Summary of Findings

CHAPTER FIVE

CONCLUSIONS

This chapter first presents a summary of findings concerning the effects of two correction methods on students’ writing quality and accuracy, and their attitudes toward the treatments. Then pedagogical implications for EFL researchers and high school writing teachers are provided. Finally, the limitations of the study and suggestions for future studies are discussed.

Summary of Findings

This present study investigates the effects of two error correction methods on the EFL high school students’ performances at narrative writing. Two groups of 45 high school female students participated in the study. The direct group received direct error correction with the teacher’s model answers provided for their errors, while the code group received code error correction with the types of their errors identified and labeled. After the treatments, students’ improvement in overall writing quality and accuracy was measured, and their attitude’s toward the error treatment they received was also explored. The findings are summarized and presented as follows.

First, although both groups made significant improvement in their posttest writing in terms of their writing accuracy and quality, neither of the two groups outperformed its comparable group. The two correction methods were equally effective in improving students’ writing quality and accuracy.

Second, when comparisons were made at two different levels of writing

proficiency, the effects of the two treatments were different. Students of higher

proficiency benefited more from code error correction than direct error correction in

terms of their improvement in writing quality and accuracy. As to students of lower

proficiency, the two correction methods were equally effective in improving their

writing quality and accuracy.

Third, based on students’ responses to the attitude questionnaire, students from the code group were found to perceive more improvement in their writing autonomy than students from the direct group. In terms of writing and editing ability, more proficient students from the code group perceived more improvement than their counterparts from the direct group. On the other hand, less proficient students from the direct group perceived more improvement in their writing ability than those from the code group.

Fourth, with regard to their attitudes toward correction responsibility, more than half of the students preferred code correction and most of the students believed that students had responsibility for correcting their own errors.

Finally, when exploring students’ preference for correction methods, difference was found between students of different proficiency levels. Higher-proficiency students preferred code error correction to direct error correction whereas students of lower-proficiency expressed their preference for direct correction. More proficient students liked code correction better because it demanded cognitive efforts from them which in turn facilitated their memorization of the correct linguistic forms of their errors. Less proficient students, on the other hand, liked direct correction because it reduced their frustration arising from their failure to find the correct answers for their errors. They liked their teacher’s model answers which gave them more confidence in producing comprehensible writing.

Pedagogical Implications

The findings derived from this study show that both code error correction and

direct error correction are effective correction methods to help students improve their

writing quality and accuracy. Hence, apart from the dominant method of direct error

correction used by senior high school teachers, code error correction is an equally

effective and feasible alternative. However, when taking students writing proficiency into consideration, effects of the two treatment methods differ on students of different proficiency. Suggestions are offered if writing teachers intend to practice error

correction in their writing instruction.

First, a less explicit method, such as code error correction conducted in this study, might be sufficient to provide higher-proficiency students with prompts for their self-correction in the rewriting process. Code error correction, compared to direct correction method, better tailors to these students’ needs since it helps internalize students’ grammar knowledge and raise their grammar awareness.

Second, the practice of direct error correction is recommended to writing teachers with a group of lower-proficiency students. For these students, the

implementation of code error correction might be risky for excessive efforts spent in locating the correct forms of their errors might consume students’ patience and result in their frustration. A more explicit method like direct error correction is considered to better serve their needs.

Limitations of the Study

There are some limitations in this study. First, the experimental design of this study did not incorporate a control group which received no feedback concerning grammar errors from the teacher but self-edited their own writing. The lack of the no-feedback group might bring about doubts on the participants’ progress in writing quality or accuracy, which could be attributed to other factors such as the instruction of regular English classes.

Second, since the participants involved in this study were only two classes of

senior students in a senior high school. They were all females and there were limited

samples. Therefore, it might be bold, if not possible, to generalize the results of the

study to all the EFL writers in senior high schools in Taiwan.

Third, although this study measured students’ progress over time by comparing the students’ pretest writing with their posttest writing, it did not examine how successfully students edited their texts in the short-term. Therefore, it was not clear that how much of the cognitive investment of editing one’s tests after receiving teacher’s error feedback was transferred to the long term improvement in writing accuracy.

Fourth, since the error treatment in this study lasted only four months, the long term effectiveness of the two treatment methods might not be well addressed. A study that requires students to write over a longer period of time, for example, one academic year, may reveal different results. To address this complicated issue more fairly, longitudinal studies are needed.

Suggestions for Future Research

The current study adds to the previous research base on the effects of direct correction and code correction on EFL students’ writing ability, writing accuracy and attitudes toward the treatments implemented. Yet many questions related to the research on error correction methods remained unanswered. Therefore, suggestions for future studies are given as follows.

First, it is recommended that future studies consider adding a control group which receives no feedback concerning grammar correction in the study design. In this way, it can be better claimed that students’ gains in writing quality and accuracy are attributed totally to the effectiveness of the error treatment they received.

Second, future studies can also consider including the measurement of students’

improvement in their content of writing. In this study, although improvement in both

writing quality and accuracy are measured, the improvement or the deterioration in

the content of writing is not observed. To fully address the influence of two treatment

methods on every aspect of writing, an inclusion of measuring the content is

recommended in future studies.

Third, it is also recommended that future studies include a direct method to examine students’ ability of editing from one draft to the next after receiving feedback.

Such measure will allow students’ comprehension of error feedback and their uptake of correction to be fully assessed.

Fourth, students should be asked to write compositions in different genres to fully explore the effectiveness of the treatments. The four topics that students wrote on in this study belonged to one genre only—they are all narrative essays. How different genres influence students’ progress in writing quality and accuracy has not been investigated. Therefore, different genres such as description, argumentation or exposition are recommended to be included in the future studies.

Fifth, the participants recruited in this study were only female high school students. To generate the results to a greater population of high school students, a larger sample of different genders can be considered when further studies are conducted.

Finally, difference in writing proficiency is the only learner variable investigated

in this current study. Many other learner variables such as different learner styles,

different motivations, and different anxieties may also influence the outcomes of the

study. Therefore, the interaction between error correction methods and other learner

variables can be further explored in the future studies.