Taiwanese Student Attitudes Towards Error Correction and Written Corrective Feedback in L 2 Writing Classes

Error correction and corrective feedback in L2 writing has long been an important issue for EFL teachers and researchers, as well as EFL students. This study aimed to examine a population of Taiwanese EFL writing students to assess their attitudes and preferences for corrective feedback in their English writing. This study utilized a 31question self-report survey to gather data from 79 EFL writing students. Results indicate that most students felt it was very important for the teacher to provide comprehensive direct or indirect coded feedback on all types of errors, but mostly on grammatical errors. The findings hold important implications for EFL writing teachers in Taiwan. EFL writing teachers should have open dialogues with their students in order to determine the most beneficial form of feedback for each class.


Introduction
Accuracy is important for learners that want to achieve target-like proficiency in language learning, and written corrective feedback remains one of the most popular methods to motivate learners to discover their writing errors and correct them.English teachers in Taiwan have long given feedback to their students to improve their accuracy and help them overcome difficulties in English writing.Corrective feedback is also used to repair learners' erroneous uses of English, or, as Ellis (2009) stated, it is "the form of a response to a learner utterance containing a linguistic error" (p.3).Thus, corrective feedback is important not only for focusing a learner's attention on linguistic forms but also for prompting them to correct their errors.
Since Truscott (1996) declared corrective feedback (CF) in L2 writing to be ineffective, harmful, and a waste of time, it has been a contentious issue for many educators and researchers.Since then, much research has shown it to be effective in certain educational contexts (Bitchener, 2008;Bitchener & Knoch, 2008, 2009;Ellis, Sheen, Murakami, & Takashima, 2008;Sheen, 2007;Sheen, Wright, & Moldawa, 2009;Van Beuningen, de Jong, & Kuiken, 2008, 2012).Written error correction continues to be the most widely used form of feedback in L2 writing classrooms.Moreover, L2 writing students have consistently shown that they both want and expect to receive some type of feedback in their writing (Elwood & Bode, 2014;Ferris & Roberts, 2001;Lee, 2004;Leki, 1991;Radecki & Swales, 1988).Indeed, Ellis (2008), in reviewing previous research on corrective feedback concluded that "there is clear evidence that corrective feedback contributes to learning" (p.885).
This study aims to examine EFL student attitudes towards the form of error correction that they receive.Specifically, this study will examine student attitudes towards different forms of error correction and corrective feedback to ascertain Taiwanese students' attitudes towards corrective feedback in L2 writing.This study is therefore concerned with exploring "students' perceptions regarding those factors believed to enhance the learning of a new language" (Schulz, 2001, p. 244).

Literature Review
A previous study assessing student attitudes towards corrective feedback was done by Leki (1991).She found that ESL students were very interested in avoiding errors in their written work and that avoiding their request for error correction may have negative effects on their motivation.These findings were similar to those of Radecki and Swales (1988), who found that students had mostly positive or neutral reactions to receiving heavily marked papers, and that students "expressed satisfaction that their teacher had marked their papers" (p.357).Leki suggested that teachers set aside time in class to discuss with their students the preferred form of error correction and which teaching behaviors would be most helpful to students' progress.Radecki and Swales (1988) and Leki (1991) both found that students showed a clear preference for indirect error correction by the teacher, where the teacher would mark the error and give a clue about how to correct it or point to a section of the book that discussed a particular grammatical point relevant to the correction.These findings were echoed by Ferris and Roberts (2001), whose students indicated that they preferred for the teacher to mark errors and label them with error codes.Lee (2004) found that secondary students in Hong Kong preferred to receive comprehensive error correction and that "students preferred teachers to play a primarily active role in error correction" (p.295).Similarly, Elwood and Bode (2014), investigating student preferences for corrective feedback at a university in Japan found that students preferred direct, handwritten, detailed feedback on content and mechanical errors.Lee, however, found that students preferred the use of error codes as opposed to direct error correction.Moreover, students in Elwood and Bode's study preferred to have their feedback written on the manuscript rather than on a separate piece of paper.Most students were against the use of e-feedback or feedback via computer.Somewhat suprisingly, the students did little with the feedback, even though they felt it was necessary to learning and even though revisions were necessary for some of the students.Despite these findings, there is a paucity of research in investigating Taiwanese students' attitudes towards error correction in their English writing.In order to investigate this issue, the following research questions guided this research.
RQ1.How concerned are Taiwanese students with errors in their English writing?RQ2.What do Taiwanese students think are the most important features for their English writing teachers to provide error correction and feedback on?RQ3.What do Taiwanese students look at when they receive error correction or feedback from a teacher?RQ4.What or who do Taiwanese students consider the best source of help with their written work?RQ5.What form of corrective feedback do Taiwanese students prefer their English writing teachers to use when providing error correction and feedback?

Research Methodology
The population for this study were 79 students studying English in a Department of Applied Foreign Languages undergraduate degree program at a private university in central Taiwan.Students were in second-and third-year writing courses.Respondents were composed of 32 males (40.5%) and 47 females (59.5%).Respondents were between 19 and 25 years of age (M = 21).It should be noted that many of the students come from vocational high schools and have low proficiency in English; moreover, they are required only to pass the TOEIC test at the B1 level before graduating.
The instrument was a self-report survey adapted from Leki (1991).The survey was modified by the researcher and translated into Chinese by the researcher's assistant, who has experience in English-Chinese translation.Unfortunately, the study by Leki suffered from several flaws.First, the survey questionnaire asked questions not directly related to the research questions.For example, questions 17-19 in Leki asked respondents about their teachers' error-marking behavior; the discussion of these results spans two paragraphs, but does not appear to contribute to the five foci of the paper.As such, these questions have been taken out of the current survey.Additionally, Leki displayed the results as frequencies and percentages only, eschewing the standard practice of providing means for central tendency and standard deviations for variability of Likert scale results (Boone & Boone, 2012).In the current study, this has been amended, and results are displayed as both means and standard deviations displayed in tables with additional frequencies and percentages displayed in the full survey in the Appendix.Finally, Leki did not examine the reliability of her instrument.In the current study, Cronbach alpha coefficients for the various sections of the study have been calculated and the reliability of the measurements has been established.
The instrument was additionally modified by changing the 7-point scale to a 5-point scale.This was done to simplify the options for the subjects.As Dawes (2008) has pointed out, there are "no appreciable differences in terms of standard variation, skewness or kurtosis" (p.75) when examining the differences between 5-point, 7point, and 10-point scales, with the only exception being that 5-and 7-point scales tend to produce higher mean scores than 10-point scales.The instrument was also modified on the advice of a Taiwanese colleague; the scale was reversed from 1 to 7 moving from positive to negative to 1 to 5 moving from negative to positive.It was thought that this would make it more appropriate and understandable to a Taiwanese audience.
The questionnaire was administered online by means of Google Forms and made freely accessible to all students via the use of a web browser, so participants had the option of taking the survey at their leisure without time constraints or using up class time.Google Forms allows the survey to be taken one section at a time and also for questions to be marked as obligatory, so that subjects cannot skip questions and thus invalidate a submission.The survey was administered in both English and Chinese (English with Chinese in parentheses), but only the English version appears in the Appendix, along with the numbers and percentages for every answer.
The survey consisted of four parts and a total of 31 questions.Before taking the survey, demographic information such as nationality, age, and year in school were gathered for each student.In the first part of the survey, there were seven Likert-type scale questions assessing students' attitudes towards error correction.All questions were rated on a scale ranging from 1 (not important at all) to 5 (very important).The first three questions assessed students' beliefs in the importance of having as few errors as possible in their written work.These three questions were found to have an acceptable reliability (α = .802).The next four questions assessed students' attitudes about the importance of the teacher pointing out errors in their written work.Like the previous section, these four questions also had an acceptable reliability level (α = .779).The second part consisted of six questions assessing what students did with the error correction they received.All questions were rated on a scale ranging from 1 (never) to 5 (always).These six questions were found to have a high level of reliability (α = .958).The third part consisted of 11 questions concerning students' preferences for different types of feedback and took the form of checkboxes, of which the student could select an appropriate answer.The last part consisted of seven different error correction examples which the students rated on a scale ranging from 1 (very bad) to 5 (very good).The examples included direct corrective feedback, indirect corrective feedback, and no feedback, such as not indicating an error, or responding to the ideas in the paper but not the error.The study was limited to these types of errors, as it was felt that these were the only types of feedback to which the subjects had been exposed during their writing courses in their undergraduate studies.
Direct corrective feedback in this study is defined as the provision of the correct linguistic form or structure above the error.Indirect corrective feedback is feedback that indicates that an error has been made in some way without providing the correct form or drawing attention to it.This may include underlining the error or indicating where the error has occurred and what type of error it is, for example, a verb tense error (Bitchener, 2009).
This research utilized a quantitative design.The survey data from the questionnaire was collected and analyzed with SPSS 22 to obtain Cronbach Alpha coefficients as well as response frequencies and percentages for each answer.

Findings
The first three questions examined students' attitudes towards errors.For simplicity, rankings 1 (not important at all) and 2 (not very important) were condensed to one category, not important, while rankings 4 (important) and 5 (very important) were also condensed to one category, important.Ranking 3 (unsure) was kept as is.These three questions achieved an overall mean of 4.35.Results indicate that the respondents overwhelmingly felt it was important to have as few errors as possible in their written work and to have the English teacher point out any existing errors.The means and standard deviations for the first three questions can be seen in Table 1.Questions 4, 5, 6, and 7 queried respondents on their attitudes towards the importance of their English writing teacher pointing out different types of errors in their written work, namely errors in grammatical forms, spelling, vocabulary choice, and punctuation, respectively.Results indicate that students overwhelmingly felt it was important for their English writing teacher to point out all of these types of errors.Means and standard deviations for these four questions can be seen in Table 2.The second part of the survey consisted of six questions concerning students' habits and how they used error correction.These were rated on a scale from 1 (never) to 5 (always), and these six questions had an overall mean of 4.34.Questions 8 through 11 asked whether students looked carefully at the marks indicating errors in grammar, spelling, vocabulary choice, and punctuation, respectively.Responses indicated that students tended to look carefully at all these types of errors.Additionally, questions 12 and 13 asked whether students looked carefully at comments on the organization of the paper or comments on the ideas expressed in the paper, respectively.Similarly, the results showed that students look at comment son both the organization of and ideas in the paper.Statistics for section two of the survey can be seen in Table 3.
The third section of the survey consisted of 11 questions concerning students' preferences for types of feedback and how the teacher should mark errors.These questions took the form of checkboxes where the subjects could check the most appropriate answer.Responses for question 14 showed that 56 respondents (70.9%) preferred the teacher to mark all errors, major and minor, in their written work.
Responses for question 15 indicated that 58 respondents (73.4%) preferred error correction to be done with a red pen.The results of question 16 show that respondents were split, with 45 respondents (57.0%) prefer direct feedback, whereas 30 respondents (38.0%) prefer indirect feedback.Question 22 asked students where they go for help if they do not know how to correct an error.Forty-nine students (62.0%) indicated that they asked another student or friend, whereas only 18 students (22.8%) indicated that they went to the teacher for help, and 10 students (12.7%) indicated that they looked in a grammar book.Question 23 expanded on these responses by asking whose advice, out of the responses for question 22, did the students usually remember the most.Thirty-nine students (49.4%) indicated that they remembered another student's or friend's advice the most, and 30 students (38.0%) responded that they remembered their teacher's advice the best.Finally, question 24 asked students what helps them the most to learn from the errors marked on their paper and helps them to avoid making the errors again.Thirty-one students (39.2%) responded that they preferred to rewrite the sentence in which the error appeared on another piece of paper, whereas 29 students (36.7%) preferred to rewrite the entire assignment.Full frequencies and percentages for the entire survey can be seen in the Appendix.
The last section of the survey presented students with examples of seven forms of feedback and asked them to rate the examples on a scale from 1 (very bad) to 5 (very good).Question 25 gave students an example of indirect metalinguistic explanation with the error underlined and the feedback referencing the section of the book in which the correct form could be found.Student responses showed a mean average 3.18, indicating that students felt this was neither a good nor bad way to mark errors.Question 26 presented students with an example of direct corrective feedback, with the error underlined and the correct form written above it.This form of corrective feedback attained a mean average of 3.96, indicating that overall, students felt it was a good way to indicate errors.
Question 27 presented students with another example of indirect feedback.The error was underlined and an invitation to talk to the teacher about their feelings or the ideas expressed in the writing.The majority of students did not like this type of feedback, and it achieved an overall mean of only 2.66.Question 28 presented students with an example of indirect coded corrective feedback, with the error underlined and the type of error (verb tense) written above it.Overall, students liked this type of feedback and rated it at an average of 3.67.
Question 29 presented students with another form of direct feedback, with the error underlined, and an arrow showing where the word should be moved to.Student responses were spread among the ratings, and it received an average rating of 2.91.In Question 30, students were presented with the underlined error, but there was no feedback indicating how to correct it, the correct form, or where to reference the correction.Responses to this type of feedback indicated that students did not like it, as it achieved an average rating of 2.57.The final example, question 31, presented students with no feedback.The error was not underlined and there was no feedback from the teacher.Responses showed that the majority of students felt this form of nonfeedback was not good, and it achieved an overall mean of only 1.91.Table 4 displays the means and standard deviations for the last section of the survey.

Discussion
The first research question asked how concerned Taiwanese students are with errors in their written English work.Results indicate that students overwhelmingly felt it was important to have as few errors as possible in their written work.Students also perceived that it was important to their English writing teacher and their other teachers that they have as few errors as possible in their written work, so the answer to research question 1 is clear: Taiwanese students place great importance on reducing the errors in their English writing and are very concerned with learning correct grammatical forms in order to achieve this.These results are similar to those of Leki (1991).
Research question 2 investigated student attitudes towards the most important types of errors for their English writing teachers to provide feedback on.Results show that students felt it was important for their teachers to point out all types of errors, including grammatical errors, spelling errors, errors in vocabulary choice, and punctuation errors.Additionally, research question 3 asked what Taiwanese students pay attention to when they receive feedback on their English written work.The results showed that the majority of students looked closely at feedback indicating all types of errors in their written work, so the answers to research questions 2 and 3 are also clear.
Students felt it was important to get feedback on grammatical, spelling, vocabulary, and punctuation errors, and they tended to look carefully at all of these types of errors when receiving feedback from the teacher.Additionally, the results from question 18 indicate that students paid the most attention to marks indicating errors in grammar, whereas the results of question 19 indicate that students tended to remember comments on the organization of the paper the most.Again, these responses are similar to those found by Leki (1991) and Lee (2004), who also found that students preferred comprehensive error correction.Similarly, Radecki and Swales (1988) found that students who were at least semi-receptive to feedback "desired to have all their linguistic errors marked" (p.358).
Research question 4 asked who or what students consider the best source of help with their written work, and the results show that students were split evenly between asking the teacher and asking a student or friend for help.Only nine respondents preferred to check the error in a grammar book, and only three preferred to ask another native English speaker for help.Additionally, most respondents indicated that looking in a grammar book was the least helpful in understanding the error.Moreover, the survey results showed that when students were not sure how to correct an error, the majority of respondents asked another student or friend for help and tended to remember advice from other students or from the teacher the best.These findings are slightly different from those of Leki (1991), who found that students preferred to ask the teacher or check in a grammar book rather than asking other students for help.Indeed, her results showed that students felt that asking other students to explain the problem was the least helpful method of correcting errors.Her students overwhelmingly preferred to ask the teacher and remembered the teacher's advice the most.These results can perhaps best be explained by the dissimilarity between the two populations.Leki's population consisted of 100 students from 37 different countries who had all scored at least 535 on the TOEFL and thus already reached a B1 proficiency level or higher.The higher proficiency students in Leki's study probably felt that other students would not have the knowledge required to explain problems, whereas the lower-proficiency students in the current study perhaps felt that they could not understand the teacher and would turn to their fellow students for help.
Research question 5 asked what form of error correction and feedback Taiwanese students preferred.The results show that the majority of students showed a preference for comprehensive error correction (marking all errors, major and minor) using a pen with red ink.Additionally, the majority of students preferred either direct or indirect forms of error correction.When rating examples of error correction and feedback, questions 26 and 28, displaying direct correction with the correct form written above and indirect coded feedback, respectively, received the highest ratings.These results also echo those of Leki (1991), whose students preferred comprehensive error correction.However, Leki found that her students preferred indirect coded feedback as opposed to direct feedback, perhaps because the respondents of that survey indicated that their current writing teacher used indirect coded feedback; thus, the students in Leki's study had developed a preference for that type of feedback.
Although the students in the current survey were similar-their current writing teacher also used indirect coded feedback-responses were more split between direct and indirect feedback methods.This discrepancy can perhaps best be explained by the populations being examined in each study.As previously mentioned, Leki's population consisted of 100 students from several countries.For the current study, the population is more homogenous and consists of 79 lower-proficiency Taiwanese EFL learners.Most of the subjects in the current study will struggle to attain a B1 level of proficiency before graduating, so perhaps they feel that direct feedback is more beneficial with their limited linguistic knowledge.Leki's students, who had better language skills, may have preferred indirect coded feedback as a means to think about the language and the way it works rather than just receiving the correct form, so they showed a preference for indirect examples of error correction.The results of the current study are, however, similar to those of Elwood and Bode (2014), who found that "detailed feedback was desired on both mechanical issues and content" (p.338) and that "both males and females strongly preferred direct feedback and coded feedback" (p.339).Ellis (2008), in his meta-analysis of corrective feedback studies also found that "overall, the results point to an advantage for explicit over implicit corrective feedback" (p.884), although the explicit instruction was operationalized in various ways throughout the studies, it did include "simply indicating that an error has been committed" (p.884), or direct and coded feedback.
One small difference, however, is that the students in the current study expressed a preference for error correction with a red pen, whereas the majority of students in Leki's (1991) study responded that it did not matter what color pen was used, and those in Elwood and Bode's (2014) study who had received feedback in blue pen showed a preference for that color, although students showed an overall significant preference for red pen.Again, this may be explained by the more homogenous nature of the current study's population as opposed to the more international and culturally dissimilar grouping of students in Leki's study.Perhaps students in general show no preference, but become accustomed to whatever color their teachers happen to use, which would explain the results of all three studies.

Conclusion
There are several limitations in this study.First, because of the small sample size, the results are not generalizable to EFL learners or even to Taiwanese students.However, this data does provide valuable feedback to the researcher and other writing teachers working with lower-proficiency EFL students in Taiwan.Another limitation is that this was a self-report survey, so the results may not be accurate.Although students were instructed to answer honestly and accurately and were ensured that their responses would be anonymous, students may have answered in a way that they thought their teacher or the researcher would want them to answer, thus skewing the results.This is an inevitable and unavoidable limitation of self-report surveys.
Another limitation to this survey is that the survey made no mention of which draft of a piece of writing is being referred to.Most respondents in this survey are familiar with a multi-draft composition process, with only the final draft being graded.Since this survey presented only examples of writing and made no mention whether it was an ungraded rough draft or a graded final draft, the students may not have clearly understood the reference and may have answered differently than they would for a different draft.A fourth limitation to this study is that this survey directed students' attention to errors in their writing without offering any counterbalancing questions concerning the ideas expressed in the writing or the organization of this paper.Thus, the results may be skewed and interpreted in a way which makes it seem that Taiwanese students are overwhelmingly concerned with error correction and feedback on grammatical forms rather than on organization or clarity of thought.
A final limitation of this survey is that the examples provided in the last section of the survey were not accurate with regards to error correction.The same sample sentence was used for all the examples, but one of the forms of error correction, that in question 29, would have produced an ungrammatical sentence.That is, question 29 indicated that the underlined word, be, should be moved to the beginning of the sentence.While the focus was on the form of error correction, actually using this 'correction' would have resulted in an incorrect, ungrammatical sentence.Respondents may have interpreted that incorrectly and rated the example as bad or very bad because the resulting sentence would have been incorrect.
The results of this study hold several implications.First, English writing teachers should provide error correction and feedback to their students.Although the results of this survey indicate a preference for direct or indirect error correction, teachers should work with their students and ask what form of error correction they prefer in order for the corrective feedback to provide the most benefit to students.Second, while the results of this survey indicate that students prefer corrective feedback on grammatical errors, the results may be skewed due to the emphasis on grammatical correction in the survey.Teachers should provide error correction on grammatical errors, but they should perhaps discuss with their students whether or not to provide feedback on the organization of the paper or the ideas expressed in the paper.While students may benefit from corrective feedback on grammatical errors, organization and clarity of thought are also important in L2 writing and may be considered more difficult by many students, whereas grammatical errors can often be corrected by learning the appropriate linguistic rules.
The results of this research show a clear preference among the population for comprehensive error correction of all errors in L2 student writing.This is not surprising and adds support to previous research.There were also findings divergent from previous research, but these were found to most likely be due to the population studied in this research as compared to populations studied in previous research.
As previously mentioned, L2 writing students have consistently shown a preference for error correction, and this study found similar findings.Moreover, previous research has shown that corrective feedback in L2 writing is beneficial to students learning grammatical forms.Clearly, L2 writing teachers should continue to provide feedback to L2 students, but they should also be cautioned to find the optimal form of feedback for their students and to ensure that their students use it properly.
As more research continues to investigate student preferences for corrective feedback and optimal feedback strategies for various populations, L2 writing teachers can continue to mold their teaching styles and strategies to ensure that their EFL students can learn English writing in the most efficient way while also maintaining a balance between teaching and correction.Ultimately, this would result in a balanced teacherstudent interaction that will lead to better teachers and better learners.

Table 1 :
Student Attitudes Towards Errors

Table 2 :
Students' Attitudes Towards the Importance of Pointing Out Errors

Table 3 :
Student Attention to Error TypesQuestion 17 asked students how carefully they looked at the marks made by the teacher on their written work.Fifty-five students (69.6%) indicated that they read every error carefully, whereas 20 students (25.3% indicated that they looked at some marks more carefully than others.Question 18 expanded on this by asking respondents which errors they look at more carefully than others.Responses overwhelmingly indicated that students paid the most attention to grammatical errors, with 68 students (86.1%) marking that response.Question 19 asked students which marks made by their English writing teacher they remember the best.Responses were split, with 33 students (41.8%) remembering comments on the organization of the paper, 26 students (32.9%) remembering comments on the ideas expressed in the paper, and 20 students (25.3%) remembering marks indicating errors in English.Questions 20 through 24 sought to understand how students deal with errors and understand what they did wrong so they can correct it.Responses to question 20 showed that students preferred to have another student or friend explain the marked errors (34 responses, 43.0%) or have their teacher explain the marked error (33 responses, 41.8%).Alternatively, responses to question 21 show that looking in a grammar book (30 responses, 38.0%), having another native speaker explain the problem (21 responses, 26.6%), and having another student or friend explain the problem (22 responses, 27.8%) were the least helpful methods to understand errors.

Table 4 :
Student Attitudes Towards Different Forms of Corrective Feedback