Online Small Groups Talk in English Collaborative Prewriting Phase Viewed from Social Presence Frame

,


I. INTRODUCTION
During the COVID-19 constraints, teaching and learning activities have transitioned to distance learning, also known as school from home or learning from home.According to the Ministry of Education and Culture of the Republic of Indonesia [1], regulation number 3 for 2020, states that the interaction between students and lecturer-students is mediated through ICT-based media.Synchronous or asynchronous ICT or e-learning components, such as forums, chat, e-mail, blogs, and social media, are utilized in electronic tutorials.Synchronous or asynchronous ICT has largely superseded physical presence by facilitating communication, connection, and cooperation [2].Many lecturers familiarized synchronous platforms to improve writing skills regarding online writing learning.They often used platforms (real-time) to minimize students' constraints in communicating with one another, which confined students' capacity to participate effectively.Also, it is even used for collaboration, especially in language learning [3].In Indonesia, practically, students learn English writing passively and need more effort to improve their English writing personally [4].Students must be provided with learning activities to help them write as well as possible in English.Therefore, a lecturer's primary duty is to develop an environment that encourages students to discuss a topic in English [5].The most common used is WhatsApp application.
To successfully learn writing courses, the WhatsApp application facilitates virtual collaboration.Students are engaged in dialogue and teamwork with their peers [6].Additionally, virtual collaboration builds cohesion through small groups in an academic context [7].From a socio-cognitive perspective, a group has a significant benefit in which students can develop writing skill throughout process-oriented approach with peer support [8].Specifically, the social presence framework is a sense of connectedness where participants ‗feel the presence of others online,' such as affection.Also, it is a category of identity or self-performance [9].
Commonly, some studies employ social presence adopted from the community of inquiry framework.One of the frameworks, social presence analyzes the affective and open communication, and cohesive elements in virtual collaboration.Garrison [10] found social presence developed interpersonal and affective relationship in an academic context creating a climate of critical thinking.Wu et al. [11] have already investigated the effectiveness of combining online learning with English language instruction to promote group writing activities.The findings indicated that online writing learning enabled students to compose better group research essays, acquire satisfaction in the learning experience, and be committed to the learning process.The perspectives of graduate students on social presence as an element of learning in online composition courses were also examined by Stewart et al. [12].The findings revealed that students perceived social presence, social comfort, and social learning, as well as their attitudes toward online learning.In addition, the CoI writing survey revealed that ongoing conversation through social presence could make students feel comfortable.Chen and Liu's findings showed that the students were divided into small groups with three to four people to discuss assignments successfully; it depended on the frequency of social presence of other group members and multifaceted talk within small teamwork [13].Chasteen [14] investigated collaborative activities in online basic writing courses.The results demonstrated that community and cooperation could help to lessen the isolation and lack of confidence that novice writers experience while beginning their college writing.Peer review, in other words, is a social activity that fosters comfort and community.These findings supported the importance of online collaborative learning in writing courses has a favorable effect through social presence.
It is still rare that collaborative writing activity is reviewed from Galley et al. framework [15].This framework focuses on social dimensions in creating cohesive online learning communities.Communication in virtual learning classrooms demands the (co)presence of potential interaction partners.However, some researchers [10][11][12][13][14] investigated social presence adopted from the community of inquiry framework to formulate student-teacher (or peer-to-peer) written-based discussion around the academic goal, analyzing the affective, interaction, and relational elements in virtual collaboration.Other researchers [3,[16][17][18][19][20][21][22], also employed online learning to facilitate writing activity on WhatsApp.Thus, to fill in the gaps left by the earlier investigations, this study will look into what students talked about while doing academic tasks and their responses to the collaborative prewriting process at a private institution in Indonesia viewed from Galley's social presence framework.
This current issue is characterized by group members' contributions to cohesion, participation, and individual identity.Cohesion is portrayed as mutual support and tolerance, turn-taking and response, humor, and emotions within a group.Participation is portrayed as exploring ideas and thinking together about an informative, knowledge issue.Finally, identity is portrayed in terms of the roles played by group members through written chat during the collaborative writing activity.Therefore, the current study aims to explore social presence a).The nature of small group discussions during the prewriting phase in aspects of cohesion, participation, and identity via WhatsApp, and b) their responses regarding the prewriting process.

A. Online Prewriting Phase
In the prewriting phase, collaborative tasks through learning together is a proper choice for EFL learners and highlight dynamic discussion between learners with diverse capabilities and background knowledge [23].Discussion creates a social environment where students can practice critical thinking; the more practice they receive, the higher their performance.The more one discusses, the more ammunition one gain for the writing struggle [24].As suggested by the researchers [25][26][27], to make the discussion interesting, students can watch the videos, observe images, or conduct a strategic interview to assist them in talking about ideas, planning content, and organizing essays and language styles jointly.McDonough et al. [28] analyzed five categories of collaborative prewriting process including content (topic, main ideas), organization, language (grammar, and vocabulary), task management, and off-task talk.Compared to a collaborative prewriting study, Li et al. produced key findings: discussing the ideas, vocabulary, and organization of students' writing, solving linguistic puzzles, sharing responsibility for the writing task, expressing learners' emotions, and maintaining learners' harmony in each group [29].Another study examined about the interactive prewriting approach in improving writing skills.The finding revealed that structured group prior-writing assignments prompted students to talk about content and organization.Also, there is an adequate link between these prior writing talks and the writings [30].Nevertheless, the collaborative prewriting process yielded a higher grade than solitary ones, and students engaged in critical thinking.

B. Social Presence
In socio-constructivist principles, Vygotsky [10] states that learning is interactive and collaborative.Learning leads to the development of knowledge and internalization of social activity.In other words, social presence (SP) is a crucial mechanism for social engagement in virtual learning environments.Learners engage in social understanding and L2 knowledge sharing with more than one person to complete a task jointly through synchronous, asynchronous, or blended learning in any educational discipline.According to Rourke et al. [31], the three forms of social presence include cohesion (phatic, vocative, or general talk), affection (emotion, humor, self-disclosure), and participation/interaction (starting or continuing thread, asking a question).The description shows that social presence is a core vehicle for social interaction in the sense of -being there‖ and -being together.‖Galley, Conole and Alevizou [32] developed social presence within Community Indicator Framework (CIF).Three of the four dimensions of social presence were used in the current study: (a) cohesion, (b) participation/interaction, and (c) identity to support online learning communities.The three dimensions have some indicators (mentioned in next section) used to analyze the synchronous interaction.
The first dimension is cohesion.It relates to community members' responsiveness and turn-taking, willingness to listen and learn, demonstrated by language use [33].Also, cohesion reinforces mutual relationships and encourages one another through actions where all participants work together.The indicators of cohesion include phatic, collaborative floors, and affective talk.Phatic is about greeting and general talk.A collaborative floor was used to construct discussion through backchannelling, initiating, asking questions, or holding the floor.In WhatsApp, talk or discussion is represented by patterns of written conversational floor posts, comments, and responses for academic purposes [25].Affection is related to emotions, humor, and mutual support.
The second dimension is participation-students' interactive activity means thinking together about knowledge.Participation indicators indicate students talk about resources (links from the web), negotiate tasks, and explore detailed academic courses.Negotiation is divided into procedure and meaning.Negotiation describes students' responsibilities together when students agree on how to do or figure out a task together.In addition, according to Hartono and Ihsan [34], one of the meaning negotiation is an expression of a non-understanding notion.In instant chat messaging, meaning negotiation appears when one (or more) interactants do not understand their talk.Another interactant responds to their non-understanding in which there is a negotiation to deal with a solution.After the nonunderstanding has been solved, the flow of the talk may proceed [35].To sum up, meaning negotiation allows one to negotiate what is related to their knowledge until other interactants understand it.Next, according to Donanci [36] and Polo et al. [37], exploratory talk means that partners give input, present reason, analyze International Journal of Information and Education Technology, Vol. 13, No. 12, December 2023 problems, give possible suggestions and corrections, and reach decisions jointly.
The third dimension is performing an identity online.Students are encouraged to establish a sense of belonging [33].Similarly, identity may emerge through written discourse in the synchronous platform and be considered as interpersonal ties.In Galley et al's framework [32], identity presents a student's awareness, commitment to a joint task, and willingness to listen in a group.When students are active in learning, they can ascribe and position their attributes in a certain period.Certain positions become more dominant in one's self.In Table I

A. Design
Cohen [38] argues that a quantitative descriptive design is a non-experimental research focused on processes and interactions.Multiple data-gathering methods use observations in natural settings.This study employed the quantitative descriptive design to investigate what students talked about academic tasks, performed and responses to the collaborative prewriting process at a private institution in Indonesia.We conducted the study in the academic year 2021-2022.

B. Participants
The participants consisted of 15 sixth-semester students enrolled in a course as convenience sampling who were involved in the study [38].Due to the pandemic, the number of students decreased from 25 to 15.They were EFL students with nine females and six males who learned an academic writing course.Their ages ranged from 20-25 years.The participants pursued graduate degrees in Indonesia's English Language and Education Department.The participants were at the upper-intermediate level.In the syllabus, to achieve undergraduates' competencies in the academic writing course, the participants may master the skills to write academic essays with good organization and language accuracy and use experts' references at the end of the semester.

C. Data Collection
The data include two categories: observation or written chats and questionnaires [38].The observation was conducted via WhatsApp.Observation data were collected from 15 participants in four groups of learning communities.As we collected, written communication were analyzed.Galley's notions help to explain the nature of conversational interactions.We identified the types commonly written on groups' chatrooms: welcome chat, supportive interchanges, humor and emotion, sharing resources or links, engaging in discussion, asking questions, and making other statements.
At the end of learning, the students were given a questionnaire form to find out the students' responses to the prewriting process via WhatsApp.We wrote all the participants' names in pseudonyms.

D. Data Analysis
Cohen [38] asserts that texts are all written communicative contents: particular phrases, sentences, and paragraphs intended to be read, studied, and understood its meaning.In this study, we used the content analysis of textual data from four groups.
Before analyzing the first research question, textual data from WhatsApp were converted into Microsoft Word without changing the content, spelling, or language.We uploaded textual data into Nvivo12 software to assist the coding system.We thoroughly read and re-read the data to comprehend the meanings and critical issues [39].Then, we coded particular phrases, sentences, and paragraphs under the available indicators by a deductive approach.To ensure the data findings, we conducted the observability and interpretive validities [38].The observability validity was to provide quantitative data by frequency and percentage.All written communication were presented in the form of evidence, shreds, excerpts, or extracts of talks as valid data.Meanwhile, to ensure qualitatively identity data, the researchers used interpretive validity to interpret the students' attributes during pre-writing activities based on written chat.The attributes were categorized into five parts: -Having the sense of belonging‖, -Having an initiating nature‖, -Having an accepting nature‖, -Having an explaining nature‖, and -Having a negotiating nature‖.
Regarding the second research question, before being distributed to the students, the questionnaires were validated using content and construct validity with two expert judgments.The questionnaires have been tried out to students instead of being a sample of this study.Then the researchers analyzed the tryout result to determine the test items' validity and reliability.Twenty-one questionnaire items were divided into three constructs: cohesion (seven items), participation (10 items), and common purpose (4 items).We used a Likert scale with intervals ranging from 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.We counted the questionnaire results into descriptive data using SPSS.Then, we created the measurement scale of the questionnaire to explain the interpretation of data and assess the results through mean and Standard Deviation (SD).The scale for questionnaire result shown in Table II

E. Procedure and Materials
Before conducting the learning process, one of the researchers contacted the English academic writing lecturer asked permission, and joined the class online via WhatsApp.The researcher acted as a passive participant in group chatrooms to observe the students' prewriting activity when completing assignments.
At the core of the activity, the lecturer created class chats and group chats (Groups 1, 2, 3, and 4).Group chats were used to discuss collaborative tasks.The lecturer distributed the resources taken from the internet, like using pictures, learning videos, and lists of questions.The collaborative writing tasks were set up in four meetings.In each meeting, the learners were allocated 1 h and 30 min to complete one task.The essays had at least 750 words.
The lecturer also distributed worksheets that were posted in the WhatsApp group.The worksheets given as a guide for discussing and writing the main topic and sub-topic, thesis statement of an opening paragraph, main ideas of each supporting paragraph, closing paragraphs, and language.On worksheets, the lecturer wrote instructions as a trigger to discuss, such as what do you think of this pictures/video?What is the background of the pictures/video?When you see pictures/videos, what learning or teaching methods are applied?What is your argument for this?
In Task 1, each group got different pictures to determine a topic and title.Group 1 talked about -The advantages of using blended learning‖, Group 2 talked about -The benefit of reading English literacy in collaboration‖, Group 3 discussed -Learning English during the COVID-19 period‖, and Group 4 discussed -The teacher's role in teaching English in elementary school‖.In Task 2, each group watched different videos to determine a topic and title.Group 1 talked about -Stimulating young learners to write‖, Group 2 talked about -Dictogloss: A strategy in English teaching‖, Group 3 discussed -Cooperative strip paragraph‖, and Group 4 discussed -The benefits of collaborative writing‖.In task 3, the groups got lists of questions to lead them to determine a topic and title.With the same issue, four groups discussed -The advantages of an online collaborative essay‖.Finally, in task 4, the groups discussed an essay with their topics.Group 1 talked about -The quality of education in college‖, Group 2 talked about -The impact of cellphone radiation‖, Group 3 talked about -Good and bad environments for studying‖, and Group 4 talked about -Reading e-books and paper books‖.In the pre-writing phase, each group discussed the content, organization of the introductory paragraph (thesis statement and topic sentences), body paragraphs (argumentation with references or sources), and concluding paragraphs.The examples of two learning videos were taken from YouTube for the collaborative tasks as in Fig. 1 and the example of group discussion as in Fig. 2  The links were like https://youtube.be/UUVPeFVbTK4,https://youtube.be/H7hWLjBvCOI,https://youtu.be/QLaS429LAJA, https://youtu.be/UDUAM4hLrqE.Then, four small groups discussed their essays as the below examples:  The analysis presented in Table III shows the frequency of each indicator in the four groups across the four academic essay tasks.Eight indicators are identified during discussions.In the cohesion category, the collaborative floor (N = 352) contributed as the most significant number indicator and was followed by an affective talk (N = 306).At the same time, phatic talk (N = 139) produced the smallest.In the participation category, the small-group talks were driven by organization talk (N = 408), followed by content talk (N = 316), language talk (N = 307), task negotiation (N = 296), and share resources (N = 83).

1) Cohesion dimension
The results as presented in Table IV below.Based on Table IV, the average number of phatic (e.g., greeting) was 5.87%, the lowest of the collaborative floor and affective indicators.Group 1 had the highest percentage (9.90%),followed by Group 4 (5.37%),Group 3 (5.23%),and the last was Group 2 (4.35 %).The excerpts of greetings, -good morning‖, -How are you‖, or -hello‖ or addressing others like -where is our friend?‖,-Where is Ros?' and continued talking general issue, -Wait, I have internet problems‖, -It is okay, Ros, we will try to understand you.‖Another example appeared from Group 1 e.g., -Frankly, yesterday I almost forgot there was an assignment, so I am in a hurry writing on tablet -and answered -I forget at all the task also, bro.I have just made a half.I think copy paste sounds easier‖ Another example -yes, finally it finished.Lucky me.‖The facts indicate that the four groups show politeness like greetings, addressing others, and simple social talks, even though it is very little apparent.
The average number of collaborative floors was 14.32%.In the initial talk, four groups discussed a new topic by giving signal attention: posing questions, maintaining the talks, and showing comprehension.The talks across the groups ranged from Group 4, 19.56%, next to by Group 3 (16.02%),Group 2 (15.06%), and Group 1 (12.53%).The sign of posing questions were communicated, Mita asked members to work together, e.g., -Can we start now?‖, -What is the background of this information?‖Another excerpt, -What video do you see?‖, -What do you think, Ros?‖In Group 3, Feti introduced a new topic -Hi intan can we start?‖-What do we start to discuss?‖ and -So, what do you think about the topic?‖ and -So, what title is about?‖-What about learning English during the COVID period?‖Based on these findings, an initial conversation was started with questions by the first peer initiating the response.Other peers responded, e.g., -Let's we discuss it first …,‖ or -Please continue..,‖ and -Yes, it may, please keep going.‖These words indicated that learners intended to foster discussion.The attention to comprehending talk appeared when Group 2, Mita proposed a topic about Dictogloss.A male student comprehended her talk, -Ok, I understand,‖ and Ayu answered, -Okey for this, too.‖Others said -yes,‖ and -agree, go ahead.‖These phrases showed the backchannel of the partner's talk after each member read/listened to an explanation.In other words, each group attempted to hold a floor and take turns.
The average number of affective talks was 15.82%, the highest result than the collaborative floor and phatic.The talks across the groups ranged from Group 2 (16.06%),Group 1 (15.34%),Group 4 (12.02%), and Group 3 (11.26%).In their talks, positive responses frequently emerged from groups compared with humor and emotion.The excerpt of positive expressions when each partner felt satisfied with their peers' work or motivated each other in Task 2 came from group 2. They supported a member who finished the task well -Great,‖ -So excellent,‖ -Perfect, so far everything is good.‖Also, they encouraged a member who had less confidence, e.g., -spirit,‖ -don't give up,‖ and gave the emoticon -topic 1.‖In Task 1, group 1 also showed positive responses when Mul responded to Anto's title -I like this title‖ Likewise, emotion was apparent only in Group 2 when performing Tasks 2 and 3.For example, when a member reacted slowly or without a reply, Mita was angry, -Oiii, where are you.I talk my self!.‖ Another excerpt, -I have no patience anymore!.... -Please reply!still had no response from peers.We need to finish this essay this day,‖ and -She does not even spare some time replying to this group.Ros, please, if you do not understand, ask me, do not be silent!.‖ Humor was apparent from group 1 by giving emoticons …. when the students commented on a funny thing.when Ros said -it's puzzle for me.‖Others were laugh -Ahahahhaa.‖These findings show that simple general talk, humor, and conflict interactions are few apparent in authentic online learning.In comparison, mutual support conversation indicates that the groups were more comfortable communicating their positive feelings.

2) Participation dimension
The results, as displayed in Table V presented below.Based on Table V, the average number of shared resources was only 7.74%, the lowest talk.Group 3 shared references or links for 4.40% of their essays.Group 2 used references 3,34%, Group 1 used references 3,83%, and Group 4 used references 3,40%.For instance, Group 1 discussed suitable references to support their argument.In Task 1, Anto recommended a link written on chat about the five benefits of blended learning, -this source seems good, http://www.imaginelearning.com, it was written by Sherri Walker in 2018.‖He also asked his peers to put another source -Don't forget to cite this theory about young learners in learning English from Brown in 2000.‖Group 2, Mita proposed reference in Task 2, -how about the dictogloss technique, … cited from David Nunan in Azies and Alwasilah (1996:58)?we used dictogloss approach from Gibbons, 2015.When you cite these sentences, please paraphrase or summarize the text, and put the expert's name as our lecturer says, oke?‖ Group 3, Hani shared sources of the link.-Let me add the link oke? www.republika.id/posts/12145/sekolah_tatap_muka_Beresiko.‖The excerpts from Group 4, -Please cite this sentence of learning effectiveness… from Hammer in 1991.‖Likewise, in Task 4, Group 4 talked about the link -Please friends, see this source, https://twosides.info,about the positive and negative sides of reading e-books and paper books, is it suitable?‖These findings show that the groups attempted to find sources for providing facts to support arguments in essays.They utilized the internet to search the links or references, but the intensity of talking sources was limited.The use of WhatsApp helps ease the accessibility of ICT-based resources.The resources could be looked, uploaded, opened by clicking easily, downloaded and kept the links on their WhatsApp.They can read and cite the parts that would be written, paraphrase the texts and give credit to the writers they had borrowed the words.However, the groups need to learn more about properly citing text.
The average number of task negotiation talks was 13.63%, the second-lowest result of the five indicators.The total talk across all groups was: as followsGroup 4 (14.62%),Group 3 (14.08%),Group 2 (13.00%), and Group 1 (11.59%).Researchers have identified two types of negotiations.In the first type, the groups negotiate tasks with their peers on how to carry out assignments.The excerpt was taken from group 2 i.e., -We have three paragraph bodies to share.‖Mita responded positively, -Okay, I choose paragraph 2. Next, Ayu offers her part by saying, -I choose paragraph 1... Ros, you get the rest of paragraphs 3 and Teuku paragraph 4.‖ Group 3 also shared responsibility, for example Feti said, i.e., -To write strip essay, Hani and Zahra will make first paragraph as the introduction, and bodies 1 and 2 about function, and I will make bodies 3 and 4 how to write an strip essay.So our paragraphs must be already written before Faisal write a conclusion.‖The findings indicate that procedural negotiation appeared when the groups built joint paragraph sharing to agree on assignments among all members mutually.
The meaning negotiation started when Group 2 discussed a title as its trigger.Mita felt the task was not easy i.e., -Oiii, it is dizzy this time.It's still an introduction.‖Ayu, Rose, and Teuku responded without giving any new input -Me too.I have not understood yet.‖Then, Mita found a solution and negotiated whether they accepted what was said, -eh by the way, do all of you agree our title is reading as a hobby?Isn't it?‖-or do you want to use the topic from the lecturer?-Rosnegotiated, i.e., -If the topic about reading is a hobby, is it more personal for a student, isn't it?‖Mita answered, -I don't think so, because our topic has a positive side, we can highlight its benefit of reading books, or we choose a title from Ms. Sr, deal?‖These talks indicate the lexical problems like -dizzy this time‖ and -I have not understood, yet.‖These facts of words signify the nonunderstanding notion so that they negotiate to find a solution to deal with other peers.
The average number of content talk across five indicators was 14.36% as the third highest indicators.Each small group demonstrated by Group 2 (15%), Group 3 (14.78%),Group 4 (13.97%), and Group 1 (13.47%).The examples of the typical content-based conversation appeared when Group 1 discussed Task 1 about the theme of Small Group Interactive Writing in Kindergarten.Anto said, -On that video, the object is the young learners means they are learning how to write….‖Then, Mul reacted, -So, the keywords are the direct method and young learners, right?‖It was answered, -It is for an introductory level.It is good for our topic because the target is young learners.So, we have to discuss more about the direct method.how about you, Ir?‖ In the following example, with different videos, Group 2 discussed a new topic and title.-I think, if we choose that title, we should put many things in our essay, starting with the definition, how to apply it in the classroom… and its pros and cons… hmmm, what do you think?‖The evidences of content show that they shared viewpoints on topics and gave arguments and reasons.Then, they selected alternative ideas and topics and confirmed them with one another.In other words, the students seek concepts for a final one which set out as a major point for composing an opening paragraph.
As shown in Table V, the average number of organization talks was the highest among the five indicators (17.89%).The most frequent communication was from Group 2 (19.00%),Group 3 (19.00%),Group 1 (18.86%), and last Group 4 (17.20%).The groups talked about introductions; how to write the thesis statement, talked about body paragraphs; how to write an example, fact, and detailed information, talked about concluding paragraphs; how to summarize or conclude paragraphs.In this excerpt, Group 2 discussed the opening paragraph of Dictogloss: How to teach writing, e.g., -Owghey, gonna send the intro first,‖ and -An introduction is ready…Please read and check it.Maybe there is a mistake or need to add…or is it okay for an introduction?Please, comments are very welcome.‖In another extract, Group 3 shared on chat, -this is thesis statement… when COVID started being a red zone situation, all public activities stopped, especially face-to-face learning activities.‖In Task 1, Group 4 discussed body paragraphs such as -I have paraphrased paragraph three like this.. to increase the effectiveness of learning, a teacher also takes part in teaching intending to be able to see his students develop directly and interact directly with their students... oh ya, the sentences were cited from Harmer 1991, p.241, is it okay?Please, correct it if it is wrong.‖Group 4 also discussed the concluding paragraph, -This is the conclusion, I try summarizing all paragraphs, the function of teachers is to help students learn by imparting knowledge…Please, correct it if it is wrong.‖From these facts, we found that the students received corrections and gave suggestions for thesis statements or paragraphs from their partners.They attempted to organize paragraphs from the intro to the conclusion.However, the current finding reveals that checking organization is the most often discussed, as instructed by the lecturer.But, the students seldom check the paragraphs whether they are bonded or not to each other.The average number of language talks was 14.41%, the second-highest result among all indicators.The language aspect was reasonably talked about in four groups.Group 2 had the highest number of talks (15.44%),Group 4 (13.76%), group 1 (12.93%), and Group 3 (12.50%)had a similar percentage.Students discussed language, including grammar, vocabulary, word transition, and spelling, as parts of checking coherence.For instance, in Task 1, Group 2 discussed linguistic problem-solving in an essay, -The Benefits of Learning English in Collaboration.‖One student asked, -What about the tense?‖ another suggested, -I think it needs revision and the word ‗we' can be changed to be students/ pupils because, in the introduction, already used students, -for this sentence, you should kurangkan penggunaan and or delete ‗and' okay.‖ These findings indicate that Group 2 has more effort to correct tenses, meaning, choice of words, spelling, and word transition to be written/ talked in their essays.These facts reveal that language correction is crucial, so they attempted to eliminate the problems of tenses, linking words, and coherences which frequently occurred.

3) Performing identity dimension
Group identity reflects the students' roles in cohesive and participative talks from the language used by group members.Students in each group showed their identity performance in learning as exposed in Table VI.
The table shows students' identities in virtual learning and depends on each member's identity.Table VI in synchronous discussions, the talks are seemingly controlled by several students at a time.Eight students in four groups performed dominantly in showing identity as long as they finished the task.For example, In Group 2, Mita and Ayu positioned themselves as the initiator, negotiators, and source person by giving input, suggestions, revisions, and recipients in accepting peers' opinions such as -Nice, now we will proceed to body 1‖, or -It is oke.We will translate together,‖ -This is International Journal of Information and Education Technology, Vol. 13, No. 12, December 2023 our essay, and let us arrange the sentences first because some are still unrelated.‖Two students in Group 3 also positioned themselves as the initiator, negotiator, and source person, e.g., -oke, we decide points of impact, then we find sources.‖The initiating nature means inviting other peers to start a discussion on every occasion, like -Where are you, friend?‖ or -Ready all?-Through the use of words ‗we' and ‗us,' the dominant students showed belongings.The accepting nature indicated that they welcomed input from others.At the same time, the explaining nature demonstrated that the dominant students have more knowledge to correct the content of the essay, words, or grammar and revise paragraphs.In the negotiating nature, the student agreed on task responsibility and bargained the meaning of knowledge.In summary, the students within each group take their roles with different contributions to form their identity that emerged during the online discussion.

B. Learners' Responses on Prewriting Activities
The questionnaires showed that the students agreed that cohesion, participation, and common purpose were significant to succeed in the tasks.The results as seen in the Table VII.The findings of means and standard deviations are presented in Table VII.The measurement scale from the three critical social presence concepts scored in the -agree‖ category.The average result (M = 3.010; SD = 0.607) fell within the -agree‖ range.The response with the highest mean score is -Provide multiple means of participation‖ (M = 3.103; SD = 0.582).It indicates that they agreed to talk about features of writing skills in an online collaborative task.The group member with lower communication benefited more from the collaborative writing task.The second-highest mean score is for the response -Provide multiple means of common purpose‖ (M = 2.962; SD = 0.619).According to their answers, every student should perform and contribute well.They concurred to work together.They felt at ease using WhatsApp, avoided feeling alone, exhibited awareness, and agreed on shared tasks.The lowest mean score is for the response -Provide multiple means of cohesion.‖(M =2.809; SD = 0.621).To sum up, they accede that cooperation and assistance from one another are crucial for finishing the responsibilities.

V. DISCUSSION
The study showed that four groups constructed social presence with different contributions.In performing identity online, eight students in each group took more roles than others.However, overall, the students responded positively at the ‗agree' level about cohesion, participation, and common purposes.

A. Academic Discussion
There are findings from this study supported and corroborated by previous research, but some differ from earlier results.The previous researchers [25,26] suggested that generating ideas through watching videos assisted students in having thoughts, planning content, and organizing essays and language styles jointly to stimulate writing ideas.
Collaborative students' talk in the prewriting phase promotes the social presence of cohesion, participation, and identity.In cohesion, the groups work together by constructing collaborative dialogue.This finding has linear with a previous study of digital interaction via WhatsApp.Collaborative floors were presented from text messages by asking questions, talking about a new topic, and understanding what other people were saying, indicating it was more dynamic [40].A collaborative floor enables students to control their turn-taking, create collaborative spaces, and develop their interactive abilities, particularly in the early phases [15].
In the affective talks, the students most frequently demonstrated positive feedbacks in Groups 2, Group 3, and Group 4. Nevertheless, it is very little apparent in which the four groups make jokes or humor.Meanwhile, emotion is apparent when one learner within Group 2 shows incidents of negative feelings because of the slow response from a peer.The findings parallel with previous studies [7,41].Mutual support was more apparent when learners encouraged each other to foster a positive group atmosphere, and to support productive collaboration among members.The preceding study by Li et al. [29] found that humor appeared as a part of social relationships.In comparison, Isohatala's study discovered no incidents of negative talk, which would have resulted in disputes and harmed collaborative learning [42].However, it may have occurred that feelings during prewriting activities emerged, such as overruling and underestimating, may manifest as negative interaction [20].
The students built the negotiation of procedure and meaning on tasks to get the deal among the members.Nevertheless, three groups dealt with tasks procedurally.Meanwhile, only some are involved in the meaning of negotiation, which starts up non-understanding, continues by clarification or request, and ends by having the deal.In a past study [34], meaning negotiation showed non-understanding notions that could mean channel or lexical trouble.Another researcher found negotiation sessions bargained their related tasks by anticipating, perceiving, resisting, or reacting to other group members [32].Mobile-assisted collaborative writing groups made swift progress into and out of the procedural bargaining stage.In Yeh and Chen's finding [43], personal negotiation was used less frequently to reach a consensus among all parties than meaning negotiation.
Sharing links to references posted via WhatsApp makes the students are easier to open and view materials.As found in the past study, sharing links on a platform should make the learners quickly open e-materials by clicking and International Journal of Information and Education Technology, Vol. 13, No. 12, December 2023 downloading links to e-journals, videos, and the web [44].
The study found four groups frequently communicated organization, beneficial support, language, content, collaborative floor, task negotiation, shared resource, and phatic/off-task talk.In contrast, the results differed from previous studies.The small group students mainly discussed content, task management, organization, language, and off-task talk [28].Likewise, Li et al.'s study [29], which analyzed joint writing discussions, also found that small groups' prewriting process was most frequently talked about the content, followed by language, task management, and language phatic, emotion, and organization.
In the current study, performing online identity depends on each member contributing to cohesion and participation.They take turns taking in the discussion.The dominant students in each group show more of their position in initiating, negotiating, explaining, accepting, and belonging than other members.Also, the dominant students often used ‗We' and ‗Us' as collective contributions.Indeed, we also see a silent student wich less communicated with other students over certain meetings.The earlier study corroborates this finding that the dominant students position themselves in talk.They co-create and reshape their multiple identities.Furthermore, becoming a silent student is also an identity and one of the various identities a student has [45].In collaborative learning, bilingual repertoires may provide greater possibilities for meaning-making, identity creation, and joint learner communities [46].
As Tang and Hew [47] state that the use of instant mobile messaging can facilitate social presence in online communication, and the students feel more positive when engaging in social and academic discussions with classmates.
In summary, the collaborative prewriting phase fosters a learning community by emphasizing learners' contributions.Moreover, prewriting enables pupils to join online groups as authoritative members.Small-group students talk in the collaborative prewriting phase synchronously can be used for fostering social presence: (a) the expression of phatic and affection, (b) the development of collaborative dialogue (cohesive social interaction), (c) the expression of agreement, negotiation, (d) the development of student's skills to write experts' sources, discuss content, language, and organization to proceed with groups' academic essays, and (e) the performance of students' identity in online learning.

B. Recommendation
The study's recommendation is aimed at university teachers, researchers, institutions, and government.First, for university teachers, WhatsApp is suitable for interactive discussion and transferring sources, links, and e-material.Still, it should be supported by another writing tool for complex essay tasks in which the learners require an extended and more prolonged composition of academic writing.Collaborative learning can be chosen as the strategy to promote teamwork in online learning presence, especially in writing courses.Second, researchers can study further research by Galley et al's framework to analyze interaction by using other platforms, which is still rare in the context of English writing learning.Third, universities provide facilities, technology standards, additional digital assets, and human resources for distance learning to promote learner-learner interaction, classroom structure design.Fourth, educators faced sudden virtual learning because of COVID-19 with limited preparation.Therefore, the government must promote learning based on technology in the curriculum as one of the compulsory teaching strategies that universities and colleges should implement to avoid technological stammer in using applications.

C. Limitations of the Study
The current study contains limitations that should be considered, even though the results may assist writing instructors in making decisions regarding creating and implementing collaborative prewriting activities in their online writing courses.For future studies, bigger participants from various institutions, programs, and disciplines can be involved.

VI. CONCLUSION
The current study reveals the small-group student talks during the virtual collaborative prewriting phase in Indonesian tertiary in line with the social presence frame.The discussion in small groups concern cohesion, participation, identity, and the students' perceptions of the prewriting process.Students discussed four fundamental discoveries in small groups: (a) in cohesion aspect, provided students to work together, to give motivation, or to express their emotions without intervention.Mutual support is the highest episodes emerged than the collaborative floor, and humor and negative expression in their talks; (b) in participation aspect, provided students to generate content, language, organization, negotiation, and shared resources.The organization talk is the highest indicator among the five indicators.Language talk is the second-highest result; content talk is the third high incidence; the two rests are the fourth and the fifth results; (c) an individual can strengthen the group's identity by positioning themselves in taking role more.Their identities can be assigned whether they resist, accept, initiate, bargain, or explain something related to the course; (d) regarding the student's responses, they agreed that the virtual collaborative strategy stimulates them to construct communication.The group member with lower communication may get benefit more from their tasks.They also agreed that students should have common purposes in completing writing tasks together.
Finally, the study has theoretical and practical implications.The theoretical implication of these findings has a beneficial effect on scrutinizing the writing learning field.The social presence in virtual collaboration helps to frame cohesion by identifying reciprocity and maintaining trust and solidarity in interaction.Participation helps to facilitate the negotiation and cognitive process through discussion.Identity theory helps to categorize and examine the moment-by-moment construction of identities by positioning themselves as initiators, acceptors, and explanatory; by increasing common purpose and self-directed learning.Therefore, individual identity is essential to strengthen teamwork in online learning.
In the practical implication, these findings show that small groups during the synchronous collaborative prewriting phase helps students improve critical thinking abilities in academic International Journal of Information and Education Technology, Vol. 13, No. 12, December 2023 writing course.The lecturers can observe not only knowledge inquiry but also observe learners' behavior, who are active regularly with others or not much active from the beginning to the ending discussion.Further, through this framework, the findings can give new insight and add research literature with different nuances.

TABLE II :
MEASUREMENT SCALE FOR INTERPRETATION OF THE

TABLE III :
FREQUENCY OF GROUPS' TALKS

TABLE V
SR: share resource, TN: task negotiation, CT: content OG: organization, LN: language

TABLE VII :
THE SCORE RESULTS OF STUDENTS' RESPONSES ON THREE MAIN PRINCIPLES