quantitative study by Hartatik and Rahmah (2016) in an Indonesian context also found that learners felt more relaxed to participate in class activities when authentic materials were applied. That learners could place themselves in real-life scenarios in role-plays (Russell & Shepherd, 2010) may help them figure out what to do during role-plays and this could lower their anxiety.

Theme 3: Role-play preparation in groups reduced learners' fear of failure whilst performing
The present study found that role-play preparation in groups helped them lessen their nervousness. Wiwi a learner with high anxiety, for example, reported: "By doing the role-play in a group, I feel more prepared and less anxious because I know my friends will help me if I make mistakes." [Wiwi/F/HA] In line with Wiwi's comments, Surya, a medium anxiety learner, also mentioned: "Roleplay in a group helps me to work together with my friends and prepare the things that we want to show. If we are prepared, I feel not so scared because I already know what to do" [Surya/M/MA] As seen in the excerpts, learners felt more comfortable because they could prepare themselves with their group to lessen the possible mistakes that might happen. The group preparation helped learners lessen their anxiety and made them feel comfortable to perform. The present study's finding was in line with that of a quantitative study by Cotterill (2014) reporting that preparation for performance is a crucial action in helping the learners to cope with the pressures of the performance. Related to the similar finding, though not specifically about role-plays, a study by Mak (2011) in a Chinese context generally found that learners needed to be given preparation time before speaking. The finding further reiterated the important role of preparation time for learners before they made any language performances.

CONCLUSION
The study has at least two limitations. First, the results of the observation where we observed anxiety behaviours may inherently become a limitation. What we perceived as anxiety behaviours may not be the case if seen from the observed learners. Hence, the categorisation of learner interviewees into high, medium, and low-anxiety learners may to a certain extent be slightly compromised. Secondly, this study was unique in its context. Hence, the findings may not represent participants from other learning contexts. This present study had several implications and contributions. First, as this study found that role-plays were unable to make all learners voluntarily participated in the activity, teachers could prepare other class activities as alternative plans to engage learners, for example, playing online games, having fun talk shows, and creating show-and-tell activities. Secondly, as learners reported fear of making mistakes when performing role-plays, teachers could at times offer learners two choices, performing live role-plays or submitting recorded role-plays. Recorded role-plays could allow the performers to redo parts where they make mistakes and submit the best versions. Doing so in groups could also provide the chance to realise peer support and make them feel more prepared. Furthermore, this study could pave a way for future studies about role-plays in online speaking classes in the Indonesian context.
There are several recommendations for future studies. First, conducting a quantitative study to find the possible interaction between speaking anxiety and attitude towards role-plays may be worthwhile. Second, some excerpts of the present study indicated that some participants were afraid of making wrong pronunciation and it leads to speaking anxiety. Regarding this, it is recommended to investigate how certain activities can be utilised to deal with students' fear of making wrong pronunciation.