Oral And Written Feedback To Improve Writing English Language Essay

Oral And Written Feedback To Improve Writing English Language Essay This study is an investigation of the perceptions about effectiveness of oral and written feedback on writing of thirty-seven Cambodian English-major students at the National University of Management (NUM). Two instruments were used to collect data from the oral feedback group (N=19) and the written feedback group (N=18) before and after the two-month treatment: questionnaires and student paragraphs. Results indicate that the two groups equally delivered better performance on holistic writing although oral feedback was viewed as preferable to written feedback. While the former positively impacted on both the micro-aspects (i.e. grammar, vocabulary, and mechanics and spelling) and the macro-aspects (i.e. content and organization), the latter encouraged revision only in language and organization. The study suggests that student writing improve, regardless of feedback method; that preference may not associate with revision; that reading be integrated into L2 writing classes; and that re vision may correlate with feedback intake which depends on learner-focus and teacher-student interaction. Introduction Since the late 1950s, attitudes towards the role of corrective feedback have changed along with the evolution of language teaching methodologies grounded on theories of both educational psychology and second language acquisition with the aim of enabling learners to acquire the target language effectively. In the late 1950s and 1960s, the Audiolingual Method (ALM), based on behaviorism and structuralism, was very popular in second and foreign language classrooms. Error correction was seen as helping learners to form good habits by giving correct responses instead of making structural mistakes. In the 1970s and 1980s, Communicative Language Teaching (CLT), developed from nativism, was commonly practised to equip learners with communicative competence in terms of function over form or comprehensibility over grammaticality. It infers that formal correction should be discontinued since it was deemed as interfering rather than facilitating the acquisition of the target language. In the ear ly 1990s, the Interaction Approach (IAA) emerged, and it entailed such three dimensional phases as learning through input, production of language, and corrective feedback that comes as a result of interaction that arises authentically. Since the mid-1990s, the position of feedback, with the dominance of CLT, has been debated among the theorists, researchers, and practitioners in the fields of second language writing and second language acquisition. In 1996, Truscott, for example, claimed that feedback on student writing should be discarded because it is ineffective and harmful. Ferris (1997), on the other hand, argued that feedback is virtuous as it enables L2 students to revise their own writing and assists them to acquire correct English. Because research evidence was scarce in support of feedback, both Ferris and Truscott called for further research into questions about the impact and provision of feedback on L2 student writing (Bitchener Knoch, 2009). Accordingly, a great body of research has been conducted with a look into teacher written feedback: correction strategies (e.g., Bitchener, Young, Cameron, 2005; Ferris, 1997; Ferris Roberts, 2001; Lee, 1997; Sugita, 2006), feedback forms (e.g., Hyland Hyland, 2001; Silver Lee, 2007; Treglia, 2008), feedback foci (e.g., Ashwell, 2000; Ellis, Sheen, Murakami, Takashima, 2008; Sheen, Wright, Moldawa, 2009), students attitudes toward feedback (e.g., Alamis, 2010; Lee, 2004, 2008a; Saito, 1994; Treglia, 2008; Weaver, 2006), and teachers beliefs about feedback (e.g., Lee, 2004, 2008b). These studies suggested that feedback plays a pivotal role in helping L2 students improve the accuracy and quality of their writing. This finding is in line with the Vygotskyan model o f Zone of Proximal Development (ZPD) which claims that learners need to be provided with scaffolding to be capable of reaching a stage of autonomy and accuracy (Patthey-Chavez Ferris, 1997). However, many of the studies have design flaws in terms of the small sample size or of not having a control group. Other studies explored the effectiveness of other feedback techniques: oral feedback or teacher-student conferencing (e.g., Hedgcock Lefkowitz, 1992; Hyland, 2003; Marefat, 2005; Sheen, 2010a, 2010b), peer feedback (e.g., Kamimura, 2006; Rollinson, 2005; Tsui Ng, 2000), reformulation (e.g., Hyland, 2003; Santos, Lopez-Serrano, and Manchon, 2010), audio-recorded feedback (e.g., Huang, 2000; Jordon, 2004), and computer-mediated commentary (e.g., Ferris, 2003; Hyland, 2003; Hyland Hyland, 2006). However, most of the studies failed to examine which feedback mode was more effective in improving student writing. Even though some of them were comparative in nature, the studies were conducted solely with a group of English-as-a-second-language (ESL) learners. As a result, conclusion is hard to be drawn with regard to the effectiveness of each feedback strategy when it is applied in another classroom setting where English is in the Kachrus (1985) expanding circle or where English is taught as a foreign language. As Ferris (2003) put it, What is preferable cannot be equated with what is effective, and what is effective for one student in one setting might be less so in another context (p. 107). In light of the aforesaid insightful and encouraging premise, this current quasi-experimental research attempts to compare teacher oral and written feedback in terms of perceptions and efficacy among Cambodian English-major students at the National University of Management (NUM henceforth). Definition of Terms: Oral Feedback and Written Feedback According to Rinvolucri (1994), the term [feedback] originates in biology and refers to the message that comes back to an organism that has acted on its environment. In biology it describes a neutral process, a link in the chain of action and reaction. (p. 287) In second language writing, feedback can be defined as input from a reader to a writer with the effect of providing information to the writer for revision (Keh, 1990, p. 294). Simply put, the teacher suggests changes that will make the text easier for the audience to read, or that help the writer to be more aware of and sensitive to his/her reader. When the writer of any piece of writing gets the perspective of the reader, then that writer is able to see more clearly where any points of confusion exist. As Keh (1990) elaborates, The writer learns where he or she has misled or confused the reader by not supplying enough information, illogical organization, lack of development of ideas, or something like inappropriate word-choice or tense (p. 295). In this study, feedback can be operationalized in terms of oral and written feedback (Berg, Admiraal, Pilot, 2006; Hedgcock Lefkowitz, 1992; Hyland, 2003; Hyland Hyland, 2006; Patthey-Chavez Ferris, 1997; Sheen, 2010a, 2010b). Oral feedb ack (OF) refers to the provision of feedback on errors and weaknesses in content, organization, and language (i.e. grammar, vocabulary, mechanics and spelling) through face-to-face conferencing lasting about five minutes for each student-writer. In so doing, the teacher gives comments (in the forms of questions, imperatives, praises, and suggestions), provides correct forms or structures in faulty sentences, tells the location of errors, makes recasts, and gives prompts in the forms of elicitation, clarification requests, and repetition of errors. Written feedback (WF), on the other hand, refers to the correction of errors and weaknesses in content, organization, and language through writing on student paragraphs. In this regard, the teacher makes use of direct versus indirect correction, coded versus uncoded feedback, and marginal versus end comments, in the forms of corrections, questions, imperatives, praises, and suggestions. Literature Review Written feedback A number of studies have been done to examine what to be commented on for substantive revision. For example, Ellis (1994), reviewing several studies on what effect formal corrections have on language acquisition, concluded that the learners whose errors are corrected improve the accuracy of producing existential structures (i.e. There is/are). However, the Ellis-reviewed studies entail only focused feedback, meaning that only one linguistic feature is targeted. Kepner (1991), in a comparative study of feedback on content and grammar, found that students who receive content feedback produce writing that has better content than those who receive grammar feedback. He also found that students who receive formal feedback do not produce fewer errors than the uncorrected group. In another study, Leki (1991) asked 100 ESL freshmen to complete questionnaires to examine how effective feedback was and how they reacted to the positive and negative comments on both form and content. He found that correcting errors in both form and content is beneficial since good writing is viewed as equated with error-free writing. Moving a step away from what to be commented on, several studies have been carried out to investigate how errors should be corrected to improve student writing. According to Ellis (1994), formal feedback is helpful to L2 acquisition only if problems are corrected implicitly or only if the errors are induced and then corrected. In a similar vein, Weaver (2006) explored how 44 students in the Faculty of Business, Art and Design perceived written feedback and if the feedback that they received showed a student-centered approach to learning. In light of interviews, questionnaires, and feedback content, he found that teacher comments are useful only if they are specific and clear, give sufficient guidance, focus on positive points, and are related to assessment criteria. Ferris (1997), examining over 1,600 marginal and end comments written on 110 first drafts by 47 university ESL students, found that marginal comments are more immediate and easier for students to locate errors and revise, whereas end comments can be more useful for writing development since they summarize major problems. Marginal comments are also deemed to be more motivating since the reader is actively engaged with the writers text (Goldstein, 2004, as cited in Hyland Hyland, 2006). In a related vein, much research has focused on whether comment types influence revisions and which of them are more, if not the most, effective. In so doing, Sugita (2006) analyzed 115 revised papers by 75 EFL students at a private university in Japan. He found that imperatives are more effective than statements and questions. In contrast, Conrad and Goldstein (1990, as cited in Hyland Hyland, 2006) found that imperatives, declaratives, or questions were less effective than the type of problem in the feedback. They further explained that problems related with facts and details were successfully revised by 50%, while those dealing with argumentation and analysis were successfully revised only by 10%. Treglia (2008) interviewed two teachers and fourteen students in a community college in the United States to examine how the students reacted to the feedback given by the teachers in the forms of mitigation and unmitigation. This study showed that the students saw both mitigated and dir ective comments easy to revise, but they liked the feedback in the forms of acknowledgements, suggestions, and choices. Alamis (2010) investigated the reactions and responses of 141 students at the Faculty of Arts and Letters of the University of Santo Tomas towards teacher written feedback. In light of questionnaires and student essays, Alamis found that praises are superior to criticisms and that content feedback should entail suggestions rather than questions, direct corrections, and indirect corrections. However, this study is a result of opinion-based responses, so it may be hard to conclude that its findings were valid. Many other researchers have moved farther to find out the extent to which teacher written feedback should be made explicit and sufficient in order to encourage comprehension and revision. In so doing, Enginarlar (1993) used 20-item questionnaires to examine the attitudes of 47 freshmen at Middle East Technical University to coded feedback and brief comment in English Composition I class. This study revealed that the participants like the two feedback types, seeing review work as a type of co-operative learning in which the amount of work and responsibility is shared by students and teachers. Ferris and Roberts (2001) also explored how explicit error feedback should be to help students to self-revise their papers. By analyzing papers written by 72 university ESL students, they found that the treatment groups outdo the control group in relation to self-revision, but the coded feedback group is not statistically different from the uncoded feedback group. Ferris and Roberts also conclude d that less explicit feedback seems to facilitate self-revision just as well as corrections coded by error type. Ferris (2003), in her review of three key studies, suggested that comprehensive feedback (i.e. all errors marked) is preferable to selective one (i.e. only some errors marked) and that indirect correction (i.e. coded and uncoded errors) is more effective than direct one (i.e. teachers making the corrections for students). Lee (2004) analyzed teacher error correction tasks and used questionnaires to and follow-up interviews with teachers and students to examine their perspectives on error correction practices in the Hong Kong secondary writing classroom. Like Ferriss (2003) reviewed studies, this research showed that comprehensive error feedback encourages substantive revision and that students depend on teachers to correct their errors. Oral feedback The effectiveness of oral feedback for improving student writing has still got very few answers (Hyland Hyland, 2006). As such, several studies have been done to examine teacher-student dialogue, and they found that successful conferencing rests on the interactive nature. For example, Hyland (2003) claimed that conferencing is fruitful when students are actively involved, asking questions, clarifying meaning, and arguing instead of accepting advice. Johnson (1993, as cited in Gulley, 2010) did a qualitative study and concluded that the question, a tool often used by teachers and tutors during a writing conference, can be ineffective in eliciting a meaningful response from students (p. 13). By contrast, Carnicelli (1980, as cited in Gulley, 2010), in his qualitative study among English-major students at the University of New Hampshire, showed that conferencing is more preferable to in-class teaching. He also noted that conference might fail if the teacher does not listen to the student, if the student feels insecure, or if the student does not remember the teachers comments (p. 13). However, this study has a design flaw in terms of not having a control group, so it is hard to conclude if such a preference is a result of conferencing, instruction, or practice. In his response to Carnicelli, Keh (1990) did his article review and pointed out that conferencing fails when the teachers take an authoritarian role, dominate the conversation, and pay no attention to what their students ask during the dialogue. He also noted that teacher-students conferencing is more effective than teacher-student conferencing since the former allows them to learn ideas and problems from one anot her. Moving a step away from the teacher-student interaction, several studies have been conducted, focusing on students-related variables that may affect the substantive revision of student writing. In so doing, Marefat (2005) examined the perception about the efficacy of oral feedback on the improvement of writing among 17 male and female Iranian students of English as a foreign language. He found that males could write paragraphs better than females, whereas females outperformed males in essay writing. He concluded that the students can produce pieces of writing with better quality, regardless of the feedback technique. Patthey-Chavez and Ferris (1997, as cited in Hyland Hyland, 2006) investigated how four writing teachers did conferencing with poorer and better students. They found that however useful teacher suggestions were for revision, the poorer students seemed to use advice more often than their counterparts. Better students were more self-confident, and they often used teacher suggestions as a base to revise their own writing. The co-researchers suggest that in the case of less capable students, conferences may be harmful if they entail appropriation rather than intervention. In another study, Goldstein and Conrad (1990, as cited in Hyland Hyland, 2006) noted that the L2 learners having cultural or social inhibitions about engaging informally with teachers are most likely to passively and unreflectively use teacher advice to revise their writing. The co-researchers found that only students negotiating meaning well in conferences were able to perform revision more successfully. This finding was similar to that of Williams (2004, as cited in Hyland Hyland, 2006), claiming that students were successful in using advice when teacher-suggestions were direct, when students actively engaged in negotiating meaning, and when they took notes of teacher comments, during the dialogues. Williams also added that negotiation is a precondition for revising higher-level texts, although her research suggested that conferencing has greater impact on correcting local errors (as cited in Hyland Hyland, 2006). However, the findings of all the four studies are based on the sm all sample size, so it is unclear if conferencing strategies and other contextual factors play a part in improving student writing. In line with the studies grounded on L2 writing theory, a number of studies have been done based on the theories of second language acquisition to investigate the impacts of indirect and direct corrective feedback, focusing on single linguistic structures. For example, Ellis, Loewen, and Erlam (2006, as cited in Sheen, 2010b) did an experimental study to examine whether implicit or explicit feedback is more helpful for adult ESL learners to acquire regular past tense. They put the students into three groups: the group with implicit recasts, the group with explicit metalinguistic feedback, and the group without any corrective feedback. The findings showed that both implicit and explicit feedback does not have any impact on the immediate posttests, but the latter is more effective than the former on the delayed posttests. In another study, Sheen (2007, as cited in Sheen, 2010b) found that explicit corrective feedback is superior to implicit corrective feedback in terms of formal acquis ition in both the immediate and delay posttests when the former is provided in the form of metalanguage and the latter in the form of recasts. Several other studies have also been done to compare input-providing feedback in the form of recasts with output-prompting feedback in the forms of elicitation, clarification requests, repetition of error, and metalinguistic clues. Lyster (2004, as cited in Sheen, 2010b) did a study with a group of fifth-grade French learners to examine whether recasts or output-prompting feedback methods encourage more accuracy of using articles agreeing with the gender of nouns. The study revealed that the output-prompting group alone outdid the control group on all eight measures of acquisition. Ammar and Spada (2006, as cited in Sheen, 2010b) investigated if recasts are more effective than prompts on the acquisition of possessive pronouns among six-grade learners in intensive ESL classes. They found that prompts were more helpful only for students with pretest scores below 50 percent, whereas recasts and prompts together were less effective for those whose scores were below 50 percent. However, t hese studies entail only focused corrective feedback, meaning that only one linguistic feature was targeted. Therefore, the results are hard to be generalized since the effects of recasts and prompts might be different if multiple-linguistic features are corrected. Research Questions As can be seen, no research had been conducted before to explore the comparative effectiveness of oral and written feedback in improving student writing in the context where English is in the Kachrus (1985) expanding circle. Accordingly, the present study sets out to look for answers to the following two research questions: How do Cambodian English-major students at NUM perceive oral and written feedback? Which feedback strategy, oral or written, is more effective in improving student writing as measured by writing performance? Methodology Participants Thirty-seven students participated in the present study, 19 of whom were males and 18 were females, with an average age of 22.59 (SD=.62) years. They were English-major students at NUM, and they had been learning English since Grade 7 of Cambodian Secondary Education (G7CSE) under the Ministry of Education, Youth, and Sports (MoEYS). The subjects were selected from each English class of the university based on the pre-treatment scores of 150-word paragraph writing. Based on this criterion, 19 of them were put into the oral feedback (OF) group, and 18 were filtered into the written feedback (WF) group. A control group was excluded from this study for two main reasons. First, it is believed that feedback is an essential element, so to get student to write without feedback would be unfair to them. Second, it is claimed that one of the things that students expect from teachers is feedback, so to deny them feedback would be unethical. Instruments Two instruments employed in this study were questionnaires and student paragraphs, both of which were used for data triangulation. The questionnaire, so-called Affective and Effective Response Feedback (AERF), consists of three sections with a total of 22 items: Section A (A1-4 for Demographic Data), Section B (B1-9 for Effective Responses), and Section C (C1-8 for Affective Responses). A five-point Likert scale (1=Strongly Disagree, 5=Strongly Agree) was utilized for the 17 items in the latter two sections (B1-9 and C1-8), and several items thereof (i.e. B1, B3, B5, B8, C2, C5, and C8) were reverse-ordered to reduce response set bias. A statistical validity analysis showed that EARF was reasonably reliable with the Cronbachs Alpha value of .853. The student paragraphs were collected before and after the two-month treatment, and they were inter-rated by three well-trained teachers each with more than four-year experience of teaching writing skills to English-major university students. The scoring was performed based on the researcher-formulated criteria divided into content, organization, grammar, vocabulary, mechanics and spelling, each of which earns equal marks (1=Very Poor, 5=Excellent), with the total score of 25. The reliability of the inter-rated scores employed by the present study was .789 for the pre-treatment scores and .806 for the post-treatment scores, using Cronbachs Alpha coefficient of internal consistency. Procedures Before this study, letters were sent to the Chair of Foreign Languages Center (FLC) of NUM, where it was conducted, and finally to the Rector of the university as well. Once approval letters were received, the researcher went on to select classes for both groups (OF and WF) and sent out informed consent forms. There were informative meetings with both groups of participants to let them know about the study and to receive signed informed consent forms. It was also made clear to the subjects that this study would not affect their course grades. They responded to the questionnaires anonymously, and those who mastered more than 80 percent of the total scores in paragraph writing would be awarded with Certificate of Recognition in order to motivate them to write and incorporate the feedback they had received from one week to another into their revision process. Data collection was conducted in the following steps. First of all, the students were asked to write a 150-word paragraph about the person whom I admire in my life. The paragraphs were then collected and inter-rated by three well-trained lecturers who had been teaching writing skills for more than four years. Based on the results, the participants were divided into two groups of similar size (OF=19, WF=18) and overall equivalent writing competence. An independent-samples t-test revealed that the overall mean score of the oral feedback group constituted 16.47 (SD=3.042) and that of the written feedback group was 16.46 (SD=3.045). Then, the treatment was conducted for two months with single-draft feedback provided on each of the three paragraph types taught during this experimental period: narrative, process, and compare-contrast. The topics included my happy story, how to make a nice cup of coffee, and rural life and city life. The feedback on each topic was comprehensive and targeted all aspects of writing: content, organization, grammar, vocabulary, mechanics and spelling. In this regard, various feedback strategies of each commentary mode were employed to ensure that both groups would receive similar treatment condition and that they would provide more authentic responses to the research questions. It is worth noting that the treatment (i.e. the delivery of feedback) was undertaken with specific reference to the operationalized terms at the very beginning of this study (Please refer to pages 3-4.). Soon after the two-month study, the participants were again asked to write a 150-word paragraph about one of the three topics (i.e. my bedroom, my house, or my favorite place), complete the questionnaires consisting of both close- and open-ended items. Finally, the data obtained from the questionnaires and student paragraphs were coded and input into SPSS 19.0 with the utilization of one-sample t-test, independent-samples t-test, and paired-samples t-test for data analysis, using the test value of 3.5 and the significant level of .05. Findings and Discussion Research question 1: How do Cambodian English-major students at NUM perceive oral and written feedback? A one-sample t-test was employed to provide descriptive statistics by comparing the mean scores and standard deviations of the oral feedback group and the written feedback group with the test value of 3.5 rather than with those of the written feedback group and the oral feedback group, respectively (i.e. oral feedback group vs. written feedback group, and vice versa). Table 1 shows that the students had highly positive attitudes towards oral feedback in the forms of detailed correction (M=4.42, SD=.838, p=.000), comprehensive suggestion (M=4.26, SD=.806, p=.001), and sincere praise (M=4.00, SD=.816, p=.016), which thus enabled them to write with increased confidence (M=4.26, SD=.452, p=.000). This preference was due to the fact that oral feedback was perceived as the cornerstone of building closer bonds (M=4.16, SD=.765, p=.001) between the student and the teacher who always paid special attention during each dialogue (M=4.58, SD=.507, p=.000). However, no statistical differences wer e significant in motivation (C5. It encouraged me to work harder on my revision) and sufficiency (C8. It was helpful enough for my revision), the p-values of which constituted .137 and .497, respectively. Table 1 Descriptive statistics for affective responses of OF group M SD t df p C1. It made me feel I had a more personal and human relationship with my teacher. 4.16 .765 3.750 18 .001 C2. I did not feel more confident about my writing. (Reverse-ordered) 4.26 .452 7.353 18 .000 C3. It gave more details about the errors in my writing. 4.42 .838 4.793 18 .000 C4. It gave more details about how I can improve my writing. 4.26 .806 4.129 18 .001 C5. It discouraged me from working harder on my revision. (Reverse-ordered) 3.84 .958 1.556 18 .137* C6. Praise was helpful for my revision. 4.00 .816 2.669 18 .016 C7. I got special attention from my teacher. 4.58 .507 9.271 18 .000 C8. It was not helpful enough for my revision. (Reverse-ordered) 3.68 1.157 .694 18 .497* * p > .05 (not significant) As can be seen in Table 2, students preferred written feedback in the forms of comprehensive correction (M=4.39, SD=.698, p=.000), detailed suggestion (M=4.39, SD=.608, p=.000), and sincere praise (M=4.22, SD=.647, p=.000), to make them feel more confident about their writing (M=4.00, SD=.594, p=.002). A one-sample t-test also indicates that statistical differences were significant in attention (C7) [M=4.22, SD=.808, p=.001], but not in relationship (C1, p=.655), encouragement (C5, p=.055), and sufficiency (C8, p=.080). Taken Tables 1 and 2 together, oral feedback, unlike written feedback, builds closer bonds between the teacher and the student because the former tends to be more interpersonal in terms of reciprocal attention during the dialogue. While written feedback, if it includes encouragement and personal, text-specific comments, can also strengthen teacher-student relationships, it is not the same experience as sitting down face-to-face for negotiation and questions. Table 2 Descriptive statistics for affective responses of WF group M SD t df p C1. It made me feel I had a more personal and human relationship with my teacher. 3.61 1.037 .455 17 .655* C2. I did not feel more confident about my writing. (Reverse-ordered) 4.00 .594 3.571 17 .002 C3. It gave more details about the errors in my writing. 4.39 .698 5.404 17 .000 C4. It gave more details about how I can improve my writing. 4.39 .608 6.206 17 .000 C5. It discouraged me from working harder on my revision. (Reverse-ordered) 4.00 1.029 2.062 17 .055* C6. Praise was helpful for my revision. 4.22 .647 4.738 17 .000 C7. I got special attention from my teacher. 4.22 .808 3.790 17 .001 C8. It was not helpful enough for my revision. (Reverse-ordered) 3.00 1.138 -1.87 17 .080* * p > .05 (not significant) Table 3 presents the descriptive statistics of the perceptions about the impact of oral feedback on improving student writing. A one-sample t-test was performed with the test value of 3.5 and the p-value of .05. The results show that oral feedback was viewed as effective in encouraging substantive revision of organization (B4) [M=4.32, SD=.671, p=.000], clarity (B1) [M=4.05, SD=.780, p=.006], content (B5) [M=4.00, SD=.577, p=.001], and grammar (B2) [M=3.95, SD=.705, p=.013]. Significantly, oral feedback was also seen as enabling students to use specific linguistic features in conformity to different genres or text-types (M=3.95, SD=.705, p=.013). Such an improvement was strongly confirmed by the