Investigating willingness to communicate in synchronous group discussion tasks: one step closer towards authentic communication

1.1 Willingness to communicate

The notion of WTC was originally introduced with reference to first language (L1) communication to explain why some people are more willing to talk than others (McCroskey and Baer 1985), based on Burgoon’s (1976) work on unwillingness to communicate. As a “non-observable psychological construct” (Peng 2020, p. 143), WTC entails an intention to initiate communication which is often predictive of communicative behaviour (MacIntyre 1994). While L1 WTC was perceived as a personality-based, trait-like disposition (McCroskey and Richmond 1991), WTC in a second language (L2) was reinterpreted as a situational variable, open to change across time and situations and defined as an intention “to enter into discourse at a particular time with a specific person or persons, using a L2” (MacIntyre et al. 1998, p. 547).

MacIntyre et al. (1998) schematized L2 WTC in a pyramid model and conceptualized a range of social, contextual, psychological, and linguistic variables that converge to shape WTC in L2. In the past decade, this model has been extensively tested in diverse informal and instructional contexts with young learners and adult students (Cao and Wei 2019). Empirical research conducted in the last decade has identified causal relationships between L2 WTC and several learner variables, including communication and language-related factors of perceived communicative competence, perceived opportunity to communicate, perceived L2 proficiency, L2 fluency, and language experience; and psychological factors of language anxiety, perceived communicative competence, personality, motivation, attitude, learner belief, emotion, gender and age (Cao 2011, 2014; Fushino 2010; Khajavy and Ghonsooly 2017; Nematizadeh 2019; Zhou et al. 2020). Elahi Shirvan et al.’s (2019) meta-analytical study examined the overall average correlation between L2 WTC and the most frequently identified predictors, including perceived communicative competence, language anxiety, and motivation. They found that these variables moderately correlated with L2 WTC with perceived communicative competence having the largest effect.

1.2 Trait and situational WTC

Similar to other individual difference (ID) factors in SLA including anxiety, personality, motivation, and emotion, WTC in L2 also has dual characteristics: trait and situational WTC. Trait WTC prepares individuals for communication by creating a tendency to place themselves in situations where communication is expected; situational WTC, on the other hand, influences the decision to initiate communication in a particular situation (MacIntyre et al. 1999). While much of the quantitative research investigated L2 WTC as a trait variable, Kang’s (2005) qualitative study was the first attempt that offered insights into L2 WTC as a situational variable. She found that situational WTC was a multi-layered construct that could change from moment to moment in the conversational context, under the joint effect of the psychological conditions and situational variables.

L2 WTC as a situational construct has been investigated in experimental, quasi-experimental, and classroom settings (e.g., Cao 2011, 2014; MacIntyre and Legatto 2011; Pawlak and Mystkowska-Wiertelak 2015). In an earlier classroom-based study on L2 WTC, Cao and Philp (2006) compared English learners’ self-report of WTC and their WTC behaviour in class in New Zealand and found that interactional settings (i.e., whole class, small groups, and dyads) exerted a significant influence on WTC. Cao’s (2011, 2013, 2014 longitudinal study on L2 WTC in the same context used a multiple case study approach and employed an ecological framework. The key participants were followed over an academic year and their classroom WTC was examined as lessons unfolded each week. This study revealed dynamic changes in learners’ WTC on lesson-to-lesson, task-to-task, and moment-to-moment bases. Methodologically, this study contributed to validating a classroom observation scheme of WTC behaviour in class.

Peng (2014) examined Chinese EFL learners’ WTC as situated in university English classrooms and found a complex interrelationship among individual, contextual, and affective predictors of WTC. Her studies also highlighted the cultural influence on L2 WTC. Yashima et al. (2016) intervention study in a Japanese EFL class over one semester revealed that differences in the frequency of self-initiated turns in discussion sessions emerged through the interactions of enduring individual characteristics such as personality and proficiency and contextual influences including interlocutor’s reactions and group-level talk-silence patterns. Zhang et al. (2018) conducted a systematic review of 35 research studies on the situational antecedents contributing to variation in WTC. This study highlighted the role of the learning situation and concluded that the situational variables could be organized into a multi-layered framework to explain the dynamic processes underpinning WTC. Al-Murtadha (2021) compared self-reported WTC in L1 and L2. He found no significant correlations between self-reported and observed L1 WTC or between self-reported and observed L2 WTC. However, significant and positive correlations were found between observed L1 and L2 WTC.

1.3 Complex dynamic systems (CDS) approach and WTC research

The application of the theory of complex dynamic systems (CDS) in L2 research, and ID variables in particular, has recently gained ground producing new insights into uncharted dimensions in SLA. With its principal focus on the phenomenon of change, CDS has secured a move away from the use of reductionist quantitative approaches (Gaddis 2002) and the perception of causality (Larsen-Freeman and Cameron 2008). Specific to ID research, Dörnyei (2009) urged a reconsideration of two fundamental issues concerning individual-level variation: the conceptualization of a dynamic system of multiple interrelated ID components; and the feasibility of distinguishing different types of ID factors. He further argued that the interconnected, constantly changing, and environmentally sensitive IDs system serves as an example of a complex, dynamic system. He proposed that individual variation in the system is not so much a function of the strength of any individual determinants such as motivation or aptitude but rather of all the relevant factors working together.

Dörnyei et al. (2015), in their edited volume on motivational dynamics, argued that the theory of complex dynamic systems looks at the dynamics that emerge from the interaction between the interdependent components of a variable. Due to this interaction, CDS tend to be dynamic and nonlinear, meaning that they, sometimes, change continuously or discontinuously (Larsen-Freeman and Cameron 2008). Such variability is not captured through correlational means.

The CDS theory has been employed in several studies to investigate dynamism and moment-to-moment fluctuation in WTC in experimental and classroom contexts. In the Canadian context, MacIntyre and Legatto (2011) employed an idiodynamic method to investigate dynamism in L2 WTC in a laboratory setting. They monitored changes in participants’ WTC across and during eight communicative tasks and measured WTC and anxiety as variables. They found an interaction among the affective, cognitive, and linguistic systems and this interrelationship triggered changes in WTC. Failure in retrieving vocabulary and unfamiliarity with the discussion topics influenced the participants’ WTC. MacIntyre (2020) reported on his research group’s work on L2 WTC dynamism on a timescale of a few minutes and the “complex interactions and rapid changes in participants’ psychological conditions” (p. 111) when approaching and avoiding communication.

The situated L2 WTC was also investigated in the Polish context and the findings confirmed that learners’ WTC was in a state of flux and influenced by a host of individual, linguistic, and contextual variables (Pawlak and Mystkowska-Wiertelak 2015; Pawlak et al. 2015). For example, Pawlak and Mystkowska-Wiertelak (2015) observed dynamism in L2 learners’ WTC during communicative tasks. Through the use of self-rating, questionnaires, and interviews, they found that contextual factors (e.g., topic, interlocutor, planning time), linguistic resources, and individual differences led to shifts in WTC. This study hinted at an interplay among linguistic, affective, and cognitive factors. This study confirmed MacIntyre and Legatto’s (2011) finding that participants’ WTC levels dropped when they struggled at lexical and conceptual levels. Failure in vocabulary retrieval caused an increase in language anxiety and loss of focus, which in turn negatively affected WTC.

Pawlak et al.’s (2015) subsequent classroom-based study investigated dynamism in students’ WTC during different classroom events. Apart from students’ self-rating and questionnaire data, teachers’ comments on students’ WTC were also used. They observed between-group changes to WTC due to interactions between contextual and individual factors and some of these interactions were unpredictable. The results also confirmed the influence of contextual factors (e.g., task, topic, and interlocutor) on WTC changes.

Employing an idiodynamic method, Nematizadeh and Wood (2019) examined the affective and cognitive dynamics of L2 students’ WTC. They explored how the affective and cognitive variables interact with temporal measures of speech. They found that the affect, as an underlying layer of WTC, and cognitive processes underlying speech fluency interacted in an unpredictable, multidimensional, and interconnected manner. The factors affecting WTC included vocabulary retrieval, preparedness, topic familiarity/transition, self-monitoring and inaccurate language. These factors also simultaneously affected speech fluency. The findings accounted for some features of complex dynamic systems including complexity, dynamism, and nonlinearity.

Similarly, Sato (2019, 2020 employed a CDST approach to investigate the relationships between WTC and its underlying factors, particularly the language proficiency of the speakers or their interlocutors. He used post-task stimulated recall activities, requiring participants to mark changes in their WTC on a transcription or a scale. The findings from Sato’s studies point to the soft-assembly nature of WTC when differing factors affected students’ WTC at lower-intermediate and advanced proficiency levels. These findings are in line with previous research (Cao 2011, 2014; Kang 2005) that interlocutor’s proficiency can promote or prohibit WTC (Ducker 2022).

The most recent attempt at investigating L2 WTC using CDST was Ducker’s (2022) research that examined WTC antecedents, ongoing WTC ratings, and spontaneous communication in an intact classroom. A unique finding from this study was the crucial role of listening comprehension as a prerequisite for participation and listening as a primary focus of learner behaviour. Ducker (2022) identified multiple engaged listening stances, including listening as a distinct turn-taking behaviour, listening carefully while observing a policy of non-interruption, listening while strictly observing allocated turns, and listening as a stance or approach to the activity (in this case, to learn about partners and topics).

Our review suggests that aspects of dynamism and interrelatedness are already recognized to some extent in the existing literature on L2 WTC in face-to-face interaction. However, there has been little attempt at investigating its dynamism in online synchronous communication.

1.4 WTC construct in online tasks

A handful of studies have explored WTC in online communication and teaching. Freiermuth and Jarrel (2006) compared Japanese EFL students’ WTC in face-to-face and online environments. This experimental study revealed that online chatting provided a more comfortable environment for the students and enhanced their WTC. In a subsequent study, Freiermuth and Huang (2012) examined Japanese EFL students’ motivation in chatting with Taiwanese EFL students using online synchronic software. They found that pronunciation difficulties and linguistic and paralinguistic deficiencies could lead to an increase in anxiety, lowering their WTC in face-to-face group interaction. In contrast, an online environment facilitated safe opportunities to communicate. Al-Amrani (2013) also compared Omani EFL learners’ WTC in synchronous, asynchronous communication, and face-to-face interaction, and reported less WTC in face-to-face communication compared to online communication due to fear of making mistakes, feeling of embarrassment, and loss of face in face-to-face contexts. Al-Amrani’s (2013) study confirmed findings from Freiermuth’s research that online communication is a more favourable mode than face-to-face interaction.

Research has also explored the positive impact of educational technologies on L2 learners’ WTC. For example, the role of massively multiplayer online games (MMOGs) in L2 learning was reviewed and synthesized by Jabbari and Eslami (2019), who found that L2 learners demonstrated a lower level of anxiety in the safe environment provided by MMOGs. The games presented many opportunities and socialization activities to enhance L2 learners’ communicative competence. Kruk’s study (2020) observed two advanced English learners’ experience in an online virtual reality game and explored their fluctuating trajectory of WTC, motivation, boredom, and anxiety. Results revealed that trajectory changes were influenced by both positive factors including interesting topics and cooperative interlocutors and negative factors such as unpleasant L2 users, monotony, boredom, and prior negative experience. Ebadi and Ebadijalal (2020) explored the use of the Google Expeditions VR tool and the results indicated that the students’ general knowledge, motivation, confidence, and cultural awareness were promoted with the help of technology and hence resulted in an improvement in their WTC and oral performance.

Lee’s (2019) study examined factors affecting Korean EFL learners’ WTC in the extramural digital environment and found that socio-political, contextual, and individual variables interacted with each other to influence participants’ L2 WTC in the extramural digital context. Lee and Lee’s (2020) study examined the extent to which the affective factors including motivation, self-confidence, risk-taking, L2 speaking anxiety and grit, and virtual intercultural experiences were linked to L2 WTC in in-class, out-of-class, and digital settings. Results suggested that demographic and affective variables as well as involvement in virtual intercultural activities played distinct roles in influencing EFL students’ WTC in three different L2 communication contexts. In particular, younger individuals who had a higher level of L2 confidence and engaged more frequently in virtual intercultural experiences had higher L2 WTC in digital settings.

Ardiansyah et al. (2020) conducted an idiodynamic multiple case study to examine the dynamism of students’ WTC in an online discussion task. Three strands of factors were found to underlie online WTC, namely, psychological, conversational, and linguistic. This study also highlighted the difference between online and offline communication, identifying the internet as a mediator to promote or hinder the WTC level. It was suggested, however, with an adequate level of self-confidence, technological issues can promote WTC in the form of asking for clarifications.

1.5 Research aims and questions

Previous empirical research on L2 WTC has revealed its trait and situational levels in face-to-face conversational and classroom contexts. Situational WTC has been found to emerge from complex dynamic interactions among social, contextual, cognitive, linguistic, and individual factors in face-to-face settings (Cao 2014; Nematizadeh 2019), but this dynamic nature of WTC has not been explored in online group discussion tasks using a mixed methods design. A further gap in previous WTC research is the lack of exploration of the relationship between observed WTC and speaking time. To address these research gaps, the current study aims to gauge the WTC of intermediate Farsi-speaking ESL speakers in online synchronous discussion tasks. The present study aims to address the following research questions:

1. Are there any correlations between self-reported trait WTC, trait WTC, dynamic WTC, observed WTC, online WTC, and speaking time?

2. To what extent does the participants’ WTC fluctuate during synchronous group discussion tasks?

3. What factors do the participants attribute their WTC fluctuations to?

2.1 Setting and participants

Upon receiving ethics clearance, the study recruited Farsi-speaking ESL speakers (2 males and 5 females) using two inclusion criteria – proficiency level (recent IELTS score of 6.5 and 7 – or C1 on CEFR scale) and length of residence (less than 2 years) in Canada. Previous research has found that the length of residence affects language confidence (Lopez and Bui 2014) which is a predictor of L2 WTC (MacIntyre et al. 2001). Language proficiency has also been suggested to affect self-confidence and WTC (Clément 1986; Yashima 2002). Therefore, the inclusion criteria were adopted to offset the effect of these variables.

Recruitment messages were posted in the Telegram channels of Iranian communities across Canada, and interested individuals contacted the first author for further details. Seven participants, who met the above criteria, received and signed online consent forms. Table 1 provides the profile of the participants. The participants did not know each other prior to the study.

2.2 Discussion tasks facilitators

An invitation to take part in the study with information on the study goals and procedures was sent to potential discussion tasks facilitators who were experienced ESL teachers at a Canadian post-secondary institution and already familiar with synchronous teaching and the Zoom application. Those interested (1 male and 5 females) later received specific instructions (Appendix 1) on facilitating the discussion tasks online to ensure consistency.

2.3 Instruments

2.4 Data collection procedure

The participants were introduced to the data collection procedure in advance using a video tutorial in Farsi. This tutorial offered tips for using and downloading Zoom, completing online WTC Google Forms, using the WTC rating scale, and jotting-down notes during the stimulated recall stage. Participants were also asked to complete an online version of the trait WTC instrument on Google Forms.

Then six thirty-minute Zoom sessions were scheduled based on the participants’ availability in August and September 2020, each moderated by a different facilitator. Six previously piloted topics (Nematizadeh 2019), which included food, technology, traveling, online versus on-campus education, crime, and advertisement, were used in the study. The first author attended and recorded all the sessions, but did not participate in the discussions. The discussion tasks started with a quick introduction to the topic and a beeping sound to prompt the participants to input their initial WTC level on a scale between −5 and +5 using the WTC rating scale. The sound continued at intervals of 5 min. The rating scale also provided space for the participants to jot down notes that later assisted them with recalling the factors that triggered WTC shifts. Immediately after, in a stimulated recall procedure, the researcher played the recording of the session and shared the screen to help the participants with recalling their thoughts while they were performing the task and elaborating on the factors triggering changes. Stimulated recall is an introspective procedure whereby participants are prompted to “recall thoughts they had while performing a task” (Gass and Mackey 2000, p. 17). It should be noted that the stimulated recall sessions did not involve any group discussions and were done silently as the participants reflected on and recalled their state of mind while completing the tasks. Following this, participants completed the online WTC scale via Google Forms, where the participants responded to items that compared their online WTC during each given task with their WTC if they were to complete these tasks in traditional face-to-face sessions. This concluded the data collection in every session.

2.5 Data analysis

3.1 Correlating trait WTC, dynamic WTC, observed WTC, online WTC, and speaking times

The first research question aimed to examine if there were any correlations between the participants’ trait WTC, dynamic WTC, observed WTC, online WTC, and the speaking times in the online discussion tasks. To this end, multiple Pearson correlation tests were performed. As Table 2 shows, trait WTC during discussion tasks and online WTC were found to be moderately positively correlated, r(37) = 0.35, p = 0.031. This means that those reporting higher levels of trait WTC during discussion tasks also tended to display higher WTC in online discussion tasks. The participants’ online WTC also correlated closely with their actual dynamic WTC rating during the tasks, r(37) = 0.484, p = 0.002. This was a meaningful finding as the participants’ general perceptions of their online WTC experience were found to be consistent with their actual rating of WTC during the online discussion tasks in this study. However, no correlations were found between trait WTC versus dynamic WTC, observed WTC, or speaking times, a finding that was expected because the trait WTC was an independent measure that involved the participants’ general WTC prior to the discussion tasks, as opposed to dynamic WTC, observed WTC, and speaking times, which were measured in the context of the tasks.

A moderate negative correlation was found between online WTC and observed WTC, r(37) = −0.40, p = 0.01, and no correlation was found between dynamic WTC and observed WTC. These results indicate that the participants’ general perceptions of their online WTC measured through the same indicators used in the observation scheme did not manifest in their observed WTC behaviors.

Speaking time was positively correlated with dynamic WTC at a moderate level, r(37) = 0.33, p = 0.04, and with observed WTC at a strong level, r(37) = 0.51, p = 0.001. This means that those speaking more during the tasks both rated their WTC higher and exhibited more of the WTC indicators.

3.2 WTC variations

The second research question aimed to examine the extent to which the participants’ WTC fluctuates during synchronous discussion tasks. To this end, a between-subject repeated measures ANOVA was performed to examine whether the per-five-minute WTC changes were statistically significant. The results showed significant changes in WTC over five-minute intervals [F(6, 216) = 2.247, p = 0.04, partial-eta squared = 0.59]. Since we were interested in whether WTC changes were noticeable and not in which minute of rating WTC changed more significantly, no post hoc tests were conducted.

Table 3 shows the participants’ mean dynamic WTC and the standard deviation per 5 min per task. Figure 1 also illustrates the per-five-minute changes of group WTC per session. As can be seen, while the group displayed high initial WTC in all the tasks, there are noticeable variations, which range from a low of 0.83 in T5 to a high of 3.16 in T3. The SDs, or the extent of dispersion, also show that the group WTC levels varied considerably upon the start of the tasks, with a SD of 1.60 recorded in T3 and a SD of 3.10 in T6. What is commonly observed across most of the tasks (except in T3) is the gradually rising WTC in the first half of the discussion tasks, with T2 and T5 reaching their peak WTC in the first 5 min, while in T1, T4, and T6, group WTC continued to rise to minutes 10 and 15, respectively. After the initial rise, the group WTC underwent a declining trend at some point with T2 and T3 in minute five, T5 and T6 in minute 10, and T1, T4, and T6 in minute 15. Except in T4, group WTC went on an upward trend toward the end of the tasks.

Figure 1:

Group mean WTC per five minutes per task/session.

The second part of this research question aimed to investigate the factors that the participants attributed their WTC changes to during online discussion tasks. In two cycles of coding, a total of 414 codes were assigned, and nine categories of factors, to which the participants attributed their WTC changes, emerged.

The codes revolve around nine categories, including factors concerning: (1) topics or discussion questions; (2) the individual factors; (3) the possession of supporting ideas; (4) interlocutors; (5) the facilitators; (6) social or environmental considerations; (7) task dynamics; (8) perceived performance; (9) organizational factors or technical issues. Due to space limitations, a few representative samples will be presented.

3.3 Topics or discussion questions

This category involved the participants’ perceptions of the topics or questions. The participants expressed concern about not understanding the topics or the facilitators’ questions, which seemed to increase their anxiety as they tried to contribute relevant inputs. In support of this, Zaki mentioned: “a new question was posed and I was trying to fully understand it so as to avoid giving irrelevant answers”. Understanding the topic or the questions proved to be very important to all the other participants, too. For instance, Dorsa and Nafiseh lost WTC as they expected more clarifications from the facilitator, with Dorsa noting: “the question was ambiguous and difficult to answer”, or Nafiseh ranting: “the facilitator was silent and would not explain the topic”. There were instances, however, where the participants’ WTC gradually rose as they developed an understanding of the topics. Parvin, for instance, noted: “as the topic was starting to click, I felt more comfortable to engage in the discussion”.

Other factors involved the participants’ overall topical experience or preparedness, which improved the participants’ WTC. For instance, Parvin mentioned that she had “relevant experiences to the topic”, or Aman stated that he “had lots to say about the topic as it was relevant to his field of study”. WTC also fluctuated as the facilitators posed new questions or when the participants perceived the discussions as tedious or questions as repetitive. New questions sometimes triggered a shift of focus in the discussion, which required some processing and preparation time on the part of the participants. As the discussions were ongoing and the participants struggled to retrieve ideas to contribute, new questions or topic shifts sometimes led to a reduced WTC, or in one case, Zaki gradually lost interest as the discussion moved on to some other subject. On the other hand, on several occasions, new questions promoted WTC mainly in cases where the participants had already contributed and had nothing to say. Therefore, posing new questions provided new opportunities for task engagement.

3.4 Individual/personal factors

This category involved the participants’ personal interests, views, experiences, or old memories, or the participants’ emotional or physical states. The ‘interest’ refers to whether something in the discussions and ideas appealed to the participants temporarily, as opposed to the general interest in a topic discussed in the previous section. WTC tended to rise when the participants related the discussion to their personal interests, experiences, or memories, while it dropped due to a lack of interest. For instance, Aman found the discussions interesting on a number of occasions: “personal interpretations have always been interesting to me”, while Zaki reported a low WTC, noting: “the discussion concerned the food decorations which I had no interest in”. This category also involved factors that appeared to provide opportunities for the participants to relate the discussions to their personal life (e.g., experiences, views, etc.) thus facilitating the generation of ideas and engagement.

In some situations, the participants felt distracted, stressed, depressed, or physically fatigued, all of which lowered their WTC. As an illustration, Aman, Dorsa, Nafiseh, and Masrez all showed low WTC and stated that they “had a long hectic day and were tired” during the sessions which were all scheduled for 6:00 pm. Nafisah further added that she “was unable to focus on the discussion and felt distracted all the time”. These factors seemed to sidetrack the participants and minimize their engagement in the discussion tasks.

3.5 Content or supporting ideas

The third category pertained to the possession of supporting ideas, which evoked a feeling of security to engage in the tasks. WTC typically rose as the participants generated or retrieved ideas relevant to the ongoing discussion. For instance, Dorsa mentioned: “after listening to Parviz, I had many more new ideas to share and my WTC rose”, or Parvin noted: Parvin: “I felt I had funny ideas regarding the topic and this made me want to talk, or I felt like I could joke about the topic”. On the other hand, a lack of supporting ideas sometimes resulted in low WTC. For instance, Dorsa noted: “while I was listening to others, I had nothing to say and this lowered my WTC”, or Peyman said: “I did not have strong ideas to make my argument”.

3.6 Interlocutors

The participants’ WTC fluctuated depending on factors associated with the interlocutors – or other participants. These included the interlocutors’ opinions and participation, interactions or reactions, and their level of familiarity or proficiency as perceived by the participants.

Many of the participants indicated that sometimes they enjoyed listening to others expressing their opinions and therefore had little WTC to engage. Aman, for example, noted: “I admired one of the participant’s ideas and was encouraged to listen rather than talk”. Dorsa also mentioned: “I found the opinions of other participants more interesting and thus had little WTC to engage”. Sometimes, WTC fluctuations depended on the quality of interlocutors’ ideas or whether or not the participants agreed or disagreed with the interlocutors’ opinions. For example, Masrez lost WTC as she “found another participant’s response too long for a 30-min session”. In two instances, Masrez and Zaki lost WTC as they were surprised by other participants’ responses. On the other hand, there were two instances where disagreeing with an interlocutor’s opinion promoted WTC. There were also a few cases where the other participants’ participation improved the one’s WTC. For instance, Parvin said: “it was a heated debate and I wanted to engage and express my thoughts”.

The interactions with or reactions of the interlocutors also affected WTC. Peyman’s WTC dropped due to an interlocutor’s reaction to what he said. He explained: “if others agreed with me, I would generally have lower WTC to engage in discussion”, which was because he preferred to listen carefully so he could concentrate and generate more ideas to contribute. Masrez and Dorsa indicated that challenging co-participants created a motivation that improved their WTC.

Familiarity or unfamiliarity with interlocutors also impacted their WTC. Nafise felt “anxious about talking to new people”. On the other hand, Masrez reported higher WTC as she got acquainted with the participant over two sessions. In addition, Zaki and Peyman indicated that perceiving the interlocutors’ lower or equal proficiency level improved their WTC.

3.7 The facilitators

This category concerned the facilitators’ support, feedback, or reactions during the discussion tasks or the way they managed, engaged, or contributed to the discussions. Aman, for instance, mentioned: “what I noticed facilitators were really important and affected the way I reacted during the tasks”, or Masrez’s gained WTC was because of a facilitator’s “nice and respectful” manner. On the other hand, the participants’ WTC declined due to the facilitators’ lack of support, interest, or feedback. Zaki noted: “the session lacked the energy and engaging force, in fact, not receiving feedback for the facilitator lowered my willingness to participate”.

The facilitator’s management of the tasks influenced their WTC. This involved whether or not the facilitators prompted or encouraged the participants to engage, engaged in the discussion themselves, provided clarifications when the participants were confused, or created equal opportunities for everyone to engage. For instance, Masrez noted that the “facilitator was directing the task very well by picking good and timely questions”. It turned out that the facilitators’ engagement also promoted the participants’ WTC. Parvin stated: “when the facilitator started to talk, I felt better because I expected him to say something or give examples that would eventually help me generate ideas and talk”.

Many of the participants were in favor of having the freedom to talk whenever they felt ready while Dorsa preferred to be given turns to talk. One of the facilitators chose to direct questions to the participants despite being instructed to not do so, and Dorsa noted: “I was used to this because it gives me the opportunity to talk and then listen to the others. This allowed me to focus on the quality of my responses rather than on when and how to get a chance to talk”. However, Zaki showed no interest in the way the facilitator managed the task.

3.8 Social or environmental considerations

This category mainly involved how the participants reserved the right for the other participants to engage and contribute to the discussion tasks. In some cases, the participants’ WTC dropped as they had already participated and thus avoided talking to allow the other interlocutors opportunities to talk. In other cases, some participants had low WTC because of having to wait for their turn. Dorsa, for example, lost WTC due to “interrupting Peyman and feeling embarrassed for this”. In the second session where the facilitator prompted the participants one by one, Parvin and Zaki exhibited low WTC because they felt they had to wait for the facilitator to directly engage them rather than volunteer to talk.

Some of the other instances involved the participants getting acquainted with the group and thus feeling more comfortable. Nafise, however, indicated that despite being forced to talk in one instance, she lost WTC because of the dominating silence and not feeling comfortable. Nafise also felt anxious and uncomfortable in session one because of “meeting new people and being in a new environment”.

3.9 Task dynamics

The dynamics of the discussion tasks also influenced the participants’ WTC. This involved the group dynamics and the participation opportunities that the participants perceived to arise from such dynamics. Some of the participants lost WTC upon finding the discussion irrelevant to the main topic, or if the session was coming to an end. For instance, Nafise mentioned: “I lost track of the discussion because of the irrelevant discussions, only one person talks, and this should have been managed by the instructor”. Presumably, Nafise had prepared ideas to share, but the discussion had taken a different direction and her ideas were no longer relevant. Silence was also found to improve WTC. Some chose to talk as silence dominated the sessions, with some perceiving it as an opportunity to engage while others, like Nafise, perceived it as an awkward moment that she had to put an end to.

Almost all the participants reported moments where they struggled for opportunities to talk, resulting in reduced WTC in all the cases. In Dorsa’s case, this also led to some anxiety: “I was slightly worried and losing WTC because others jumped in and wouldn’t let me talk”.

3.10 Perceived performance

This category mainly involved the participants’ perceived communication incompetence or lexical issues. The perceived communication incompetence pertained to situations where the participants found themselves unable to communicate their points, which lowered their WTC in all the cases. For instance, Aman noted: “I do not see the ability to make my point. It is terrible that the English language is so difficult for me after studying it for such a long time”.

3.11 Organizational factors or technical issues

This category involved moments of low WTC due to arriving late, experiencing technical issues typical of online sessions, or getting distracted due to background noise originating from other participants’ microphones.

The participants lost WTC due to being late, and this was mostly because they had missed the facilitator’s introduction to the topic and/or the showing of the images which would help with generating ideas. The low WTC was triggered by the anxiety of not being aware of what was going on and having to wait and learn more about the ongoing discussion. In support of this, Zaki mentioned: “due to arriving late, I had a high level of anxiety and needed some time to catch up. Due to stress, I felt I was unable to comprehend what the others were saying, and it was difficult for me to concentrate”.

Technical issues also had an impact on the participants’ WTC. For instance, Dorsa mentioned that “poor internet connections increased her anxiety giving her the impression that she might not speak well”, which lowered her WTC. Masrez also experienced multiple connection issues, which caused her to lose track of the discussion and lose WTC as a result. Nafise also mentioned that she had lost the speaker’s voice on multiple occasions which had sidetracked her and lowered her WTC as she had no idea what was being discussed.

In two more cases, Aman and Nafise mentioned that the background noise originating from another participant had disrupted them. While Aman did not lose much WTC and only mentioned it was annoying, Nafise lost WTC, noting: “I wish there weren’t distractions behind people like family members of the participants. I lost track of the discussion because of this”.