Collaborative writing: Product, process, and students' reflections PDF

Preview

Summary

Journal of Second Language Writing 14 (2005) 153–173 Collaborative writing: Product, process, and students’ reflections Neomy Storch Department of Linguistics and Applied Linguistics, School of Languages, 5th Floor, Arts Centre, The University of Melbourne, Parkville, Victoria 3010, Australia Abstra...

Description

Accelerat ing t he world's research.

Collaborative writing: Product, process, and students' reflections Yachiyo Ohta

Cite this paper

Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers

Download a PDF Pack of t he best relat ed papers 

Pat t erns of Int eract ion in ESL Pair Work Neomy St orch Collaborat ive writ ing t asks in t he L2 classroom: Comparing group, pair, and individual work Ana Fernandez Dobao More LREs Bet t er Writ t en Product s-L2 Learning t hrough Collaborat ive Dialogue Eunhee Park

Journal of Second Language Writing 14 (2005) 153–173

Collaborative writing: Product, process, and students’ reflections Neomy Storch Department of Linguistics and Applied Linguistics, School of Languages, 5th Floor, Arts Centre, The University of Melbourne, Parkville, Victoria 3010, Australia

Abstract Although pair and group work are commonly used in language classrooms, very few studies have investigated the nature of such collaboration when students produce a jointly written text. This study set out to investigate collaborative writing. The study was classroom based, and the participants (23) were adult ESL students completing degree courses. Students were given a choice to write in pairs or individually. Although most chose to work in pairs, some chose to work individually. All pair work was audiotaped and all completed texts collected. All pairs were also interviewed after class. The study compared texts produced by pairs with those produced by individual learners and investigated the nature of the writing processes evident in the pair talk. The study also elicited the learners’ reflections on the experience of collaborative writing. The study found that pairs produced shorter but better texts in terms of task fulfilment, grammatical accuracy, and complexity. Collaboration afforded students the opportunity to pool ideas and provide each other with feedback. Most students were positive about the experience, although some did express some reservations about collaborative writing. # 2005 Elsevier Inc. All rights reserved. Keywords: Collaborative writing; Second language writing; Learners’ attitude to pair work

The use of small group and pair work in classrooms, particularly in second language (L2) classrooms, rests on strong theoretical and pedagogical bases. From a theoretical perspective, the use of small groups/pairs accords with a social constructivist view of learning. The roots of social constructivism are based on the work of Vygotsky (1978). According to Vygotsky, human development is inherently a socially situated activity. E-mail address: [email protected]. 1060-3743/$ – see front matter # 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jslw.2005.05.002

154

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

A child’s (novice) cognitive development arises in social interaction with a more able member of society. The more able member (expert), by providing the novice with the appropriate level of assistance, stretches the novice beyond their current level towards their potential level of development. Such assistance is now commonly referred to in the literature as scaffolding. However, as a number of researchers have shown (e.g., Donato, 1994; Storch, 2002), scaffolding can also occur among peers when working in group/pair work. Thus, from a social constructivist perspective, learners should be encouraged to participate in activities which foster interaction and co-construction of knowledge. From a pedagogical perspective, the use of small group and pair work is further supported by the communicative approach to L2 instruction and its emphasis on providing learners with opportunities to use the L2. However, the use of small group/pair work in writing classes seems quite limited. It tends to be limited to the beginning stages (brainstorming), or more commonly, to the final stages of writing—the peer review stage. In this final stage, students review each other’s written text and make suggestions on how it could be improved. A number of researchers (e.g., Ferris, 2003) have noted the benefits of such peer reviews. Foremost among these benefits is that peer reviews are a way of raising students’ awareness of audience considerations (Leki, 1993), and at the same time, they may help learners develop analytical and critical reading and writing skills (Nystrand & Brandt, 1989). One of the drawbacks of peer reviews, however, is that the focus is often on the product of writing rather than the process of writing. In L2 contexts in particular, a number of studies (e.g., Lockhart & Ng, 1995; Nelson & Carson, 1998; Villamil & de Guerrero, 1996) have shown that when students are asked to peer review, they tend to focus on errors at the sentence and word level. Thus, the process of writing remains a private act, where writers are left to their own devices when making important decisions about their text (Hirvela, 1999). A number of scholars, writing about developing L1 writing skills (e.g., Daiute, 1986; Wells, Chang, & Maher, 1990), argue that students should collaborate throughout the writing process. Such collaboration means that learners have joint responsibility over the production of the text. This may promote a sense of co-ownership and hence encourage students to contribute to the decision making on all aspects of writing: content, structure, and language. Ede and Lunsford (1990) refer to this as a singular text/plural authors approach. Research findings on collaborative writing have been positive. Research conducted in L1 settings (e.g., Higgins, Flower, & Petraglia, 1992; Keys, 1994) has shown that collaborative writing is a way to foster reflective thinking, especially if the learners are engaged in the act of explaining and defending their ideas to their peers. Research conducted with L2 learners (e.g., Donato, 1988; DiCamilla & Anton, 1997; Storch, 2002; Swain & Lapkin, 1998) has shown that in the process of co-authoring, learners consider not only grammatical accuracy and lexis but also discourse. Furthermore, and depending on the kind of group/pair dynamics formed (see Donato, 1988; Storch, 2002, 2003), collaborative writing may encourage a pooling of knowledge about language, a process Donato termed collective scaffolding (Donato, 1988, 1994). However, despite the support in the research literature for collaborative writing, as a language teacher, I have often noticed that when I ask students to work in pairs (or small

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

155

groups) on tasks which require written output, some students seem reluctant to do so. They seem to prefer to complete such tasks individually. Such observations have been reported by other L2 teachers as well (e.g., Peretz, 2003). Thus, when faced with some opposition by learners to group/pair joint writing activities, the question facing language teachers is how strongly should they encourage learners to work jointly on writing tasks. Are compositions produced collaboratively better than those produced by students writing individually? What is the nature of the writing process that pairs engage in? These are some of the questions that the present classroom-based study set out to investigate. It should be noted that studies on collaborative writing have, by and large, focused on the attention to language and the socio-cognitive processes evident in such interactions. The nature of the writing process and of the written text produced have received scant attention. There are very few studies that have compared compositions produced collaboratively with compositions produced individually. A study by Storch (1999), which compared individual and pair performance on a range of grammar-focused exercises, including a short composition, found that exercises completed in pairs were generally more accurate than when completed individually. However, in that study the same students performed the exercises in pairs and individually. The other issue that needs further investigation is learners’ views concerning collaborative writing. Results of studies on students’ attitudes to group/pair work in general are mixed. Some studies report that learners have predominantly positive attitudes to pair and group work (e.g., Mishra & Oliver, 1998; Roskams, 1999), while others report that learners have reservations about pair and group work (e.g., Hyde, 1993; Kinsella, 1996). However, it should be noted that most of these studies rely on surveys rather than on interviews conducted with students immediately after experiencing a collaborative activity. Moreover, these studies have examined learners’ attitudes to group/pair work in general, rather than to the activity of collaborative writing. Thus, the current study set out to investigate the product, process, and student reflections on collaborative writing. Specifically, the study set out to compare the texts produced by learners individually with those produced by learners working in pairs. For learners working in pairs, the study also investigated the nature of the composing process. Finally, the study attempted to elicit the learners’ reflections on the activity of collaborative writing.

1. The study The study was conducted in an ESL writing class offered for credit at a large Australian university. Two levels are available: ESL 1 and a higher level ESL 2. The study was conducted in ESL 1. Student placement in these classes is determined by a universitydeveloped diagnostic language test. Students who are advised to enroll in these classes are considered in this context to be of intermediate language proficiency. That is, although these students have the required ESL proficiency to be accepted for study at the university (an average of 6.5 on IELTS or 233 on the computer-based TOEFL test), their scores on the diagnostic test indicate that they require additional attention to their academic writing skills and grammatical accuracy.

156

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

Table 1 Participants Participants

Age

Country of origin

Writing proficiency score (on in-house diagnostic test) a

Individual participants Ayako Arthur Ulia June Jeffrey

22 20 24 29 20

Japan China Indonesia Korea Indonesia

6 5 6 6 6

Pairs Charley Mai Ryan Noriko Angela Jasmine Victor Tanako Yong Ed Olivia Shirley Howard Sam Anne Pam Maria Lee

25 42 20 29 20 22 21 21 26 23 19 20 21 20 20 20 25 27

Thailand Vietnam Taiwan Japan Malaysia Taiwan Vietnam Japan South Korea Japan Hong Kong Indonesia China Thailand Thailand Thailand Bosnia China

5 6 6 6.5 5 6 5 6 5.5 6 5 5 5.5 6 6 6 5 5.5

a

Score is given on a scale of 9.

The task used in the study was a data commentary text. Students were given a graphic prompt and asked to compose a short (one to two paragraphs) text. The prompt showed the language proficiency of two groups of migrants (Vietnamese and Laotians) before and after coming to Australia (see Appendix A). The task was presented as class preparation for the upcoming assignment. The task was not graded, but it was collected at the end of the class and subsequently returned with feedback comments.1 The study was conducted in two parallel in-tact classes. The researcher taught both classes. Data collection occurred in week 4 of the semester, thus the students were familiar with each other. Given my previous observations that some students may be reluctant to write in pairs and reported survey findings (e.g., Hyde, 1993), students in both classes were given the choice of working in pairs or individually. Of the 23 students present on the day of data collection, 18 chose to work in pairs and five chose to work individually. As can be seen from Table 1, the participants2 in the study came from a range of language backgrounds. The majority were international students from Asia. The average 1 It should be noted that although the task was not graded, and hence of low stakes, these students are mature age students who tend to take all class tasks fairly seriously. 2 Pseudonyms are used for all participants.

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

157

age of the participants was 23. Given that the students were placed in the class on the basis of the diagnostic ESL test results, they were fairly homogeneous in terms of their L2 writing proficiency. Their scores on the writing subtest of the diagnostic test tended to range from 5 to 6 on a 9-point scale. The pairs were self selected and consisted of two male pairs, four female pairs and three male/female pairs. Students who chose to work in pairs were given tape recorders to record their talk as they completed the task. All completed compositions were collected. Then, within 1–4 days after the class, students who worked in pairs were interviewed individually (in the researcher’s office) about the experience of writing collaboratively. The interviews were tape recorded.

2. Data analysis Pair dialogues and interviews were transcribed verbatim. Thus, the data set included the completed compositions and transcripts of the pair talk and of the interviews. Each data source was analyzed separately. It should be noted that the composition data of two pairs had to be discarded from the data set. The data of one pair had to be discarded because the tape recorder was not activated correctly until the last few minutes of the task. In the case of the other pair, the two students had difficulties reaching agreement, and their joint text contained a number of instances in which two versions were written side by side. Given the nature of the analysis undertaken (elucidated below), it was not possible to include the text produced by this pair in the analysis. 2.1. Analysis of the compositions The texts completed by the students were analyzed using both quantitative and qualitative measures. Quantitative measures included measures of fluency, accuracy, and complexity. Fluency was measured in terms of the total number of words. Accuracy and complexity measures were based on a count of T-units and clause analysis. AT-unit is defined by Hunt (1996, p. 735) as ‘‘one main clause plus whatever subordinate clauses happen to be attached to or embedded within it.’’ This measure, despite concerns expressed by Bardovi-Harlig (1992), is the most commonly used unit of analysis of both written and oral discourse (Foster, Tonkyn, & Wigglesworth, 2000). In order to measure for complexity and accuracy, the compositions had to be analyzed for clauses, distinguishing between independent and dependent clauses. An independent clause is one which can be used on its own (Richards, Platt, & Platt, 1992); a dependent clause must be used with another clause in order to form a grammatical sentence in English. There is some disagreement among researchers as to how to code for clauses, particularly dependent clauses. In this study, following Foster et al. (2000), a dependent clause was one which contained a finite or a nonfinite verb and at least one additional clause element of the following: subject, object, complement or adverbial. This definition seemed particularly suitable for this data, given that the learners often omitted the subject element in a clause (see Appendix B). To measure accuracy, two measures were used: the proportion of error-free clauses of all clauses (EFC/C) and the number of errors per word. Admittedly, as Bardovi-Harlig & Bofman (1989) point out, such measures do not distinguish between type or severity of

158

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

errors. However, a measure of the number of errors per word does at least account for the distribution of errors in relation to the production unit (words in this case). Errors in this study included syntactical errors (e.g., errors in word order, missing elements) and morphology (e.g., verb tense, subject–verb agreement, errors in use of articles and prepositions, errors in word forms). Errors in lexis (word choice) were included only when the word used obscured meaning. All errors in spelling and punctuation were ignored (see Appendix B). In analyzing texts it is important to consider not only grammatical accuracy but also complexity. This is because accuracy may be achieved as a result of a learner not taking any risks in their writing and relying on simple, well-controlled forms. At the same time a trade off may exist between complexity and accuracy. The more complex the sentences produced, the more likely they are to contain errors (Foster & Skehan, 1996). Complexity reflects the writer’s willingness to engage and experiment with a range of syntactic structures, moving beyond coordination to more complex structures which include subordination and embedding. One measure of complexity is the proportion of clauses to Tunits (C/T). Foster and Skehan (1999), based on their previous research, conclude that this is a reliable measure, correlating well with other measures of complexity. Another measure of complexity is the proportion of dependent clauses to clauses (DC/C), which examines the degree of embedding in a text (Wolf-Quintero, Inagaki, & Kim, 1998). In order to check for inter- and intra-rater reliability in coding, and following the advice of Polio (1997), guidelines were formulated stating clearly what constitutes a T-unit, a clause, and an error (see Appendix B). Then, a random sample of four compositions (forming approximately 30% of the entire data set) were coded by a second rater and recoded by the researcher 2 days after the initial coding. Intra-rater reliability for T-unit and clause identification was 99%, and inter-rater reliability was 97%. Inter- and intra-rater reliability for error counts was lower (86%). Discussion between the raters resolved all disagreements. Because of the small sample size and the fact that data was collected in intact classes, the non-parametric Wilcoxon–Mann–Whitney test (Hatch & Lazaraton, 1991) was used to determine the statistical significance of all quantitative results. A qualitative evaluation of the written text took into consideration the content and structure of the text and task fulfillment. A 5-scale global evaluation scheme3 was developed which took into consideration the kind of elements identified by Swales and Feak (1994) that a data commentary type text should include (see Appendix B). Inter-rater reliability on the global evaluation of five randomly selected texts was checked, and the differences between the two raters seemed small. There was complete agreement on three compositions. On the two compositions where the two raters differed, the difference was only half a band score. 2.2. Analysis of pair dialogues In the first instance, all dialogues were examined for any distinct phases of writing: planning, writing, and revision phases. The time spent on these phases was noted. Then all 3 The ability to award half points makes this a much larger scale (in line with the 9-point scale used to assess IELTS writing).

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

159

talk was segmented into episodes. An episode varied in length from a single turn to a number of turns. Each episode was coded for what the learners seemed to focus on. Seven focus areas were identified. These are set out with examples from the transcripts in Fig. 1. All episodes were timed (in minutes and seconds), and the total time spent on these aspects of writing was calculated.

Fig. 1. Coding of episodes.

160

N. Storch / Journal of Second Language Writing 14 (2005) 153–173

2.3. Student interviews Recorded student interviews (of all 18 students who chose to work in pairs) were transcribed. The students’ attitudes to pair work in general and to collaborative writing were collated and summarized.

3. Results Bef...