A Study of the Quality of Feedback Via the Google Classroom-mediated-Anonymous Online Peer Feedback Activity in a Thai EFL Writing Classroom *

The current study investigated the quality of feedback produced via anonymous online peer feedback activity in a Thai EFL writing classroom. It also explored how the students perceived the anonymous online peer feedback activity. Peer feedback tasks, questionnaires, and an interview were used to collect the information. The results from the peer feedback tasks showed that the quality of peer feedback significantly improved. The results from the questionnaires showed that the students agreed that the online peer feedback activity helped them to improve their writing although there were some problems that should be improved. The follow-up interview revealed that the anonymity of the writers or the feedback givers would not affect how most students would react to the writings or feedbacks. It could be noted from the study that sufficient training must be provided before implementing the anonymous online peer feedback activity. The findings provide new evidence for scholars and instructors who are interested in implementing an anonymous online peer feedback in a Thai classroom.


INTRODUCTION Background and rationale of the study
In Thailand's educational system, English is taught as a foreign language at primary, secondary, and tertiary levels. English is then regarded as one of the most important languages of education, politics, and commerce (Srinon, 2011). With this regard, Thai students' competence in English is expected to be improved. Srinon pointed out that English writing is also regarded as one of the skills that is important for all students both in English as a second language (ESL) and English as a foreign language (EFL) context. Academic writing is important for education of all levels, and in the context of Thai universities in the upper northern region, English major students study courses about academic and research paper writing in their junior and senior years.
Academic writing is a complicated process. The courses normally integrate the knowledge students have learned in the previous writing courses and requires them to write in academic fashion. Basically, the courses require the students to apply advanced reading and writing skills to produce written academic work which is suitable for academic and professional publications. It also requires students to think, analyze and reflect on their work. According to the nature of the course, a threecredit course requires 45 hours of class time, a three-hour of lecture weekly for 15 weeks, but three hours a week may not be enough for practicing writing skills. Chrisman and Crandall (2007) pointed out that an average person who takes a writing course requires at least 100 hours of teaching to move up to the next level of proficiency. However, as pointed by Kibler (2005), in traditional, teacherfronted EFL writing classrooms, it was unlikely that teachers would have enough time to attend to students' needs and to support students to fully engage in the writing process; however, with computer and online peer feedback activities, teachers could maximize the amount of time for the students to engage in the writing process.
Peer feedback is a common practice of writing process in American ESL situations, but it is not commonly practiced in an EFL context (Taylor et al., as cited in Rosalia, 2010). In the past, most studies regarding peer feedback targeted only the improvement of the writing output (Lundstrom & Baker, 2009). Although there might be similar peer feedback activities that were done earlier, the contexts of the current study was different from the previous studies in terms of locations, participants, proficiency in English, the activity of writing, readiness of facilities, and the number of personnel. In addition, few studies, if any, were conducted in the context in which the writing tasks and peer feedback activities were conducted 100% online, and writers and reviewers were anonymous to one another. According to Chuenchaichon (2014) who reviewed 48 EFL writing research studies that were conducted in Thailand during 2004-2013, it was found that there were only five studies that dealt with writing feedback, and among the five studies, only one dealt with online writing and identifiable peer feedback. More importantly, none dealt with the study of the quality of feedback.
Investigating the quality of feedbacks through the anonymous online peer feedback in the context similar to the present study is rare. In addition, although the effects of concealed identity have been studied in many disciplines (Scott, Rains, & Haseki, 2011), it has rarely been studied in the field of education studies because it had few application to the traditional face-to-face classroom (Lin, 2018). Nevertheless, due to the invention of online communication technology allows anonymity to be implemented in classrooms (Scott, 2004), the implementation of anonymity and communication technologies in education has drawn researchers to study its effectiveness in the contexts of education (Lin, 2018;Yu, 2012;Yu & Liu, 2009). Therefore, it was expected that when combining anonymity with the online peer feedback activity, the findings of the current study would add to the literature regarding the quality of feedback produced in the anonymous online peer feedback activity in a Thai EFL writing classroom, and would provide an option to improve the students' written feedback in Thailand's EFL writing classroom with the use of online activities that help facilitate the teaching and learning of EFL writing.

Purposes of the Study
The current study aimed to examine the quality of the feedbacks (measured from the components of the feedback itself and the interaction between peers) given by peers in the context of anonymous online peer feedback activity: the interest was on the quality of feedback that the feedback givers gave to the writers. The other focus of the current study was to examine the students' perception toward the online peer feedback activity.
Therefore, the purpose of this study was to examine the effects of online peer feedback in two areas of the study: (1) the quality of feedback that the students produce, and (2) the students' perception toward the online peer feedback activity.

Research Question
The research questions stated in the current study were intended to investigate the quality of feedbacks that the students produce throughout the online peer feedback activity and the students' perception toward the peer feedback activity. To that end, the general research questions of the current study examined the students' learning outcomes after they were engaged in the peer feedback activity while addressing the following questions: 1. To what extent did the anonymous online peer feedback activity improve the quality of the feedback that the feedback givers provided to the writers?; and 2. How did students perceive the anonymous online peer feedback activity?

LITERATURE REVIEW The Writing Process Approach
A simple definition of the writing process would be the steps a writer would normally follow when carrying out a writing project, such as pre-writing, drafting, revising, editing, and publishing. Furthermore, these steps cannot be carried out without a context. The role of audience and audience participation are essential. Hyland (2003) asserts that encouraging student writers to actively participate through negotiation of the writing process through peer feedback makes them more independent, more critical, autonomous writers. Peers can carry out this function, and the role of the peers will change depending on the steps being carried out (Coffin, Curry, Goodman, Hewings, Lillis, & Swann, 2003).
can gain a sense of audience. It has exponentially increased the audience that students have to share their writing with (Ito et al., 2008). Students can write, edit, and publish work to be read to a wide range of audiences almost instantly. The use of online blogs, wikis, and collaborative writing software, such as Google Docs, are all a part of this digital writing structure that students are tapping into. This brings them a larger, more magnified feeling of audience. Some websites act like writing clubs. They have forums where students can submit their writing and read each other's works. Fellow members then rate, edit, and make comments.

Peer Feedback
Peer feedback can be valued as an essential technique to teaching process writing. It allows students to play the role of the authors and reviewers whose task is to give comments to their peers' written work (Hansen&Liu, 2005;Lam, 2010). It provides students with opportunities to communicate and interact with their peers, and generate a source of information, negotiation of meaning, and a sense of group cohesion. Rollinson (2005) states that peer feedback creates a real audience as the reader will let the writer know if the message needs clarification. Moreover, it encourages collaborative dialogue, interactions and cooperation with peers, which is in line with what Vygotsky's (1987) suggestion on how learning happens through the guidance and response of others. In an attempt to improve the teaching and learning of English language skills in EFL context, the social constructivist approach was introduced into the EFL teaching (Dueraman, 2012). According to Vygotsky (1962), an individual's development happens from the interaction with people and environment. This kind of interaction results in higher order thinking skills (O' Donnell & Hmelo-Silver, 2013) and a person's own perspective of the world. English language users need socially engaging reasons and communities to improve their writing. The interactions between the writers and readers can enhance learning as suggested by the Sociocultural theory. Therefore, writing for real readers makes the writing activities more purposeful.

Quality of Peer Feedback
The literature on assessment in higher education argues that peer feedback should lead to learner autonomy. Nicol and Macfarlane-Dick (2006) characterize good feedback as that which: "(1) clarifies what good performance is, (2) facilitates self-assessment, (3) delivers high quality feedback information, (4) encourages teacher and peer dialogue around learning, (5) encourages positive motivation and self-esteem, (6) provides opportunities to close the gap and (7) uses feedback to improve teaching" (p. 203). They propose that this type of feedback is important for supporting and developing self-regulation in students. Similarly, Liu and Carless (2006) define good peer feedback as the processes to develop skills such as critical reflection, listening to and acting on feedback, and sensitively assessing and providing feedback on the work of others; in addition, Keppell, Au, Ma, and Chan (2006) suggest that interactions between reviewers and writers constitute a good feedback because the interactions and negotiations between them eliminate the power that one might hold over the other by the virtue of position and responsibility. Accordingly, autonomy is about learners supporting and challenging each other. It is a capacity that is fostered through not only asking students to take responsibility for their own learning or to work independently, but also to help them see how their learning is based on collaboration and critical friendships (Ciekanski, 2007). Elements of autonomy or self-regulation, critical reflection, and reciprocal interactions should inform a definition of quality feedback.
As the number of learners using computers continues to rise, definitions of quality peer feedback ought to be informed by the literature on Computer Assisted Language Learning (CALL). The feedback must be carefully organized so that it gives students concrete ways to see their own and each other's progress. Chapelle (2001) argues for a CALL materials approach that addresses language learning potential, learner fit, meaning focus, authenticity, impact, and practicality. In other words, the product and process of peer feedback should include high quality comprehensible input (it should not be full of language errors); be at the right difficulty level considering Zones of Proximal Development (Vygotsky, 1978); communicate meaning; link to "real life" outside the classroom; be a positive technology and learning experience in and of itself; and work without mechanical failures or excessive cost.

Attitude toward anonymous online peer feedback activity
Anonymity, according to Marx (as cited in Lin, 2008) is a condition in which the identity of the people in their communication units, e.g. messages, is concealed. Understanding the students' attitudes toward the online peer feedback has been considered important when designing learning activities that aimed at improving learning (Liaw, Chen, & Huang, 2008). The study of Yu (2009) reveal that, the anonymity of the participants was seen as useful and supportive to their performance in the peer feedback activity. In addition, the study of Yu and Liu (2009) also showed similar results. The majority of the participants in their study expressed that they preferred to be anonymous when doing the peer feedback. The results of these two studies seemed to suggest that being anonymous could result in the participants' having positive toward the online peer feedback activity.

Context and Setting
The current study took place at one university in the north of Thailand. The participants were an intact group of 21 students who enrolled in Academic Writing course in the first semester of academic year 2017. None of them had experience doing any kind of peer feedback activity before.

Research Design
The current study required both quantitative and qualitative methods to investigate the effects of the anonymous online peer feedback activity on the quality of the feedback and the students' experience from doing such activity. The quantitative investigation, especially quasi-experimental study, has the capacity to present convincing numerical evidence of skill improvement, and the qualitative investigation allows for the understanding of students' perspective on the activity (Ellis, 2008).

Instructional instrument
Online Peer Feedback Activity. The Google applications, Google Classroom and Google Docs, were be used in the present study as a platform for the students to share their writing pieces. The Google classroom served as a place where the students submitted their writing, while the Google Docs was used as a platform where the students produced their works and gave feedbacks to their peers.

Research instrument
Peer feedback. The peer feedback activity in the current study comprised two steps. The first step was giving feedback to the writing. In this step, the feedback giver gave feedback on the writing of their fellow students. The second step was commenting on the feedback. After the feedback giver had given a written feedback on the writing, another feedback giver reviewed the feedback of the first reviewer and gave comments for the first reviewer to improve his/her comments before giving it to the writer. The quality of the feedback in each step was determined using a modified version of Rosalia and Llosa's (2009) 'Product' and 'Process' rubrics.

Students' Experience toward the Anonymous Online Peer Feedback activity
At the end of the study, a questionnaire was distributed to the students to investigate their perceptions and experiences toward the anonymous online peer feedback activity. The questionnaire consisted of two sections. Section one was the rating scale, and its main focus was on the students' perceptions toward the anonymous online peer feedback activity. Section two was the open-ended questions, and its focus was on the students' experiences with the anonymous online peer feedback activity.
In addition, a follow up semi-structure interview was conducted at the end of the peer feedback activity. The questions in the interview aimed to probe into the students' perceptions towards whether the anonymity of the writer and the feedback giver would affect them.

Instructional Training
In the current study, the peer feedback activity was a part of the instructions. Because the peer feedback activity was new to the students, a training was necessary for them. According to Lam (2010), the purposes of the training are: to equip the feedback givers with sufficient skills for reviewing tasks and to raise their awareness in providing effective peer feedback to the writers. In the current study, the peer feedback training followed the three-tier peer review training introduced by Ricky Lam (2010). The training was be divided into three stages: modeling stage, exploring stage, and consciousness-raising stage.
The modeling stage prepares the students for the peer feedback session. The heart of the modeling stage is the four-step procedure that helps the feedback givers to give a sound comment to the writer. The four steps include: clarifying, identifying, explaining, and giving suggestions.
The exploring stage allows the students to practice the four-step procedure. In this stage, a guidance sheet for feedback givers will be given to the students to follow. After the students finished reviewing the example writing, they will give feedbacks, exchange them with peers and discuss.
The consciousness-raising stage aims at enabling the students to analyze the peer feedback that they give to writers. Similarly, as writers, the students will be trained to analyze and respond to the feedbacks properly.
The peer feedback training was carried out once a week for three weeks (week 1, week 2, and week 3). After the peer feedback training finished, the students were introduced to the Anonymous Online Peer Feedback.  Practice analyzing the type of feedback  Practice evaluating the feedback whether to be adopted in the revision

Anonymous Online Peer Feedback Process
To conduct the Anonymous Online Peer Feedback activity, the instructor had to facilitate the process of sending writings to reviewers, and sending the reviewers' feedbacks back to the writers. The reviewer for each writing was randomly selected from the students in the section.
In the current study, the anonymous online peer feedback process was be facilitated by the researcher. The steps of the peer feedback process are shown below.
1. After the students finished each draft, they submitted their draft in Google Classroom.
2. The researcher, working as a facilitator, sent each draft to two reviewers.
3. The reviewers gave feedback and returned the draft to the researcher. 4. The researcher had the reviewers review the feedbacks of each other and make give feedbacks.
5. The reviewers returned the comments on the feedbacks to the researchers.
6. The researcher sent the comments of feedback to each reviewer to improve their feedback on the draft.
7. The reviewers improved their feedbacks and sent their revised feedbacks to the researcher.
8. Finally, the researcher sent the revised feedbacks to the writer.
All steps were done online, and the writers and reviewers were anonymous to one another.

Data collection Procedure
The study began in the first semester of academic year 2017. The quantitative data collection of the writing scores began as a part of the regular classroom environment for the participants who were involved.
The quantitative data from the students' writing assignments and feedbacks were collected throughout the semester. The anonymous online peer feedback activity allowed the researcher to collect the quantitative information of each writing assignment which include: the drafts, the revised drafts, the feedbacks, and the responses between the feedback givers. This collected information allowed the researcher to investigate the quality of feedback.

The Quality of Peer Feedback
The quality of the written feedback. To determine the quality of the feedback itself, each peer feedback was graded using the modified version of Rosalia and Llosa's (2009) Product Rubric. For each writing assignment, the students produced a feedback. As a result, the paired samples t-test was used to analyze the change of the quality of feedbacks in each assignment.
The quality of the peer feedback process. The interaction between the reviewers was graded using the modified version of Rosalia and Llosa's (2009) Process Rubric. The score for the interaction between the reviewers of each assignment was collected throughout the study and finally the scores were analyzed using the paired samples t-test.

Students' Experience toward Online Peer Feedback
The quantitative results from the questionnaire were calculated and analyzed using descriptive statistics, while the qualitative results from the open-ended questions were analyzed and categorized using the thematic analysis.
The interview data were transcribed and categorized according to the emerging themes using the thematic analysis.

RESULTS
The results in this section are presented according to the order of the research questions.

Research question 1:
To what extent did the anonymous online peer feedback activity improve the quality of the feedback that the feedback givers provided to the writers?
In the current study, the quality of peer feedback was measured by two criteria: the product score and the process score. The product score was given to the feedback (written comment) made by the feedback giver, and the process score was the score given to the interaction between the feedback givers who gave feedback on the same piece of writing. The students engaged in two peer feedback activities; therefore, there were two sets of product scores: Product1 and Product2; and two sets of process scores: Process1 and Process2.

Difference between Product1 and Product2 Scores
To investigate whether there is a difference in the quality of feedback (written comment) of the students in group two between the Product1 scores and the Product2 scores, the paired samples test was performed. Note. * p < 0.05 Table 1 shows that there were 16 students who completed the Product1 and Product2. The mean score of the Product1 was 5.00 and the mean score of the Product2 was 8.34.
The result from the paired samples test in Table 1 shows that there was a significant difference between the Product1 and Product2 scores (t 15 = -8.104, p < 0.05). On average, the Product2 score was 3.34 points higher than the Product1 score (95% CI [-4.22, -2.46]). Therefore, it can be said that the quality of feedback of the students increased.

Difference Between Process1 and Process2 Scores
To investigate whether there is a difference in the quality of feedback (the interaction between feedback givers) of the students between the Process1 scores and the Process2 scores, the paired samples test was performed.  Table 2 shows that there were 11 students who completed the Process1 and Process2. The mean score of the Process1 was 4.50 and the mean score of the Process2 was 6.82.
The result from the paired samples test in Table 2 shows that there was a significant difference between the Process1 and Process2 scores (t 10 = -6.84, p < 0.05). On average, the Process2 score was 2.32 points higher than the Process1 score (95% CI [-3.07, -1.56]). Therefore, it can be said that the quality of feedback of the students increased.

Research question 2:
How did students perceive the anonymous online peer feedback activity?
The results from the questionnaire. The perceptions toward the online peer feedback activity of the indicated that the students agreed that the online peer feedback activity could help them in many areas. The results of the questionnaire as showed in Table 3 showed that the average scores given by the students was 3.18 (SD = 0.16). The responses from the question one could be put into five different categories: 1) knowing the mistakes from the feedbacks; 2) using the feedbacks to improve the work; 3) learning the mistakes from friends' work; 4) learning to give feedback in academic writing; and 5) helping the students to review their past knowledge.
The responses from the questions two could be put into four different categories: 1) not proficient feedback givers; 2) cumbersome process; 3) inappropriate feedback; and 4) irresponsible feedback giver.
The responses from the question three could be put into three different categories: 1) in-text feedbacks; 2) summary of the first feedback giver's feedback; and 3) the revised feedback.
The responses from the question four could be put into three different categories:1) steps of AOPF; 2) feedback givers; and 3) facilitation.
The results from the interview questions. The interview questions mainly focused on the anonymity of the writers and the feedback givers. The questions were: 1) As a writer, how would you have reacted differently if you had known the person giving feedback on your writing?; and 2) As a feedback giver, how would you have reacted differently if you had know the writer of the work you commented on?
Regarding the interview question one, most students said that nothing would change, but some of them said that knowing the feedback giver might affect how they reacted to the feedback.
They who responded that nothing would change said, "We have to value others' opinion. And if it is useful, we will use it because it is an assignment." The students who responded that it might affect how they reacted to the feedback said, "It depends on the ego of the person. It depends on the closeness. If we are very close, I may pay more attention to the feedback;" "I may not take the feedback into consideration if that person is weaker;" "It might affect my feeling. It is better not to know;" "If the feedback giver is better than me, that's good. But if the feedback giver is poorer, well, you know." Regarding the interview question two, the majority of the students responded that they would do the same, while some of them reported that they might react differently.
The students who responded that they would react the same way reported that, "I would do the same: I want to help the writer by giving feedback to help the writer improve their writing." The students who responded that they might react differently said that, "If I did not know, I would comment honestly. But I knew my friends' topic, so I gave feedback that would satisfy that them;" "If I knew that the writer did not pay attention to the work, I would not pay attention to giving the feedback because the writer would not pay attention to reading the feedback as well;" "If I was close with the writer, I would write honestly. If not, I would write what is necessary;" "Some might take it personal."

DISCUSSION
The current study investigated the quality of feedbacks that the students produce throughout the online peer feedback activity and the students' perception toward the peer feedback activity.

Research questions one
The findings revealed that the anonymous online peer feedback activity enabled the students to improve the quality of the feedbacks. The findings correspond with the work of Rosalia (2010) whose study showed that the students produced better quality feedback according to all dimensions in the grading criteria. One possible reason that contributed to the increase in the scores based on the grading criteria could be because of the training on how to give feedback (Berg, 1999;Hu, 2005;Min 2006).

Research Question two
Quantitative results. The quantitative results from the students showed that they were positive toward the Anonymous Online Peer Feedback activity. This quantitative result corresponded with MacLeod's (1999) study in which almost 80% of the participants positively accepted the anonymous online peer feedback activity. MacLeod's findings were also echoed by the study Lin, Liu, & Yuan (2001) which revealed that significantly more students were satisfied with the anonymity of the online peer feedback activity.

Qualitative results
Open-ended questions. According to the results from the open-ended questions, the students found that the anonymous online peer feedback allowed them to know their mistakes, learned from the friends' comments and writing, and use the feedback to improve their work. The results provide support to the findings of many researchers (Bhalerao & Ward, 2000;Nilson, 2003;Lu, 2005) that the peer feedback activity exposed the students to various opinions, which helped assure them the correctness of the comments. The result was also in line with Mangelsdorf's (1992) study in which the participants reported that the peer feedback activity helped them revise the content of their work.
The results also revealed the drawback of the peer feedback activity. The students reported that some feedback givers lacked responsibility. This finding supported Lu's (2005) study that some students raised concerns about the lack of responsibility shown by some students. Another finding concerning the drawback of the peer feedback activity was that some feedback givers were not proficient enough to give feedback. The issue of unqualified feedback givers was also reported by Lu (2005). She continue to point out that the unqualified feedback givers produced poor quality feedback that made the writers hesitate to take their peers' feedback seriously.
Regarding the suggestions for the improvement of the anonymous online peer feedback activity, it was found that the students suggested that the process and the feedback givers be improved. The researcher found that the problems that happened with the process and the feedback givers came from insufficient training. The students in the study never did any kind of peer feedback before, so when they encountered an online peer feedback activity, they had to learn two different tasks: peer feedback and online activity which require sufficient training to achieve the goal of the activity.
Interview questions. Most of the students responded that they would do the same if they knew who the writers or the feedback givers were, while some students reported that knowing who the writers or the feedback givers were might affect how they would react. The responses of most students contradicted past study. For example, Lu (2005) showed that in the anonymous peer feedback activity, the participants were less anxious because their identity was protected; hence, free of fear of being wrong or hated. Despite the fact that the results of the current study were not align with past study, it is worth mentioning that the students in the current study had zero experience with any kind of peer feedback activity; therefore, they did not know the experience when doing another kind of peer feedback activity. So, the finding of the current study contributed to the literature in this area.

CONCLUSION
The current study aimed to investigate the quality of feedback in the anonymous online peer feedback activity and the perception of the students toward the anonymous online peer feedback activity.
The results from the analysis feedback showed that the quality of feedback improved significantly. The results from the questionnaires and interview showed that they students agreed that the anonymous online peer feedback benefit them in many areas of writing. However, there were some problems about the process of the activity and the feedback givers and they should be improved. The results from the interview showed that most students said that they would do the same if they knew who the writers or the feedback givers were.
It was found from the study that training is an essential part of the success of the peer feedback activity, especially when implementing with students having no experience with such activity.