OF THE INTEGRATION OF PEER-ASSESSMENT IN EFL WRITING CLASSROOM (CASE OF

ABSTRACT


Introduction.
Writing skills are considered to be the most complex and challenging skills among all 4 macro skills for L2 language learners to master. Learners for whom English is lingua franca, adopt the skills of speaking, reading, and listening much easier than that of writing. Close look at the problem reveals that to address this issue, adequate pedagogical approach to teaching writing is of vital importance. The researchers emphasise the hindering effect of the 'outdated skills of pedagogy' (Dar, Zaki, & Kazmi 2014). For instance, several decades ago, so called 'product approach' to writing was very common. EFL/EAP teachers used to give students ready-made summaries, essays, and etc. and the learners were expected to use them as the templates for writing, which proved to be ineffective, as learners could not understand what criteria and qualities made the paper good or fair (Shamim, 2008;Reinholz, 2015). The assessment system was also failing. The most problematic was the lack of writing practice and absence of proper feedback.
In this study we will focus on peer assessment, as one of the most effective techniques for developing writing skills by giving feedback and assessment. Peer assessment has been the subject of close observations for several decades at school, tertiary and university levels. TESOL experts and instructors have been discussing different models and format of its implementation in the classroom. The results of numerous surveys suggest that peer assessment is the efficient tool to provide timely, constructive, and individualized feedback to the L2 learner ( Finch, 2014) and failures of this class technique in the implementation process in EFL classes. Such challenges can include language proficiency level, social and cultural factors, the whole class participation, individual factors (selfconfidence, acceptance of criticism) and etc.
This research, based on the self-reflective classroom observation practice, learners' feedbacks, as well as on a small-scale quantitative and qualitative survey implemented in the EFL classroom, is one of the first attempts in Georgian L2 classroom to highlight some of the major challenges faced in the process of implementation of peer assessment and suggest recommendations for its successful integration in teaching academic writing.
Despite the fact that this research was implemented in EFL writing class, the research outcomes and the suggested model of the peer assessment can be beneficial for any other writing class context, including ESP, EAP or academic writing course delivered in mother tongue.

Literature Review.
Peer assessment has become the subject of increasing interest among TESOL and EFL professionals during the last 10 decades. The researchers examine and explore the effectiveness of the method for teaching writing, especially in L2 classrooms. The studies explore the efficiency of the peer assessment as a form of formative assessment ( -Creates flexible teaching atmosphere.
-Promotes students' participation in EFL classes.
-The best class technique to support the 'process approach to writing, '' (giving the opportunity to the learners to master writing by preparing different drafts that is also very motivating for them as it gives the students the possibility to improve their grades).
-Helps to connect the process of reading and writing (Liu& Hansen, 2002) -Promotes social interaction and sharing responsibilities among peers (Spear 1987; APA, 2013). However, peer-assessment, as any type of class techniques, has its disadvantages and challenging aspects that also have long been the subject of the studies and numerous surveys implemented by the researchers and TESOL professionals around the world. Summative list of the drawbacks of this assessment techniques identified by majority of researchers and TESOL professionals (Spear, 1988 Ober and Flores (2020) suggest several 'instructional strategies' to address the challenges listed above that may arise during peer-assessment process. We will cite the classification offered by them unchanged below: "…. Decide when to schedule the peer review, ideally well enough in advance so that students have time to revise their work.
• Provide assessment tools (e.g., grading rubrics) in advance so that students have a framework for understanding the purpose of the assignment.
• Think about the structure of the peer review. Will the peer review take place in the classroom, or online? Will students be working in pairs or in small groups; how long will students have to collaborate with each other as reviewer and reviewee; when will you know to nudge students to switch roles? Consider online peer review as a viable option if classroom time is limited.
• Several tools can make this possible including a content management systems Wiki or Word or GoogleDocs with the track changes/suggestions feature turned on.
• Provide example checklists and clear instructions to guide students in providing feedback.
• Help students prioritize certain forms of feedback over others (e.g., content over mechanics, or vice versa).
• Monitor and make sure to spend approximately the same amount of time with each student or group of students.
• Encourage students to incorporate feedback, both in terms of scheduling due dates for drafts, but also by giving some form incentive for revising their work. Is attending and active participation in the peer review part of their grade or are students expected to make a substantive revision as part of their grade?
• Discuss with students' different approaches for incorporating suggestions and feedback into the final work or being prepared to provide a justification for declining feedback" (Ober &Flores, 2020).
Opinions differ whether peer editing should be soly written, oral, or of mixed character, whether it should be anonymous or not. Some researchers argue that in an anonymous peer assessment process, assessors are fairer in their feedback than when they know who the authors of the paper are (Rotsaert et al. 2018;Vanderhoven et al. 2015). However, recent qualitative reviews (Panaderro & Alqassab 2019) found out that in majority of experiments anonymity showed little benefits. Moreover, some studies argue that anonymity limits the collaborative nature of peer assessment (Stribros &Wichmann 2018). The researchers are still exploring the scenarios which would be the most flexible to make peer-assessment process more acceptable, as well as the most effective for development of writing skills in L2 learners.
Based on the literature review offered above, we consider, there is still the room to cover regarding effective implementation of peer assessment methodology in class. In this research, we decided to focus on the major factors for successful implementation of peer assessment in Georgian EFL classroom. The research question raised in this study is: RQ: what are the most significant factors for successful implementation of peer assessment in EFL classroom in Georgia?

Research Method.
The research methodology applied in this study is based on quantitative and qualitative data analysis. Quantitative data were analyzed with t-paired test and thematic analysis was applied to qualitative data collected through the participants' feedbacks and classroom observation practice.
Building upon the finding in the previous sections of the presented paper, we developed two different scenarios of peer assessment to experiment in EFL classroom of a Georgian University. The participants were ordinary students, first year students of EAP classroom, 3 groups, 102 students in total. All of them were BA students, of the same English language proficiency level. after passing Unified National Exams, they were graded and grouped based on the Internal placement test. The author of this research is a professor leading the course of EAP. Initially the students did not know that they were part of the experiment, as far as this course was a compulsory course they had necessarily to cover. Another reason for that was that being part of the experiment could course their uneasiness and worries. However, at the end of the course, and the experiment, which was going on for 3 months, they were informed that they were participating in the research experiment and they gave their consent to make the data collected through the experiment public. Still, any confidential information (such as names, surnames) is not provided in this research.
Having in mind all the finding and broad experience shared by the researchers on different models and scenarios of peer assessment, 2 experimental scenarios were developed.

Scenario 1-description.
Peer assessment was included in the course syllabus as a part of summative assessment. Participants were trained in a peer assessment task, given clear instructions and guidelines regarding the process; peer assessment was not anonymous, not to limit collaborative skills between participants with feedbacks and assessments provided in writing. The students in the role of the assessors (will be referred to as assessors or draft 'authors' below) were given rubrics and criteria, upon which they assessed the writings. The deadline was set for both parties. The Assessees (authors of the paper) had to submit the draft electronically to the professor's mail and the professor forwarded them for peer assessment to the predefined group of assessors. The assessments and feedbacks were provided in the special checklist and forms created by the professor for peer assessment purposes (see the appendix 1 below). The group of assessors had to submit one copy to the assessee and another to the professor. It is noteworthy that as a part of the experimental scenario, neither assessors' group nor the assessees knew that the professor was going to check each writing herself to compare the results (points) given by the peers to each other. The papers were checked by the professor using the same criteria and rubrics as the peers did.
The next stage of the experiment was to analyze the quantitative data collected through this scenario i.e., to compare the assessments provided by the assessors to those, assigned by the Professor using t-test. This type of analysis, as well as comparison of the points to detect the rate of correlation between the points assigned by peers and by the professor is tried and tested practice among EFL researchers (Double, McGrane & Hopfenbeck, 2019; Dar, Zaki, & Kazmi 2014). However, in this research we adopted this quantitative method of analysis to detect the difference between 2 scenarios of peer assessment with different variables, so as to decide which scenario proves to be more effective in Georgia EFL classroom. The table 1 provided above shows the analysis of the assessments assigned by the peer assessors and by the Professor to 102 participants. The maximum points the students could gain was 10 and minimum 0. From the statistical data given in the table above, we can see that means points given by the peer assessors and that of given by the professor are different, which denotes that there is a statistical difference between them. Range of points shows that variation of scores is higher in professor's assessments, minimum points assigned by the peer and by the professor are also considerably different 2:0.

Scenario 1 class observation and participants' feedbacks.
After the first experimental scenario of peer assessment, Qualitative data were collected based on the participants' feedback and classroom observation. As a result of thematic analysis provided after the first peer assessment process described above, the following circumstances were identified: -The learners were shy and reluctant to assess their friends and especially give points to each other. They though that these points were the final ones having impact on their classmates' final grades.
-Assessed students expressed the desire of oral interaction with assessors. They lacked the chance to defend their positions and comment on the feedbacks provided by the peer assessor.
Scenario N2-description. Experimental scenario No. 2 was designed based on the results of qualitative and quantitative analysis of the first scenario of the peer assessment presented above. All steps were the same as in Scenario 1, but taking into consideration the feedbacks provided by the participants, we introduced following variables in the process: Variable 1. Before the second cycle of peer assessment task, the learners were informed that their assessments are not final, i.e., the final points are assigned by the professor. Moreover, each assessor would be assessed on reviewing skills. Variable 2. We made the process of giving feedbacks more democratic, adding to it the possibility of oral interaction, between the reviewer groups and the authors. It gave the assessed students the possibility to express themselves, approve, accept, or reject the feedbacks suggested by their peers. If they could provide a logical argument, they could even reject the integration of the feedbacks provided by the peer in their next draft.
For more clarity, we'll cite below the assessment scheme for a student reviewer described in our own course syllabus: "...3 Peer-assessment (1 point each, 3X1=3) is done on seminar hours (week 5,8,12). During the whole course, the student must participate in peer editing process of 3 essay drafts presented by other students in the class. The class will be divided into several groups of 3-5 students. They will read each other's drafts, mark the checklist in the 'Peer assessment Sheet document' (attached to the course syllabus) provided by the Professor. It is timed process (1 hour). When the time is up, each 'author' student will make a short outline of his essay and editing group members will make comments.
The author of the draft again will comment on peer assessment checklist filled in by the group of assessor students. The author can refuse or accept their suggestions and mark corresponding box in 'Peer Assessment Sheet document.' However, the authors must present their arguments for their positions. Checklist is prepared and signed in 2 copies. One is kept by the student and the other will be kept by the professor.
Attention: Since peer assessment is a bilateral process and it involved at least 2 parties: each party (student-author and student assessor), will be assessed. After the peer assessment and turning in the rough draft, instructor will check it, give back to students with the final feedbacks. Then students will finalize the draft and turn it in directly to professor for the assessment and feedbacks." The next stage again was the same as in scenario 1. _ to analyze the quantitative data collected through scenario 2, i.e., to compare and statistically analyze the points assigned by peer assessors and by the professor. T-test was applied to analyze the quantitative data.
The Results of Quantitative and Qualitative Analysis -Scenario N2. The table 2 provided above shows the assessments assigned by the peer assessors and by the Professor to the same 102 participants in the second peer assessment task (scenario 2). The maximum points the students could gain was 10 and minimum 0. From the statistical data given in the table above, we can see that means points given by the peer assessors and those given by the professor are already slightly different (in 0.2) that means that there is no statistical different between them. Range of points shows that variation of scores assigned by the peer and by the professor is the same -8, minimum points assigned by the peer and by the professor are also the same -2.

Scenario 2 class observation and participants' feedbacks.
Qualitative data were collected based on the participants' feedback and classroom observation after the second cycle of the experiment. As a result of thematic analysis of the data collected through class observation and feedbacks, provided after the second cycle of peer editing process described above, the following circumstances were identified: -The students were no longer reluctant to assess each other and put the marks, because they knew that it was a professor who would assign a final assessment.
-They enjoyed the process of oral discussion with their assessor(s). It gave each party better understanding of their mistakes, was an effective step towards improving their writing skills and selfconfidence.
-They considered the peer-assessment task interesting, effective, and even enjoyed it. -

Discussion.
Judging from the table 1 above, which provide the data of the first scenario, we can argue that assessment assigned by the peers seems to be more loyal than the assessment carried out by the professor. Consequently, it cannot give us fair picture about the learner's writing skill, which mean that peerassessment process carried out by scenario N1 presumably can fail to be equitable. Thematic analysis of the feedbacks provided also proves that the main reasons for such kind of outcomes are the social factor of friendship and lack of oral interaction on the feedbacks provided by the peers.
On the other hand, experimental scenario N 2 of the peer-assessment with 2 variables defined (1. The final points/grades are assigned by the professor and not peer 2. possibility of oral interaction, between the reviewer groups and the authors before the final points are assigned) showed close correlation between the points assigned by the peer and assessor (table 2). Participants' feedbacks confirmed the outcomes of the quantitative survey. The students seemed highly motivated to participate in the peer-assessment that made the process fairer, more equitable and effective.
As we have already mentioned above, this study answers the research question _ 'What are the most significant factors for successful implementation of peer assessment in EFL classroom in Georgia?', however, we recognize that besides the factors identified in this study, there can be more, depending on the level of language proficiency, age and culture of L2 learners. We consider that the outcomes of this research can be beneficial for further studies of this problem.

Conclusions.
This study found out that the most significant factors that hinder successful implementation of peer assessment techniques in EFL/EAP classroom in Georgia, are the social factor of friendship and the lack of oral interaction between the peer reviewer groups and the reviewees before the final points are assigned.
Respectively, adoption of a two-fold model of peer-assessment that includes: 1. peer-assessment provided in writing, with final assessment assigned by Professor and 2. oral interaction between the assessor(s) and assessed students, will make the peer-evaluation one of the most effective techniques for developing writing skills in EFL classes in Georgia. In addition, it will enhance transferable life skills of critical thinking, discussion, argumentation, acceptance of the mistakes, respecting deadlines as well as the sense of belonging to the same community.