Patterns of Pair Interaction in EFL Dyadic Talk: A Study of Peer Responses

Introduction

Social psychology and general education have extensively dealt with the nature of group interaction. In education, for example, there is a large volume of research on cooperative groups (Johnson & Johnson, 1990, 1994; Sharan, 1990; Slavin, 1990). This research area in education has identified essential elements of cooperative groups, the merits of cooperative learning, how learners relate to each other within the groups, and the factors—such as academic status or gender—that affect these relations (e.g., Cohen, 1994).

In that line of research, second language (L2) settings have also dealt with group work, and more commonly on pair work. These L2 studies have primarily focused on the linguistic interactions that take place between participants, that is, “negotiation of meaning” (Long, 1983) and the factors that affect the quantity of these negotiations (e.g., Gass & Varonis, 1985; Pica, Holliday, Lewis, & Morgenthaler, 1989; Pica, Holliday, Lewis, Berducci, & Newman, 1991). However, the above studies fail to address that negotiation of meaning or topic and learning outcomes may be subjected to the peer relations as well (Banbrook, 1999). Thus a growing number of studies have begun to examine more closely the dynamics of group and pair behavior in L2 contexts (e.g., deGuerrero & Villamil, 1994; Donato, 1988). These studies have shown not only that there are differences in the patterns of pair behavior but, more importantly, they suggest that some patterns are more conducive to learning than others. As a case in point, Nelson and Murphy (1993) found that learners who interacted in a cooperative manner were more likely to use peer suggestions to revise their writing than those who interacted in a defensive manner.

In addition, a small number of studies have investigated the peer–peer interaction across proficiency levels (Kowal & Swain, 1994; Leeser, 2004). The significant questions are whether more proficient ones benefit from working with less proficient peers or less proficient learners benefit from receiving more proficient learners’ help. Do they benefit from peer–peer dialogue if they are in the same proficiency level? Is this consistent across different groups? Accordingly, research studies need to examine how the learners having thesameproficiency levels interact with higher, equal or lower proficiency peers. This study was conducted in an attempt to provide insight into underlying dynamics of this under-explored area.

When learners engage in pair or small group work, they take more control over their own behavior, and each learner interacts with others on the basis of his/her own incentives, learning styles, academic experience, purposes of the classroom and so on. Moreover, a pair/group inevitably consists of multiple persons, who will have multiple goals or orientations to the task. Thus, it is essential to study the interaction not only from a cognitive viewpoint, but also in its totality, including learners’ agency and the context the learners are involved in to study the complicated language learning processes in the classroom.

An important issue of relevance in this study is that EFL students with a common language and culture may behave differently from the ESL students. As there still is a paucity of research in the Iranian EFL context, this study yields both practical and theoretical results and can be applied to fields of language teaching particularly writing skill.

Theoretical Framework

Peer–Peer Interaction from a Sociocultural Perspective

The main premise of sociocultural theory is that learning is social and a semiotic mediated process, happening first on the interpersonal level and then on the intrapersonal level (Vygotsky, 1978). The concepts of mediation and Zone of Proximal Development (ZPD) are criterial. Human cognition is mediated especially by language as the semiotic tool and learning within ZPD is perceived to occur in the expert–novice dialogic interaction, in which an expert provides contingent and graduated assistance to a novice. Lantolf (2000) claims that ZPD is ‘‘more appropriately conceived as the collaborative constructions of opportunities for individuals to develop their abilities” (p. 17). A good number of studies such as Donato (1994), Foster and Ohta (2005), Ohta (2000a), and Storch (2002a, 2003) have so far investigated how learners support each other within their respective ZPDs.

The collaborative effort leads to the co-construction of knowledge and meaning in the situated social context. Donato (2004) considers collaboration beyond simple input–output representations and admits the significance of the collective learning and mutual relationships among learners. In peer–peer dialogue, learners collaboratively ‘‘engage in problem solving and knowledge building” (Swain, 2000, p. 102). The interaction has thus a dialogic and dynamic nature through which we can discover how the gap between language use and language learning is bridged with the two co-occurring in the same activity. Or, as Swain (2000) maintains, peer–peer collaborative dialogue mediates L2 learning.

Activity Theory and SLA

Vygotsky’s concept of activity theory states that to understand the action that takes place perceiving the underlying motives is essential. According to Engestrom (1987), since each activity is joined by collective components, and all components influence each other, and as motives are realized in goal oriented actions, the activity is dynamic. Hence, the activity will vary based on the motive, goals and specific conditions.

Activity theory observes social activity as a framework for understanding the dynamic relationship between individual, social, cultural, historical, and institutional contexts. Thus, language learners’ individually-differentiated behavior in the classroom can be analyzed from an activity theory perspective.

In a study, Coughlan and Duff (1994) demonstrated how a task is seen in a different way by each individual and becomes a different ‘activity’, when it is actually performed by each learner. Gillette (1994) also upholds the Vygotskian principle that the initial intention of activity defines the outcome of engagement in the activity. Learners are quite active in their learning, and act in ways that cannot be predicted from the tasks which they are assigned to perform. Accordingly, Breen (1987) describes how through their reinterpretations during engagement with the task, learners turn ‘task–as-work plan’ into the ‘task-in-process’ in terms of its objectives, content, procedure and learning situation.

Peer–Peer Interaction and Patterns of Pair interaction

Of the main studies into the nature of pair interaction in an adult ESL classroom, Storch’s (2001a; 2001b; 2002a; 2002b) longitudinal investigation can be named. Based on her pair talk data, four patterns of interaction can be identified amongst pairs. In the collaborative pattern, both learners work together throughout the task completion process and help each other. Dominant/dominant pairs, on the other hand, show reluctance or inability to involve with each other’s contribution. Dominant/passive pairs involve a dominant participant who controls the task with a disciplinarian stance, and a passive peer who maintains an unassertive role. Finally, in expert/novice pairs, the more knowledgeable learner (expert) actively encourages the less knowledgeable learner (novice) to engage in the task.

While collaborative pairs assist each other throughout the task, dominant/dominant pairs are unlikely to value each other’s contribution. Dominant/passive pairs involve one learner taking control while the other largely concedes, and finally the last type of pattern is present in expert/novice pairs which consist of a more knowledgeable learner helping a less knowledgeable partner. Among all the four pattern types, collaborative pair dynamics and expert/novice pair dynamics may be more facilitative of L2 learning.

Language Proficiency and Peer Interaction

Previous researches have looked into the impact of proficiency on a range of interactional features, including negotiation of meaning and focus on form, scaffolding, and patterns of communication. One of the earliest studies of the role of proficiency in learning was Gass and Varonis's (1985) study of negotiation of meaning in peer interaction which included dyads that were mixed and matched for proficiency. Their findings showed that it was not the level of proficiency that influenced negotiation of meaning, as both high and low proficiency dyads had similar amounts of negotiation. Having high- and low-proficiency learners in the same dyad, however, led to more miscommunication and more attempt to resolve miscommunication through negotiation.

In Yule and Macdonald's (1990) study of peer interactions, communication was strongly influenced by whether the high- or low-proficiency learner was cast in the dominant role for the task. Their task required one learner to give map directions to their interlocutor. When high-proficiency learners were the direction givers, there was little negotiation, and the low-proficiency learner participated very little in the conversation. On the other hand, when the low-proficiency learner gave the directions, both interlocutors contributed more, as they collaboratively engaged in solving the conflicts built into the task. Iwashita (2001), in her study on modified output in interactions, predicted that the high-low proficiency group makes the most modified output, but contrary to her hypothesis, the results showed that the high-high group produced the most modified output.

Some research on scaffolding indicates advantages to mixed-proficiency groupings; others observe that these may stifle the less proficient learner. Ohta (2000b) discovered that more proficient learners helped less proficient ones in coping with the linguistic demands of the task. However, other evidence indicates that as the proficiency gap between high- and low-level students widens, the stronger student often excludes the weaker student from the interaction (Kowal & Swain, 1994). Lee's (2008) study of scaffolding on peer interactions among higher and lower proficiency learners reveals that the frequent use of scaffolding behaviors assisted learners in generating structures they could not generate on their own.

The contradictory results of previous studies demonstrate the inherent complexity of the social phenomenon of peer interaction, where different factors interact to shape the communication and necessitate careful examination to enlighten if different groups of students interact in similar ways during peer response sessions. The present study attempts to find out which patterns of interaction are present in intermediate EFL students paired with partners having similar or different language proficiency levels when engaged in peer reviews. More specifically, the purpose of the present study is to explore patterns of dyadic interaction between intermediate learners paired with partners of different (L2) proficiency levels in an EFL writing classroom.

The Study

Participants

Based on the proficiency test results (TOEFL), the beginner, intermediate, and advanced participants in the essay writing classroom were identified (Table 1). At first, the participants were administered the paper based TPEFL and through a conversion process designed in Canada the scores were transformed into the equivalent levels as can be seen in the table below. VEC represents Vancouver English Center scale used to produce the TOEFL equivalency table.

The participants were all females and their ages ranged between 20 and 27. They were classmates and knew each other and thus, seen eager to work with their partners. They were allowed to self-select their partners so that they would feel at ease in negotiating over the completion of the given tasks. As the focus was on intermediate partnership, the data of 12 intermediate students (i.e. 6 pairs) interacting with advanced, intermediate, and beginner partners were studied. They were required not to use their L1 while performing the task.

Table 1. Participants

*Names are all pseudonyms

Advanced Writing Classroom

This study was carried out in a second-year essay writing classroom in a non-state university. The chosen course, essay writing, meets part of the university’s requirement for EFL students after having passed the introductory course of writing. The objective is to help students of English to write clearly and concisely in an academic setting for a variety of purposes. Writing, rewriting, and sharing writing with others, learning strategies and techniques (e.g., editing and analyzing errors) for taking responsibility of the quality of written work are emphasized and the fundamental principles of writing a good essay (organization, support, unity, coherence) are practiced.

As this study was supposed to deal with the interactions in the classroom, this writing class provided an opportunity for them to discuss their feedbacks on their writing. As such, students carried out writing assignments throughout the semester which allowed them to practice writing, and revising in different patterns of development. Upon the completion of their first drafts, students would participate in peer response sessions, receiving feedbacks to incorporate in their subsequent drafts.

Procedures

This case study aimed at carrying out an in-depth analysis of interaction processes as a natural phenomenon in its natural context, drawing on the perspective of the participants involved. First, one of the researchers who was also the instructor of the course communicated her expectations about peer responses, and guided the students in a training session to practice giving feedback on writing.

As for the study, peer response data were collected and observation notes were taken during one week to the end of the semester. Students exchanged papers and silently read their peer’s work, while making brief notes on the drafts. When pairs were ready to begin discussing their papers, they turned on the recorder of their cellphones that was placed on the desk between them, and were asked to speak as they normally would during a peer response session.

Students negotiated whose paper to discuss first, and then switched roles. While they were verbally discussing each other’s papers, the instructor circulated around the room, quietly listening to their conversations to ensure that they were on task. Occasionally students asked the instructor questions on the accuracy of the language used in papers. In these cases, the instructor encouraged students to arrive at the answer by working together.

The peer response session lasted for almost the duration of the eighty-minute class time: five minutes preparing students to the task, twenty minutes for reading and making notes, and fifty-five minutes for discussing the feedback on each other’s paper. However, students were allowed to move from one part of the procedure to another as they were ready, so there was variation in the amount of time spent reading and reviewing each paper.

Immediately after the peer response session, the recordings were transcribed checking students’ writing drafts and noting the segments that seemed important for further understanding of how students shared control over the direction of the peer response task. It was attempted to represent the talk as it occurred. The guide to transcribing symbols by Storch (2002a) was employed in this study (see Appendix A). The transcripts, as the main source of data, were approached with the general dimensions of mutuality and equality as guidance, the two areas of Storch. For example, discussions of a revision that seemed collaborative, arguments or disagreements about revisions or about the task, and areas where the writer did not seem to understand the suggested revision were marked. Any parts of the recording unnecessary to listen to especially the parts where participants discussed things other than the peer responses were excluded.

Then during the next two days, they had their stimulated recall interview of approximately 45 minutes with the researchers. In addition to transcriptions of peer response sessions, stimulated recall interviews with participants provided an additional data source. These interviews prompted participants to recall thoughts they had while performing a task or participating in an event.

How they felt about giving or receiving feedback at particular segments and whether they agreed with the partner’s suggestions for revisions were the issues discussed. The entire peer response recording was played during the interview. To ask them questions about the peer response session, the audio clip had to be paused, but they could also pause the recording any time they needed to comment.

Results

Data analysis consisted of two stages. In the first stage, pair talk data were analyzed for the patterns of dyadic interaction and the salient traits that characterized these patterns. In the second stage, the consistency of the results obtained from pair talk data were compared with the results of the interview data gathered from individual students in order to trace the effects of the assumed types of patterns of dyadic interaction.

Patterns of Dyadic Interaction

In the first stage of data analysis, Storch’s (2002a) patterns of dyadic interaction were used as a guideline. The coding scheme that emerged from examining the pair talk in Storch (2002a) describes pair interactions based on the extent to which learners engage with each other’s suggestions (mutuality) and the extent to which they share control over the direction of the task (equality). As Storch (2002a) notes, categorizing peer interaction is, by its very nature, imprecise.

Students’ interactionswere reviewed and assigned one of four patterns of interaction (collaborative, dominant/passive, expert/novice, or dominant/dominant). During the analysis phase, the transcripts were explored to determine how students shared control over the task (equality) and engaged with each other’s suggestions (mutuality).

Here instances of the transcripts are provided to demonstrate the observed patterns. In the first excerpt, Ahoor, the advanced learner, and Armina, the middle intermediate one, were reviewing Armina’s writing paper together. Although there were proficiency differences, Ahoor tried to make the response session stress-free and this friendly manner assured her partner to take part comfortably. Thus, Armina tried to build confidence and show higher ability in language. We saw the evidence manifesting the expert-novice pattern.

Excerpt 1

• • Ah: OK. First let’s explain mine. It was difficult to choose a topic. I thought about it and chose nursing home. And to limit more, I chose Mercy Killing...
• • Ar: what does mercy killing mean?
• • Ah: It means that when a guy is in a very bad pain he should have the right to stop his life… He should have the right to ask someone to help him to stop his life... How the nurses who do not give this right add to [the problem
• • Ar: You mean] there are a number of nurses who participate in killing.
• • Ah: Yeah exactly. Doctors and nurses. Well [done!
• • Ar: The story] that if someone is in pain kill [him
• • Ah: Yeah.] Now your paper. I understood yours. But write a title which shows your [effect.
• • Ar: Why?] It was the title of the textbook…
• • Ah: It was very broad. Different from yours…
• • Ar: OK…
• • Ah: you put good supports… You talked about relaxation. There were positive points of mental… You should say my favorite music genre mentally elevates me or mentally positive [4]
• • Ar: yeah
• • Ah: or is mentally positive to me… That’s it…
• • Ar: Right! I made a mistake. I wrote the title of the textbook.
• • Ah: Of course you had changed it a little. You said my favorite music genre. But I said that the effect needs to be added. [5]
• • Ah: In a paragraph it is better to see agreement…
• • Ar: OK…
• • Ah: There are lat. A lot [what?
• • Ar: It] refers to kinds.
• • Ah: OK but it should be mentioned. [4]
• • Ar: There should come something. [5]
• • Ah: yeah… Something should come to complete it. After that, it was very good. [5] Here was good. As a motivator. You gave a formal stance to your article…
• • Ar: Here that I said as a musician?
• • Ah: Yeah. You give formality to your writing. It means that you are not talking in air.
• • Ar: (Laugh)
• • Ah: You have worked a lot as a musician. (aha). Here is a very good support. Great. That is why I said use it in the title too.

As observed above, from the beginning, the friendly atmosphere is established. Ahoor first started to talk about her own writing and shared the difficulty she had in choosing the topic (line 1). When Armina asked about the meaning of mercy killing, Ahoor in a very friendly manner described the term (line 3). The expert-novice pattern continued and then after this warm beginning, Ahoor was starting to discuss the partner’s paper. In line 8, Armina tried to seek reason for having some changes in her title proposed by Ahoor. Having given a positive reaction (line 11), she encouraged the partner not to guard against her viewpoints, to admit mistakes (line 14) and thus actively participated in the talk.

Again in the following lines the same pattern was going on. In a language indicating warmness, positive reinforcements were given and then the needed revisions were softly suggested. Ahoor tried to explain a linguistic point (line 16) and encouraged the partner to participate (line 18). Positive reinforcement was recurrently provided (lines 22, 24, 26) when the problematic parts were mentioned (lines 16, 20). It can be said that Ahoor contributed to all aspects of the text: generating ideas, suggesting how to organize the ideas and deciding on how to best express them.

In Excerpt 2, when Armina tried to review Ahoor’s writing, the pair showed signs of the collaborative pattern. They worked together on all parts of the task and were willing to offer and engage with each other’s ideas.

Excerpt 2

• • Ar: OK. About your essay. The first point I found was that the title, the effect was controversial. Right? The title, I think, does not convey the effect…
• • Ah: yes, the effect might be [5] Nursing homes suppressing euthanasia increasing or creating more problems...
• • Ar: Yeah...
• • Ah: They are making more problems. Instead of saying that suppressing euthanasia is part of a problem. So what? [4] It does not show that there is an effect. What is your idea if I say that it is making more problems?
• • Ar: OK. Another point… Quotation. For example, about Mark. Here Reference is needed...
• • Ah: Yeah. I haven’t put them at the end. There is also another citation. [5]
• • Ar: OK. So here you put dash.
• • Ah: This is not dash. It was typed incorrectly. It should be corrected.
• • Ar: Ok. This one that you put for this sentence. They usually put dashes when the sentence …. is unrelated. [4]
• • Ah:  somehow [if
• • Ar: Like] this
• • Ah: (Aha). You are right. For example, it was possible not to bring it here.
• • Ar: you can omit it.
• • Ah: yeah. Or I can say it in another way. [7]
• • Ar: Yeah. It was very good. It took time to understand yours. Because you wrote in advanced level. For me, at least, it was more. I couldn’t find problems. The points I mentioned do not decrease its value. It was informative. You wrote well. About layout
• • Ah: Oh, title. I should have paid attention to the title.
• • Ar: Yeah. The title. More… (um) … In layout… Here there is a title in this page and you should have used the enter key to put it in the next page. It was left here.

In the second excerpt, Armina tried to play the same role that Ahoor had in reading her paper. She expressed her idea of the title, and Ahoor warmly welcomed the comments (lines 1, 2). During the review, it was the help of Ahoor and her positive feedbacks that motivated her to confidently participate in talk. When they were about to finish, Armina herself confessed to the complexity of Ahoor’s paper affirming its being beyond her level (line 15). Anyway, she attempted to show good ability of herself.

Their talk was highly cohesive and it was created by the learners repeating and elaborating utterances or by completing ideas. They had moderate equality and mutuality and engaged constructively with one another’s suggestions. These findings suggested that the successful communication in mixed-proficiency pairs could be affected not only by the proficiency of the learners but also by affective factors.

In Excerpt 3, we had an instance of a pair work which had a middle intermediate, Niloofar, and an advanced one, Aida, paired together, reviewing Niloofar’s writing. However, one member of the dyad, Aida, established a more dominant role and as a result little negotiation took place.

Excerpt 3

• • A: Right. Support! Right. In some parts you have given unnecessary information. You have some unrelated points here. You do not need to say by people here. And, here. Use passive structure. I think it is better to omit structure. You forgot to separate your paragraphs. Why?
• • N: Laughing.
• • A: Sometimes you forgot capitalization. It has many grammatical errors. Study the paper carefully. I think you shift a lot in this writing. You have many different things. As a reader, I like to see, know what it is about. Separate paragraphs. Here, in the case, add ‘the’ and here, apples and quinces which may be. Add ‘which’. It needs to be revised grammatically. Here delete ‘structure which’. You do not need to say it. ‘Taste’ is the verb in this sentence. Shift to the next paragraph here. You didn’t use enough examples. It needs many examples. Omit this part. Before the last paragraph.
• • N: mmm… In the paragraph of the conclusion
• • A: OK. Let me see. Supports. Shorten your supports.
• • N: OK.
• • A: shorten your supports. If it comes after introduction... you can join it with the introduction. Your introduction is good but …. See you can replace this one with the other one. Write more about similarities. In most of the paragraphs you have the description of the physical structure. Talk about the chemical structure too. Very good. It is a good clincher. I prefer this. Right. You didn’t have any citations. Make the changes. OK?
• • N: OK.

The above excerpt is typical of the talk found in dominant/passive pairs. Aida took control of the task as an authoritarian stance; Niloofar, as the passive participant, contributed very little to the task. Aida expressed her view on the content and structure of Niloofar’s writing. Her partner was not inclined to express any idea and it appeared that Aida did not try to involve her either, thus, she continued with examining the rest of the paper. Later, they reviewed Aida’s paper. Both kept their same roles. Aida remained dominant and Niloofar passive.

Unlike excerpt 1 where Ahoor, as the more proficient learner assisted the less proficient learner in coping with the linguistic demands of the task, this was not the case in excerpt 3 where we saw the same proficiency gap between the high- and intermediate students, but Aida disregarded her partner. Having had a dominant role, she impeded scaffolding and hampered collaborative learning opportunities. On the other hand, peer interactions among Ahoor and Armina revealed frequent use of scaffolding behaviors which motivated Armina to try to use language similar to her higher proficient partner.

In Excerpt 4, Hani as a middle intermediate and Mozhan as an upper beginner were reviewing Hani’s writing paper together. The learners seemed unwilling to consider each other's suggestion. The talk showed high levels of conflict. It could be identified as an instance of the dominant/dominant pattern. Although both participants contributed to the task, there was a high level of disagreement. The two learners frequently rejected each other's suggestions. There was no evidence of collective scaffolding in the pair talk, having behaved as if they had responsibility only over their own contributions. Although both members of the pair contributed to the task, they did not engage or were unwilling to engage with each other's contributions.

Excerpt 4

• • M: First you should have an essay made of five paragraphs. It should have an introduction, three central paragraphs and a conclusion. Your introduction should have a motivator. It has but it does not have coherence. Your motivator is not complete. You should explain about what you want to say. You said …
• • H: About my homeland.
• • M: About what? About what of your homeland? OK. [It’s not complete.
• • H: I said in the Middle East,] Iranians are well-known for their generosity. Then I explained culture and other things
• • M: OK [so
• • H: is there] any problem?
• • M: No. But if you explained more, it was so much better. Next, you have a thesis statement. You said that it is about your homeland. About what? Which aspect of your homeland?
• • H: About ... In general, everything about my homeland. [Poets, history
• • M: So you said] about your homeland? OK? What? The effect of on your home-
• • M: Let’s move forward. We said that the thesis statement should be … what? It should include our main idea. The three main ideas of your paragraphs are as you said you are going to talk about your homeland. It is not completely true. You talked about your culture. The effect of culture on what? [It’s not clear.
• • H: I also talked about customs] here. And I also talked about the generosity of people here.
• • M: OK. But you should have mentioned the effect of what on what. And you didn’t say. OK? After that there should be a blueprint. It is not important that there is a blueprint or not. And you do not have it. [And then
• • H: OK but just that] I mentioned that I wanted to talk about for example clothes I think I have the blueprint but the point is that I have not completely talked about it. I said it very well. I said traditional clothes, behavior
• • M: These are central points
• • H: I felt that it was said to us that we can explain them in the central points.
• • M: [We will see.
• • H: I mean] what we put in several categories can be explained in the central paragraphs. We should ask the professor later.
• • M: The professor said that it is not very important if there is n-

In this episode, the main issue of disagreement seemed to be whether or not Hani had written in the pattern of cause and effect. Besides, it seemed that there was disagreement on the structure of the introduction, particularly the blueprint. Mozhan began the interaction in a confident voice and asked Hani, “which aspect of your homeland?” (line 3) and Hani responded “in general, everything about my homeland-poets, history” (line 8). Mozhan repeated her question in different wordings (line 9) and demanded explanation and at the same time Hani tried to defend herself. The explanation was inadequate.

In this excerpt, each participant clung to her own view and no agreement was reached. Mozhan continued the discussion by saying “it is not important that there is a blueprint or not” (line 12). Here, though Mozhan believed that there was no necessity for the blueprint but Hani again tried to express her disagreement. The discussion was going on while no consensus was reached. Hani ended the episode by saying that “we should ask the professor later” (line 17). These two were both engaged in trying to control the direction of the task, but were unable to reach consensus, having exhibited the high equality but low mutuality that characterize the dominant-dominant pattern. They also kept the same roles when later they reviewed Mozhan’s paper.

In the next excerpt, it was observed that Shohre, the middle Intermediate and Azar, the upper beginner were reviewing Shohre’s paper. When Shohre saw Azar unwilling to take turn, on her request, she decided to act as dominant for the rest of the pair work. In the following, they were working on Shohre’s paper.

Excerpt 5

• • A: I haven’t understood yours.
• • Sh: Nothing. You haven’t understood anything.
• • A: No.
• • Sh: I wrote that I am an English teacher. I am very happy for the job I selected. I believed that everyone after finishing school or university must find a job, must choose a job that s/he likes. Having an interest in the job is more significant than salary. I mean that if you like your job, you need no more reason to wake up soon in the morning. OK? If you don’t like your job, it can make stress or even health problems. I am an English teacher. I love my job. I teach in university and school. I have many students. OK? They are very polite students. I enjoy teaching them. Now there is a question. Why I have selected this job? There are many reasons. [4] One is that…

In excerpt 5 above, Shohre, who positioned herself as a proficient learner, was talking about her own paper and Azar, who took the passive role willingly, demonstrated little engagement. It was not clear whether Azar understood her as she made few contributions. This episode, which is an instance of a dominant/passive one, displayed low equality, as Shohre was controlling the direction of the interaction, as well as low mutuality, because there was no evidence of engagement or collaboration.

In Excerpt 6, Shohre, the middle Intermediate and Azar, the upper beginner, were reviewing Azar’s paper. As an expert, Shohre administered the task, made more authoritative utterances, while the novice partner assumed a passive or peripheral role.

Excerpt 6

• • Sh: You see; you have no subject here. You should have your subject. What was your subject?
• • A: That is the soccer is full of controversies …
• • Sh: It should have a subject first. Next, first of all, you need an introduction. First you should explain about soccer, how soccer is and the rules. OK? You don’t have an introduction. OK? [5] And again for the second paragraph. [6] OK. Punctuations. You have used punctuations well… Comma, full stop. [6] Right? [4]  
• • A: The question mark? Here.
• • Sh: I didn’t understand this part well. What did you want [to say?
• • A:    Fight], Combat…
• • Sh: They make? [5]
• • A: It makes a fight. [4]
• • Sh: ah [8] Wait. Betwe-You haven’t explained well. You have given a very short explanation. [7]
• • A: um...
• • Sh: When you talked about football, you should say that one of the reasons, for example, is they do not allow women to go. Why? [6] Because it is full of controversies. You haven’t explained. You just gave titles. [5] What is this one?
• • A: The referee’s [4]
• • Sh: What’s it? [7]
• • A: I think they are troublesome for the team…
• • Sh: You should bring it at the end. They make a lot of [4] with a lot of …for their team. What is this? [6]
• • A: um. [4]

Although Shohre offered chances to Azar to participate by questioning her (lines 1, 3, 5, 7, 11, 13, 15), her involvement in the task decreased as she became reluctant to say anything in front of her expert partner (line 16). In this pattern, the distribution of discourse contribution was not symmetrical, the expert contributed more and the novice learner became more and more quiet. This pattern has low mutuality and equality. In this study, the expert/passive pattern of interaction is considered to be a non-collaborative orientation.

In the following excerpt, Mona and Bahar are both middle intermediate learners. They were reviewing Bahar’s writing paper. Mona seemed to assume the role of the expert, led the task, and tried to make Bahar involved in the interaction.

Excerpt 7

• • M: Now about your essay. If you change it to the cause and effect essay type, it becomes a very good essay. For example, here, you talked about the educational context at that time. Very good. You can talk about loneliness. 
• • B: I tried to use the comparison strategy.
• • M: You wrote about the Style of writing of George. Yes? What is his style? [5]
• • B: Realism. I used examples from the novel of Silas Marner…
• • M: Good. You can also use this and talk about the effects the novel has on you. Have a cause and effect essay. For example, you see when someone is religious does not necessarily mean that he is a good man. Understand me?
• • B: No, I cannot understand you.
• • M: See. Everyone by any belief can be a good person. You can infer this from this novel. For example, you can talk about the effect of psychology. (Um) … or, you can write that this person became a symbol for me. I can understand many things from this person. Compare this person.
• • B: Because of the similarities?
• • M: Yes. You can mention in the introduction that this person has similarities to me and that is the reason he has effects on me. It attracts me. Right? [4]
• • B: (Um).
• • M: Very good. A conclusion has two parts: a reworded thesis statement and a clincher. What are they?
• • B: (Um)
• • M: What does a reworded thesis statement mean?
• • B: It’s clear. It rewords the thesis statement.

The pattern displayed in the above excerpt is medium equality, because the expert student had more control over the task than the other. The expert student directed the task in order to ensure that the novice understands how to implement comments during revisions. The level of mutuality was moderate. She did not impose her views (lines 1, 5, 7, 9) but rather tried to provide explanations and invited Bahar’s contributions (lines 3, 11, 13), and Baharh apparently confirmed her and repeated the suggestions made (lines 4, 8, 14).

Mona did not try to look from the point of view of Bahar, but instead tried to inject her own personal ideas into her writing even though they did not seem to be completely true. It sounds that Mona just tried to play the role of the expert and looked beyond her actual proficiency level. In this pair, despite the ongoing encouragement of Mona, Bahar’s involvement decreased over time as she became reluctant to take part. It has a non-collaborative orientation and is considered to be as the expert/passive pattern of interaction. The pair then went on to review Mona’s paper while they still kept the same roles.

In excerpt 8 below, Parisa and Saba are both middle intermediate learners. They were reviewing Saba’s writing paper. This is an example of the type of interaction found in collaborative pairs. Here, the two participants contributed jointly to the composition and are willing to engage with each other’s ideas.

Excerpt 8

• • P: So let’s start with your essay. The title has problems. First it should be functions. And what does it mean? Element devices? [6]
• • S: Oh, sorry. Electronic devices. I mean electronic devices. I wrote it wrongly. [5]
• • P: (Aha) OK… And I also think that it is better to have changes in the title. Right? But I do not know exactly what you should say. [7]
• • S: The title can be electronic devices… Just electronic devices. [5] Or I can say functions of electronic devices in different places.[7]
• • P: (Um)... Maybe yours is fine…
• • S: Yeah. I think functions of electronic devices … has [no problem
• • P: OK, fine]… In the next paragraph you talked about office electronic devices… And in this paragraph, you have electronic devices used in industry. [4] That is a very long paragraph… You can talk about agriculture in a separate paragraph.[5]
• • S: You are Right… That is very long… In another paragraph. What do you think if I have automobile industry in a separate paragraph too. [6]
• • P: Yes, I think it is better. [5] And in the final paragraph, … you were talking about super smart electronic devices…
• • S: So?
• • P: You went to another category of electronic devices... You made no example for this group. [4] You didn’t have it in [the body
• • S: No] this is my clincher. The last sentence I talked generally. [6]
• • P: (Oh). That’s right… Very good.

The pair talk produced above can be characterized as highly coherent. Cohesion was created as the participants engaged with each other and completed each other’s utterances. The participants engaged with each other’s suggestions: there was negative or corrective feedback in the form of explicit peer repair (lines 1, 7) as well as positive feedback in the form of confirmations (lines 5. 13). There were requests and provision of information (lines 4, 8). Thus the talk showed a pattern of interaction that was high on equality and mutuality. The same pattern, collaborative, was also kept when later they were reviewing Parisa’s paper.

Interviews

The next data source, the stimulated recall interview, was used to better understand participants’ interactions during peer response and their intended roles. In this kind of interview, participants are prompted to recall thoughts that they had while performing a task. This kind of interview has been used in both SLA interaction studies that apply pair dynamics (see, e.g., Watanabe & Swain, 2008), and peer response studies in the L2 writing literature (see, e.g., Nelson & Carson, 1998). This approach was chosen because it allowed us to further understand how participants experienced the pair dynamics of peer response, and how they decided to use their peer’s feedback when revising.

Each participant had an individual meeting with the researchers for a 45-minute stimulated recall interview where they listened to the peer response recording and reviewed the revised draft. After having listened to the recording, we directed the participant’s attention to the parts that her partner had asked for the change, and probed her view.

To better understand how students experienced the interactionthat they had been assigned, the researchers focused on how participants explained their pair work, how they characterized their spoken interactions with their partners, and how they felt about the task and feedback they received. Below the results of the interview for each case are given.

Follow-up interviews made it clear that Armina accurately believed Ahoor belonged to a higher proficient group and Ahoor also saw Armina as an intermediate learner. Armina was grateful for feedback she received, and her goal was to be able to gain information necessary for writing well, to communicate easily with Ahoor, and to achieve good marks in this course. She claimed in the interview that she saw an opportunity, and seized it.

Ahoor valued the joined collaboration, and according to her, ‘the goal of the activity is helping each other.’ Her enthusiastic participation in the classroom activities was observed throughout the period of the data collection. She repeatedly offered chances to her partner to participate more actively by raising questions and seeking her opinion.

Similarly, Aida thought of Niloofar as a middle intermediate learner and Niloofar considered Aida as an advanced one. However, here, Niloofar missed the opportunity to participate in the pair work discussion. She claimed that as her proficiency level was much lower than her partner, she did not have much background information about what her partner was talking about and thus became confused. She was not confident that her paper was satisfactory, but did not feel comfortable enough to ask for help. According to her, she was unable to keep up with Aida’s reading speed. In the interview she mentioned that she could not frequently asked her about where in the text they were up to or what they were talking about. Her inability to keep up with her peer caused her to be dissatisfied with her participation in the task and therefore to evaluate the response session experience negatively.

Aida considered herself as an advanced learner, having relatively higher background knowledge in English. Though she did not let her partner to get involved, and conducted the peer response session as a monologue, she thought of her partnership as very helpful for Niloofar. From the beginning, Aida actively participated in the task, but she seemed to focus more on maintaining her role as a confident reviewer, and paid little attention to whether Niloofar was in agreement with her ideas or interested in following her advice. Aida simply presented her review and there was no further discussion between them. Aida’s primary focus according to what she claimed in the interview was to fulfil the teacher’s requirements concerning the task. Her lively personality and advanced knowledge allowed her to take initiative in the peer response session.

Shohre believed Azar is an upper beginner and herself as intermediate, and Azar also agreed with her. Shohre believed that her active control of the learning process has a positive effect on her partner’s learning outcomes. Azar viewed herself less proficient compared with her partner. She did not see the connection between the task and success in her studies and hence this resulted in non-participation in the classroom. In Azar’s case, active control of the partner did not have a positive effect on her learning outcomes. Shohre’s behaviour was unhelpful for Azar, as Azar claimed that her partner’s explanations went beyond her interest and her level.

For the case of Hani and Mozhan, Mozhan thought of herself and her partner as intermediate and Hani saw Mozhan as a beginner. The dominant/dominant pair believed that arguing was enjoyable, and, at the same time, they also expressed their disagreement with their partner’s ideas. Negative views focused on the limitations of the student reviewer and the feeling of inadequacy on the part of the reviewer or themselves. They referred to the experience as useful, but believed that pairing with a higher proficient learner is more helpful.

Hani and Mozhan had different intentions in approaching the task which influenced their involvement and the relationship they formed. As their extract showed throughout the task, Mozhan tended to prove herself, and Hani’s focus seemed to be on defending the accuracy of her own writing.

Bahar did not consider Mona as a stronger writer, so she was dissatisfied that she controlled the peer review. On several occasions she was seen to be reluctant to agree with Mona and commented that it was difficult for her to understand and apply her ideas in her writing. She said the writing task was not difficult because she was familiar with the topic but was not interested to follow her partner’s advice. On the other hand, Mona found Bahar to be quiet, which she interpreted as a lack of confidence and low proficiency. The availability of the recorder also seemed to affect Mona’s participation as she said that ‘I tried to show the instructor that I had studied well.’

For the case of Parisa and Saba, both thought of themselves and their partner as high intermediate. Collaborative pairs believed that giving and receiving feedback have a positive impact on their own writing ability. They attempted to accommodate each other as they shared an understanding of the task, their needs and focuses of their partner.

Generally, students interviewed on their perceptions of and their reasons for those perceptions all specified that the peer review was a beneficial technique. They were appreciative of the social contact, the opportunity to speak, share ideas, and to learn from others. However, there were negative views which focused on the limitations of the student reviewer and the feeling of inadequacy on the part of the reviewer or themselves. It was interesting that almost all of the students even the ones having held such negative views agreed with such a pair work activity. The results are of particular interest to the present study as the participants were from different proficiency backgrounds.

In addition to the complexity of multiple goals and approaches, an individual learner’s approach and participation seems dynamic and can change at different moments, rather than being definitive and fixed. Taking Niloofar’s case as an example, her lack of knowledge, her partner’s role and inattention to her seemed to result in her passiveness and reluctance to participate in the task. On the other side, for the case of Armina and Ahoor, Ahoor invested her time and effort as an expert, arguably taking on the role of teacher, and making the learning of her peer her own goal. Accordingly, Armina became more willing to be engaged and shifted to a collaborative partner.

Summary of Results and Discussions

This section summarizes the results of analysis, reporting on the patterns of interaction found in the data of 6 pairs, discussing the effect of the proficiency level, and describing the factors that affected the patterns of dyadic interactions. The results of the first research question, the patterns of dyadic interaction that were found between intermediate learners paired with partners with different foreign language (L2) proficiency levels are tabulated below (Table 2).

Table 2. Patterns of Interaction

As the above table demonstrates, it can be claimed that the proficiency level did not decisively determine the learners’ performance in peer response. The findings show that each learner engaged with the task differently by having reinterpreted the task based on individual experiences with previous peer response sessions in similar situations they have had. The study also revealed the effects of peer influence on learners’ behavior, and highlighted the multiplicity of aspects on which learners focused during their partnership. Their partnership was shaped by what they expected and required for the event, and the goal they defined for themselves triggered their participation.

No doubt differences in proficiency could have affected the relationships learners form. Although proficiency differences may have pushed learners to specific types of patterns, they did not determine them. In agreement with Storch (2004) and Watanabe and Swain (2008), it was observed that it is not necessarily actual but perceived proficiency which may affect how learners interact. Each learner's perception of their own and their partner's proficiency shaped the interactional patterns between them. Follow-up interviews made it clear that perceptions of interlocutors' proficiency did not always line up with their actual proficiency. For example, Mona found Bahar to be quiet, which she interpreted as a lack of confidence and low proficiency.

A number of individual learner factors have been proposed to explain why learners form different relationships when working in small groups or pairs, whether on collaborative writing tasks or in peer response activities. These factors include: personality (e.g. Alvarado, 1992; Malmqvist, 2005)_, cultural differences (e.g. Lockhart & Ng, 1995; Nelson & Carson, 1998), differences in L2 proficiency (e.g. Kim & McDonough, 2008; Storch & Aldosari, 2013; Watanabe & Swain, 2007), as well as the classroom environment, including the teacher's teaching style (DiNitto, 2000).

In agreement with Kim and McDonough (2008), in pairs formed of similar proficiency learners, there was evidence of a collaborative orientation, with pairs working collaboratively or forming expert/novice relationships. Using interview, it was found that pairs who formed collaborative and expert/novice relationships had positive views about collaborative writing and valued each other's contributions. They believed that respecting each other's opinions in collaborative situations was important to them.

The participants in this studies who formed collaborative relationships viewed the collaborative task as a shared responsibility. The learners who formed an expert/novice relationship had complimentary goals. The expert wanted to provide assistance; the novice sought to learn from the activity and from the expertise of her co-participant.

It was observed that in interaction with a positive interlocutor, who was eager to have more cooperation, the style shifted. Rather than having remained passive, when they saw the partners’ collaborative preference and their warm welcome, they became more cooperative. Thus, the roles in interactions could be shifted.

In contrast, the learners who formed dominant/dominant and dominant/passive relationships saw little value in collaborative writing and the task was viewed as an individual responsibility to them. In the interviews, learners who adopted a dominant role expressed desires to display their knowledge. Such learners were often driven by a desire to complete the task quickly rather than by a desire to learn from each other. In the case of the passive participant, in the interview, it was revealed that they preferred to take a distance from the activity they were unwilling to have.

Working with a more competent speaker may increase or decrease a lower proficiency learner's opportunities to participate depending on their own and their partner’s partnership and their interactional relationship. Learners at a lower level may have more chances to learn when working with higher level learners.

When generally thinking of the higher proficiency learner teaching the lower proficiency learner, it may actually be the higher level learners that benefit most. Higher proficiency learners are more likely to take on expert roles in interaction, and learners who act as experts may learn the most (Watanabe & Swain, 2007). This can explain the relationship formed between Shohre and Azar, and Aida and Niloofar, but not the pattern between Hani and Mozhgan.

The findings suggest that it was not just proficiency levels but the relationships that learners formed and the roles that they adopted (or are assigned) which affected the interaction between learners. The difference in learners’ perceptions of the task, along with the connection between a learner’s approach and their personal goals had a role in the partnership. However, since in response sessions they had partners, a learner’s behavior was also affected by their peers. This study showed how learners’ actual participation in tasks was related to individual factors, as well as interactive relationships with group members. These relationships affect both the quantity and quality of their dialogue and thus have implications for language learning.

Conclusion

In the present study, learners from the same proficiency level, middle intermediate, are shown to form different relationship patterns paired with a mixed- or matched-proficiency interlocutor. The learners adopted different roles and some changed the patterns during the task. Individual learners' interactions were influenced by the perceived rather than the actual proficiency levels they believed themselves and their partners possessed, and also by their roles.

The results highlight the complex relationship between proficiency and interaction patterns. Perhaps an important factor, which may explain the kind of relationships learners form and in turn the nature of their engagement and contribution to the collaborative writing activity, is how the learners orient to the activity. This notion of orientation encompasses learners' attitude to the tasks given and partnership, to being asked to work in groups/pairs, as well as to the goals that drive their actions.

Although proficiency narrowly played a role in shaping interactions and the types of interactional patterns, the findings suggested that orientation to the task, partnership and affective factors were significant. Dyads that had positive feelings toward each other and toward the task they engaged participated more collaboratively and were more likely to learn through interaction than those that did not, regardless of proficiency. Of course, a limitation, to be acknowledged here is that if the participants had had more time to discuss more writing papers together, they could have shifted their interactional styles with their partners.

Studies have shown that simply assigning students to work in pairs does not guarantee collaboration. Donato (1988) argues that in order to investigate the potential of group work for language learning, the reserachers need to consider the dynamics of the group. Drawing on the work of Petrovsky (1983a, 1983b), Donato (1988) distinguishes between diffuse or loosely knit groups and groups which formed a collective, a socially cohesive unit. He analyzed the data of L2 learners of French working in groups on a writing task and on an oral classroom presentation. In both sets, evidence of collective scaffolding was only found in the talk of the groups forming a collective unit. In loosely knit groups, instances of collective scaffolding were rare, and requests for help were often ignored.

This study in agreement with the findings of previous studies suggests that mixed-proficiency pairs are not surely superior in terms of engaging in negotiation of meaning. Successful interaction is not related to the mixed or matched proficiencies but to how individual learners allow themselves and their partners to shape their interactional styles. In each interaction, the individuals themselves significantly shape how the interaction is carried out. Having had a high- and low-proficiency pair engaged in an interaction does not automatically lead to a dominant/passive interaction.

In agreement with Philp, Walter, and Basturkmen (2010) and Lazareton and Davis (2008), the findings of this paper reveals that proficiency (and perceptions of proficiency) may affect interaction in some ways but it is not the only factor which affects negotiation of meaning. The findings showed different dyads with similar features, the same proficiency levels, similar perceptions on the proficiency of themselves and their partners, who formed different interaction patterns together.

From an activity theory standing, it may be argued that in the classroom, learners involve in multiple aspects of a broader activity at the same time according to different goals, such as aiding their partners, satisfying the teacher’s expectations, attending to their own or other learners’ interests, or keeping good relationships with their peers.

Considering the variation within peer interaction that occurred because of differences among peers, it was observed that the pairs formed distinct relationships. Learning through peer interaction in the classroom is a complex and dynamic process. The purpose was to advocate a research that examines peer interaction in its own right, as a context for communication. To understand how peer interaction may impact on learning, researchers need to study the variation among these interactions. These variables have implications for how to implement pair work to maximize the language learning afforded by collaborative writing tasks.

Even for learners in the same classroom, there are inevitable differences in proficiency. The way learners construct and engage in interaction is also diverse. Studies that examine peer interaction among proficiency groups can shed light on the processes involved and group dynamics. Not all group activities provide opportunities to learn.