Towards Understanding Teacher Mentoring, Learner WCF Beliefs, and Learner Revision Practices Through Peer Review Feedback: A Sociocultural Perspective

Background. The existing literature has focused on learner perceptions or beliefs about peer review tasks over the recent decade. However, little has been known about the relationships among learner beliefs about written corrective feedback (WCF), related teacher mentoring process, and learner revision practices. Purpose. We thus aimed at addressing the gap by exploring how teacher mentoring and learner WCF beliefs may inform learner revision practices in the peer-reviewed process. Methods. We included four Chinese EFL students majoring in English as the participants and collected their WCF belief survey data. We also collected their actual practice data through PeerCeptiv , an online writing and rewriting platform. In addition, we traced the teacher mentoring practices and interviewed the participants about their beliefs and practices in the peer review and back-evaluation process. Results. Through the mixed-methods design, we reported our major findings: the student participants believed empathy and resonance was the primary advantage of peer feedback, and teacher mentoring facilitated them in understanding and performing the peer review and revision tasks; we also found the student review process consisted of evaluating, resonating, learning, and reflecting practices and the student revision process included crediting, arguing, correcting, and polishing practices. Implications. From a sociocultural perspective, we centered our discussion on these research findings by claiming that scaffolding in different forms work together enhance learner performance, and student beliefs appear in a complex manner with student actual revision practices. We also offered insights for future studies and practical implications for language teachers.


INTRODUCTION
Considerable studies have revealed the merits of peer feedback in learner writing performance (Hyland, 2003;Liu & Hansen, 2002;Sato, 2016). Specifically, peer review is an important instructional activity to raise learners' awareness of their strengths and weaknesses, and promote collaborative learning (Chang, 2012). If students master peer review skills, they can figure out peers' writing issues and also improve their own writing skills (Chaktsiris & Southworth, 2019;Law & Baer, 2020;Woodhouse & Wood, 2022). Peer review research has thus attracted much attention in literacy or writing education (Cho & Schunn, 2007), and second language writing education (Bui & Kong, 2019;Sánchez-Naranjo, 2019). However, Kim and Mostafa (2021) reported: "Within the research domain of | Research Article perceptions of written CF (corrective feedback), the examination of learners' perspectives is the least explored area." (p. 574, cf., Nassaji & Kartchava, 2021). In fact, learner beliefs about feedback activities can significantly affect how they engage in the process of giving and receiving feedback (Yu & Hu, 2017). With the perceived research gap, studies on learner WCF beliefs require scholars' attention and efforts.
In addition, studies report that teaching mentoring may inform learners' beliefs and peer review performance. Sato (2013) found that learners' active responses and willingness to give peer feedback had been greatly improved after receiving CF training. He also found teacher corrective feedback training had enhanced students' confidence in giving feedback. In fact, teacher mentoring is of great importance to both teachers and students, as divergences or inconsistencies between teachers' intentions and learners' interpretation of those intentions may result in negative effects on learning (Kim & Han, 2007;Lasagabaster & Sierra, 2005). However, there is scant research on how teacher training or mentoring assist learners to re/form their beliefs, improve their willingness to give peer back, and improve their revision practices. With all these considerations, the study attempts to address the gap by further explore EFL learners' WCF beliefs and unpack how their WCF beliefs together with teacher mentoring may inform their revision practices through peer review tasks.

LITERATURE REVIEW Learner Beliefs about WCF
Learner beliefs are defined as learners' metacognitive knowledge about learning (Wenden, 1999). Learner beliefs serve as a complex learner characteristic that greatly impacts the second language (L2) learning process (Han, 2017). The way that learners go through the task of learning is one socio-cognitive factor that determines their journey of the language learning process (Sato & Storch, 2022). While some learner beliefs are unique among individuals, certain beliefs appear to be shared in common (Campbell et. al., 1993). In second language acquisition, learner beliefs have been studied in a socio-cognitive orientation, emphasizing how learner beliefs were influenced and shaped by many factors including past learning experience, cultural background, and social and political contexts of language learning ( (Barcelos, 2003;Yasmin, 2021).
Most recent studies in the field have yielded findings on how psychological factors could shape learner beliefs. Those factors include motivation (Wang & Zhan, 2020), anxiety (Rahimi & Zhang, 2019), self-regulation (Cho, Yough & Levesque-Bristol, 2020), and learner autonomy (Yasmin & Sohail, 2018). A few other studies also explored how gender (Iwaniec, 2019), language proficiency (Wong, 2020), and strategy use (Tang & Tian, 2015) may also influence or inform learner beliefs. In addition, recent studies show that certain beliefs are common among learners, teachers, target languages, cultures, instructional settings, and age groups (Aslan & Thompson, 2021). Dörnyei and Ryan (2015) highlighted research about learner beliefs can help to explain factors behind learners' motivation and aptitude, so to understand learner belief is important for teachers and learners.
Learner beliefs about WCF have yielded some key findings. For example, Chen, Nassaji, and Liu (2016) found that learners overall expressed a favorable attitude toward error corrections and comments, especially feedback on the content and the organization of their written assignments. Their finding was consistent with studies on WCF (e.g., Amrhein & Nassaji 2010;Ashwell 2000;Brown 2009;Karim & Nassaji 2015;Lee 2008;Montgomery & Baker 2007;Schulz 2001) that show the acceptance of WCF in both ESL and EFL contexts. Chen, Nassaji, and Liu (2016) also found that the students liked explicit feedback and direct correction over indirect correction. This was in line with reported statistical comparisons of the learners' performance, which showed a clear advantage for explicit feedback over implicit feedback for the delayed imitation and grammaticality judgment (Ellis, Loewen, & Erlam, 2006). Moreover, Kong and Teng (2020) found that self-efficacy plays an important role in the peer-review process for L2 young learners. To be exact, those with high self-efficacy could follow the instructions according to the training session and learned a lot from the the peer reviews. However, students with low self-efficacy ignored the guidelines and promoted their skepticism of peer review in the end.
In addition, the existing literature yields findings on learner perceptions of different types of feedback. In general, students preferred teacher feedback due to its reliability (Abedi et al., 2015;Ertmer et al., 2007), and believed that peer-feedback is only effective in a friendly and co-operative environment (Kavaliauskiene & Anusiene, 2012). Similarly, Zhang and Rahimi (2014) investigated teacher, peer and self-correction feedbacks and showed that students value teachers' Rollinson (2005) reported peer feedback is less authoritarian and more informal, which may encourage and motivate learners to write and revise. However, peer feedback may also be lengthy, student personality-oriented and requires teachers to give direction or organize the peer feedback tasks.
To sum up, the existing literature on learner WCF beliefs derives from studies focusing on how students and teachers perceive the WCF effects (e.g., Amrhein & Nassaji 2010;Brown 2009;Chen, Nassaji, Liu, 2017;Diab 2005;Karim and Nassaji 2015;Lee 2008;Montgomery and Baker 2007;Simard et al. 2015) and also those on the comparison among different sources or types of WCF (Ertmer et al., 2007;Van den Boom, Paas, & Van Merriënboer, 2007). However, there is still a dearth of literature extending this line of inquiry. One of the reasons might be WCF is complex in nature, as | Research Article it includes different forms of interactions among tasks, individuals and writing texts (Bitchener & Ferris, 2012;Chen, Nassaji, & Liu, 2017). Therefore, further studies examining learner WCF perceptions and beliefs in the peer review process are highly needed.

Teacher Modeling in Student Peer-Reviewed Process
The importance of teacher modeling or mentoring has long been discussed in the existing literature.
Van Steendam et al. (2010) argued that instructional methods comprising modeling, practice, and feedback are ''very powerful'' and can help learners ''acquire new, complex cognitive skills, such as writing, revision and learning the different steps in a strategy'' (p. 318). Despite the persuasiveness of this argument, many scholars have focused on the effects of whole training programs, and only a few of them have explored the influences of these methods on the classroom-based peer review training. For example, Berg (1999) examined the influence of 11 peer-review training activities on peer revision and revision quality and found that the trained group significantly outperformed the untrained group in revising for meaning and improving the quality of texts. Hu (2005) reported (un)successful experiences through trial and error in a three-year peer review training program in his action research. Min (2006) investigated the effect of in-class teacher modeling and after-class individual teacher-student conference and pointed out that a positive peer review training has an influence on student writers' revision types and quality. Liou andPeng (2009), Lam (2010), and Rahimi (2013), adopting and adapting Min's (2006) peer review training procedure, explored the effect of training on Taiwanese university students' comments through web blogs, Hong Kong university students' comments, and Iranian university students' comments in traditional writing classrooms, respectively, reported similar successful peer reviewing training effects on peer reviewers' comment focus and quality. In conclusion, instructional methods play an important role in the peer-reviewed process and more explorations are needed to better understand this process. However, most of the existing literature focuses on the primary, teacher-led training rather than the peer review training or mentoring process. As the peer review process is inherently a constructivist process that follows a learner-centered philosophy, we argue that studying how teaching mentoring as a facilitating factor is necessary.

Theoretical Framework: A Sociocultural Perspective
A fit theoretical framework is highly useful to examine constructs that we propose in the study. We looked for a framework that may help us depict a general picture of teacher mentoring, learner WCF beliefs, peer review process. We then found the Vygotskian theoretical framework of socio-cultural theory (SCT) a suitable justification to explain relations among our proposed constructs. An SCT framework argued that cognitive development, a result of social interaction, can improve individual learners' competence under the guidance of a more experienced individual as a way to advance their zone of proximal development or ZPD (Liu & Hansen, 2005). The premise offers two-folded insights to study teacher mentoring and peer review process. One the one hand, teachers or instructors as experts in the writing classroom are more experienced individuals that guide and direct the students in the writing and revision process. On the other hand, learning and knowledge construction are mediated through interaction with others (Doolittle, 1997). Students who engage in collaboration during peer feedback sessions can negotiate meaning and construct their understanding of language mechanics, or local aspect, and discursive features, or global aspect (Mao & Lee, 2022).
In addition, previous studies have identified peer collaboration as a useful approach to give a strong boost in their foreign language development through interaction (Donato, 2004;Lantolf, 2000;Lantolf & Thorne, 2006). One significant finding is that collaboration among peers "allows students to use language to mediate their language learning because in collaboration students use language to reflect on the language they are learning" (Shrum & Glisan, 2005, p. 25). In second language writing, Hu (2005) argued that a collaborative activity involved "students reading, critiquing, and providing feedback on each other's writing, both to secure immediate textual improvement and to develop, over time, stronger writing competence via mutual scaffolding" (pp. 321-322).
Scholars have long acclaimed that social interaction and negotiation of meaning lay the foundation for the construction of knowledge (Doolittle & Hicks, 2003;Lantolf, 2000;Lantolf & Thorne, 2006;Vygotsky, 1978). The process of writing, giving feedback, and revising is a typical embodiment of social interactions among teachers, peers, and learners themselves. The embodiment represents an interactive process that "a more knowledgeable 'other' structures the learning experience in a way that allows the novice to overcome whatever limitations in skill might impede his or her attainment of a desired goal" (Prawat, 1996, p. 217).
Another insight that we got from an SCT perspective for the current study is that the importance of this social mediation is situated in authentic environments and tasks where the individual can interact with others and thus becoming "self-regulated, self-mediated, and self-aware through feedback received from the environment and self-reflection on their understanding and experience" (Doolittle & Hicks, 2003). Writing, reviewing, and revising process simply reflects the insight. Yu (2020) used the framework of Vygotskian sociocultural theory to investigate tutor-tutee interactions carried out in a series of tutoring sessions and the tutee's subsequent revisions to her writing drafts over the | Research Article course of a semester. Leontjev and Pollari (2022) applied the Vygotskian sociocultural theory to explore how peers' comments can inform teacher assessment and guidance of second language learners' writing in authentic environments. In conclusion, an SCT perspective is a fit theoretical framework to explore teacher mentoring, learner beliefs, and learner practices in peer review tasks that we attempt to focus on in the current study.

Research Aim & Questions
With the perceived research gap and spotted theoretical framework in the section above, we presented our research aim and questions and our rationale of the research design in this section. We also introduced our research site, participants and data collection process as well. Generally, we aimed at exploring Chinese EFL learners' feedback beliefs and how these beliefs inform their actual revision practices through peer review tasks. In the process, we also attempted to see how teacher mentoring works in mediating Chinese EFL learners' beliefs and practices. With the research aim and purposes, we set up the following research questions: 1. What is the role of teaching mentoring in the WCF? 2. What WCF beliefs do Chinese EFL learners hold?
3. How do WCF beliefs together with teacher mentoring inform revision practices?

Research Context & Participants
We carried out the study in the Chinese EFL context, where English is often taught as a required course for students in colleges or universities for different purposes. The research site for the study was supposed to be conducted through a face-to-face medium in a research-based university in a coastal city. However, due to the pandemic, the study was largely carried out through an online writing and rewriting system.

Student Participants
We included four Chinese EFL learners who majored in English in the research site, but they were in different classes or grades. As typical EFLs in China, they all came to this university through GaoKao or the Entrance Exam to College. Their average age was about 21 and overall language proficiency was about advanced level, given their entrance exam score for the English subject and their years of academic training.
As English major students, most of their courses were delivered through the English medium instruction by instructors most in applied linguistics or TESOL backgrounds. Table 1 presents biographical information of the four participants. For the participants' privacy and ethical consideration, we also provided them with pseudonyms.

Instructor
Instructor for the study graduated with a doctorate degree in TESOL, and his research interests included language teacher beliefs and practices and TESOL methodology. When we conducted the study, he had been teaching the English language courses, linguistics courses, and language teaching pedagogy courses for more than ten years.

Research Method and Design
We adopted a mixed-method design to combine both quantitative methods and qualitative methods. The rationale for choosing such a design fit the research aim and purpose.
As we attempted to explore language teacher mentoring practice and learner review/revision practices, we believed qualitative methods including observation, interviews, and inductive analysis met with the aim and purposes. In addition, our exploration of language learner beliefs about WCF would be feasible through surveys or questionnaires and interviews, entailing the necessity of using quantitative research instruments. We depicted our design in the following figure (see Figure 1).

Research Instruments & Measurements
As the study included a mixed-method design, we used a survey as a research tool to solicit language learner beliefs about WCF (see Appendix I). We adapted the survey that  et al., 2016), entailing its validity and reliability. Specifically, the survey included question items requesting the participants' biographic information, beliefs about writing, and different forms of WCF beliefs. While we used the survey to collect multiple sources of corrective feedback from the learners, we focused primarily on peer feedback by using teacher feedback data to compare the data on peer feedback. Therefore, the use of the survey was not for finding generalization purposes as typically revealed through a quantitative design; instead, we used the survey as a pilot study or prerequisite for our interviews. We got a general picture of the participants' beliefs through the survey results and then centered around some results for discussion in the interviews through data triangulation.
We then used interviews and observation methods to trace the instructor and the participants during the mentoring, review and revision process through PeerCeptiv, formerly SWoRD (Scaffolded writing and reviewing in the discipline). PeerCeptiv is an online peer assessment platform (Cho & Schunn 2007;Schunn 2016), with many features resembling features in other intelligent writing systems including but are not limited to Expertiza (Gehringer 2010), EduPCR (Wang et al. 2016), and PigaiNet (Wang, 2022). As its name indicates, the primary feature of PeerCeptiv is the systemic peer-review process that gets students involved in the writing and revision tasks (Lorretto, DeMartino, & Godley, 2016;Ruegg, 2017). Existing literature proves interviews as useful tools to find out learners' perceptions of different feedback practices (Hedgcock & Lefkowitz, 1996;Lee, 2008;Mendonca & Johnson, 1994;Tsui & Ng, 2000;Yang, Badger & Yu, 2006).

Data Collection and Analysis
We collected and analyzed data according to the mixed methods scheduled in the study. We collected the learner WCF beliefs through the survey and documented the instructor and the learners' practices through PeerCeptiv and interviews. Documentation of the teacher and the learner practices included collected texts, messages, and posts through their social media software (i.e., WeChat) and threads created in PeerCeptiv platform. Learner practices occurred in four steps: writing, peer reviews, back evaluation (of the peer reviews), and revision. Specifically, tasks were distributed and assigned in a continuum which starts with a writing task for the participated students, a peer review task that required three peers to give their specific feedback, a back evaluation task that asked for the learner beliefs about the given peer reviews, and then a revision task that helped them polish the writing. In the study, the task continuum appeared twice within one month.
In the interview stage which occurred after the task continuum, we interviewed the four sampled participants, transcribed the interviews, and then analyzed the transcripts. Using the items in the survey and documented data from PeerCeptiv, we guided the learner participants to provide rich information for their responses about their WCF beliefs, peer review tasks, and revision practices.
In terms of the data analysis, we firstly reviewed four participants' answers and compiled a profile of each participant to track their belief during the process. Secondly, we conducted a cross-case comparison to compare and integrate the findings generated from each case in order to form a deeper and fuller understanding of the participants' belief in the program. To ensure the validity and reliability of our research, we returned to our interviewees to further check their answers. The whole process was conducted in Chinese as all our participants were Chinese students and they can better express their thoughts in Chinese.
Ethical considerations were seriously taken into consideration when conducting this empirical study. Specifically, we informed the participants of the research objectives and ensured them of the confidentiality. We also let them know the participation was totally voluntary and they might leave anytime in the process. We kept all the data and transcripts confidential and acknowledged the participants' time and contribution to the study.

Figure 1
Research Design for the Study

RESULTS
As we aim at exploring how teacher mentoring and learner WCF beliefs may inform learner revision practices, we frame this section in three primary parts, that is, we report our findings and analyses on teacher mentoring and involvement in the peer review process first, and then we describe our findings of learner WCF beliefs including learners' overall WCF beliefs, and learners' specific beliefs about teacher feedback, peer feedback, and the PeerCeptiv platform. We finally report our findings of learners' review and revision process.

Teacher mentoring and involvement in the peer review process
We explored how the instructor participants designed, implemented, and evaluated the whole process through a mentoring rather than an intervention process. The instructor got involved in the peer review process through five stages, each of which serves a different function. The five stages included session & tech training, Q&A, task reminders, task completion monitoring, and the stage of encouragement, compliment, & praise (see Figure 2).
Specifically, the instructor scheduled an online training session due to the pandemic for the students involved in the process. He demonstrated how to use PeerCeptiv platform and the timeline of completing the writings, reviews, and revisions. He set up a social media chat group through We-Chat and left the students to pose questions there. Then, he checked the chat group on a regular basis, responding to the students queries about technology, tasks, and deadlines for different tasks. He also sent reminders typically prior to the deadline of each task, pushing the students towards their writing and review completion. After each deadline, the instructor checked the students' completion of these tasks. Then, the instructor credited the students for their works and kept encouraging them to move forward in the process.
All the four participants reported through their interviews that the whole teacher mentoring process greatly facilitated their writing, reviewing, and revising process. CF and QL typically reported that the teacher mentoring saved them from missing the deadlines for their tasks, especially in the second round of review tasks when they were preoccupied with other stuff or assignments. FJ also credited the instructor for the demo he made at the beginning of the process, giving her a clear picture of the whole process. QL reported that she was somewhat concerned with the mentoring at first, as she thought there might be some more work to do through attending the training session and checking the reminders. However, she ended up with acknowledging the mentoring process which facilitated her to complete all the tasks successfully and satisfactorily. HY found that the teacher mentoring process actually left her great power in sticking to the process, as she argued the instructor had already set up a good example for her to deliberate his endeavor and stamina in the process.
It is worth mentioning that the instructor worked as a facilitator and guide rather than an interventionist or lecturer in the teacher mentoring process. He organized the process in a constructivist way to enable students to analyze and complete the tasks. By doing so, the process left the peer review tasks and peer feedback as the primary constructs that we focused on, meeting with the research aim and purposes. In addition, some positive psychology tenets (Seligman, 1990) appeared in the mentoring stage. For example, the incorporation of compliments, encouragements and praises into the mentoring process saved the students from concerns and fears and stimulated them to move forward in the learning process (Seligman, 1990;Jordan & Sorell, 2019).

Figure 2
Teacher Mentoring in the Writing, Review, and Revision Process The finding can also be revealed through their surveyed question item when being asked what they expect the teacher or instructor to do for their feedback. Three of the four participants responded that they expected the teacher or instructor to correct major but not minor errors in their writing. However, even for the participants who did not hold a strong preference of one single type over the other, they also believed, teacher feedback might be their choice if only given one type of feedback.

Beliefs about Teacher Feedback
All the four participants through their interviews reported that teacher feedback was quite useful in pointing out logical, coherent, or organizational issues in their writings, which served as the first primary feature of teacher feedback. The participants understood teachers might be busy with teaching, research, or service work, so most of the participants expected teachers to give them suggestions from macro instead micro perspectives to save teachers' time.
For example, HY reported: Another feature of teacher feedback is the professionalism.
The participants believed in their instructors and hardly challenged any teacher feedback. For example, FJ argued that: "I prefer teacher feedback for its authority and professionalism. I sometimes might challenge my peers' feedback, because I really don't know his/her language proficiency." This kind of professionalism may provide the participants with a sense of security. QL, for example, reported: "The teacher's review makes me feel more secure (than my peers'), and I feel that this (issue or error) really needs to be revised or corrected".
Together with the professionalism as the teacher feedback feature comes the third feature of teacher feedback, namely, is the authority. Authority of teacher feedback, according to the participants, indicated a sense of irresistibility and undoubtedness. The participants might take it for granted all the feedback from teachers would work and they hardly challenge any feedback. With the teacher professionalism and authority, the participants actually expected teachers to correct all their errors. For example, QL stated that "If teachers may have enough time to grade our papers, I still expect teachers or instructors correct all errors in my writing. Instructors are authoritative, experienced, and professional." We also found through the participants' interviews that they were more concerned with teacher knowledge input than teacher feedback in their writing process. Specifically, they expected teachers to give more about content knowledge, writing skills or writing ethics than teacher feedback. For example, QL reported that "when I was taking practical writing courses the instructor taught us so professional and helpful knowledge on how to write resume, cover letter, or job application letters. For me, their professional knowledge shared is way more important than feedback to give."

Beliefs about Peer Feedback
Different from the participants' expectation of teacher feedback on logic and coherence, all the participants believed their peer feedback should focus on grammar errors or typos from micro perspectives. For example, CF reported that peer feedback was more "detailed". FJ also expected peer feedback to focus on "minor issues like grammar errors".
Of the three functions that HY summarized for her peer reviews, spotting minor issues including grammar errors or wording issues was the primary one.
The most important feature or advantage from peer reviews, according to the participants, was the empathy and resonance. All the participants reported their peers fully understood what had been conveyed, typically through their shared stories or the events or phenomena occurring or prevailing in their generation. For example, Reading through the assigned papers and doing the peer reviews for these participants works as a way to do mental exchanges between readers and writers. HY stated: "peer feedback is a way to exchange ideas between readers and authors. I got to know my peer reviewers really know me through their reviews. I also like reading their writings as their peer reviewer." However, unlike the authority and professionalism from teacher feedback, uncertainty is one primary feature or even disadvantage of peer feedback. The participants might be uncertain about the accuracy of their peer feedback, especially for these suggestions on content or coherence. They might also challenge their peers' suggestions or feedback, so they believed some suggestions might come from misunderstanding. For example, CF reported: "For peer feedback that I may not agree with, I may share some points or even argue with the reviewer in my back evaluation." Likewise, FJ stated "When I receive peer review, I sometimes challenge some peers' feedback. I feel that what s/he said is not fully correct, while I rarely refute it directly." Also, the quality of peer feedback might vary among reviewers due to language proficiency and comprehension. For example, in the survey item asking, "Do you think the linguistic competence of peers will influence the quality of peer review?", three of the participants chose "strongly agree" for the item. The participants through the interviews also reported that sometimes they were concerned with the reviewers' abilities or performance. For example, HX reported I wrote three paragraphs, which were progressively advanced; but the reviewer did not see it (my design) and felt that the three para-graphs were not logically organized or even redundant. He finally gave me a low score for the review.

Beliefs about PeerCeptiv Platform
The participants in the study reported that they found Peer-Ceptiv platform useful, as it provided the participants with a way to do human-machine interaction or internet-assisted communication for the review process. Instead of doing simply grammar checking or proofreading work for the submitted writings, PeerCeptiv as the participants reported, delivered real communications between the participants and their reviewers. For example, QL compared another frequently-used platform with PeerCeptiv and argued the other platform replied on the technology and algorithm too much, lack of real communications between authors and reviewers.
Another feature that the participants reported about Peer-Ceptiv was the clear rating rubric, which provided the participants with directions and items to grade upon. CF reported: "it was user-friendly, but with the mentoring from the instructor." The other participants also stated that it took them longer for the first round of reviews through PeerCeptiv than the second round. They needed to be accustomed to the platform. However, in the second round, it become easier to operate.
In addition, CF stated: "we (as authors) also enjoyed reading and commenting on others' works so that we know the difference of language proficiency among all our peers". Peer-Ceptiv in this vein, provided the participants with access to learn from their peers' writings and getting to know their peers. HY suggested: "If possible, we may go through another proofreading platform for the first round of reviews, PeerCeptiv for the second, and then instructors for the third. I know that might be time consuming."

Learner Review and Revision Practices
In the study, all the four participants were highly engaged in their writing, review, and revision process. The participants submitted their first writing assignments through PeerCeptiv. As required, the participants were writing their reviews to other participants, while waiting for their reviews. Three random, anonymous reviews were then delivered to the participants through the platform. It is worth mentioning the four participants were selected from a pool of 23 students, which indicates their anonymous reviews were not typically from each other. After receiving the reviews, they analyzed the reviews and then responded to their reviews by indicating how they had addressed issues in the reviews. The responding process was termed as back-evaluation. Then, the participants submitted their revisions through PeerCeptiv. As stated previously in Data Collection and Analysis section, the steps and procedures resemble a task continuum which occurred twice in the whole process.

Review Practices
We analyzed the actual writings and reviews of the four participants, together with the transcriptions of their interviews. We found that the participant review process generally comprised of four different practices, including evaluating, resonating, learning, and reflecting. However, the four different practices may not occur in sequence or together.
Evaluating. All the participants evaluated their randomly assigned papers through the platform as the primary task for their review process. With the mentoring and rubric from the instructor, the participants evaluated their assigned papers from dimensions of unity, support, coherence, and wording & sentence skills. For example, CF reported that the training and mentoring did help her save time for the review, as she referred to some wording and sample review sentences from the review rubric and template that the instructor had given to her. Also, FJ found the review process led her to better understand what defined a professional peer-review task and get to know how she could write a review report.

Resonating.
We found the primary feature or advantage of the peer review process was the sense of empathy and resonance generated from the participants' reading of the reviewed writings. Different from the teacher feedback that is rational, professional, and systemic, peer feedback may be emotional and personal. As reported in Excerpt 3, QL shared her example of how peers in her generation from different cities could not date or meet and suffered from pandemic successfully aroused her peer reviewers' emotion. Likewise, FJ's reported life story about her grandma's death resonated with her peer reviewers. Behind the empathy and resonance lies the fact that peer review is a way of communication and emotion exchange, connecting readers with authors.
Learning. In the study, HY reported she enjoyed learning from reading the assigned articles and believed she could learn some vocabulary and wording from the reading. She even reported she sometimes held some admiration of the authors when she read sentences and structures of great craftsmanship. Similarly, FJ also reported she did learn something from reading the assigned writings, regardless of the wording or paragraph developing techniques. Lundstrom and Baker (2009) found that learners could seize what their peers were able to do better than them and pay attention to those aspects in their own writing. In addition, Ruegg (2017) found that learners overwhelmingly saw the advantage of reading others' essays more than that of receiving peer feedback.
Reflecting. HY in the study reported that reading through others' writings helped her realize weaknesses or shortcomings in her writing. She continued: "I might be unable to see the whole picture. I mean I cannot see my own errors in the writing. however, when reading my peers' writings, I may reflect and recall if I've made similar errors or mistakes." However, this reported finding only occurred in HY's case.
When we checked the survey item that explored participants' perception about learning from the peers' writings, only HY reported she did learn from the reviewed writings. However, the other three participants either chose "I don't know" or "disagree" as the response to the item.

Revision Practices
In the study, we included back-evaluation and rewriting practices in the revision practices. Specifically, the participants, having received their peer reviews, were required to respond to the peer reviewers on how they had addressed their feedbacks. The process was a back-evaluation process. Based on the peer reviews and the back-evaluation, the participants then rewrote the paper and submitted it through PeerCeptiv. We found the actual revision process in the study may include the following four practices.

Crediting.
We found that all the participants first acknowledged their peer reviewers' contribution while doing the back-evaluation. CF and FJ reported in the interview they learned this strategy to give compliment or acknowledgment before arguing their points from the teacher mentoring. The instructor set up a good example in the sample review in which compliment and recognition precede the actual suggestions and comments. HY used different strategies to do the crediting or acknowledgement: in some back-evaluation, she expressed her thanks directly (e.g., Thank you for the advice.); in others, she confirmed the review and expressed her beliefs that the review had been quite helpful (e.g., The review is very helpful.). There were also some back-evaluations that she used a mixture of these strategies (e.g., Thanks for the suggestion! Your review is helpful and instructive.).
Arguing. As revealed from the participants' stated beliefs about peer feedback, the participants, on some occasions, may not totally accept their peer feedback. Because they did not believe some peers really understand their intentions or examples in the writing, or they believed there might be some misunderstanding or miscommunication. Therefore, when doing back-evaluation, the participants may argue with the reviewers or clarify their points with their explanations (see Table 2).
Correcting. For the correcting practices, all the participants in the study corrected their typos, grammatical errors, or wording issues in the revised manuscripts (see Table 3). However, compared to these grammatical errors or wording issues, the participants paid less attention to coherence or cohesion feedback in their revisions. This could be revealed through the participants' interviews or stated beliefs, as they reported they were not sure of the reviewers' feedback accuracy and were concerned they might give inaccurate feedback.

| Research Article
Polishing. The participants generally polished their revised writings and resubmitted them through the platform to finalize the whole process. The polishing work primarily includes the proofreading of the writings which might be done through other platforms or the participants themselves. As reported in their interviews, the participants also used other technologies or platforms (Grammarly or Pigainet) to help them with the proofreading work. From the submission of the first draft to the resubmission of the draft, the participants spent more than one month completing the whole task continuum.

DISCUSSION
In the study, we reported our findings on how teacher mentoring and learner WCF beliefs had informed the learner review and revision process through PeerCeptiv, an online peer-review platform. Generally, the teacher mentoring process was systemic and well-organized in five stages, including session & tech training, Q&A, task reminders, task completion monitoring, and the stage of encouragement, compliment, & praise. Learners' overall beliefs about teacher and peer feedback varied, as learners believed different types of feedback might serve different functions.
In addition, all the learners believed the platform was useful in helping them with their review and revision process. With the instructor's and peers' help, the learners' review and revision process went smoothly through different steps and stages.
Teacher mentoring as the appetizer, peer feedback as the main course, and revision as the dessert A sociocultural perspective may offer substantial insights on aligning feedback with learner's writing development (Yu, 2020;Leontjev & Pollari,2022). One of the typical insights is feedback provides writers with scaffolding opportunities (Nassaji, 2021). In the study, scaffolding in different forms work together may enhance learner performance. It is through such collaborative support that feedback can further learners' interlanguage growth and ability (Aljaafreh & Lantolf, 1994). In the study, teacher mentoring facilitates students to understand and perform the peer-review tasks.
All the participants reported their benefits from the teacher mentoring, either as a push to help them complete the task continuum or as a guide to simplify their review procedures. However, the teacher mentoring is not the crucial factor that leads to the completion of the whole process. At the end of the first paragraph of the article, the author put forward her topic sentence, which can summarize the content of her article well. In addition, the author focuses on her regret in the subject paragraph. At the end, the author quoted the sentence in the book to express her feelings. But to be honest, I don't think this sentence is related to the theme of the article. (Anonymous) In terms of grammar, punctuation, and style, the authors have done a good job. Only one suggestion for you is that "in order to" could be replaced with" to". (Anonymous)

Sample Participant Responses
Thank you very much for your comment. The last quoted sentence means that these lost years cannot come back, which expresses regret in a more poetic expression. Perhaps this quotation alone will make people confused, so I will carefully consider the sentence at the end of the introduction again and give a simple explanation to let readers understand it more intuitively, so as to achieve the purpose of summarizing the whole paper. (QL) Thank you for the advice. I was trying to make the words more diversified by choosing "in order to". I suppose I can try to make some changes according to your advice. (HY)

| Research Article
The second tenet we may gain from an SCT perspective is the mediation, which highlights higher-order mental activities are all socially mediated (Vygotsky, 1978). Mediation is achieved through the various forms of physical and symbolic tools and artifacts that equip people with some connection between themselves and the world (Wells, 2007). In this study, the development of learners' feedback beliefs and their revision practices was achieved through teacher mediator, peer mediator, and PeerCeptiv mediator. However, some mediators may play a more significant role than others in the process. Metaphorically, we may compare the mentoring as an appetizer to stimulate the participants' motivation, but the peer review is the main course that helps the participants complete the information exchange, communication delivery, and learning process. Having said that, student beliefs about WCF together with teacher mentoring may inform student practices of these peer-reviewed tasks.

Complex relationships between learner beliefs and actual practices towards peer reviews
In the study we found student beliefs appear in a complex manner with student actual revision practices. We found it irrational to link learner beliefs and practices in a linear way, as the two constructs are internally dynamic and may appear in different tensions (Gao, 2021;Gao, Qin, & Gu, 2022). Specifically, we found some learner beliefs about their peer feedbacks helped them correct their writings, under the condition that these feedbacks fit the participants' feedback preferences and epistemology. Writers and authors actually hold their own ideology and epistemology, and they get recognition and a sense of fulfillment from readers who get resonated with their ideology. Wang and Zhang (2020) suggested increased learner beliefs of self-efficacy and perceived value of English learning promoted learning motivation and self-regulation.
In addition, learner beliefs are often consistent with their revision practices in terms of grammatical or wording issues, which fall into linguistic or language dimensions. However, for content and logic dimensions, learner beliefs about may not necessarily be consistent with their actual revision practices, as learners may be unsure of or concerned with the peer feedback. These uncertainty and concerns may result from language proficiency gaps, miscommunication, or misunderstandings. This finding was against the existing literature typically in the 1990s entailing that learners from collectivist cultures are unwilling to be critical of each other (Allaei & Connor, 1990;Carson & Nelson, 1994;Nelson & Carson, 1998). The finding in a way indicates that EFL learners in the current generation (2020s) even in the collectivist culture have developed their critical thinking further than the previous generations. The finding is consistent with the existing literature exploring linguistic features and ideology, and critical thinking of the current generation learners in China (Gao & Zeng, 2021). Further studies may attempt to examine if gender factor would make differences in learner beliefs and revision practices, as a significant difference in the strength of links between self-regulation and instrumentality, self-efficacy beliefs, and English self-concept has already been reported in research (Iwaniec, 2019).
While Hedgcock and Lefkowitz (1994) found that learners in EFL contexts focus primarily on accuracy rather than the development of ideas. In this study, we may challenge this focus by arguing it might not be the de-emphasis on the development of ideas but be the learners' preference to take accuracy issues over the idea issues in peer review feedback. This could be explained through the existing literature reporting learner perception of teacher feedback and peer feedback. Rollinson (2005) reported peer feedback is less authoritarian but more informal, so students may believe peer feedback serves better for accuracy purposes than for conceptual or ideational purposes.

CONCLUSION
In conclusion, this paper explored EFL learner beliefs, teacher mentoring, and revision practices through peer review feedback from the sociocultural perspective. Findings show that EFL learners believed empathy and resonance was the main advantage of peer feedback, and teacher mentoring facilitated them in understanding and performing the peer review and revision tasks. Moreover, student review process included evaluating, resonating, learning, and reflecting practices and the student revision process comprised crediting, arguing, correcting, and polishing practices. Lastly, scaffolding in different forms work together may enhance learner performance and student beliefs appear in a complex manner with student actual revision practices.
We wrap up the paper by offering some research and pedagogical implications to the reader who might be interested in WCF studies or improving their actual practices. While prolific literature has been conducted over the years on feedback studies, there remains more to explore in the field. For example, we may propose a further step for future research in exploring tensions between learner beliefs about different forms of feedback and their actual writing and revision practices. We also found room for exploration on personal or contextual factors mediating learners' preferences of one type of feedback over another.
One of the curriculum design or pedagogical implications for the study is the possibility of using an ensemble of different grading platforms together with teacher feedback in writing curricula. In the study, the participants reported different platforms may serve different functions, considering writing is no longer defined simply as a language skill but a medium to exchange or communicate between readers and writers.

| Research Article
While we acknowledged the devotion from the four participants in the study, we understand the limited participant size and sampling method is not flawless. We thus expect scholars conducting similar research in the future may improving the design or sampling method of the current study.
That is how we as authors treat our peer feedback, the recurring theme of the study.