Assessing Episodes in Verbalization Process of EFL Students’ Collaborative Writing

One of the ways to engage EFL students in writing is to assign them to work collaboratively. Collaborative writing requires a verbalization process resulting in episodes related to language, texts, and scaffolds. This study examined the use of episodes in collaborative writing of EFL students set in pairs by the teacher. It identified the most productive type of episodes which include language-related episodes (LREs), text-related episodes (TREs), and scaffolding episodes (SEs). It also scrutinized the categories of episodes within each type of episode. The study involved 20 pairs of Indonesian students from the English Department of a reputable university in Yogyakarta, Indonesia. The students were given an integrated reading-writing task and asked to work in pairs. The results of the study showed that SEs were the most productive type of episode, followed by LREs in the moderate occurrence, and TREs, which were the least productive type of episode. The results also revealed that among the categories in each type of episode, lexis-focused (LREs), organization-focused (TREs), and repetition (SEs) were more productive than the other categories of episodes. These results imply that the most productive categories of episodes could be catalysts in the teaching of writing, which employs collaborative writing tasks either in pairs or in small groups. This study offers insights into creating activities to encourage writing activities that especially involve types of pairings.


INTRODUCTION
Various types of tasks can be assigned to students when teaching English as a foreign language (EFL) writing. The students can be asked to work either individually or collaboratively in the process of completing their writing activities. Unlike individual writing, which requires the students to be more active in their thinking process, collaborative writing expects students to be cognitively active and socially engaged (Deveci, 2018). Social engagement in collaboration is apparent in the verbalization process that occurs during interaction and meaning negotiation (Zhang, 2018). The verbalization process in collaborative writing is an operationalization of language learning (Li & Kim, 2016).
In collaborative writing, the verbalization process comprises several types of episodes. One of the types is language-related episodes (LREs). LREs are defined as occurrences in collaborative dialogue in which students negotiate explicitly about the language they are producing, question their language, or correct themselves or others (Leeser, 2004;Swain & Lapkin, 1998). Studies about LREs have introduced several categories of LREs that could be grouped into three major categories based on the linguistic focus that the students negotiate. They are forms of grammar, lexicon or lexis, and mechanics, and three categories are based on the quality of the problemsolving outcome, i.e., correctly solved, incorrectly solved, and unsolved (Kim & McDonough, 2008;Leeser, 2004;Niu et al., 2018). Mozaffari (2017) found that the teacher-assigned pairs produced significantly more LREs than the students' selfselected pairs, and these pairs tend to produce more off-task episodes. Several research studies have explored the categories, quantity, and quality of LREs (Kim & McDonough, 2008;Leeser, 2004;Niu et al., 2018;Storch & Aldosari, 2013;Watanabe & Swain, 2007).
Other types of episodes in the verbalization process during collaborative writing include text-related episodes (TREs) and scaffolding episodes (SEs). TREs are defined as any episodes in which learners talk about text development, including the discussion about the content and organization of the text (Neumann & McDonough, 2015;Watanabe, 2014). On the other hand, SEs are defined as the language-mediated process provided by a peer to achieve a higher-level performance that was previously beyond the learner's existing level of ability (Hanjani & Li, 2014;Hassaskhah & Mozaffari, 2015;Mozaffari, 2017;Watanabe, 2014). The present study addresses the three types of episodes in the verbalization process during collaborative writing.
This study has been inspired by the quest for unfolding the best ways in managing pairings of learners within the variety of types of episodes in the verbalization process of collaborative writing for the optimal learning outcome. Previous studies (e.g., Adodo & Agbayewa, 2011;Fauziah & Latief, 2016;Maftoon & Ghafoori, 2009;Zamani, 2016) on pair collaborative writing have concentrated on two separated areas, experimental studies on the effect of proficiency pairings on the EFL students' writing ability and descriptive studies on the collaborative process that occurs as observed through the verbalization process. In relation to the first area, the findings of existing studies (e.g., Adodo & Agbayewa, 2011;Fauziah & Latief, 2016;Maftoon & Ghafoori, 2009;Zamani, 2016) have provided mixed conclusions and, therefore, created space for further investigation, notably using different research designs that could integrate both writing outcome (after intervention) and the pair writing process. The second area also needs more exploration since the previous studies (e.g., Kim & McDonough, 2008;Leeser, 2004;Niu et al., 2018;Storch & Aldosari, 2013;Watanabe & Swain, 2007) mostly compared the collaborative writing process between pair and individual writing in strictly controlled and free writing tasks. Other comparisons that could describe how proficiency pairings (heterogeneous pairs and homogenous pairs) are negotiated in integrated writing tasks (such as in reading-writing tasks) are not yet available. Therefore, the present study attempts to fill in these identified gaps.
Studies on TREs and SEs are not as comprehensive as those on LREs. Most studies compared the TREs and SEs between individual writing and pair collaborative writing (Hanjani & Li, 2014;Neumann & McDonough, 2015;Watanabe, 2014). Therefore, the present study focused on the area which has not been adequately studied, that is, the variety of types of episodes in the verbalization process of collaborative writing. Accordingly, this study is intended to answer the following research questions: 1. Which type of episode is the most productive in the verbalization process of collaborative writing? 2. Which episode category is the most productive from each type of episode in the verbalization process of collaborative writing?

EFL Writing and Pedagogical Approaches
It has been widely accepted that writing is a complex language skill, and therefore it is expected that EFL students find it difficult, also partly because the language is not their mother tongue. Writing is complex because, first, it needs to undergo some stages (Karim et al., 2017). There are at least three stages in writing, namely, planning, execution, and evaluation (Rosário et al., 2019) and the extended version covers eight stages, consisting of prewriting, planning, drafting, pausing, reading, revising, editing, and publishing (Williams, 2003). Second, the written product should be presented in a particular structure, which includes an introduction, body, and conclusion (Oshima & Hogue, 2007), or it has to follow the generic structure of certain genres (Dirgeyasa, 2016). Finally, to produce a good piece of writing, a writer has to be aware of the components that affect readers' judgment of the writer's ability in writing, namely content, organization, vocabulary, language use, and mechanical aspects (Weigle, 2002). Due to the complexity of writing, attempts have been made to help language learners in general or EFL learners, in particular, to learn how to write. A. Susanti,U. Widiati,B. Y. Cahyono & T. I. S. T. Sharif,Assessing episodes in verbalization process of EFL students' collaborative writing | 542 The ability of EFL students to produce a piece of writing has always been a challenge. There are three main pedagogical approaches suggested in the teaching of writing, i.e., text-oriented approach, writer-oriented approach, and reader-based approach (Hyland, 2016). A text-oriented approach, parallel to a product-oriented approach (Nunan, 1999, p. 272) views "texts-as-autonomous objects" and highlights the writing product; therefore, this approach concerns more the error-free sentences and language forms, i.e., grammar, syntax, and mechanics (Hyland, 2016). Matched to this view is a quantitative measure utilizing T-unit, error-free, and word count analysis for each writing aspect, i.e., complexity, accuracy, and fluency (Issacson, 1988). In addition, a text-oriented approach also considers text-as-discourse, which corresponds to "a discourse-based approach" (Nunan, 1999, p. 287), as well as a genrebased approach (Hyland, 2003) that views writing ability as the ability to create coherent and cohesive discourses following suggested structures.
In contrast to the text-oriented approach, which concentrates on the product, the writer-oriented approach views writing as a personal expression, cognitive process, and a situated act. Therefore, writing ability in a writer-oriented approach is linear to a process-based approach (Nunan, 1999) and is defined as the ability to plan and develop ideas and then use specific revision and editing practices to finalize the draft in a provided context (Hyland, 2016;Yi, 2009). The reader-oriented approach views writing as social interaction, social construction, power, and ideology. A readeroriented approach emphasizes the reader's awareness; therefore, a reader-oriented approach considers a successful writer as one who can assume the readers' perception and expectations and balance those assumptions into the relevant structure and content of the discourse, so that the writing activity can serve its communicative purpose (Hyland, 2003(Hyland, , 2016. In the field of second language writing, the integration of language comprehension and production, particularly reading-writing, is receiving growing attention because numerous real-life writings are composed in response to a text (or texts) demanding a high degree of reading skills to integrate the input materials into the written response (Plakans & Gebril, 2012;Weigle, 2004). Some studies have compared the composition process and the writing quality between writing-only tasks (independent writing) and reading-to-write tasks (integrated writing) (Plakans & Gebril, 2012). According to this integrated approach, writing ability is defined as the ability to gather information, develop thoughts, and then write to produce an organized response that incorporates selected information from the available sources.

Collaborative Writing and Verbalization Process
Collaborative writing is understood as a joint production or co-authoring of a text by two or more writers (Storch, 2018). It emphasizes joint ownership because the writers are engaged in the whole writing process or partial writing activities such as group planning or peer editing (Storch, 2018). The effectiveness of collaborative writing, compared to individual writing to promote writing ability, has been confirmed by several researchers (Khatib & Meihami, 2015;Shehadeh, 2011). Factors contributing to the quality of collaboration and the outcome of collaborative writing have been reported, more particularly regarding language proficiency and patterns of interactions (Watanabe & Swain, 2007), the task variation (Kim & McDonough, 2011;McDonough & Fuentes, 2015), and member personality, collaboration experience and cultural values (Rezeki, 2016).
In pair collaborative writing, in particular, the quality depends on the equality and mutuality (Storch, 2005) of the interaction process between peers, which is known as the verbalization process (Watanabe, 2014), or previously known as 'languaging', as well as collaborative dialogue (Swain & Lapkin, 1998). The verbalization process is defined as the dialogue in which speakers are engaged in problem-solving and knowledge building (Swain & Lapkin, 2002). In the case of L2 learners' interactions, the verbalization process is the dialogue in which learners work together to solve linguistic problems and/or co-construct language or knowledge about language . The verbalization process mediates L2 learning since peers provide L2 learners with opportunities to engage in collaborative dialogue as they seek out and assist with language-related problems . In a controlled writing task, pairs tend to focus on grammatical issues, while in a free writing task, pairs tend to focus on lexis and discourse matters (Kim & McDonough, 2011;Leeser, 2004;Niu et al., 2018;Storch & Aldosari, 2013;Watanabe & Swain, 2007).

Episodes in the Verbalization Process
Several studies have used language-related episodes (LREs) instances of the verbalization process as a unit of analysis to examine the peer-peer dialogue as learners jointly engage in problem-solving tasks (e.g., Swain & Lapkin, 1998;Swain & Watanabe, 2012;Watanabe & Swain, 2007). An LRE is defined as any part of a dialogue where students talk about the language they are producing, question their language use, correct themselves or others or reflect on their language use (Swain & Lapkin, 1998). LREs have been recognized as a valuable construct for understanding the process and product of L2 learning and have been used in peer-interaction research studies to identify the degree to which L2 learners address language-related problems, allowing for the systematic analysis of these episodes.
Extending peer interaction episode studies, several researchers have initiated the exploration of non-LREs by analyzing the frequency and/or categories of other types of episodes appearing in the verbalization process in the collaborative writing task, namely TREs (Watanabe, 2014) or content and organization episodes (Neumann & McDonough, 2015), scaffolding/non-scaffolding episodes (Hanjani & Li, 2014;Watanabe, 2014), and on/off-task episodes (Hanjani & Li, 2014;Mozaffari, 2017).
The investigation of the verbalization process, especially in reading-for-writing tasks, needs to be conducted to provide evidence that would help predict the LREs production of pairs when they work in reading-for-writing tasks. This task is more flexible than a controlled task (e.g., dictogloss) but less free than a free task (e.g., composition task). In the EFL field, there has been a growing interest in this kind of task. It represents real-world writing in that people usually read several reading texts first and then integrate the input materials into the writings (Chan et al., 2015). These led to the present study on the occurrences of a variety of types of episodes in the verbalization process of collaborative writing. A. Susanti, U. Widiati, B. Y. Cahyono & T. I. S. T. Sharif, Assessing episodes in verbalization process of EFL students' collaborative writing | 544

METHOD
The present study explored EFL students' verbalization process reflected in three types of episodes: language-related episodes (LREs), text-related episodes (TREs), and scaffolding episodes (SEs). It also examined the categories of episodes within each of the types of episodes. Twenty teacher-paired Indonesian EFL students of the English Department of a reputable private university in Yogyakarta, Indonesia, participated in this study. When the study was conducted, the students attended the Reading-Writing course offered in the department. The students were assigned to work collaboratively and required to write a short essay on the topic of 'How to Be a Good Parent'. Before starting to write, the students had a brainstorming activity about being parents. They also read a text about 'Strict and Relaxed Parents'. After brainstorming and reading the text, they worked in pairs to write the instructed essay. During the collaborative writing process, their verbal discussion was audio-recorded. The audio recordings were transcribed and examined.
The transcripts of the verbalization process during collaborative writing were categorized according to the types of episodes (LREs, TREs, and SEs). The LREs were related to any discussion about grammar and mechanics, and TREs dealt with any discussion about the content, the writing purpose, and the organization of the writing pieces. Finally, SEs covered any discussions within the collaborative writing, including those about language and text and other aspects to support each pair member to complete the given task.
Since language is multifaceted, it is possible that an episode does not exclusively belong to one category. One piece of data (episode) can also belong to more than one category. As the data of verbalization process source were collected from the spoken discourse, the indicator of each episode utilizes para-tones such as the boundary of topic markers, which include shallow pitch, even on lexical items, loss of amplitude, and a lengthy pause (Brown & Yule, 1983;McCarthy, 1991).
In this study, the investigation of the LREs, TREs, and SEs was firstly focused on the frequency of occurrences. The result of frequency analysis was in the form of numbers indicating the number of LREs, TREs, and SEs appearing in the interactions (Leeser, 2004;Swain & Lapkin, 1998). A descriptive quantitative approach was undertaken to offer results in the form of frequency evaluation. Then, the types and outcomes of the LREs were studied. The types were categorized into three (forms, lexis, and mechanics), and the outcomes were also categorized into three (correctly solved, incorrectly solved, and unsolved). Furthermore, the types of TREs were analyzed by categorizing them into two (content and organization). Finally, the SEs were investigated into several categories (repetition, elicitation, and justification) (Kim & McDonough, 2008;Leeser, 2004;Niu et al., 2018).

RESULTS
The results of the study were based on the analysis of the audio recording transcripts of the types of episodes and evaluation of students' collaborative writing products.

The Most Productive Type of Episode
To answer the first research question about the most productive type of episode, all elicited episodes were counted and classified into their types. The results of the analysis of the three types of episodes (LREs, TREs, and SEs) in terms of occurrences and percentages are shown in Table 1.  Table 1 shows that SEs achieved more than half of the number of occurrences (51.9%), making this episode the most productive type of episode. LREs are in the middle position with almost one-third of occurrences (32.3%), which is followed by TREs with only about one-sixth of occurrences (15.8%), indicating that this is the least productive type of episode.

The Most Productive Category from Each Type of Episode
To answer the second research question, each episode was analyzed in greater detail. The results of evaluating the students' collaborative writing products revealed the categories within each of the three types of episodes.

Categories of language-related episodes (LREs)
There are three categories of LREs identified in this study, i.e., form-focused LREs, lexis-focused LREs, and mechanics-focused LREs. In form-focused LREs, students discussed verb agreement. For example, they discussed that the options either 'said', 'says', or 'say' should agree to 'some people.' Then, the anonymous decision was to use 'say'. Then they realized that 'some people' was a plural noun. Another piece of discussion was on expressing ideas in a superlative sentence using comparative 'more' or superlative 'most' as in the sentence "parent was the most important in life" or "parent is the more important in our life". They opted for the word 'more' because they found that it suits the context better. Examples of lexis-focused LREs indicated that the students discussed the use of intensifiers 'very' or 'extremely'.
Next, the adverbs of time 'nowadays' or 'recently' were also discussed. In mechanics-focused LREs, a student did not know how to spell the word 'meritorious.' The use of punctuation in a sentence, either a comma (,) or a full stop (.), was also discussed. Table 2 displays the frequency of each LRE category produced during the recorded collaborative writing task. The students consistently paid very close attention to lexical matters rather than grammar or mechanics. Table 2 indicates that the lexis-focused category is the most productive category of LREs, with a total number of nearly half of the occurrences (46.8%). Then, it was followed by the lexis-focused category with occurrences of more than one-third (34.8%). The mechanics-focused category was the least productive, with almost onefifth of occurrences (18.4%). A. Susanti, U. Widiati, B. Y. Cahyono & T. I. S. T. Sharif, Assessing episodes in verbalization process of EFL students' collaborative writing | 546 The teacher-paired students produced three categories of outcomes: correctly solved, incorrectly solved, and unsolved. Examples of correctly solved outcomes of LREs were apparent when the students discussed the use of verb agreement 'has' or 'have' and the verb form after 'to'. Word choices were also deliberated. Incorrectly solved outcomes of LREs were seen when students discussed the use of 'both of…', whether they would use 'both of the kind', 'both of the kinds', or 'both of each' and incorrectly used predicates. Another mistake is word selection between 'everybody' or 'every child,' and the incorrect choice was made (i.e., everybody). The unsolved outcome of LREs was evident when students argued on how to translate an idea 'dengan batasan-batasan yang sudah disepakati' [based on the agreed boundaries] into English. Translating words like 'bertengkar' [argue] is also left unsolved. In the end, they neither made any conclusion nor did they write a relevant sentence to those issues in their final drafts. Table 3 illustrates the frequency of the outcomes of LREs.  Table 3 shows that the majority (83%) of LREs occurrences were categorized as correctly-solved outcomes, with the lexis-focused category being the most frequent LREs, followed by form-focused and mechanics-focused. Only a few LREs were in the incorrectly-solved category (12%) and the unsolved category (5%). These findings imply the students were more concerned with lexis-related issues and less with mechanics.

Categories of text-related episodes (TREs)
Two categories of TREs were revealed from the study, namely organizationfocused and content-focused categories. Table 4 presents the occurrences of categories found when the students talked about the organization or content of the essay they wrote. Table 4 shows that the number of occurrences of the two categories of TREs was slightly different. However, the trends were almost similar, suggesting that there is a close relationship between the organization (50.74%) and content (49.26%). Examples of categories of TREs include organization-focused and contentfocused categories. Examples of organization-focused categories were apparent in the discussion of how to write the introductory, body, and concluding paragraphs. The examples of content-focused TREs were noted when students discussed the given topic, how to outline before writing the essay, and how to conclude.

Categories of scaffolding episodes (SEs)
There were three categories of SEs exhibited by the students in the collaborative work, namely repetition, elicitation, and justification, as presented in Table 5.  Table 5 shows that the category of SEs that occurred most frequently was repetition (89.5%), followed by justification (5.7%), and elicitation (4.8%). Some examples of SEs shown by students during the audio-recorded collaborative writing task include three types of SEs: repetition, justification, and elicitation. The repetition category happened when students kept repeating the same words. They mentioned 'which', 'whether', 'and', 'between', 'we need to…', '…to educate…', 'reading', and '...parenting' several times throughout the conversations. Elicitation episodes occurred during the development of paragraphs. Sentences were constructed by eliciting a response by questioning each other. An argument on the usage of the phrase 'less of attention' and 'lack of attention' in a sentence is an example of justification episodes. In the writing process, a student proposed an outline consisting of five paragraphs (introduction, contents, closing paragraph), the other students agreed, and confirmation was made.

The Frequency of Occurrences of LREs
The present study has revealed that 141 categories of LREs were produced (see Table 2). Although the links between the number of LREs and language learning achievement are still being explored, previous studies have confirmed that LREs were sources of language learning (Swain & Lapkin, 1998;Swain & Watanabe, 2012;Watanabe & Swain, 2007). The trend found in this study should be perceived with caution because the data in this study are limited only to verbalized language. Non- A. Susanti,U. Widiati,B. Y. Cahyono & T. I. S. T. Sharif,Assessing episodes in verbalization process of EFL students' collaborative writing | 548 verbal language such as signs, mimes, and expressions that might happen during pair interactions was not captured. Furthermore, Table 2 has shown the categories of LREs as a trend that students paid especially close attention to lexis-focused LREs. These findings align with the previous studies, which found that intermediate learners would pay more attention to vocabulary issues (Kim, 2009;Niu et al., 2018). Additionally, paying more attention to the lexicon seems to be consistent with the result of a study conducted by Rahayu (2020), which shows that in the collaborative writing, the interactions among Indonesian-Indonesian pairs focused on 'the language-related aspects' mostly on lexical choice and the meaning, more than the mixed pairs of Indonesian-Chinese students. The outcomes of the LREs indicate that the students' pairs could successfully resolve the LREs. This evidence strengthens the previous studies that pairs could scaffold each other during collaborative dialogues and eventually met the correct solution to their language problems (McDonough, 2004;Mozaffari, 2017;Storch, 2008;Watanabe & Swain, 2007). The other trend observed in Table 3 is that when there were unresolved LREs and incorrectly resolved episodes, pairs would prefer to avoid difficulties or matters that none of the members could decide. They resolved these problems by not continuing their discussion about the problems.

The Frequency of Occurrences of TREs
The TREs were the least episodes produced after SEs and LREs. Previous studies have reported that the students' TREs depend on the writing topic and the nature of the writing task (Neumann & McDonough, 2015;Niu et al., 2018). In the current study, all pairs were given the same topic when they did the collaborative writing. The number of TREs was less than the other types of episodes since, in this study, the task was reading-for-writing. The pre-writing assigned reading texts much facilitated the students' content of the writing.
Students paid relatively equal attention to the organization-focused and contentfocused categories of TREs (see Table 4). This finding contradicts the previous findings that collaborative pairs produced more content-focused TREs than organization-focused TREs (Neumann & McDonough, 2015). The contradiction might be a result of the difference in the pairing techniques and the writing task. In the previous study by Neumann and McDonough (2015), the pairing technique was not explicitly explained, and the students were assigned to produce free writing, while in the current study, the students were paired based on their proficiency and were assigned to complete integrated reading and writing tasks.

The Frequency of Occurrences of LREs
There were 227 occurrences of categories of SEs produced by the students (see Table 5). In L2 learning, especially in collaborative tasks, scaffolding could be initiated by teachers and peers. In the current study, the teacher gave general feedback to students at the end of the lesson. During a pair of collaborative dialogues, however, the students talked to each other by producing scaffolding episodes. In addition to the effectiveness of teacher scaffolding (Vonna et al., 2015), peer scaffolding facilitates L2 learning (Donato, 2004;Watanabe & Swain, 2007). Scaffolding could increase the students' motivation as they get immediate and relevant feedback from their peers (Dabao & Blum, 2013;Rezeki, 2016).
The most frequent category of SEs was repetition, followed by justification and elicitation. This is in line with the previous studies, which showed that repetition was the most frequent category of episodes during pair collaborative writing (Watanabe, 2014;Yoshida, 2008). The students in this study frequently repeated their peer's words or phrases. Repetition could be used as a cognitive strategy to acquire new words and memorize unfamiliar terms (DiCamilla & Anton, 1997). Justification and elicitation were less productive categories of SEs. Yet, the appearance of justification and elicitation categories of episodes showed that the pairs were engaged, and they complemented each other (Storch, 2008;Zhang, 2018). Such engagement could support the intensiveness of peer feedback. Previous studies have reported that peer feedback is effective for the learning of EFL writing (e.g., Cahyono & Amrina, 2016;Kusumaningrum et al., 2019).

CONCLUSION
To summarize, the results of this study indicated that students who were set in pairs mostly produced SEs, followed by LREs, and the least productive counterpart was TREs. In addition, the form-focused category of LREs, the organization-focused category of TREs, and the repetition category of SEs were the categories of episodes that were mostly produced during collaborative writings. Collaborative writing tasks have enabled the students to scaffold each other, emphasizing language forms and text developments and sharpening their collaboration strategies. The most frequent category of scaffolding episodes was repetition, followed by justification and elicitation. This may suggest that students need repetition in lesson delivery, too.
Students seem to have paid close attention to lexis-focused LREs. Therefore, teachers may consider teaching students more vocabulary acquiring strategies. As students could successfully resolve the LREs, teachers may focus on areas containing incorrectly resolved LREs and further encourage them to attempt the unsolved LREs. If left unchecked, students might continuously make the same mistakes or feel demotivated to seek answers to unsolved matters. Finally, students paid attention to the organization-focused and content-focused TREs in the context of a reading-forwriting task.
Overall, the study has unfolded the type of episodes which was the most productive in the verbalization process of collaborative writing and which categories of episodes were the most productive in the verbalization process of collaborative writing. The results could be a catalyst to a much-guided EFL teaching approach. In addition, the results could be valuable, especially for EFL teachers who are preparing sets of instructional materials and techniques to develop EFL students' writing skills.
This study results in some issues that need to be examined in the future. The small number of research subjects challenges the generalization of the research findings. Accordingly, future researchers could find accessible research sites that have more students in their writing classes. While the students in this study were all of a relatively similar age and had the same courses in the previous semesters, this study did not collect the students' perceptions and attitudes toward collaborative writing or make notes of their collaborative writing and their previous writing courses to weigh