The role of dialogic interaction in EFL writing assessment: A sociocultural perspective

This study investigated EFL writing processes through collaborative assessment from a Vygotskian sociocultural perspective. This study focused on assessing essay writing of one intact group of Iranian EFL learners at the intermediate level of language proficiency (N = 15) based on the placement test administered at the institute. Both learners and their teacher assessed essays based on an analytic scale during seven sessions. Moreover, being engaged in collaborative dialogue, the learners were given the opportunity to discuss existing differences between their self-assessed scores and those assigned by the teacher in order to reach an agreed score. Moreover, all their dialogic interactions were audio-recorded for later transcription and micro-analysis. Findings of the microgenetic analysis revealed that benefiting from the mediatory role of the collaborative dialogue and the checklist, the group under investigation gained insight into their capacities, weaknesses, and their metacognitive awareness was raised as the result of conducting CA. The findings suggested that learners took advantage of the collaborative dialogue in which they were engaged while assessing their writing tasks. In particular, the dialogic interaction afforded the learners the chance to present, discuss, and test their ideas and consequently enhance their awareness of the writing tasks. The collaborative assessment could help the learners gain a better insight into their strengths and weakness; further, it led to their metacognitive awareness about components of a good piece of writing.


INTRODUCTION
Language assessment has witnessed a radical shift from traditional forms of assessment, known as Assessment of learning (AOL), measuring students' achievements for administrative and reporting purposes (Black & William, 2009, 2018William, 2001), to new forms of assessment dubbed as assessment for learning (AFL) by the UK Assessment Reform Group (2002), where learners are at the center of attention and "teachers' evaluation of student learning, their feedback, feedback from peers, and students' self-assessment play an important role in mediating students' learning and knowledge construction" (Lee, 2017, p. 9). Following Lee (2017), the focus of classroom writing assessment needs to be shifted from standardized tests administered to measure lexical knowledge of test-takers to the current assessment trends, which view writing as the ability to produce a coherent piece of text. It needs to be done so that it can ensure the enrichment of teaching and learning activities through feedback provided in the process of assessment. In fact, in AFL, learners take on responsibility as a result of their involvement in a self/peer assessment that Bachman and Palmer (2010) believe it is urgent for assessment to yield satisfactory results and in its own right improve learners' "metacognitive and self-monitoring abilities" (Lee, 2017, p. 10).
Under such conditions, assessment is conducted for improving learning rather than for gauging it (Lee, 2017(Lee, , 2019Lee & Coniam, 2013). In other words, students' promotion in learning hits the top of the AFL agenda (Black, 2003). Moreover, unless students are provided with feedback by others in assessing themselves and others, they would not make any noticeable improvement in their learning process (Black, 2003). That is to say, students should have a clear-cut outline of the goals they are moving towards, a thorough understanding of the criteria they are assessing against and should be engaged in self-and peer-assessment to enrich their work (Jones, 2010;Lee, 2017). However, AFL has not received considerable attention in the L2 writing assessment (Lee, 2007(Lee, , 2017(Lee, , 2019. In other words, the L2 teachers have still used the traditional tests, which serve summative functions; and therefore, their focus has remained on the L2 learners' final achievements (Lee & Coniam, 2013).
Concerning advantages that AFL brings about to both learners and teachers and the fact that AFL has not received considerable attention in L2 writing (Lee, 2007), drawing on Fahim et al. (2014), this study is set out to examine collaborative assessment (CA), a type of AFL, to gain a better understanding of learners' involvement in the assessment.
In the current study, both learners and teachers are involved in the CA, with the purpose of engaging the "student, their peers and tutor in a thoughtful and critical examination of each student's course work" (McConnell, 2002, p. 43). According to Chau (2005), the critical characteristics of collaboration are "mutual goals (working towards a mutually acceptable assessment grade), dynamic exchange of information (presenting, defending and elaborating views on the grade by a tutor and student) and role interdependence (emphasizing individual accountability for a meaningful exchange to take place)" (p. 27).
Accordingly, through CA, both teachers and students move towards a common goal, which is assigning a score acceptable to both. Furthermore, both sides have equal rights to express and defend their views on the final score. As McConnell (2002) asserts, when students have the opportunity of evaluating their own and others' performances, they will adopt a different perspective on learning from those who only undergo the teacher assessment. Fahim et al. (2014) set two levels for implementing CA, namely teacher-student and student-student levels. At the teacher-student level, the teacher, on the one hand, assesses students' performance; on the other, students are provided with an opportunity to assess their performance according to a set of established criteria that are acceptable to teachers and learners. Afterward, assigned scores are compared. If there is a discrepancy between the scores calculated by the teachers and learners, as Fahim et al. (2014) noted, they sit together and discuss over the points of difference to reach a consensus over one average score. In student-student collaborative assessment, the two assessors are students; students are initially engaged in assessing their peers' performances based on clear-cut scoring procedures; then, the students will be asked to assess their performance according to the scoring procedures. Finally, the two sides are given a chance to sit together and negotiate on the detected points of difference. More specifically, student-student CA encompasses both peer-and self-assessment. More relevantly, in the CA paradigm, learners are engaged in selfassessment and provided with peers' feedback, which, as stated by Shepard (2000), are "central part[s] of the social processes that mediate the development of intellectual abilities, construction of knowledge, and formation of students' identities" (p. 4). In essence, CA pays particular attention to the sociocultural aspects of assessment (McConnell, 1999) and reflects the main Sociocultural foundation stone; that is, development takes place in the social medium with the help of others (Vygotsky, 1978).
Under the scenario of sociocultural theory (SCT), learning is conceptualized as an enterprise shaped and reshaped through social interaction on the premise that the human mind is always mediated through interaction with self or others (Lantolf, 2000). Put it another way, SCT is rested on the idea that higher forms of thinking are formed and enhanced through interaction in a social context, and then they are transferred from the social to the individual level (Vygotsky, 1978). This dynamic process of learning and development best occurs within a metaphoric space referred to as 'Zone of Proximal Development' (ZPD), the distance between the actual developmental level (i.e., what an individual is able to do without others' support) and the potential developmental level (i.e., what an individual is able to perform under the guidance and support of others) (Vygotsky, 1978). In fact, the ZPD forges a relationship between a person who is expert enough in performing tasks and one who has a limited capacity but has capabilities to take part in the process of task performance (Poehner & Lantolf, 2010); the expert plays the role of a supporter who helps learners compensate for the aspects of tasks which they are unable to consider and perform on their own (Anton, 1999). More specifically, the expert temporarily scaffolds the novice so that they can carry out a task that is beyond their current level of abilities. In principle, the mediator provides the learners with appropriate calibrated aids through a mutual dialogue since the accurate diagnosis of learners' capabilities cannot be made but through dialogic interaction between the expert and novice (Aljaafreh & Lantolf, 1994).
Administering teacher-student CA defined by Fahim et al. (2014), this study set out to illustrate the microgenetic interaction between EFL learners and their teacher while they are engaging in assessing EFL learners' pieces of writing to monitor changes that may occur in the writing process.

METHOD Design
This study was aimed to explore the possible effects of CA on improving the writing proficiency of EFL learners. To this aim, following Fahim et al. (2014), the CA was administered in a writing class and taking a sociocultural perspective. The study in particular was attempted to capture the interactions between the teacher and her learners through microgenetic analysis dealing with the process of learning based on the view that focusing merely on the products may cause the minimal changes in the process of learning to be overlooked (Vygotsky, 1978).

Setting and participants
One intact female group of 15 intermediate EFL learners in an Iranian private language institute participated in this study. It should be noted that all students enrolled in the institute were female. The institute was planned to hone L2 learners' communicative skills and covered all language skills. Participants attended the classes two times a week, and their ages ranged from 17 to 23. Further, they all have been studying English in this institute for about two years; however, they did not have any experience of being engaged in the CA prior to the study based on the data collected from them.

Tasks
Seven writing tasks were chosen from the participants' regular textbook, Four Corners, (Bohlke & Richards, 2012) on the following grounds. First, it was hypothesized that task demands were in line with the participants' level of L2 writing proficiency. Second, selecting the tasks from the participants' regular textbook was taken as a strategy to enhance the ecological validity, assessing learners "in situations which more closely resemble actual working conditions" (Gardner, 1992, p. 91). Third, it was postulated that the participants would do the tasks more seriously.

Scoring procedures
An analytic 100-point scale, designed by Jacob et al. (1981), was used for assessing writing tasks. In comparison with the holistic scoring in which a single score is assigned to a piece of writing, the analytic scale assesses writings according to five main aspects of writing. It gives different values to these categories: 'content,' the extent to which the topic is elaborated (30 points), 'language use,' the extent to which grammatical points are addressed (25 points), 'organization,' the extent to which paragraphs are coherently written (20 points), 'vocabulary,' the extent to which new and correct vocabularies are used (20 points), and 'mechanics,' the extent to which spelling, punctuation, capitalization, and paragraphing are considered (5 points). Therefore, such an analytic scale allows teachers to gain a more nuanced understanding of language learners' strengths and weaknesses. Moreover, it affords learners an opportunity to reflect on their writings to overcome their weaknesses and boost their strengths (Brown & Abeywickrama, 2010). Further, analytic scoring lays the groundwork for more points of difference and, consequently, more discussions between the two sides of CA.

Semi-structured interview
To inquire into the metacognitive knowledge of the learners and their views about the nature of the study, a semi-structured interview was arranged between the teacher and CA group at the end of the treatment sessions. Holding an interview in a friendly manner in which learners feel comfortable with the situation to express their opinions freely might afford the researchers with the possibility to evaluate the outcomes of the investigation from the 'emic' perspective; that is, the learners' perceptions of the study (Mackey & Gass, 2015) which might otherwise be difficult to elicit. Simply put, the interview was conducted to reflect the personal attitudes of CA learners towards their experience of being involved in the CA sessions. Further, it should be noted that the interviews were conducted in Farsi, learners' first language. They were then translated and transcribed verbatim and reviewed several times by the researchers to elicit the recurrent and repetitive patterns. This process is called microgenetic analysis.

Microgenetic analysis
Microgenesis, one of the genetics of the genetic model proposed by Vygotsky (1986), is defined as "the moment-to-moment co-construction of language and language learning" (Gutierrez, 2008, p. 2). The genetic model is premised on the fact that the comprehensive understanding of the higher, culturally organized levels of human mental functioning is only achieved through the study of the process rather than the product of development (Vygotsky, 1978). In line with the previous argument, most sociocultural research (Aljaafreh & Lantolf, 1994;Davin, 2013;Poehner & van Compernolle, 2013) perform the microgenetic method since focusing merely on the products may neglect the genetic relationship between the elementary and higher levels of the mental activity and may not provide the researcher with the inner nature of mental development (Vygotsky, 1978). By conducting a microgenetic analysis, teachers can observe the subtle changes that may go unnoticed in a particular course of learning when the students go through the learning process. Moreover, the circumstances precede and follow a change, and the change itself is brought to light through direct and intensive observation in the microgenetic method (Siegler & Crowley, 1991). As a result, microgenetic analysis can assist teachers in diagnosing learners' needs and consequently to tailor their methods of teaching to learners' requirements. In this study, microgenetic analysis is conducted with the aim of examining the internal nature of the dialogic interaction between the teacher and the learners when they joined to share their knowledge of L2 writing throughout the CA.

Procedures
In the first session, the participants were provided with an introduction on how to organize, develop, and support their ideas logically. After giving this briefing, the teacher afforded the group copies of the analytic scale checklist according to which the students were supposed to assess their writings. For the students to learn the procedures of the CA in practice, the teacher provided the students with a sample of writing scored based on the analytic scale. Then, the learners were given a chance to assess the sample based on the same scale and negotiating it with the teacher if there was a difference between the grade given by the teacher and one assigned by them. The learners were asked to write about the topics assigned for each session at home and to assess them based on the analytic scale prior to class. Afterward, the teacher scored the learners' essays according to the checklist. Then, the students were given a chance to compare their appraisal with the teachers'. If learners detected a discrepancy between these two scores, they negotiated it with the teacher to reach a mutually agreed score. The learner was encouraged to argue according to the score distribution in the checklist. This strategy was undertaken to offer the students some incentive to obtain a more-in-depth insight into the qualities of writing proficiency. It is worth mentioning that this procedure continued for seven sessions, and all interactions between the teacher and the learners were audio-recorded and transcribed verbatim in order to track the trend of the learners' awareness about different aspects of the writing proficiency and its possible effect on gains in writing as a result of collaborative dialogues.
It should be noted that the participants were notified that they were taking part in a study, and written consent was obtained from them. Moreover, in order to remove the researchers' effects and bias, the researcher randomly selected the group of participants, not from the classes she taught. In this way, learners were invited to take part in a one-toone dialogic interaction with the researcher out of their regular class.
Also, all the interaction was audio-recorded and analyzed to examine the whole process for any detailed changes that might occur.

FINDINGS AND DISCUSSION
This study focused on analyzing the naturally occurring interactions between EFL learners under investigation and their teacher while undertaking the CA in the classroom context. However, since space does not permit demonstrating the in-depth analysis of the recorded data, a few brief episodes are presented here to document the process of knowledge and awareness co-construction on the social level.
Episode 1 (Extract 1 below) is an example of the interaction between the teacher and one of the learners, Sadaf (a pseudonym), while they were involved in the CA of Sadaf's first composition on "how to study better" (See Appendix for the transcription key). This episode indicates that the learner is initially reluctant to articulate her argument. This might be since she has been grown up in a teachercentered and product-oriented educational background in which learners are not allowed to partake in assessment processes (Fahim et al., 2014). In fact, using an imperative, comparing the scores, and putting a strong emphasis on 'can' (CAN), the teacher, in an earnest attempt, tries to solicit Sadaf's contribution. Simply put, when the teacher assures her that she has the authority to change her scores if she bases her argument on the scoring checklist, Sadaf ventures to express her opinions and challenge the teacher's scoring. As viewed, the mediatory tools helped Sadaf coconstruct the understanding that there was a rationale behind the teacher's scoring of her writing, and it was not a haphazard undertaking. As showcased, Sadaf's incapacity to get the meaning across arose from her difficulty in choosing appropriate words. Through dialogue, the teacher tried to draw her attention to this weakness. In such a situation, language functioned as a mediatory tool in the process of knowledge formation. As Swain (2006) mentioned, the problems were resolved through languaging, 'the process of making meaning and shaping knowledge and experience through language' (p. 98). Precisely put, noticing discrepancies between the scores, the teacher and Sadaf made use of language to join their resources together and remove them.

Extract
Episode 2 (Extract 2) is another example of a small portion of the collaborative dialogue between the teacher and Sadaf on her fourth writing. In Episode 2, the microgenetic analysis showed how knowledge is co-constructed through collaborative dialogue. Sadaf, who was once unwilling to discuss discrepancies with the teacher in the first session, is now more willing to initiate a discussion with the teacher over a discrepancy she has noticed. As seen in turn 3, she tries to seek the teacher's rationale underlying the assigned score and then continues to develop her argument based on the scale. Finally, her convincing argument persuades the teacher to reconsider her initial sore and consequently change it.
Moreover, the following episode (Extract 3), which is a part of a mutual interaction for the composition in the seventh session, attests to Sadaf's promotion in writing, especially in content and organization, by engaging in problem-solving interaction between a more knowledgeable person and a novice. It came to light that Sadaf has managed to rather appropriate the knowledge coconstructed in the previous dialogic sessions. This resonates with the view that involving learners in the dialogic interaction, in which they perform a prominent role 'in shaping the verbal agenda of classroom discussion, can help them to secure improved attainments in outcome' (Skidmore, 2006, p. 511 As vividly portrayed in these episodes, because of the knowledge co-built through the first session where the teacher highlighted Yalda's mechanics error and articulated the reason of grading the mechanics section 4 (turn 9 in Extract 4), Yalda's awareness of the language she produced was raised in the last session illustrated in Extract 5 (turn 1). Additionally, regarding the language use section, the scores indicate Yalda's progress in grammar. T: 3 scores ((Laughs)).

Extract
This episode (Extract 5) can be taken as Yalda's advancement in writing due to conducting the CA from a sociocultural perspective. In principle, getting involved in the dialogic interaction and providing with the assisted scaffolding help (Swain, 2000), Yalda arrived at a richer understanding of the problems with her performance, gained a greater insight into her abilities, and developed a greater awareness of the language produced. The most striking is that although Yalda showed a strong tendency towards getting the best mark under the influence of the product-oriented perspective and consequently selected the scores from the highest ranges of the checklist in the first session, she gradually came to know that the learning process takes precedence over the learning product in the CA. Consequently, the discrepancy between the score given by Yalda and that given by the teacher was decreased in the last session.
To put more flesh on the bones of our argument, some episodes from the teacher and Mina (pseudonym) are presented. The extracts are included to show the microgenetic growth of the student. Extract 6 is provided to display the conversation between the teacher and Mina on her first performance.  This extract displays that Mina backed up her position by relying on the checklist and her writing. Like other extracts, this extract opens a window into the collaborative dialogue in which the teacher and Mina collaboratively created opportunities to resolve the existing problems by using language. In fact, the language was mediating to construct knowledge and awareness. Further, the teacher, by using language, aligned the pedagogical and evaluative purposes and promoted learner involvement. In fact, in line with AFL tenets, the teacher tried to provide a pleasant environment for Mina to express her views and, in so doing, occupied an influential role in the process of learning. Extract 6 is drawn from the conversation of the teacher and Mina on her fourth attempt.
Extract 7 is another case for the interaction between the teacher and Mina on her last composition.

Extract 7 1 M:
The only point of contrast is the score of the language use section. Extract 7 indicates that although there were some minor problems in her essay, Mina developed some control over her writing ability after being involved in a seven-session procedure within which she had been given the chance of reflecting on her performance, expressing her idea about the assessment procedure, taking control over her learning, and sharing the knowledge she gained through conducting self-assessment. More importantly, she had been provided with the teacher's calibrated mediations on her errors. Regarding the language use, what was the potential level of Mina at the first session became her actual level at the last session. It appeared that Mina profited from engaging in the collaborative dialogue established through CA.
To recap the discussion, from a sociocultural perspective, the study's microgenetic findings offered some evidence in favor of Vygotsky's (1978) statement that development takes place in the social encounters as a result of the mediation of more capable others and its transmission to the individual level. In particular, getting engaged in the dialogic interaction with their teacher, it seemed that the EFL learners made use of the checklist as an objective meditational tool and the language (i.e., both social and private speech) as the most powerful mediatory tool to achieve an understanding coconstructed with the teacher through the moment-tomoment verbal interaction while conducting the CA. Worthy of note is that, through performing CA, it appeared that the learners moved towards selfregulation and took more responsibility. They obtained opportunities to internalize the co-built knowledge and metacognitive awareness constructed by mediatory tools such as the scale and dialogic interaction employed in this study.
Additionally, the findings can be justified based on the literature on collaborative dialogue (e.g., Kowal & Swain 1994;Swain 2000Swain , 2006. As pointed out by Swain (2000Swain ( , 2006, language can be used to deepen our knowledge and awareness of tasks at hand. More precisely, engaging the learners in collaborative dialogues with their teacher sets the ground for the learners to render their initial thoughts to actual words; then, they are benefited by an occasion to reflect upon their initial hypotheses about their writings under the guidance and support of a more capable teacher and by resorting to the provided scale in order to test out and refine the hypotheses. Furthermore, it should be mentioned that the points of difference between the learners and teacher's positions on various aspects of the writing tasks are worked as cognitive conflicts 'in which different views, theories, and beliefs are raised to a conscious level through talk' (Tocalli-Beller & Swain, 2005, p. 23). In reality, establishing a dialogue in which conflict perspectives are presented, the teacher provided learners with the opportunity to weigh their ideas against different perspectives and raise their metacognitive awareness of their language problems. As Limon (2001) claims, the collaborative dialogue is construed as a necessary factor for the cognitive conflict to promote learning.
More to the point, as Lee (2017) pointed out, taking advantage of the AFL benefits, it seemed that learners obtained the metacognitive awareness of the tasks at hand, an insight into the components of good writing, an understanding of their weaknesses, an understanding of their abilities to self-assess their performance based on reliable and reasonable criteria, and implementation of effective strategies to get on with the task of writing. By way of illustration, Sadaf confessed that The experiment was remarkably effective. Previously, I thought that I was not able to write even some successive sentences.

CONCLUSION
Findings obtained from the microgenetic analysis suggested that learners benefited from the collaborative dialogue in which they were engaged while assessing their writing tasks. The dialogic interaction afforded the learners the chance to present, discuss, and test their ideas and consequently enhance their awareness of the writing tasks (Swain, 2000(Swain, , 2006. Additionally, learners first used calibrated assistance, the scale, and the language to co-build knowledge with their teacher. However, their dependence on the mediational tools gradually decreased over time. They were better able to self-regulate the writing assessment (Vygotsky, 1978). Especially worthy of note here are the results of the interviews conducted after the treatment sessions. In principle, in pursuit of involvement in mutual interaction, learners' metacognitive awareness of the writing tasks, their abilities and weaknesses in writing, and the strategies which could be used to grapple with their problems increased.
Despite the reported findings, some limitations regarding the study need to be acknowledged. First, the number of participants was limited; therefore, the generalizability of the findings should be done with some caution. Consequently, further investigations on the CA need to be done with a more significant number of participants and learners of other proficiency levels. Second, this study did not attempt to examine the effects of CA on specific components of writing provided by the analytic checklist developed by Jacob et al. (1981). Thus, further research can be directed at investigating how different aspects of writing might benefit from engaging in CA procedures.

REFERENCES
Aljaafreh, A., & Lantolf, J. P. (1994). Negative feedback as regulation and second language learning in the zone of proximal development.