ESTELA ENE
SENTA E. GÖRTLER
KARA MCBRIDE
University of Arizona, Tucson
Abstract:
This paper investigates the impact that a teachers virtual presence or lack there of has on students chat behavior with regard to error correction, uptake, target language use, and on-task behavior. The data come from beginning German students engaged in pair and small-group chatting activities at a major American university. Transcripts from chat sessions in a first-semester German class and a second-semester German class were analyzed. The data were triangulated with student surveys and teacher interviews. Results suggest that the teachers participation styles had a greater influence on learners chat behavior than simply whether or not the teachers were present and that the form-focused participation style of one of the teachers had an apparently inhibitory effect on learner participation.
INTRODUCTION
To date, computer-mediated communication (CMC) research has focused on student participation patterns and their effects on second/foreign language acquisition. Teachers' roles, on the other hand, have not received as much attention in CMC studies. The present article aims to fill this gap in the research by investigating teachers' roles in synchronous CMC and by determining the importance of teachers' presence in a first- and second-semester foreign language classroom chat. In particular, we investigate the differences between teachers' chat room participation styles and their effect on the learners' opportunities for language acquisition. Input, output, error correction, and uptake—all assumed to enhance language acquisition—were the means of determining the impact of teachers' participation styles on learners' experience.
LITERATURE REVIEW
Interaction between learners as well as between learners and language teachers, native speakers (NSs), or more advanced learners can enhance second language acquisition (SLA) (Long, 1996). As pointed out by Gass and Selinker (2001), output embedded in interaction plays several roles in SLA: (a) it helps learners create knowledge through syntactic and semantic processing (Swain, 1985), (b) it allows learners to practice or apply existing knowledge, (c) it fosters automaticity, (d) it provides an opportunity to elicit further input, and (e) it offers the opportunity to test hypotheses and to receive feedback about previously formed target language (TL) hypotheses. Other researchers such as Swain and Lapkin (1995) and Izumi, Bigelow, Fujimara, and Fearnow (1999) have also confirmed the important role of output in SLA.
Within the interactionist framework, feedback that is embedded in interaction is claimed to facilitate SLA. A modification of the learner's TL grammar can be triggered by several factors such as attending to input (especially modified input such as foreigner talk), noticing differences between hypotheses formed and actual NS production, testing hypotheses during output, and receiving implicit and explicit feedback in response to language production. Within this framework, negative evidence provided during interaction is crucial. Long (1996) argued that “negative feedback obtained in negotiation work or elsewhere may be facilitative of SL development at least for vocabulary, morphology, and language-specific syntax, and essential for learning certain specifiable L1-L2 contrasts” (p. 414).
The focus of much discussion in SLA has been whether to provide corrective feedback explicitly or implicitly. Explicit negative feedback includes overt error correction (i.e., the correction is clearly identifiable, and a correct form is supplied) or rule explanation. Overt, or explicit, error correction, “provides explicit signals to the students that there is an error in the previous utterance” (Panova & Lyster, 2002, p. 584). Implicit negative feedback includes recasts (simple and complex), as well as negotiation moves following a communication breakdown such as confirmation checks, clarification requests, and repetitions (Long & Robinson, 1998).
While explicit negative feedback clearly interrupts the flow of the conversation, implicit negative feedback is less disruptive. Within the perspective of implicit negative feedback, Lyster and Ranta (1997) found recasts to be the most frequent form of implicit correction. Nicholas, Lightbown, and Spada (2001) summarized the essential features of recasts as acknowledging content, providing positive affect, and modeling the correct form. Recasts also provide positive evidence for the reformulation and are claimed to promote noticing by contrasting learners' ill-formed utterance to the teacher's reformulated TL utterance (Saxton, 1997). Recasts can be accompanied by overt signals (e.g., tone of voice and facial expression in face-to-face interaction). When this happens, the combination of overt signals and recasts could arguably be construed as explicit negative feedback (Nicholas et al., 2001). Although there is evidence to suggest that both implicit and explicit forms of corrective feedback can help learners produce more accurate utterances (Carroll & Swain, 1993), recasts result in student self-correction less often than more explicit forms, such as metalinguistic feedback (Lyster & Ranta, 1997).
RESEARCH QUESTIONS
This study examines the effect of a teacher's presence in chat rooms by investigating the following research questions:
1. How are TL use, error correction, and on-/off-task behavior affected by the teacher's presence and actions in a chat room in which beginning FL learners are interacting to complete a task?
2. What are the students' perceptions about the teacher's role in and effect on chat room interactions?
3. How do these perceptions compare with observed effects of the teacher's presence in the chat room?
The first research question will be answered with the help of an analysis of the chat transcripts from two first-year German classes. Chat transcripts from two classes were separated into categories characterized by the teacher's presence or absence. TL use and on-/off-task behavior were determined based on the number of turns produced in the TL and the number of turns produced in on-/off-task posts. Error correction was measured by the teacher's corrective feedback moves in response to students' errors; the effectiveness of the teacher's corrective feedback was measured by the degree of students' uptake. Students' perception of the teacher's role were elicited in a student survey administered at the end of the semester. The results from the survey were triangulated with the results from the chat transcript and analysis of interviews with the two teachers involved in the project.
RESEARCH DESIGN
Setting and Participants
The research was conducted at the University of Arizona in Tucson. The participants were recruited from a first- and a second-semester German class that engaged in weekly chatting activities. The classes were taught by two female teachers. The first-semester teacher was also one of the researchers. Both instructors were experienced Graduate Teaching Assistants/Associates who had been using chat in their FL courses for more than 2 years.
First-semester Class
The composition of the first-semester class was rather unusual. It was dominated by nontraditional learners (i.e., graduate and undergraduate students proficient in other languages and a high school student). Fourteen students were enrolled in the class, including six nonnative speakers of English (four Spanish speakers and two Norwegian speakers). All students used English in their normal everyday interactions. The ages of the students ranged from 16 to 35 years of age. The students did not express or otherwise display anxiety about typing or using a computer in general, and the instructor did not perceive any resistance to working with computers.The students engaged in chat interactions for 10 to 20 minutes at the end of class once a week, and the chat activities were either information-gap or role-play activities (see sample activity in Appendix A). Students chatted in pairs or groups of three, following grouping practices of the traditional classroom. The students and the teacher were present in the computer lab at the same time. Although the students were not supposed to talk to each other in person, they sometimes did. Some students also availed themselves of online dictionaries while chatting. The teacher of the first-semester class used a chat program produced by the College of Humanities at the University of Arizona. The participants saw their own chat room on their screen. Each computer was used by one participant only, and the total number of chat rooms varied between five and seven. The chat room software contained a large window for the discussion at hand, a small window listing the participants in the given chat room, and a bar displaying German characters with diacritics (see Figure 1).
Figure 1
College of Humanities University of Arizona Chat Program
The chatter saw only the entries of the other participants in the specific chat room with the most recent entry appearing at the top of the discussion window. The teacher decided whether or not to be visible in the chat room. Although the teacher could observe only one chat room at a time, the full text of the discussion appeared on the teacher's screen upon entering the chat room.
Second-semester Class
The second semester class was a more traditional class of 20 students with most having limited experience in studying a FL. The students (7 females and 13 males) were between the ages of 18 and 24, except for one who was in his late 30s. These students chatted in pairs or groups of three for about 10-20 minutes at the end of class during their weekly lab hour. The chat activities were for the most part information-gap activities. The teacher used the IRC Français chat program (see Figure 2).
Figure 2
IRC Français Chat Program
In contrast to the chat program developed by the College of Humanities at the University of Arizona, this program allows the teacher to be in four chat rooms at the same time, but the teacher is always visible to the students in those rooms.
Chat Activities
In both classes the chat tasks were either directly taken from the textbook Kontakte (Terrell, Tschirner, & Nikolai, 2000) or slightly adapted to include a web quest. The two task types from Kontakte that the teachers considered appropriate for chat delivery were role-play and information-gap activities. The first-semester teacher used more role plays, whereas the second-semester teacher used more structured information-gap activities. While the specific tasks between the two classes differed, the nature of the tasks was comparable. Both task types were designed with the intention of creating meaningful discourse on a specified topic encouraging the use of the language structures of the chapter. An effort was made in both classes to relate the activities to the textbook as closely as possible because, as Müller-Hartmann (2000) and others have stressed, it is important to integrate CMC components directly into the regular classroom.
DATA COLLECTION
The data were collected using three instruments: chat transcripts, student surveys, and teacher interviews. Each is described in more detail below.
Procedures
For the purposes of this study, the students in the first-semester German section were randomly assigned chat partners in the first lab session during the second week of the semester. The students chatted once a week for 10-20 minutes for an entire semester. Pairs 1, 2, and 3 were always observed, while the other pairs were observed only on a few occasions. Due to a high fluctuation in attendance, the groupings were not as consistent as we would have liked, but they provided sufficiently consistent data for our purposes. In this class, the transcripts were corrected for grammatical mistakes and periodically returned to the students. However, most students reported to the teacher at the end of the semester that they did not consult the corrected transcript. The chat time was counted as 10% of the final grade; grades were assigned based on task completion and TL use, not on grammatical accuracy. In the second-semester class, the teacher randomly assigned new chat partners for each lab session. The teacher observed four pairs at a time and rotated through the rooms so that she participated in all chat sessions for a short amount of time. Even though the second-semester students did not receive feedback on their chat transcript, the students still had the option of printing the transcripts for their own use. The students in this class were aware that the teacher would eventually read the transcripts to grade the assignments, which also counted as 10% of the final grade. This portion of the grade was also determined by task completion and TL use after the teacher reviewed the transcript.
Instruments
Transcripts
The first four weeks of chat transcripts of the semester were excluded from the study to allow for a familiarization phase with the medium. Due to technical issues with the chat server, transcripts could not be retrieved for two additional chat sessions in each class. Hence, out of the 14 weeks of chatting during the semester, eight chat transcripts per class were collected and analyzed. Transcripts included posted messages and the screen names of the participants. Both chat programs automatically sort transcripts by person, which means that the transcripts for two partners can look quite different. For example, if Student 1 logs on before Student 2, Student 2's transcript will not reflect any turns completed before his/her entry in the chat room; however, Student 1's will. Another commonality was that both programs automatically displayed messages indicating when another student entered or left the chat room. However, only the IRC program indicated the teacher entering and leaving the chat room. An additional difference is that the chat program developed by the University of Arizona indicated the time of posting. For sample transcripts, see Appendixes B and C.
Student Survey
The student survey was distributed to the participants after completion of all eight chat activities (see student survey in Appendix D). The survey was administered by one of the researchers not involved in the teaching of the class. The survey elicited students' perceptions about the usefulness of the chat activities and how they saw their own and others' roles in the chat sessions. Page 1 had Likert-type scale items with room for comments, and page 2 had open-ended questions.
Teacher Interviews
The two teachers who participated in the project were interviewed in order to elicit their perceptions about their participation style and influence on the chat room dynamics and the students' behavior during the chat sessions. Field notes were taken during the interview.
DATA ANALYSIS
Transcripts
examine parts of speech or speech acts, which would have obligated the researchers to take into account various grammatical and discourse features as criteria for defining a turn. On a few occasions, code switching occurred within a posting. In these instances, the part in the TL was counted as a half turn (.5 in the tables below), and the part in the non-TL was counted as another half turn. In the first semester class, corrective feedback by the teacher was mostly implicit and all except for once in the TL. Modeling (9 occurrences) and recasts (22 occurrences) made up 94% of the corrective feedbackOverall, a similarly high and positive response came from students in both classes. Reported use of the TL (Item 1) was the same in both classes, with a mean of 3.7 and a mode of 4.
Student Surveys
The quantitative results of the survey allowed for a numeric comparison of the two classes. Differences and similarities in the student surveys highlighted points of focus that could then be compared to the open-ended answers from the surveys and the results from the transcripts and teacher interviews.
Teacher Interviews
Teacher interview field notes were qualitatively analyzed and compared with the findings from the transcripts and student surveys. The underlying teaching philosophies were also extracted from the teacher interviews.
SUMMARY OF RESULTS
Transcripts
First-semester German Class
On/Off-task Behavior and TL Use
Whether the teacher was present or not, the students produced about the same number of turns (802 with the teacher present and 822 without the teacher present) (see Table 1). The actual time spent on chatting was also equal. In both conditions, the students were on task more than five times as often as they were off task (84% of the turns with the teacher present were on task; 83% without the teacher present). When the teacher was present, most of the off-task behavior was greetings (12% of turns), which, when produced in the TL, were a welcome off-task behavior that the teacher considered a natural part of the conversation. When the teacher was not present, there was more off-task behavior of other sorts. On-task behavior involved more task negotiation when the teacher was not present. The differences in the two kinds of on-task behavior and two kinds of off-task behavior when the teacher was present in the chat room versus when the teacher was absent were significant, yielding a c2 of 11.82 (p < .01, df = 3).
Corrective Feedback
The frequency of implicit error correction used by students increased as the semester progressed. Self-correction occurred three times with the teacher present and only once without the teacher present.
Second-semester German Class
On/Off-task Behavior and TL Use
The second semester students spent most of their turns on task, both when the teacher was present (76%) and when the teacher was not present in their chat room (71%), and greetings constituted a large portion of off-task behavior. There were only 8 tokens of task negotiation when the teacher was present, compared to 66 when the teacher was absent.
Survey
A survey was given to the students during a class period following the completion of the chat activities analyzed in this study (see student survey in Appendix D). Each question had four choices to keep the students from defaulting to neutral answers. The responses were coded on a scale of one to four, with four indicating the most positive response. Averages were computed and t tests were run to check for significant differences. Ten of the 14 students in the first-semester class completed the survey; 18 of the 20 students in the second-semester class completed it. Although not all the students in the two classes completed the surveys, all students' transcripts were nevertheless analyzed since the survey was intended to provide insight into the students' perceptions. Furthermore, due to the anonymous nature of the survey, it was impossible to identify the students who did not complete it. Therefore, the surveys included in the analysis here have to
be treated with some caution because not all of the chat participants' perceptions are reflected in the results.
Interviews
The first-semester teacher, one of the researchers in the project, was interviewed about her motivations and goals with chat by one of the other researchers. The teacher's comments indicated that much of her inspiration for using chat is that it is a novel activity that adds variety to class. It is also an alternative forum for conversation. For this teacher, conversational flow is of the highest importance during chat and other communicative activities in the classroom. She strives to get her students to work largely independently, as opposed to relying on her as the teacher to keep an activity going, and she prefers not to correct students' errors while they are concentrating on how to express meaning. As she said, “If my grandmother could understand it, then I don't have to correct it.” By this she means that she limits her correction during chat to errors that cause breakdowns in communication, or, occasionally, errors that others in the class can understand but that are too far from the TL as to pass the “grandma” test.
This teacher does not believe, however, that there is no place for error correction. On the contrary, she believes that the opportunity to later correct printed chat transcripts in a detailed fashion is one of the primary benefits of using chat in a FL classroom. By correcting the transcript later, as opposed to correcting the output at the time of its utterance, the teacher can attend to details without interrupting the original flow of conversation. Nevertheless, review of her students' transcripts showed that more than half of her responses address the content of what students say rather than their grammar or spelling.
DISCUSSION
In this study, effects of the teachers' presence and participation style were analyzed by examining differences in TL use, on- and off-task behavior, and error correction. These are discussed in more detail below.
TL Use
In both classes, non-TL use was lower when the teacher was present than when she was not: 9% versus 15% in the first-semester class and 1% versus 3% in the second-semester class, and, in both classes, the differences were statistically significant. Although these percentages may intuitively seem lower than they would be for face-to-face communication, this result corroborates and refines findings of other studies. For example, Ortega (1997) stated that “There is also anecdotal evidence that backsliding into the L1 is minimized in electronic discussion (Beauvois, 1992; Chun, 1994; Kelm, 1992; but see Kern, 1995)” (p. 87). It is also possible that since the students have more time to formulate and can erase errors before sending the message (and even access online dictionaries), actual turns on screen can be more easily posted in the TL (see also Ortega, 1997).
On-/Off-task Behavior
The students in the first-semester class were on task 84% of the turns in comparison to 73% of the turns of the students in the second-semester class. The less structured activities allowed the first-semester students to express themselves or socialize within the limits of the task, while personal expression in the second-semester class was likely to be off task. Most of the off-task turns in both classes under both conditions were greetings, and the variation was not significantly different (c2 = 1.62, p = .20, df = 1). In the survey at the end of the project, students also reported staying on task almost all of the time, which is supported by the teachers' impressions and the analysis of the transcripts.
Interestingly, on- and off-task behavior was almost the same in both conditions in the first-semester class: 84% on task with the teacher present and 83% without the teacher present, the difference being insignificant (c2 = 0.002, p =.96, df = 1). However, in the second-semester class students stayed on task better when the teacher was present (76%) in comparison to when she was not (71%), and this difference was significant (c2 = 4.51, p <.05, df = 1).
Error Correction
Error correction by peers and self-correction, with or without the teacher present in either class, was very infrequent. Error uptake was comparable under both conditions in the first-semester class and more frequent without the teacher present in the second-semester class. Since the students in second semester showed a higher tendency to engage in uptake processes in general, they may have been more trained to be more sensitive to differences between their language production and that of another person. When the teacher was not present and, therefore, not able to correct mistakes, error uptake was a distinct possibility. In the context of the high number of total turns produced by the students in the TL in both classes and under both conditions, error uptake was low. The number of correction uptakes was lower than the number of error uptakes in the first-semester class. However, not all errors were corrected, and the number of errors far exceeded the number of corrections. Analysis also revealed that some apparent error uptake, especially in the first-semester class (where students relied on third personal singular verb forms found in the model task instead of conjugating the verbs to first or second person), may have been interlanguage errors that the students were producing because they were at the same developmental stage. These realizations should weaken worries about error uptake expressed by others (Beauvois, 1992; Kelm, 1992; Storch, 2002). However, the issue deserves more attention in future research.
CONCLUSIONS
The original purpose of this investigation was to determine the importance of teachers' participation in a chat. As illustrated above, not only did teachers' participation turn out to be important, but even more so, the teacher's participation styles proved to influence learners' participation in chat. The conclusions with respect to the study's research questions are detailed below. n terms of the effects of a teacher's presence on learners' interaction, the study revealed that the teacher's presence in or absence from the chat room correlated with systematic differences in student behavior only in terms of turns produced, as will be explained below. Peer-to-peer error correction, self-correction, correction uptake, and error uptake were limited in both classes and in both conditions. The most notable differences between the teacher-present and teacher-absent conditions was the total number of turns produced by the students. In the second-semester class the students produced almost twice as many turns when the teacher was not in the chat room. This dramatic difference was not observed in the first-semester class. When total turns produced with the teacher present and the teacher absent are compared between the two classes, the difference is significant and strong (c2 = 58.11, p < .00001, df = 1). Hence, the identified silencing effect can be attributed not to the teacher's presence in the chat room but rather to the teacher's participation style. In this situation, the more form-oriented teacher of the second-semester class seemed to cause some performance anxiety in the learners. The students' perceptions about the teacher's role in and effect on chat room interactions were overall positive and matched both the teachers' self-descriptions in the interviews and the findings based on the transcript analysis. In general, the students appreciated receiving input and corrective feedback from their teachers in that particular teacher's style. However, some second-semester students remarked that their teacher's presence and participation styles made them feel self-conscious.
IMPLICATIONS AND RECOMMENDATIONS
Generalizations cannot be made on the basis of this study due to certain limitations such as the small number of participants, differences in class profiles, and differences in chat software and implementation. Future adaptations of this study should attempt to control the above mentioned variables. A further interesting investigation would be to analyze the roles of the teacher when chatting from emote locations without a physically close presence.
Generally, it was clear that students enjoyed chatting and saw it as a valuable part of the language classroom. For example, one student in the first-semester class expressed the following opinion: “The advantages [of chatting] would be practicing and self-thinking sharpens ones [sic] responses when involved in a dialogue. Rather than just absorbing. The disadvantage, well I cant [sic] think of any right now.” Similarly, a student in the second semester wrote: “Advantages: helps to improve one's ability to communicate, switches things up a bit. Disadvantages: NONE!” Remarks such as these support recommendations that chat be used in beginning-level FL classes such as the ones included in this study.
From a pedagogical standpoint, this study suggests that language teachers involved in chats with their beginning-level students might want to adopt a conversational style in order to stimulate student participation and output production, a beneficial factor in SLA (Swain, 1985). At this level, corrective feedback seems to be least inhibiting when implicit. The effectiveness of one corrective feedback style versus another needs further investigation with better measurements of effectiveness than uptake. Once such a measurement can be identified, other issues can be analyzed. One intriguing question that remains is how the visual modality in chat alters the nature of error correction.
Tidak ada komentar:
Posting Komentar