Pre-service EFL Teachers’ Conceptions of Learning about SLA through online Discussion Forums and WhatsApp in Blended Learning

Introduction

Computer-mediated communication (CMC) has been widely used as a tool for learning and teaching in higher education. According to Wikle and West (2019: 206), the most widely used CMC tool for student-student and student-instructor communication ‘is asynchronous online discussion forums’, hereafter ODFs. As established by these authors, ODFs are often implemented in courses that include both online and face-to-face components, or blended-learning courses. A key assumption underpinning the use of ODFs in education is that dialogue among peers facilitates collaborative knowledge building, which results in deeper, more meaningful learning. Other benefits of ODFs may include encouraging language learners’ participation and self-expression (Cantor Barragán, 2009) and encouraging teachers to reflect on their syllabi and to design new tasks (González Moreno, 2011). However, studies have also shown some of the negative effects of ODFs, such as leading to misconceptions (Yap and Chia, 2010), not engaging in collaboration (Delahunty, 2018), and perceiving ODFs as purposeless and boring (González Moreno, 2011).

In a meta-analysis, Lucas, Gunawardena and Moreira (2014) found that most studies using the Interaction Analysis Model or IAM (Gunawardena et al., 1994) to look at ODFs and learning showed limited collaborative knowledge building. Lucas et al. (2014) conclude that more research is needed on students’ perspectives on what constitutes collaborative knowledge building in ODFs. In particular, studies of Asian teachers have found that these teachers are reluctant to evaluate and challenge others’ claims (Sing and Khine, 2006; Hou et al., 2009, 2011).

In language teacher education, a number of studies have addressed teachers’ perceptions of the use of ODFs in their training (Arnold et al., 2006; Arnold and Ducate, 2006; Biesenbach-Lucas, 2003; Delahunty, 2018; Delahunty et al., 2014; Lord and Lomicka, 2004; Lord and Lomicka, 2008; Sardegna and Dugartsyrenova, 2014). The findings of these studies, like those of studies focusing on students, are mixed. For example, some studies have found that ODFs are useful in creating a sense of belonging to a community of practice (Lord and Lomicka, 2004; Lord and Lomicka, 2008), helping develop pre-service teachers’ autonomy, and providing good models for future technology use (Sardegna and Dugartsyrenova, 2014). Yet some pre-service teachers prefer face-to-face activities to ODFs (Sardegna and Dugartsyrenova, 2014), or perceive ODF interaction as forced, unnatural, and burdensome (Biesenbach-Lucas, 2003).

Studies focusing on both perceptions and online discourse have found that building positive interpersonal relationships and the instructor’s presence are key in favouring pre-service ESL teachers’ collaborative learning (Delahunty et al., 2014). Students’ active attention to the posts of others is also a key factor; its absence brings about negative consequences for motivation and learning (Delahunty, 2018). Studies analysing in-service FL teachers’ messages in ODFs have generally found positive attitudes (Son, 2006) and some knowledge building (Nami et al., 2018), although they have also highlighted the need for more collaboration among teachers.

Against this backdrop of mixed results, we agree with Shin and Kang (2018) that more research is needed on the use of online tools for language teacher education. Focussing on pre-service teachers’ experiences with such tools, and in particular ODFs, is important because the quality of such experiences is likely to influence their future decisions on whether and how to use ODFs in their own teaching (Dooly, 2009; Guichon, 2012). Accordingly, understanding pre-service teachers’ experiences of learning aspects of language teaching with the help of ODFs can have meaningful implications for pedagogical designs that address pre-service teachers’ learning needs. This is a worthy goal in light of recent calls to interrogate ways to maximize the impact of language teacher education on language teachers’ thinking and practice (Kubanyiova and Feriok, 2015). It also aligns with calls to increase language educators’ awareness of ‘the constraints and possibilities for making meaning and communicating in online venues’ (Guichon and Hauck, 2011: 193).

This article addresses these needs by presenting the results of a phenomenographic case study carried out with a group of Mexican undergraduate pre-service EFL teachers enrolled in a blended-learning SLA course. The study aims at describing participants’ conceptions of learning SLA concepts through ODFs in a blended-learning SLA class using Moodle. The focus is not on how well they learned concepts. Instead, we focus on their conceptions of ODFs as learning tools.

Research Questions

The following questions were addressed:

Our method was phenomenography, which is gaining popularity in language studies (Chang, 2016; Nohriza Mohd, 2010; Polat, 2013; Norberg et al., 2018). Although it has been used to investigate e-learning in other disciplines (Ellis et al., 2006; Tsai, 2009), phenomenography has not been used widely in investigating ODF-mediated language teacher education.

Methods

Phenomenography is both a theoretical perspective on learning and a methodological approach. According to Harris (2011: 110), ‘within the phenomenographic research approach, conceptions are the central unit of description about people’s experiences’. Conceptions are defined as ‘different ways of understanding’ (Marton and Pong, 2005: 335). Conceptions are the minimal units of experience, understood as the self-aware internal perception and interpretation of external phenomena (Feldon and Tofel-Grehl, 2018). According to Marton and Booth (1997), for any given phenomenon there will be less and more complex ways of experiencing it. The more complex ways include more aspects of the phenomenon (i.e. more conceptions) that bring it into sharper focus.

The goal of phenomenographic studies is to produce a hierarchical grouping and arrangement of conceptions. Such groupings are called categories (Tsai, 2009). Their hierarchical arrangement is called an outcome space (Marton and Booth, 1997). Besides the categories, the outcome space may include the themes that are perceived across the categories (Marton and Booth, 1997); these may be called dimensions (González, 2009), or critical attributes of variation (Heinone et al., 2017). The intended use of the categories in the outcome space is ‘to guide teachers to approach learning from their students’ perspectives, or at least to understand why their students might be thinking or doing certain things’ (Polat, 2013: 114). The categories in the outcome space should be plausible, distinct, parsimonious, and hierarchically related (Marton and Booth, 1997). There are different ways of presenting outcome space. Some include quantifications of the number of participants holding a conception and of variant and invariant conceptions. We follow the purely qualitative approach for presenting the outcome space in Gonzalez (2009; Gonzalez, 2012) and Norbert et al. (2018) because it is a good fit for our data and design.

According to Marton and Booth (1997) learning experiences have what they call referential and structural aspects. Phenomenographers vary in the ways they understand and apply these terms, and it is therefore recommended that scholars define them explicitly (Harris, 2011). For us, referential aspects relate to how the students understand the object of learning (i.e. their learning of concepts), while the structural aspects relate to how they approach learning (how they understand the act of learning and the strategies they use to learn) and why they do it that way (González-Ugalde, 2014). In accordance with our goal to focus on conceptions of ODFs as a way to learn, rather than on the concepts being learned, this article thus centres primarily on structural aspects. Referential aspects are addressed only inasmuch as they are essential to understand structural ones.

The Context, Participants and Data Collection

The study was conducted in the context of a five-month SLA course taught in the fourth term of the Bachelor of Arts degree in English course at a small, urban, public university in Mexico. This course was selected for theoretical and practical reasons. Among the theoretical reasons are our interest in the connections between SLA theory and teaching, which is a longstanding topic in language teacher education. Another theoretical reason is that it presents an interesting case because the professor who teaches it, unlike many other teachers who teach SLA courses in Mexico, has a great deal of expertise and international recognition in the field of applied linguistics. He earned both his Master’s degree and his doctorate at American universities under the supervision of renowned American and British scholars. The practical reason is that this professor is a long-time friend of the lead author, who facilitated the team’s access to the course. The professor was approached in the Fall/Autumn of 2017 to explain the specifics of the study and to ask for his permission to conduct it in the following year. Upon his request, the study was registered with and approved by the university’s Research Office, with the lead author as the principal investigator.

Twenty-seven pre-service teachers were enrolled in the SLA course in the Fall/Autumn of 2018, when data collection took place. At the beginning of the course, all the participants were invited to participate by the lead author. The lead author came on the first day of class, presented the study to them, asked if they agreed to participate and asked them to fill out and sign informed consent letters. All of them did so. The sample for this study consists of 16 of them as this was the number at which we achieved saturation of categories of description. Note that small samples are common in phenomenographic research (González, 2009; González-Ugalde, 2014), and our sample complies with the minimum of 16 participants recommended by Trigwell (2000) to capture variation. The participants were 11 females and 5 males, aged 19–20 years old. They were all students of the second year of the university’s BA in English, which focusses on ELT. They did not have any prior experience with blended learning, ODFs, or the online platform Moodle. We aimed for saturation as a sampling strategy rather than a census sample because of time constraints inherent to the project.

Three semi-structured interviews were conducted with the participants, one at the beginning of the instructional period in focus, another one in the middle, and the last one at the end. The shortest interview was 7 minutes and 5 seconds long; the longest interview was 19 minutes and 3 seconds long. The interviews took place in the lead author’s office and were recorded on a Motorola G6-plus mobile phone. They were transcribed manually by an assistant. Interviews are the most common technique of phenomenographic data collection (Walsh, 2005). The interview protocols were designed following the guidelines in Booth (1997), Marton and Booth (1997), and Åkerlind (2005) and included ‘how’ and ‘why’ questions. They were validated using the Interview Protocol Refinement Framework (IPRF, Castillo-Montoya, 2016). All the interviews were conducted in Spanish, the participants’ native language. Interview excerpts in this article are translations.

The SLA course was taught using a blended-learning approach: the class met once a week for two hours and then students completed activities using the learning platform Moodle. This study focusses on only one unit, that on interlanguage development, which lasted five weeks. The online activities for this unit were five ODFs. Table 1 presents relevant information about them. This information was gained from an interview with the professor conducted by the lead author and from screen captures of the Moodle platform. All the ODFs had the following features in common: it was mandatory to contribute at least twice, once with a new contribution and at least once with a response to someone else’s posts. Messages were not graded for quality; students were encouraged explicitly to ask questions and express and negotiate disagreement; the outcome of each ODF except four was a paragraph that summarized the team members’ conclusions. There was no word limit on individual posts or the final paragraph; the professor intervened when students’ showed misunderstandings or misconceptions. The instructions for each ODF stated its purpose, listed the relevant concepts, established the desired outcome, and outlined the participation requirements. No explicit etiquette guidelines were provided beyond statements like ‘it is okay to ask questions and disagree’. The participation structure was in teams of four students. Participation was kept intra-team except for ODF 3, which had both an intra-team and an inter-team component.

OPEN IN VIEWER

Table 1.

The Five ODFs.

Week	Title	Purpose	Relevant concepts	Outcome
1	Feelings about correctness	To share and reflect on feelings about ‘correct’ and ‘incorrect’ L1 features, prescriptivism and descriptivism	Prescriptivism and descriptivism	Posts with team members’ feelings and reflections and a paragraph summarizing collective reflections
2	First concept connection activity	To connect a set of concepts	Interlanguage, error, mistake, non target-like forms, input, output, L2, second language, foreign language	Posts negotiating connections among concepts, concluding paragraph expressing agreed-upon ideas
3	Second concept connection activity	To connect two sets of concepts, to evaluate another team’s work	Set 1: Interlanguage development, prescriptivism, descriptivism, target-like and non target-like forms, standard language Set 2: Developing system, input, output, intake, noticing, accommodation, restructuring, mind, second language, foreign language	Posts negotiating connections among concepts, concluding paragraph expressing agreed-upon ideas about set 1, concluding paragraph expressing agreed upon ideas about set 2, concluding paragraph expressing agreed upon ideas about the connection between both sets of concepts. Posts discussing another team’s paragraphs, paragraph showing the agreed-upon overall evaluation of those paragraphs*
4	Assessment task	To grade a test by a hypothetical learner with multiple non target-like forms taken into account what had been learned about interlanguage development and assigning points accordingly on this scale: 0-.25-.5-.75-1	All of the above plus self-referential assessment (Larsen-Freeman, 2014)	Posts discussing each answer’s score and the rationale for it, a table with the agreed-upon score assigned to each answer
5	Reflection on assessment task	To reflect on the rationale for grading the way they did	All of the above plus self-referential assessment (Larsen-Freeman, 2014)	Posts with team members’ reflections, concluding paragraph expressing agreed-upon ideas

Data Analysis

We followed a discovery-based, team approach. In a discovery approach, categories emerge from the data; prior theories on the phenomenon are not used and the researchers keep their own conceptions at bay, at least initially (Walsh, 2000). Phenomenography can be conducted by a scholar working alone (Åkerlind, 2005), or by teams of scholars (Bowden, 2005). Our process was iterative and involved moments of both individual and collective work and constant checking of emerging categories against the data, which are all typical features of phenomenographic data analysis (Walsh, 2000; Bowden, 2005). The steps we took are explained below.

Results

Learning through ODFs was conceived as providing different kinds of learning opportunities. These depended on how participants conceived their roles in the tasks, the role of the face-to-face classes, and their relationship with the teacher in the context of the ODFs. These different ways of experiencing learning through the ODFs are captured by the three categories below:

As mentioned above, the categories emerged from learner’s conceptions about various task and contextual features. These were arranged in the following dimensions:

Discussion and Conclusions

Research question one focussed on the participants’ conceptions of learning through ODFs. We found three categories (groupings of conceptions): Going over the content, Expanding knowledge metacognitively, and Negotiating knowledge. The second question enquired about the strategies for learning. These include 1) re-reading content, 2) reading others’ posts and the teachers’ posts attentively, 3) searching for additional information, and 4) using WhatsApp as a parallel ODF site to negotiate logistics and potential interpersonal friction. The third question focussed on reasons for using strategies. The reasons for using the first strategy are mostly connected to memorization, although the content was also re-read in order to adjudicate claims. The reasons for using the second and third strategy had to do with curiosity and a desire to check understanding. The reasons for using the fourth strategy had to do with logistics and, more interestingly, with the need to follow politeness norms.

The parallel discussions on WhatsApp were conceived as a way to reduce potential interpersonal friction that could otherwise have emerged in the Moodle ODFs. This implies that the Moodle ODFs were thought of as potentially face-threatening, and participants were therefore reluctant to make face-threatening acts in the Moodle ODFs. Interestingly, the fact that participants were willing to express critiques and negotiate disagreements on WhatsApp suggests that the face-threatening potential lay not in the acts themselves but in the teachers’ witnessing of those acts. This reluctance to engage in unmitigated, non-choreographed debate mirrors López-Islas’ (2001) finding that Mexican college students in several social science disciplines avoid expressing and negotiating conflict in ODFs. The need to first negotiate potentially conflicting meanings and then stage the negotiation may stem from Mexican politeness conventions. Earlier studies with Mexican EFL learners have found that they are reluctant to disagree with or criticize their peers because doing so would be face-threatening and go against cultural politeness norms (Mugford, 2011; Sandoval-Cruz et al., 2019). Similarly, limited ODF engagement in evaluation and negotiation of claims by Asian teachers and learners have been attributed to cultural conventions (Biesenbach-Lucas, 2003; Lucas et al., 2014).

In parallel, studies with Master of Arts TESOL students have found that cultivating good interpersonal relationships and feeling safe when expressing critiques and disagreement assisted collaborative knowledge building in ODFs (Delahunty, 2018; Delahunty et al., 2014). The role that discursive cultural norms play in learning through ODFs can therefore not be ignored, and should be addressed more pointedly by future studies. Teacher trainers should also be aware of the need to teach pre-service teachers discursive strategies for evaluating and negotiating claims. This can be done using interventions of the kind developed by Meskill and colleagues (Meskill and Anthony, 2007; Meskill, 2009; Meskill and Sadykova, 2011). Examples include microanalysing and reflecting on simulated conversations (Meskill and Anthony, 2007; Meskill, 2009), and the Moodle fishbowl technique (Meskill and Sadykova, 2011), with a specific emphasis on experts’ modelling and pre-service teachers’ observing and analysing the expression and negotiation of disagreement. These tasks should address politeness norms and their role in facilitating or impeding learning through ODFs.

Future studies might also explore conceptions of WhatsApp and other social networks as parallel spaces of meaning negotiation and knowledge construction. Nevertheless, such research would not come without risks. It is possible that students think of and use such spaces to escape teacher surveillance, which in principle might afford them opportunities for more meaningful, authentic exchanges. Therefore, any attempt to deepen our understanding of social network use as parallel sites of knowledge building must be undertaken with great care, lest we kill precious spaces where the kinds of learning we want to promote might already be happening on their own.

The categories we found are somewhat similar to those reported for engineering students by Ellis et al, (2006). The less advanced e-learning categories in that study were Checking ideas and Acquiring ideas, and the most advanced ones were Developing ideas and Challenging ideas. Our category of mere repetition, or Going over the content, was absent in their data. Their categories of Checking, Acquiring, and Developing ideas seem to be encompassed by our second category, Expanding knowledge metacognitively. Our and their most advanced category seem to be equivalent, as they both involve a focus on debating ideas.

Crucially, our study has shown that the more advanced category, Negotiating meaning, includes a different way of attending to peers, one that involves not just reading but also responding to others’ messages, and writing with the goal of eliciting a response. This finding echoes that of Delahunty (2018) on the importance that attending to others’ messages, and having one’s messages attended to, holds for collaborative knowledge building to happen in the context of ODFs. Nevertheless, even in the more advanced category, the material and the professor were not challenged but remained the ultimate resource for adjudicating claims. Although this is perhaps appropriate at the undergraduate level, it might be desirable to find ways to promote critical questioning of the material and the professor.

The least advanced category in our outcome space, Going over the content, can be said to represent a shallow, fragmented approach to learning. In this category, learning through the ODFs is conceived of as offering opportunities for memorization through repetition. This conception is not very conducive to autonomous learning or deep learning. As a result, interventions need to be designed in order to scaffold changes from this conception to more advance ones. Besides the techniques outlined above (e.g. the Moodle fishbowl), such interventions should include work on metacognition, self-regulation, beliefs about learning, and beliefs about collaborative knowledge building.

A limitation of this study is that we did not analyse the Moodle or WhatsApp posts. Such analysis might reveal possible instances of knowledge expansion and meaning negotiation in participants who are not aware of their engagement in such practices. Further, the fact that all our participants were unfamiliar with Moodle ODFs implies that a different sample of more experienced participants might yield a different outcome space. A strategic decision for this study was to exclude the referential aspect of learning, which in this case had to do with key SLA concepts such as interlanguage, and attitudinal shifts from prescriptivism to descriptivism as defined by the professor. Future studies should therefore examine ODFs in connection with interview transcripts to characterize referential conceptions and probe possible connections between those and the structural aspects reported here. Future studies should also examine pre-service teachers’ conceptions of the face-to-face classes. In this study, the face-to-face setting was conceived of predominantly as a source of truth to adjudicate claims, but more focussed interview questions with a different sample might cast further light on its role vis-a-vis the ODFs.