E-peer feedback training on L2 writing: A case study of inexperienced learners

Online peer feedback has been widely applied to teaching L2 writing because of its potential benefits. Accordingly, training is one of the prominent factors contributing to the effectiveness of the activity, especially for inexperienced learners. Acknowledging this significant issue, the current study aims to develop specific online peer feedback training for inexperienced learners based on their performance naturally during the online peer feedback process, followed by an assessment of its effectiveness. Using the qualitative method design that draws upon the sources of data, including students’ online written comments (Google Docs), oral feedback (Zoom recorded), observations, their drafts, and their reflections and responses through semi-structured interviews, this study presents an in-depth analysis of students’ performance toward peer feedback (without training), how the training is developed, and the impacts of training on the students’ practice. Finally, the study discusses pedagogical implications regarding the instructions and training for online peer feedback in teaching L2 writing, which can be applicably adapted for similar EFL or ESL contexts and learner backgrounds.


Introduction
Peer feedback is a pair or group activity in which students read, evaluate, and comment on each other's writing (Tsui & Ng, 2000;Zhang & Hyland, 2023).This collaborative activity has been widely studied in varied EFL and ESL contexts for decades and has been encouraged to be applied in teaching L2 writing (Vorobel & Kim, 2014;Yan & Tang, 2023).Notably, online peer feedback has become prevalent because of the technological advantages (Li & Hebert, 2023;Ma, 2022;Ting, 2023).Nonetheless, learners have been reluctant to do it due to their limited knowledge in L2, low writing proficiency, and inexperienced skills (Ahmed, 2021;Do, 2023a;Ferris & Hedgcock, 2014;Hyland, 2019;Irwin, 2019;Kim, 2022;Saeli & Cheng, 2021;Yu & Lee, 2016).Or students did not use techniques that would optimize their learning (Hubbard, 2013).At the same time, inadequate planning and understanding of students' difficulties associated with online peer response by instructors may lead students to prefer teacher feedback (Yu, 2019).Ahmed (2021) found that uncertainty (lack of confidence) and inexperienced skills were the main problems with peer feedback for ESL students in his study.To solve these problems, proper training that provides learners with the knowledge and skill base to do the task efficiently and effectively is suggested (Hubbard, 2013).Furthermore, instructors are advised to monitor the process of peer feedback and assist students over a period of time (Ahmed, 2021;Cheng et al., 2023;Hubbard, 2013;Lam, 2010;Min, 2005;Tuzi, 2004).
Indeed, instructions and training for peer feedback in L2 writing classrooms have been discussed by L2 researchers (Lam, 2010;Min, 2005).In particular, Min (2005) provided the four training steps for peer review, including clarifying the writer's intention (1), identifying the problem (2), explaining the nature of the problem (3), and making specific suggestions (4).The training significantly improved the peer review performance of EFL Taiwanese students as they provided longer and more specific feedback as opposed to short and vague feedback in prior training.Similarly, Lam (2010) found this training effective for his Cantonese-Hong Kong students.It is worth noting, however, that this training was developed and used for handwritten comments (pen-and-paper) and face-to-face interaction, which may be different from online peer feedback as it "is not a simple alternative to face-to-face feedback" (Guardado & Shi, 2007, p. 458).Thus, the training for e-feedback is considered necessary if teachers want to apply the mode of computer-mediated communication to peer response in teaching L2 writing.Since this concern has been rarely studied in L2 writing research (Casinto, 2023), this poor background needs to be addressed to enrich the body of literature on the training of online peer feedback.The most recent study conducted by Zang et al. (2023) was about online peer feedback; however, students' experience with this mode of feedback was not the focus of their study.It goes without saying, therefore, that this is the central motivation and necessary rationale for the current study to be conducted to address problems authentically emerging during the implementation of online peer feedback, which is helpful for developing proper training for target students.Such a study can yield useful implications that help language teachers, especially novice teachers, and students with practical knowledge about the application of online peer response activities to L2 writing, which could lead teachers to better instructions and students to effective collaborative work.The study aims to answer the three following research questions: 1. How do inexperienced learners perform the task of online

Context and participants
As a language teacher teaching in the Vietnamese context, where a learnercentered approach receives less attention (Do, 2023b), the researcher recognizes that the majority of students usually depend on teachers for learning and are passive in cooperative activities with friends, such as peer feedback.This could be the reason why language teachers hesitate to incorporate peer feedback activities into the instruction, which is also mentioned by Yu (2019) in the EFL Chinese context (Macau and Mainland China).In order to help EFL Vietnamese students become autonomous, independent, and competent writers, the researcher wants to integrate peer feedback into writing classrooms, where students discuss and interact with each other in a cooperative environment.The role of the teacher in this view is to assist students in expressing their own meaning and understanding the process of composing writing as a facilitator.As mentioned earlier, nevertheless, students are not familiar with the learner-fronted approach, which could challenge them to do the activity of peer feedback.Thus, training is considered significant.This project was not implemented in an actual classroom teaching context because the focus of this study is online peer feedback.The researcher wanted to explore how students do online peer review (electronic comments) rather than handwritten comments (pen-and-paper mode), as they might do in traditional classrooms.Autonomy is also a goal for this study when students can do it at home with their friends or whenever they want to switch their essays to each other for peer feedback after this project (lifelong learning).Therefore, the researcher invited volunteer EFL Vietnamese college students to join the project through Zoom.After sending out the invitations, two EFL Vietnamese students named Hanh and Sau (pseudonyms), aged 24 and 25, respectively, were willing to participate in the project because both wanted to gain experience with peer feedback and improve their writing abilities.
These two students have learned English for 10 years and three years at college, where they have been taught by different non-native language teachers (Vietnamese) (grammar, writing, and reading) and English native teachers (British) (speaking and listening skills).They are third-year college students majoring in English language at a private university located in Ho Chi Minh City, Vietnam.There are four required academic English writing courses, from paragraphs to essays, in the first two years of the English language program.Hence, these two students had finished taking all these writing courses in the English program before attending this project.They were unskilled learners in terms of providing e-feedback.As for English proficiency, Hanh is intermediate (IELTS 4.0), and Sau is high-intermediate (IELTS 5.0).Regarding the personal aspect, Sau seems to be a quiet person, and Hanh is a talker.When it comes to writing skills, Sau is better than Hanh in terms of organization (idea arrangement).Hanh has abundant ideas, but she has trouble arranging those ideas logically (this assessment is based on their first writing drafts).They are both not confident enough to speak English, especially Hanh.These students do not usually use 100% English in the English classrooms, except for some activities that require students to use English or teachers who are native.As a result, the researcher used Vietnamese to communicate with the students to avoid misunderstandings.

Data collection
According to Liu and Edwards (2018), students "should be given ample opportunities to familiarize themselves with not only what to comment on but also how to make comments" (p.114), the two issues about online peer feedback training have been called into questions: What should be covered or included in the instructions for online peer feedback?And how to provide e-feedback?It is critical to find out what technical skills and techniques students need, which arise during the implementation process, to help them know how to provide e-feedback appropriately (Casinto, 2023;Liu & Edwards, 2018).To do so, there are two rounds of data collection, which include students' electronic feedback on Google Docs, their revisions, their oral feedback on Zoom (recorded by students and sent to the researcher), the researcher's observations from recorded videos, and semi-structured interviews (questions provided in Appendix A).Tables 1 and 2 present these two rounds, followed by detailed descriptions.Of note is that a consent statement was obtained before the process of data collection started (week 1 of the first round).The whole process was conducted for almost three months (February 2023 to May 2023).In the first round, the students were asked to write a short essay (350-500 words) on this tourism writing topic: "Tourism has negative environmental impacts on tourist areas.What are these impacts?How can damaging impacts be lessened?"This writing topic was selected based on the students' favorites.They had one week to write it on Google Docs (week 2) and then were asked to provide written feedback (online comments) in the following week (week 3).
Then they shared feedback with each other and revised their drafts (week 4).
In this round, the researcher wanted to see how the two students performed in a natural way (without detailed teacher's guidance), so the researcher did not provide instructions except for the general rubric adapted from Weigle (2002), which includes three elements, namely content, organization, and language use (Appendix B).After having the students' essays and their feedback, the researcher looked at how they provided feedback to each other (e.g., what they commented on, how they provided feedback on Google Docs) and their revised essays (first and second drafts).Then the researcher listed questions and asked them in the interviews (week 5).The interviews were conducted in Vietnamese with the students individually.In the second round, the researcher developed the instructions (week 1) and provided training for e-peer feedback (the rubric was modified) (Appendix C) (week 2).The students in this round were asked to write another topic: "Some people claim that e-learning has many benefits that will replace face-to-face education soon.Others say that traditional education is irreplaceable".Again, this topic was chosen based on its familiarity with the students (week 3).Like in the first round, the students wrote the first draft on Google Docs and shared it with each other for online peer feedback (week 4).According to their performance in the first round, oral feedback was added after students received written feedback, which will be explained in the result section.Then the students revised their essays (week 5).Finally, the interview was the last step.Data collected from this round was used to answer the third research question: To what extent is the training helpful?

Data analysis
To answer the research questions, the students' feedback (written and oral), their drafts, observations, and interviews from the two rounds were analyzed.
In particular, students' comments from two rounds of feedback would be collected and counted for analysis.Their comments were coded as content, organization, and grammar.In addition, the researcher looked at how the students provided feedback (technical skills) on the Word files based on the elements of being visible and readable.Then the first drafts were compared to the second drafts (revisions) to examine whether the activity was effective.When it comes to observations, the researcher watched the students' oral comments (the time of video recorded by students was 39.25 min) and listed what they discussed with each other.After that, the researcher transcribed important information in Vietnamese and translated it into English.This is the supplemental data that supports the main data of the interviews, which will be presented next.
Regarding the information collected from the interviews, which were recorded by Zoom (Table 3), the researcher listened to the records, transcribed, and translated them from Vietnamese to English when appropriate.Students' responses and reflections toward the process of proving feedback and the impacts of online peer feedback from semi-structured interviews were analyzed in parallel with their performance of online feedback on the Word documents and their revisions.In other words, these sources of data were analyzed simultaneously so that the students' responses to the activity of online peer feedback would be compared and examined with their performance.In round one, for example, the students' reflections about their struggles with proving e-feedback would be examined via how they provided feedback to each other.Similar to the second round, benefits and concerns about online peer feedback were figured out (data was coded into specific themes based on students' responses, such as benefits and practical concerns).Additionally, as noted, there are some notes taken from observations watched from recorded videos while the students provided oral feedback (round 2) that were used for analysis to strengthen the credibility and validity of the study.The following sections present how the students experienced online peer feedback without training (1), how the instructions and training were developed (2), and the impacts of the instructions and training for online peer feedback (3).

How do inexperienced learners perform the task of online peer feedback without training?
Although the two students provided feedback to and helped each other on surface errors (Figures 1 and 2), the activity seemed unhelpful due to four reasons: technical skills (how to provide e-feedback), inexperienced peer response skills (what to provide feedback), engagement, and language use.
First and foremost, the students seemed to lack technical skills in providing online feedback, which was confirmed to have never been done before: "At first, I found it a bit difficult because this is my first time doing peer feedback on Google Docs.So I did not know how to do it.We have never done this activity before, just mainly received feedback from teachers" (Hanh).Indeed, the students provided feedback on the right margin of the file in a long paragraph, which was hard to see and read (Figures 1 and 2).In addition, Hanh seemed confused about where to revise her essay.The researcher reminded the two students to revise their papers on the next page and highlight the revisions in red.Since Sau followed the instruction, Hanh instead revised right in the margins of the file (Figure 3), which made it hard for the teacher (researcher) to read her revisions and compare her changes.Second, the students were uncertain about what to comment on.They provided general and vague feedback, which was considered unhelpful.On the one hand, Hanh provided general feedback, as Sau reflected: "She provided very little feedback, and her comments were very general, so my revision did not significantly differ from my first draft" (Sau).In Figure 1, Hanh wrote 99 words of comments, and most of them were positive and general comments.Sau, on the other hand, provided vague feedback, which confused Hanh.Accordingly, Hanh wanted to know where "ideas are not clear, ideas are logical but not enough, use some wrong sentence structures, wrong tenses, inaccurate, and repetitive words" from Sau's feedback (Figure 2).As a result, the second drafts of Hanh and Sau still contained grammatical errors and writing problems.To shed light on these arguments, Table 4 shows some examples of ineffective revisions between the two drafts of the two students.Some basic errors that remain the same between the two drafts of the students' work are underlined.

Students
Draft 1 Draft 2 Hanh "Nowadays, Tourism is a highly developed industry.Almost everyone in the world like to go to travel everywhere after a long time with work or study.This is a piece of good information for tourism, but the bad thing is the impact bad to the environment, which makes the environment seriously impacted.
So We need to have some measures to lessen damaging impacts." "Nowadays, Tourism is a highly developed industry.Almost everyone in the world likes to go to travel everywhere after a long time of hard work or study.This is a piece of good information for tourism, but after trips that the bad thing is throwing away trash leads to the impact bad on the environment, which makes the environment seriously impacted.So We need to have some measures to lessen damaging impacts.In this text, I will give some effects and solve this issue." Sau "In conclusion, rapid economic development tends to cause a bunch of negative influences on the environment nowadays.The tourism industry has been affecting tourist places for two reasons: air and water pollution and overexploitation of resources.However, I believe that If we implement the above solutions, It is that we can enhance the quality of tourist areas as soon as." "In conclusion, rapid economic development tends to cause a bunch of negative influences on the environment nowadays.The tourism industry has been affecting tourist places for two reasons: air and water pollution and overexploitation of resources.However, I believe that If we implement the above solutions, we can enhance the quality of tourist areas as soon as." Similar to the finding of Min (2005), EFL Taiwanese in his study were confused about how to comment even though they were provided a guidance sheet.In the same vein, Cantonese-speaking Hong Kong students in Lam's (2010) study acknowledged that vague feedback and short comments, such as "well done, need to improve grammar", were not helpful for their revisions.The Chinese students in Fan and Xu's (2020) shared the same feeling that they received general feedback from peers, which could not help them improve their writing.
One of the main reasons for this phenomenon is that inexperienced students are not sure about what to do while giving feedback (Liu & Edwards, 2018).As for the technical skills, the students did not take advantage of the useful tools in Microsoft Word while providing feedback.Since these tools facilitate students pointing out the writing problems or errors right on the edge of the file, the students in this study failed to do so, as shown in Figures 1 and 2. This problem has been emphasized by Hubbard (2013), who found that students did not use techniques that would optimize their learning.Third, both students used "she" instead of "you" to refer to their partner in their comments, which might reduce their efforts for cooperation to help each other improve their writing.This problem may be because the students were asked to share the file with the researcher for the observations, so they might have thought that this was the assignment rather than the learning activity between two of them (they did it for the teacher, not for their partner), and waited for the teacher's appraisal.The meaning and value of peer feedback may change if students have this thought when doing peer review.This could be the reason why they did not spend much time giving feedback to each other in this first round, which was evaluated as perfunctory.As mentioned, autonomy learning is one of the main goals of this activity in teaching L2 writing, so that students will be able to provide feedback to each other outside classrooms.Based on the performances of the two students, they seemed to be passive in this first round of feedback.
Lastly, language proficiency could be a potential problem, which inhibits the students, especially Hanh (lower level), from expressing her opinions or thoughts on Sau's essay in a comprehensive way.It could be seen in Figure 1 that Hanh had difficulty expressing her evaluations in English.In line with the result found by Yu and Lee (2014), EFL Chinese students in their study (a similar EFL learning context where students usually use L1 in the language classroom) felt more comfortable using L1 to provide feedback than L2.
Due to those problems and mistakes, the feedback in this round had little impact on their revisions (their revisions still contain the same errors and writing problems compared to their first drafts; see Table 4).It is clear that the students found it challenging to give feedback, even though they were provided with a writing rubric.This is similar to the claim of Lam (2010), based on his own classroom teaching, that his Hong Kong pupils (Cantonese speakers, using English only for academic studies) struggled to provide feedback.The findings support the statement of Liu and Edwards (2018) that students with negative experiences with peer review can decrease their level of engagement and minimize their productivity.To help the students better reap the effects of the online peer feedback activity, specific and proper training is developed and discussed in the next section.

What issues are necessarily involved in online peer feedback training for inexperienced learners?
Acknowledging those issues mentioned in the previous section, the researcher developed the instructions for online peer feedback and provided a 90-minute training for the students with the hope of helping them do the task appropriately and effectively in the second round.Adapting the training's steps developed by Min (2005), the researcher first provided guidance for what students should do for peer feedback, such as identify and explain writing problems or errors (content, organization, and language use), and provide suggestions.Then the researcher modeled how to provide e-feedback on Google Docs (e.g., margin comments and sandwich feedback).Based on the performances of the two students in the first round, additional information was added, which will be described in detail as follows: As for the technical skills, the students were advised to provide feedback in appropriate places.Accordingly, marginal comments are usually used for detailed comments about content, tense, grammar, sentences, or word choice that students want to point out right on the file (Ferris & Hedgcock, 2014).Meanwhile, general feedback should be added at the end of the paper to remind the writers of their main writing problems and encourage them with positive remarks.
Regarding the content of the feedback, the rubric should be revised to be more specific.Poor writers usually follow the rules and instructions given by the instructor (Yan & Tang, 2023), which could be the reason for the phenomenon that happened in the first round.The students followed the general rubric for giving feedback (Appendix B), which resulted in vague and general comments.In addition to being inexperienced with online peer feedback and at low levels, the students needed specific rubrics that emphasized what they should look at in the essay through detailed questions.For instance, students should analyze whether the content is relevant to the writing topic or not and whether the writer provides strong topic sentences for each paragraph (how does the topic sentence cover ideas of the paragraph?).Similar to organization, the students were advised to look at the connections between sentence and sentence, paragraph and paragraph to see whether they are well organized.In the same vein as grammar or language use, the students were advised to carefully consider grammatical errors such as tense, word choice, etc.They were suggested to double-check those comments by looking them up in dictionaries or researching on internet sites.In addition to gaining experience with vague and general feedback in the first round from both students, they were asked to provide extra oral feedback via Zoom after they had made electronic written comments (students might have questions after receiving online feedback or ask for suggestions).The detailed questions and oral feedback may challenge students to read for a thorough understanding, which may prevent them from providing general and vague comments.
When it comes to language use, the students were allowed to write comments in L1 or L2, whichever they felt preferable and confident with (if L2 prevented them from providing clear feedback, then L1 was allowed to be used).Lastly, peer feedback is a learning activity between the two students rather than an assignment that should be sent to the teacher for the purpose of grading.Hence, using "you" when students provided feedback was suggested, even if it was then shared with the teacher.As noted, this could increase learner autonomy in learning L2 writing.
To sum up, the guidance for peer feedback activities should be secured with the following rules, as emphasized by Hyland (2019) and Prilop et al. (2023): Be respectful, be conscientious, be legible, be encouraging, be formulated, and be specific.These are necessary for students who are inexperienced in doing peer feedback (Liu & Edwards, 2018), especially in the format of online feedback.Those notes were thoroughly provided, explained, and modeled in the training via Zoom before the students attended the second round of writing and peer feedback.Details of the training, guidance, and revised writing rubric are provided in Appendix C.

To what extent is online peer feedback training helpful?
"I am now confident to provide feedback".It is clear to recognize that the two students performed better than in the first round, as they provided more comments in the second round, which echoes Min's (2005) findings on EFL Taiwanese learners.Importantly, the students in this study avoided the previous four problems in the first round of peer feedback without training.These findings shed light on the statement of Liu and Edwards (2018) that "if students are trained and given guiding support, the interaction in the peer response group is useful, and the comments generated can be constructive in helping students revise their drafts" (p.40).The following paragraphs describe these positive outcomes in more depth based on the performances of the two focal EFL Vietnamese students.
First, the students provided detailed and understandable feedback on the margin of the paper (Figure 4).Students looked at their partner's essay in detail to help each other find writing problems and errors that had not received much attention in the first round.Surprisingly, Hanh (lower level) figured out an inappropriate comma in this sentence of Sau's essay and reminded her of this error: "Students often do not want to answer the questions teachers ask, and do not desire to share their ideas".Hanh shared: "I first looked at her paper in a general way to know her ideas, then I moved to specific areas such as sentences, words, grammar, and punctuation".In contrast to the "general" feedback in the first round, Hanh's feedback seemed more precise and explicit in this second round.In the same vein, Sau found some mistakes when she looked closely at Hanh's essay.She advised Hanh, for example, to use linking words like 'when' to replace 'that' in this sentence: "That studying at home, we always feel distracted by food", and delete "so" in this sentence: "Some people claim that e-learning has so many benefits that will replace education soon".Sau's comments became more direct and clearer, which were different from her "vague" feedback in the first round.Sau shared: I just read Hanh's essay generally in the first round, so I did not provide clear feedback.Based on the instructions, I read her essay carefully in this second round.I first read the whole essay, and then read paragraphs one by one, looking at sentence by sentence, word by word.By doing so, it is easy to see her writing strengths and weaknesses, and I can provide more feedback.(Sau) These experiences support the statement of Ferris and Hedgcock (2014) that marginal comments force learners to read and provide feedback carefully and specifically.Additionally, the training is significant for students to provide helpful feedback, as Taiwanese students in Min's (2005) study said that they (peers) focused on not only grammar but also ideas and organization while doing peer review after the training and having a conference with their teacher.In other words, the support and specific guidance of the teacher are necessary for students, especially inexperienced students.Second, the students effectively provided general feedback at the end of the essay.Accordingly, they knew how to provide "healthy feedback" to each other by writing some notes at the end of the paper reminding them of some errors and supporting their partner by mentioning some positive feedback on what they had done well, which encouraged their friend to sustain their commitment to revising and practicing writing (Figure 5).As a result, both students did well on this point, as Hanh shared: I know how to support and motivate my friend by providing some praise about her writing strengths.Indeed, she motivated me when she complimented me on the fact that my essay was better than the previous one, my ideas seemed relevant to the topic.These comments motivated me to continue revising the essay.(Hanh) Third and fourth, the students used "you" (Figure 5) and languages (L1 and L2) (Figure 6) appropriately to deliver feedback.When the students were trained to provide feedback, they became more comfortable and confident providing comments to each other than in the first round.Looking at this comment from Sau: "You should use 'think' instead of 'thought'.Using past perfect tense is not reasonable" in Hanh's essay: "In the modern world, society is growing, many people thought that digital education via the internet had become popular".Although this sounds direct, Sau seemed confident to use "you" to refer to Hanh, which increased her engagement with peer feedback.The students provided longer comments and suggestions for the writer.This positive change is also mentioned by Min (2005) with EFL Taiwanese students after the training.
Regarding the choice of language used for delivering feedback, Sau continued using L2, but Hanh chose L1.This permission helped Hanh (lower level) provide feedback easily and effectively in a comprehensive way, better than using L2.Similar to the EFL Chinese learners in the studies of Yu andLee (2014, 2016), students used L1 to provide feedback on the content and organization of classmates' writing and argue with each other rather than L2.Consequently, L1 is considered helpful, as it mediates students providing feedback effectively in EFL learning contexts.Enhancing the problems raised in the first round, their feedback is helpful to each other in specific ways, such as word choice, grammar, language use, and idea development (mentioned above) (Appendix D shows more examples of this improvement).Furthermore, students tended to recognize their own mistakes in their essays when doing peer review and gained the skills of analyzing and revising their own writing.Berggren (2015), Lam (2010), and Yu (2019) also found this significant outcome that students become conscientious writers and reinforce their audience awareness after critiquing their peers' essays.
I recognized that I did the same thing as Hanh when I commented on her essay that she wrote many single sentences in one paragraph and suggested to her that she might write some compound or complex sentences to make the paragraph more advanced.Also, I have the opportunity to re-learn previous knowledge as well as accumulate new knowledge to rewrite my essay better (for example, I have to check the structure or look up words in the Oxford dictionary to give feedback).From this experience, I understand the two-way benefits of peer feedback.I spent time providing feedback for her, and at the same time, I learned for myself.(Sau) Likewise, Hanh did not know that "on the one hand" could be used at the beginning of the paragraph when she commented on Sau's paper that "you should not use this phrase here".She realized this mistake: "I learn some new knowledge via peer feedback activities like the use of on the one hand, on the other hand, linking verbs, and tenses" when Sau (higher level) explained to her why she used this phrase through oral comments.This reflection also illustrated the benefit of oral comments (Zoom) after the online written feedback, which gave the students another chance to explain what they were not sure or confused about.The necessity of oral feedback was also shared by Sau: When I give feedback, I do not know if she understands what I mean, and sometimes I am afraid that she misunderstands my comments.Thus, an additional oral feedback task helps us understand what we mean in our comments.(Sau) Hence, it could be helpful for students if the oral feedback task was added right after written peer feedback, as Liu and Edwards (2018) indicated that peer response is more beneficial and optimized if more than one mode of peer response is used.By doing so, students can negotiate with peers about revisions to gain a deeper understanding of peer comments (oral feedback) rather than just accepting or rejecting them with confusion (after written feedback).
Generally speaking, the rubric with guiding questions and the training exert their effectiveness in helping students provide helpful feedback, reducing apprehension, and gaining their confidence: "The important thing is that I realize I am now confident to provide feedback, which I previously thought I could not do" (Sau's reflection).This is a meaningful reflection since L2 students, especially those at low and intermediate levels, are usually not confident to provide feedback due to their insufficient linguistic knowledge and previous peer feedback experience (Zhang & Hyland, 2023).This statement may encourage other inexperienced and low-level students to actively participate in online peer feedback in learning L2 writing.
The intervention of the teacher for better revisions.As mentioned, the students' revised papers were better in terms of grammar, word choice, and ideas, which results from the comments on these writing problems.Another important point, nevertheless, that attracted the teacher's attention (the researcher) is their comments on organization.Indeed, the students had comments on the organization of ideas, but their comments on this issue were sketchy, and they seemed not to revise effectively in this sub-writing skill.This phenomenon was also found with EFL Chinese students in Fan and Xu's (2020) and Yan and Tang's (2023) studies, that students provided form-focused feedback by a large percentage.Based on the observations and interviews, this problem does not lie in the students' engagement; it lies in their writing abilities, their knowledge of L2, and previous experience of teacher feedback, which were also mentioned in Do and Phan (2023).Sau indicated that: I think I can now provide feedback, but my knowledge still has limitations in a certain way.We still need the teacher to help us with what we are unable to recognize or provide suggestions.(Sau) This is a significant finding that helped the teacher understand the students' knowledge of their L2 writing abilities, which was illustrated through the feedback process.In other words, it helps the teacher determine which specific areas students are able to do and are unable to go further for feedback.Then the teacher will assist them in a timely manner as a facilitator.The appropriate time for the teacher to be involved in the process of assisting pupils could be right after their written or oral feedback.Conferences or meetings between students and the teacher at this time could be necessary.Since both students were reluctant to provide comments on this sub-writing skill, the teacher's clear explanation may help them revise their essays more appropriately.
Furthermore, the improvements were limited to the flow of the sentences, which could be due to the negative transfer from L1.Some sentences were not written smoothly because the students applied the structure of L1 writing to L2.This can be seen clearly in the case of Hanh (lower level) when she translated from Vietnamese to English to express her ideas, so her ideas were not conveyed smoothly: "People believe that e-learning has many benefits to help teachers and students a lot.Firstly, studying via the internet helps us have flexible time, or in other words, we can learn on our own time without the following schedule" (revised essay).Less than Hanh, Sau still had this problem: "...this leads learners to gradually become more passive in social interaction as well as in themselves developing".Although peer response is effective, students' writing abilities and knowledge have limitations in certain ways, especially those with low or intermediate levels of proficiency and inexperienced writers.Therefore, teachers are advised to follow up on the peer review process to help them at the right time.The role of the teacher in this situation is to provide suggestions and explanations for which students are unable to help each other.Teachers could find this helpful, as they know when they should be involved to help learners as a necessity and monitor the online peer feedback process in a large class effectively.

Conclusion
Based on the in-depth analyses of the performances of the two students in this study regarding online peer feedback, the findings show that the students tend to provide helpful feedback and enjoy working together if they are given explicit instructions with guiding questions and training.This attention and preparation may prevent students from giving vague and general comments to peers.This practical knowledge would be helpful for other practitioners in teaching L2 writing regarding online peer feedback.In other words, providing writing rubrics and asking students to do online peer feedback do not always guarantee a way of improving learning outcomes; the instructions should be developed based on students' experiences and challenges regarding peer review activities, then model how to provide feedback through training.Once students have the skills to provide feedback and understand the effectiveness of peer feedback, they will be able to know how to cooperate with the activity affectively, behaviorally, and cognitively.Otherwise, the activity could be predicted to be worthless and time-consuming.This seems to be understandable; nonetheless, not many teachers have adequate planning and fully understand students' difficulties associated with the activity of providing peer feedback (Ahmed, 2021), which could lead to ineffective activity.In addition, following up on peer review sessions to provide feedback or assistance appropriately at the right time is necessary, as Ferris and Hedgcock (2014) emphasized that "teachers can structure peer review activities in numerous ways, which should differ depending on the stage of the composition process" (p.259).
Because of the small number of participants and the limited proficiencies of the students, the generalization of this study should be considered with caution.These issues therefore open the door for future researchers to enhance the generalizability of research findings in different contexts and with different sizes of participants.Furthermore, the formative evaluation of the training can be considered by future researchers as they will be able to modify the instruction as a continual process (more feedback rounds will be explored).Despite these limitations, the experiences of online peer feedback in L2 writing from Hanh and Sau could be an example to encourage L2 students (in similar learning backgrounds) to engage with this cooperative learning activity and help them know how to provide feedback appropriately, which could prevent them from having the same problems as the students did in the first round.For writing teachers, those experiences of the students also encourage them to explore their students' struggles with online peer feedback, which could help them tailor suitable instructions and careful preparation for their target students (each student may have different challenges or problems).Additionally, teachers may consider some ideas in the instructions for online peer feedback in this study (Appendix C) for their guidance with modifications and simplifications regarding affective issues, cognitive strategies, sociocultural competencies, and linguistic issues based on the needs of students and their backgrounds, as Hyland (2019) stated that "flexibility is an essential element of all planning and delivery" (p.80).Generally speaking, peer feedback is a useful task for L2 writers; nevertheless, it should be used in conjunction with an array of conditions or considerations that language teachers may be aware of before introducing it to their writing classes.

How to provide e-feedback
► Write detailed feedback in the margins of the paper if possible (e.g., tense, grammar, sentences, content, or word choice).► Write general feedback if possible at the end of the essay to remind the writer of some important points of writing."Sandwich" feedback is recommended.Accordingly, you will write encouraging remarks at the beginning and end of the note and point out some critical feedback in the middle.
Note* ► These processes should be modeled on a text written by someone unknown to students to let them know how to give comments on the Google Docs (e.g., give comments in the margins of the essay and general comments with a balance between praise and constructive comments).► Students are welcome to ask friends or the teacher if they have any questions or struggle with providing e-feedback during the process of peer feedback.► Regarding what to comment on (e.g., the content, organization, and language use), students can consider the following specific guiding questions.

What to comment on
Students are advised to focus on the content first to understand the essay, followed by organization and language use.Provide general and detailed feedback and suggestions, if possible.

Content Questions
Introduction: Address the writing task effectively.
-Does the writer address the writing task effectively?>> Is the content relevant to the writing topic?If not, provide explanations or suggestions, if possible.
Body: Address clear ideas or explanations.
-Does the writer provide strong arguments (topic sentences) for each body paragraph?-Does the writer provide specific facts or examples to support his or her argument (topic sentence) for each paragraph?Are they relevant to the statement or argument?
Note: A topic sentence tells the reader the main idea or thought that the writer is trying to express.It is a one-sentence summary of the entire paragraph.Each sentence that follows helps to develop the idea presented in the topic sentence.
Conclusion: Summarize the ideas.
-Does the writer summarize the ideas that are relevant to the ideas presented in the introduction and body parts?Why or why not?

Figure 4 .
Figure 4. Sau provided detailed feedback on the margin of Hanh's essay

Figure 5 .Figure 6 .
Figure 5. Sau provided general feedback at the end of Hanh's essay

Table 2 .
The process of round two

Table 3 .
Information of interviews