Mobile Edition
Print
Get the Mag
Weekly UpdatesResearch Questions
The following research questions guided this study:
1. To what degree will NNS EFL teacher candidates perform autonomously as they attempt to correct their own and others' grammar errors in a long-term collaborative writing task?
2. How accurate will they be in making these peer and self corrections?
3. What can these postings tell us about students and long-term web-based collaborative writing?
Students and Instructor
The 40 students in this class were all in their final year of a BA program in English Language Teaching in Mexico. They ranged in age from 21-23 years old. They all received passing grades and participated extensively throughout the class. All of the students in this class were at the same level of second language proficiency according to a series of inhouse English proficiency exams. The students demonstrated comfort and familiarity with the technology. They were required to complete an integrated online orientation to the course management system that provided the researcher a detailed (mouse click by mouse click) record of their performance through the students' automated orientation to the course site, course expectations, and varieties of media and links they would be expected to utilize. Records of their progress illustrated a high level of user capability. However, there were a small number of students who demonstrated slight difficulty accessing some resources. These students participated in a videoconference learner training session with the instructor that guided them in the use of these resources. Learner training is important for successful use of instructional technology (Hubbard, 2004).
This study introduced wiki technology in the first week and required students to locate and report on information related to culture, the class' area focus, that they had found through Wikipedia. Students collected data and reflected on the accuracy of this information by referring to external sources of information. All of them were familiar with wikis (in the form of Wikipedia) before it was introduced in the class. The instructor (who is also the researcher) teaches in an intensive English program with 20 years of experience teaching ESL and EFL in varied settings, including numerous experiences using technology in the language classroom and at a distance.
The Task
The Wiki collaboration task was treated very differently from other tasks in the course. Other tasks included an extensive amount of presentation, feedback, and interaction between students and the teacher, including feedback regarding grammatical accuracy. It was not unusual for discussions to reach eight layers of embedding, suggesting active ongoing interaction. The wiki task was initiated by the teacher and left up to the students. It was their responsibility to collaboratively construct the wiki as a reflection of what they had learned in the class as a community. There was no intervention from the teacher. This was intentional to determine if the autonomous constructivist activity would enable students to establish a sense of responsibility for the ongoing maintenance and revision of the document. The wiki itself was constructed within the content management system, resulting in a safe password-protected environment for students to share ideas and take risks with their language. The intent of this study was for this group of students to be solely responsible for the construction of knowledge. They would not receive feedback, updates, revisions, or elaborations from anyone other than their immediate peers, thus encouraging a sense of responsibility.
Students were encouraged to cull what they perceived to be the highlights of the other activities into the collaborative construction of this wiki. Four times throughout the quarter (Weeks 1, 5, 9, and 13) they were sent a simple set of instructions to guide them to participate:
These instructions were intentionally left brief and free of information related to the topic in order to allow for the observation of student behavior without undue influence from the prompt.
Data & Analysis
Most previous studies of collaborative writing have focused on the face-to-face, or CMC-based, metatalk of students as they progressed through collaborative writing tasks. The present study relies upon the data provided by the wiki itself rather than face-to-face observations. When a student chooses to alter a portion of the wiki text, she must select the portion that she will work with. Portions are typically divided into sentences, paragraphs, or a series of paragraphs in a particular subtopic. In this study only the portion of the text that the student demonstrates attention toward at the editing level is taken into consideration. Thus, an error that is overlooked at the beginning of a portion chosen for editing is counted, but an error in a previous sentence or paragraph is not. This approach only holds students accountable for the portion of the test they choose to edit, ignoring the interaction with any peripheral text. Language related episodes (LREs; Swain & Lapkin, 1995) were used to identify learner attention to form throughout the construction of the wiki. Typically LREs refer to the metalinguistic attention to discourse between students rather than actual language use. For the purposes of this study an LRE is defined as any language oriented contribution to the wiki. Alterations that were not considered germane include alteration of visual style and inclusion of hypertext links, images, and other media. LREs were coded according to the following:
* Form Only, Content Only, Form/Content, Content/Form
* Accurate, Not Accurate
For the purposes of this coding, Form/Content was used to refer to revisions that seemed to focus on form with some additional, often minor or extraneous, alteration to the content of the text. The label Content/Form was used to refer to revisions that seemed to focus on the content with a minor contribution to the form. Data analysis was conducted independently by two trained raters. These raters both hold Masters degrees in Linguistics. Both were given the basic preceding LRE categorization and told to identify error types as they emerged. Thus, coding was established in an ongoing manner as the raters interacted with the data. The raters negotiated three isolated LREs where they did not initially agree about coding. Thus, inter-rater reliability was (0.99). The error categories that emerged during the analysis included:
* Articles
* Coordination
* Fragment
* Part of speech
* Punctuation
* Run on sentence
* Spelling
* Subject/Verb agreement
* Word choice
Follow-up interviews were conducted with the students to gain insight into their decision-making process and develop an understanding of their willingness and ability to attend to the form errors observed in the wiki. As the autonomy framework establishes, there is a distinction between ability and willingness in the demonstration of autonomy (Littlewood, 1996). All of the participants were invited to participate in the interviews. A total of twenty interviews were conducted. Each of these was between ten and twenty minutes in length. These were done individually with the participants after the completion of the course by the instructor and were guided by the questions in Appendix. Each interview also provided an opportunity to ask students about their individual contributions and any changes they might have overlooked. Interviews were conducted using desktop video conferencing software.
FINDINGS
The overall tendency among participants was to focus on meaning rather than form. When form was central to a revision, it was nearly always accompanied by some additional contribution to the content rather than an isolated incidence of error-correction. Although the students were capable of achieving a level of grammatical accuracy in their more formal writing, they seem to consider a web-based collaborative activity to be less form demanding. They tended to defer to meaning, and often even design and style, rather than attend to grammatical concerns. In many cases they were willing to devote a great deal of time altering font and adding links to support the content of sentences that contained numerous grammatical errors. When asked about this observation, some responded that they had no problem understanding the meaning of the sentences in question and, thus, they did not bother to correct these errors.
A total of 233 edits were made by the students in an overall history of 160 total iterations of the wiki. This indicates that some iterations involved multiple contributions. Among these 233 edits, 169 (73%) involved LREs while 64 (27%) of the total contributions involved no LREs. These 64 contributions addressed formatting, font, or other design issues. An overview of all contributions is given in Table 1.
Of the 169 contributions that involved LREs, twenty-nine (17%) were form-only. Ninety-two (54%) were content-only. Six (3%) were form-focused with some additional contribution to content. Forty-two (25%) were content focused with some attention to form. Seventy-seven (45%) of the LREs included some element of attention to form (Form + Form/Content + Content/Form). These are the contributions upon which these findings will focus. First we will take a look at the types of errors which students were most likely to attend to.
Error Type and Attention
Student initiated attention to form was divided into ten categories. The breakdown of these ten error types is presented in Table 2.
Students initiated an attempt to correct word choice and spelling errors much more than anything else. In fact, these two error types combined accounted for 39 (51%) of the 77 revisions that attended to form. Individually, word choice accounted for 25 (32%) and spelling accounted for 19 (25%) of the 77 attempts. Coordination and Subject/Verb agreement each accounted for 8 (10%) of the LREs primarily or secondarily focused on form.
FEED