EFL learners’ perceptions and preferences of written corrective feedback: a case study of university students from Mainland China

Although a growing body of research has examined the effectiveness of written corrective feedback (WCF) for improving L2 learners’ grammatical accuracy, fewer studies have investigated the extent to which different educational settings would influence learners’ perceptions and preferences of WCF. This paper reports on an exploratory study that investigated learners’ perceptions and preferences of WCF in an EFL setting. Quantitative and qualitative data were collected from 64 intermediate, advanced-intermediate, and advanced English learners across three proficiency levels (intermediate, advanced-intermediate, and advanced) in a major provincial university of Mainland China. Through extensive written questionnaires, the study explored these learners’ perceptions and preferences of the various dimensions of WCF. The results showed that although the participants tended to have a neutral opinion on the role of explicit grammar instruction, overall they expressed a favourable attitude towards error correction. In particular, they held a strong preference for extended comments on both content and grammar of their written work. The qualitative data further indicated that the participants wanted to take more initiatives in the revision process of their writing with less interference from teachers. Overall, the findings confirm the value of WCF for EFL learners outside English-speaking countries. The findings also highlight the significance of individual and contextual factors in the ongoing debate over the effectiveness of WCF.

Dealing with learner errors is a critical aspect of second language (L2) teaching. For many language instructors, correcting errors and commenting on students’ written assignments are among the most common functions of their daily work. Yet, the effectiveness of corrective feedback for improving L2 accuracy is still the subject of much debate. In particular, the role of written corrective feedback^{Footnote 1} (WCF) has been a controversial topic in L2 writing and L2 acquisition research. Although pedagogical discussions of L2 learners’ written errors can be traced back to pre-1980 composition studies (e.g., Cohen and Robbins 1976; Shaughnessy 1977), contemporary debates on WCF have been mainly inspired by Truscott’s (1996) thought-provoking essay. In it, Truscott writes: “substantial research shows [grammar correction] to be ineffective and none shows it to be helpful in any interesting sense” (p. 327). Truscott’s strong opposition to WCF has faced numerous challenges and received critiques from researchers, who, through empirical research or other scholarly synthesis, have argued that grammar feedback is essential for second language acquisition (SLA) and should remain an important component of L2 instruction^{Footnote 2} (e.g., Guénette 2007; Ferris 2004; Hyland and Hyland 2006). Clearly, there are several points of contention over WCF. There is, however, widespread consensus that: WCF is a complex, multivariate subject on which there is inadequate research adequately (Bitchener and Ferris 2012). Fortunately, there has been a steady growth of WCF research over the past two decades, with more and more studies moving from descriptive analyses toward experimental designs. The objective of these studies is to investigate WCF’s effects on L2 learners’ acquisition of particular grammatical structures (e.g., Bitchener 2008; Sheen 2007).

The growing research on WCF is also spurred by the internationalization of higher education in English-speaking countries. For instance, the total enrolment number of international students in Canadian universities increased from 4 to 8 % from 1992 to 2008. In 2009, international students accounted for 33.4 % of Canada’s total graduate student population (Statistics Canada 2011). There are similar trends in the United States, Australia, and New Zealand. Such dramatic demographic changes have significantly altered the dynamics of academic writing instruction at English-speaking universities. Students’ academic performance is largely determined by how well he/she is able to write. This is particularly the case in the humanities and social sciences. Increasingly, university instructors are faced with the challenge of responding to and evaluating the work presented by students for whom English is a second language. Decisions about WCF are not just about improving (L2) accuracy (Brunton 2009). They should be carefully considered within the overall goals of writing instruction since a grammatically flawless thesis can be still regarded as an unsuccessful essay for improper conceptual and organizational issues. In this regard, the pedagogical implications of WCF research are beyond L2 acquisition and L2 writing.

One particular area that has attracted increasing attention is how students and teachers perceive the usefulness of WCF (e.g., Amrhein and Nassaji 2010; Brown 2009; Diab 2005; Karim and Nassaji 2015; Lee 2008; Montgomery and Baker 2007; Simard et al. 2015). Learner perception presents a key variable influencing the role of WCF for two major reasons. First, if students may construe instructional techniques in different ways than the teacher may have expected. This disconnect can impair learning effectiveness (Amrhein and Nassaji 2010). However, students’ general favourable attitude toward WCF, as evidenced by previous studies, informs instructional best practices. This observation offers more supporting evidence in terms of the value of WCF in the ongoing academic debates (Ferris 2012; Saito 1994; Schulz 2001). In spite of the increase in perception-oriented research on WCF, there are, however, still many unknowns with respect to this topic. One of them is whether there are differences in students’ perceptions of WCF within EFL and ESL contexts. Although contextual factors such as learning contexts have received some attention in previous research on oral corrective feedback, they have been largely neglected by WCF research (Goldstein 2001). In particular, previous studies on how students and teachers perceive WCF have been mainly conducted in the context of ESL programs for international students in English-speaking countries. A noticeable missing piece in the picture is EFL programs in developing countries where cultural and classroom dynamics are drastically different from those found in English-speaking countries.

This paper reports on an exploratory study in which English-as-a-Foreign-Language (EFL) learners’ perceptions and preferences of WCF were investigated. Quantitative and qualitative data were collected from 64 EFL learners. This happened by means of written questionnaires. The participants were recruited from a major provincial-level university in Mainland China. The students were intermediate, upper-intermediate, and advanced The remaining sections of this paper will take the following organization. Section Review of literature provides a brief review of previous WCF research and explains the significance of contextual variables within the ongoing debates on the role of WCF for L2 acquisition. Section Methods introduces the study’s data collection and data analysis procedures. Section Results reports the results. Finally, Section Discussion discusses the study’s theoretical and pedagogical implications.

The proliferation of WCF research is driven by its significant theoretical and pedagogical implications for SLA research and instruction (Bitchener and Ferris 2012; Ellis 2010; Ferris 2010). Theoretically, competing SLA theories have made different claims about the roles of error correction in language acquisition. WCF research has served as an effective means of testing these competing claims. While theories derived from Chomsky’s universal grammar (Schwartz 1993) tend to reject the role of corrective feedback in facilitating language acquisition, cognitive interactionist theories such as interaction hypothesis (Long 1991) and noticing hypothesis (Schmidt 1994) propose that error correction assists language acquisition by helping learners to establish form-meaning mappings. WCF research has also been motivated by practical and pedagogical concerns (Bitchener and Ferris 2012; Nassaji and Fotos 2010; Nassaji 2015). Error correction plays a central role in classroom discourse; its amount and form are closely related to the efficiency of L2 instruction. This is particularly true for WCF since many L2 writing researchers are concerned with whether WCF is able to assist L2 students to improve the overall effectiveness of their writing (Ferris 2010).

Debates on WCF are mainly fuelled by Truscott’s (1996) thought-provoking essay. In it, Truscott questions the efficacy of error correction as an instructional tool for L2 writing. The key arguments of Truscott’s original essay (and its follow-ups) were as follows: (1) grammar correction practice goes against SLA theories; (2) existing evidence suggests that WCF has very little potential benefit for student writers; (3) the practical problems faced by teachers and students negate the usefulness of grammar correction; and (4) grammar correction is time consuming for both students and teachers (Truscott 1996, 1999, 2004, 2007). Truscott’s remarks have received strong rebuttals from many researchers. Ferris (1999, 2004) acknowledges that Truscott’s critiques highlight the complexities of WCF activities and the practical issues associated with them. However, he says that Truscott’s dismissal is unfounded since it simply neglects many empirical studies supporting the effectiveness of WCF. Based on comparative experimental studies, other researchers have demonstrated that WCF, if well designed and executed, is able to noticeably improve learners’ performance on certain grammatical structures (e.g., Bitchener 2008; Bitchener and Knoch 2010; Chandler 2003; Ellis et al. 2008; Hartshorn et al. 2010; Liu 2008; Nassaji 2011; Sheen 2007; Shintani et al. 2014; van Beuningen et al. 2012). A few recent examples of this research strand and their findings are as follows. Shintani et al. (2014) explored the effect of direct and metalinguistic feedback and found a stronger effect for the former than the latter. Bitchener (2008) reported that the students who received direct corrective feedback with written and oral meta-linguistic feedback, along with the group that received direct corrective feedback with no meta-linguistic feedback, outperformed the control group who did not receive any feedback whatsoever. Similarly, Abualsha’r and AbuSeileek (2014) found that the students who received corrective feedback delivered via computer about error types while writing essays performed significantly better than those who did not receive any corrective feedback whatsoever. Mawlawi Diab (2015) found that students receiving direct error correction and metalinguistic feedback outperformed students receiving only metalinguistic feedback. For example, the students who received direct error correction made fewer pronoun and lexical errors.

However, the mounting evidence on the effectiveness of WCF does not fully discredit Truscott’s doubts on WCF. To date, many studies trying to validate the effectiveness of WCF have been small and short-term accounts, and they have tended to offer close-up treatments to relatively straightforward grammatical structures such as article and past-tense usage in English (Liu and Brown 2015). There are also considerable discrepancies among WCF studies. For instance, while some studies have reported a major effect of indirect WCF that only identifies grammatical errors (e.g., Chandler 2003; Ferris 2006), others have reported similar or more positive effects offer by direct WCF in error identification and correction (e.g., Bitchener 2008; Bitchener and Knoch 2008). Similarly, some studies (e.g., Ellis et al. 2008; Sheen 2007; Yang and Lyster 2010) have found that WCF focused on a single linguistic feature is most effective. Conversely, other studies have advocated the use of WCF on a variety of linguistic challenges (e.g., Bitchener et al. 2005; Evans et al. 2010).

Thus, recent research has is increasingly geared toward multivariate and disparate factors that might mediate the effectiveness of WCF. Contextual analysis is playing an increasingly significant role in WCF research (Ferris 2010). As it turns out, the role of contextual variables is generally de-emphasized in many experimental and quasi-experimental studies. That is because these studies have tended to come up with for broad, overarching results (Brunton 2009; Ellis 2010). As Goldstein (2001) points out, “the (WCF) research has been largely non-contextual and non-social” (p. 77).

Contextual variables in WCF can be divided into three broad categories: learner, methodological, and situational variables (Evans et al. 2010). The learner variables refer to students’ individual differences. These include their L1 background, their perceptions of WCF, and their individual learning style. The methodological variables include different designs of WCF activities. Situational variables are associated with the institutional context – curriculum design, class size, frequency of class meetings, and teacher variables. Seedhouse (2004) made a distinction between “form-and-accuracy contexts” and “meaning-and-fluency contexts”. His study suggested that error correction is more likely to occur in form-and-accuracy contexts.

Among the variable categories, the situational variables may be the most important yet the least studied variables. Although previous studies have demonstrated that WCF is effective in FL, L2, and immersion settings, we do not yet know to what extent situational contexts are able to mediate the effectiveness of WCF in various contexts. Specifically, little is known about the interaction among situational, methodological and learner variables and how, collectively or individually, each of these factors mediate the effectiveness of feedback. For example, factors such as curriculum design and class size may constrain how WCF activities are implemented. Their effects, however, may be closely related to learner variables.

Learner variables, such as motivation, are dynamic phenomena driven by learners’ personal emotions. These variables are also driven by instructional settings. In their interviews with EFL learners from Mainland China, Han and Hyland (2015) illustrated how learner engagement with WCF is mediated by the interactions between individual and contextual factors. Situational contexts may also influence how learners and instructors understand WCF activities. Previous studies have extensively shown the discrepancies between students’ and teachers’ beliefs vis-à-vis WCF. For instance, in their examination of ESL learners’ metalinguistic reflections, Simard et al. (2015) found that some error corrections may lead to learners’ erroneous hypotheses about the intent of their instructors. Similar results are also reported in other studies (e.g., Brown 2009; Diab 2005; Montgomery and Baker 2007). One recommendation offered by these studies is that a shared understanding of WCF needs to be reached among students and instructors. Although previous studies have underscored learners’ positive attitude vis-à-vis (or in regard to) WCF, these studies have also found that students have their own preferences and opinions about certain types of WCF. Some studies (e.g., Ferris 1995; Lee 2005; Radecki and Swales 1988; Schulz 2001) have found that students prefer a grammar-based approach. In such an approach, there is a large amount of accuracy-oriented WCF., Other studies (Amrhein and Nassaji 2010; Ashwell 2000; Lee 2008) have demonstrated that students prefer content-based correction. Here, attention is paid toward writing content as well as grammatical errors.

Research on how WCF is perceived has, however, rarely been linked to situational variations of WCF activities. We know little about whether there are differences in how students perceive WCF between EFL and ESL contexts. Lee’s (2008) recent research presents an exception. She drew on her teaching experience in Hong Kong to address how teachers’ feedback practices are constrained by cultural and institutional factors, and a lack of adequate teacher training. Lee’s focus on situational variables has been echoed by research on intercultural variations in writing instruction (e.g., Loewen et al. 2009; Nelson and Carson 2006). In this research, EFL students from collective cultures such as China and Japan have been found to be uncomfortable in peer review sessions. This is because breaking group consensus is against their cultural preferences for harmonious communication. Given the importance of situational variables in WCF activities, examining the issue of corrective feedback in different contexts seems to be a research topic worthy of further exploration.

Based on the above observations, the present study was designed to examine students’ perceptions and preferences of WCF in an EFL context. Using written questionnaires, the study surveyed a total of 64 EFL learners studying at a major provincial university in Mainland China. We focused on this particular learner category for two reasons. First, previous studies on learners’ perceptions of WCF have been mainly conducted in ESL classrooms of English-speaking countries where language instruction in which language instruction tends to be both meaning- and form-focused and also occurs in contexts where learners use English in their daily lives outside their classrooms. This is different from the EFL context in Mainland China in which form-focused instruction is the primary mode of teaching and learners also have few opportunities to use English outside classroom contexts. Second, EFL learners from Mainland China have become the second largest group of international students in English-speaking countries. Abetter understanding of their perceptions of WCF has important pedagogical implications for language instruction in countries such as Canada and the United States.

Overall, the present study addressed the following research questions:

1. 1.

   How do EFL learners at tertiary level in Mainland China view the role of grammar instruction and written corrective feedback in their writing classes?

2. 2.

   What amount and types of WCF do theses learners prefer and why?

3. 3.

   What types of WCF do theses learners regard as effective and why?

4. 4.

   Does these learners’ level of language proficiency affect their preferences of WCF, and if so, how?

Quantitative and qualitative data were collected using a written questionnaire that elicited participants’ perceptions about WCF and the reasons behind such opinions. The quantitative data were collected by means of close-ended questionnaire items with multiple choices or Likert scale formats. The qualitative data were collected with (or “using”)open-ended questions. The quantitative data were collected to analyse (or “understand”) general patterns of the participants’ preferences for grammar instruction and WCF activities. The qualitative data were collected to explore the rationales behind their preferences. The questionnaires were distributed to 64 EFL learners in the English department of a major provincial university in Mainland China.

The institutional setting

The study was conducted in the English department of a large public university in China. The department offers both undergraduate and graduate programs in English studies and currently has approximately 350 undergraduate students. (e.g., Amrhein and Nassaji 2010; Brown 2009; Karim and Nassaji 2015; Montgomery and Baker 2007; Simard et al. 2015), The situational context of the present study is unique in several ways from previous research (or “outcomes”) on the subject (then list the sources). First, the languages of instruction in the department are Mandarin Chinese and English. Undergraduates are taught in both Mandarin Chinese and English in their first two years’ of study due to the limitation of their English proficiency level. The language of instruction, however, changes to English for the upper-level undergraduate courses. Second, the department’s teaching methods are mixed in nature. The department’s curriculum includes many foundation courses in English – focusing on grammar and language accuracy. Many traditional EFL instructional techniques such as dictation, drill exercises, and grammar translation are widely used in these classes. On the other hand, the department also has many general courses in English literature and Anglo-American cultures. In these classes, activities are more content-based and communicative. Third, the English department at this university, like many other language departments across Mainland China, suffers from large class sizes. The class sizes at the department range from 20 to 90 students, making it difficult for instructors to provide individualized learning to the students. The class size issue is even worse for non-English major university students in Mainland China. The average class size for these students can range from 100 to 300. As a result, WCF often becomes one of the few available options for learners who want to interact with instructors. Thus, our study’s situational context presents a typical EFL scenario – found in many English departments in Mainland China. This context is starkly different that of ESL language classrooms in English-speaking universities.

Participants

This study involved 64 EFL learners. There were 21 first-year, 23 s-year, and 20 third-year students. These participants were randomly recruited from various classes with the help of instructors in the English department. The uneven gender distribution of these participants (16 males and 48 females) reflects the general gender ratio of many English departments in Mainland China.

The institution requires students majoring in English to pass the Test for English Majors Level 4 (TEM-4) in their second year of study and the Test for English Majors Level 8 (TEM-8) in their third year of study. Administrated by the Ministry of Education of China, both tests are and designed to examine a participant’s overall proficiency in English. A passing level of TEM-4 is roughly equivalent to 6.5 on IELTS. A passing level of TEM-8 is roughly equivalent to 7.5 on IELTS. In the present study, all the second-year and third-year participants passed TEM-4 and TEM-8 respectively. Thus, the participants were at three different English proficiency levels: intermediate (first-year), upper-intermediate (second-year), and advanced (third year).

Questionnaire design

The questionnaire used in the present study was based on Amrhein and Nassaji’s (2010) original design and evolved extensively after three revisions (see Additional file 1: Appendix A for the questionnaire). The primary purpose of the questionnaire was to elicit the participants’ perceptions of WCF and the reasons behind their opinions. As such, close-ended questions with multiple choice or Likert scale formats were used to measure the central tendencies of the participants, followed by open-ended questions exploring the reasons behind their opinions. The questionnaire was relatively long, with two demographic questions, ten close-ended questions, and seven open-ended questions. Three of the ten close-ended questions have add-on qualitative components for the participants to justify their choices. The questionnaire was designed with the following considerations in mind: (1) a minimal usage of technical jargon to make the survey questions easy to understand; (2) an adequate number of questions that can be finished in less than 30 min; (3) an extensive use of open-ended questions to get a better understanding of the reasons behind the participants’ perceptions of WCF, especially as it relates to EFL.; (4) a focus on learning practices rather than on theoretical issues pertaining to L2. Finally, to generate the best responses, the questionnaire was also translated into Chinese; the participants could choose either the English or Chinese version.

Data analysis

The questionnaire responses were entered into an Excel spreadsheet and then exported into SPSS 22.0 for statistical analysis. For the quantitative data, both descriptive statistics and inferential statistics (the chi-square test for nominal variables and the one-way ANOVA test for the mean scores of ordinal variables) were conducted to explore statistically significant differences across the three proficiency levels. For the qualitative data, the participants’ explanatory responses were coded by the first and third authors, based upon their common themes. Different coding results between the two coders were then negotiated and in the end common responses by the participants were recorded. The following section will describe the questionnaire’s major findings.

Item (1) of the questionnaire asked the students about their general perceptions of grammar instruction in writing classes. As Table 1 shows, most students acknowledged the necessity of grammar instruction in writing classes (total average rating: 4.45/5) and believed that learning grammar would improve their English writing skills (total average rating: 4.27/5). Yet, they expressed less positive opinions on explicit grammar instruction, as shown in the results of statements (d), (e), and (f). In particular, the statement “I like studying English grammar” received the lowest mean ratings across the three proficiency levels (total average rating: 3.14/5). By contrast, the statement “I think that language practice in real contexts is more important than grammar instruction in the classroom” received very positive ratings across the three proficiency levels (total average rating: 4.41/5). Another trend was that students tended to hold less positive opinions about grammar the more proficient they got. As shown in Table 1, the mean ratings of most statements dropped when comparing the 3rd-year student group with the other two groups. The only exception occurred in statements (c) and (g) where the 2nd-year students held less positive views on the efficiency grammar instruction and the effectiveness of communicative language teaching than the 3rd-year group. Nonetheless, the One-way ANOVA test showed no significant difference among learners at different levels of language proficiency.

Item (2) explored the participants’ perceptions of WCF. As shown in Table 2, overall, the students held a very positive view regarding WCF in writing instruction, with a total average rating of 4.37 out of 5. This result was further supported by the students’ qualitative responses. There, they indicated that the importance of WCF came from the following factors: (1) WCF is able to help with the identification of recurring errors; (2) WCF provides opportunities for further improvement in writing quality; and (3) unlike spoken language, English writing requires more attention to form and accuracy.

To further examine the participants’ perceptions of different types of WCF, Item (3) asked about their most preferred error types for correction. As Table 3 shows, the responses were fairly consistent across the three proficiency levels, with organizational errors being the most popular response (42 in total), followed by grammatical errors (16 in total) and vocabulary errors (6 in total). The Chi-square test showed no significant difference among learners from different language proficiency levels. Overall, the students considered organization as the most important aspect of their writing performance.

Table 4 presents the participants’ responses to Item (4), in which their opinions on instructors’ error correction priority were examined. The overall result was mixed, with Option C (only correcting errors that interfere with communicating ideas) being slightly higher than Option A (correcting all errors) and Option B (only correcting major errors). In terms of the students’ qualitative responses, the most common reason that students chose Option A was “it is instructors’ responsibility to provide detailed and thorough feedback” whereas those choosing Options B and C expressed their concern of instructors’ working load and desire for some degree of independence in their revision processes. The Chi-square test showed no significant difference among learners at different levels of language proficiency.

Item (5) investigated the participants’ preferences of error correction techniques. As shown in Table 5, the most preferred technique among the students was “locating the error and also indicating the type of error”, with a total average rating of 4.12. A statically significant difference (p = .033) was also found in this technique among learners at different levels of language proficiency: despite its popularity, there were different opinions about this technique. Specifically, the 3rd-year students (the advanced level learners) showed a lower average rating (3.85) than the other two groups. This indicates that the more advanced the students, the less they required explicit feedback on their grammatical errors. These students, however, needed more attention to other aspects of their writings. The second most popular technique for the students was “correcting the error and then providing an explanation for the correction.” This response earned an average rating of 4.09. By comparison, the least favoured technique was “simply indicating that you have an error in the sentence by putting a cross next to it without locating or correcting the error” (total average rating: 2.97), which was also the most indirect technique on the list.

Items (6) and (7) examined the participants’ perceptions of the extended comments offered by instructors. The results are shown in Tables 6 and 7. The students regarded extended comments as an important aspect of their learning process (total average rating: 4.31) and their most preferred comment type was “comments on the writing’s overall quality” (total average rating: 4.59). By contrast, “comments on the grammar” were the least favoured comment type among the students (total average rating: 3.76). The 3rd-year students’ average ratings to all comments types were noticeably higher than the other groups’, suggesting that these students preferred comprehensive feedback to simple error correction.

In Item (8), the participants were asked about how carefully they would review their instructors’ feedback. Most students (54 out of 64) responded that they would carefully read the feedback and correct all the errors (see Table 8 for details).

Timing presents a crucial aspect of classroom activities since various teaching goals always compete for limited class time. Items (9) and (10) explored the participants’ preferences of the timing of WCF. As shown in Tables 9 and 10, the students held different views on the timing of grammatical error correction and the timing of content and organizational error correction. In terms of grammar feedback, the most popular response (20 out of 63) was feedback on second drafts. By contrast, more than half of the students (34 out of 63) preferred their instructors to provide content and organizational error correction on the first drafts. This result suggests that instructors might need to adjust the ratio of grammar and content comments depending on which stage of the writing process the students were at.

As Nassaji (2011) points out, “most studies of feedback on written errors have focused on unidirectional feedback without any student-teacher interaction or negotiation” (p. 317). To overcome such limitation, Items (11) and (14) specifically dealt with students’ attitudes toward self-correction and the potential limitations of unidirectional WCF. Item (11), “how useful do you find oral communication about written errors between the instructor and students”, showed mixed results. Twenty-three students supported the idea; 41 respondents found it ineffective. Those in support said oral communication is more direct and comprehensible. By contrast, students opposing oral communication argued that WCF made future review easier and more accessible. Item (12) asked: “do you think that teachers should ask students to identify their own errors.” Here, majority of the students (48 out of 64) believed that self-correction is an essential skill for English writing and that students should be encouraged to practice this skill. The answers to Item 13 show that students want detailed correction, along with comments by the instructors. In terms of the limits of unidirectional WCF, the majority of the students (40 out of 64) pointed out that they would like instructors to provide detailed comments on the content and organization of their written assignments and more follow-up bidirectional communication.

Finally, Items (15) to (17) addressed the students’ general perceptions of WCF. The results of these items confirmed the previous findings. The students valued WCF activities and regarded them as an important aspect of their language learning experience. However, they were unsatisfied with some aspects of WCF. In particular, they found fault with its emphasis on identifying grammatical errors. They also are critical of WCF for being insufficiently interactive.

Our study sought to illustrate how EFL learners from Mainland China perceive various WCF techniques. The study also sought to examine whether an EFL context makes learners’ perceptions and preferences distinctive (from non-ESL contexts you mean. If so, that is how I would phrase it). The survey results revealed that although a sizeable amount of survey participants held neutral or negative opinions toward explicit grammar instruction, they still expressed a favourable attitude toward error corrections and comments, especially feedback on the content and the organization of their written assignments. The results also show/demonstrate/highlight/underscore that the students preferred direct correction to indirect correction. Many students also expressed strong desires for more self-correction as well as interactive activities in the revision process. This signals a potential limitation of unidirectional WCF techniques.

The following discussions are inferential rather than definitive. This is in light of limited participants in the survey and the internal variations among language classrooms in Mainland China. Still, findings supply empirical evidence to show that the EFL learners from Mainland China held a positive belief about WCF and its role in EFL learning. which depicted a positive picture of WCF activities in their EFL writing classes of Mainland China. This conclusion aligns with (or “lines up against”) studies on WCF (e.g., Amrhein and Nassaji 2010; Ashwell 2000; Brown 2009; Karim and Nassaji 2015; Lee 2008; Montgomery and Baker 2007; Schulz 2001) that show the acceptance of WCF in both ESL and EFL contexts.

The results of this research found that overall the students held a very positive view regarding WCF in writing instruction. Such a finding is similar to what Amrhein and Nassaji (2010) reported from ESL participants. The students in their study also “thought it most useful for teachers to provide WCF on as many errors as possible” (p. 114). Our study also showed a broad conceptualization of WCF by EFL learners, in which not only grammatical errors but also content and organizational errors were considered as important potential targets for correction. This finding is somewhat different from Amrhein and Nassaji’s (2010) finding in that their ESL participants were not content when their teachers paying more attention to content errors than grammatical errors. This discrepancy might be indicative of pedagogical differences between ESL and EFL. (e.g., ESL contexts are more meaning-oriented while EFL is focused on form and accuracy). The importance of correcting content errors, however, echoes Ferris’ (2010) recent argument regarding the interdisciplinary nature of WCF research. Ferris (2010) points out that SLA and L2 writing scholars have taken different approaches to WCF research. While the primary concern of SLA studies is whether WCF would facilitate the process of SLA, L2 writing studies are more concerned with whether WCF would improve the overall quality of L2 writing performance. As shown in our study, the surveyed EFL learners were more inclined to the view of L2 writing: their primary concern was the overall communicative quality instead of the accuracy of grammar. Such an opinion should not be construed as wanting to separate the goal of accuracy from the goal of effective writing; for instance, it does not discredit the necessity of correcting grammatical errors. It does, however, invite EFL instructors to reconsider how to balance grammar instruction against content-related issues in their WCF practices. For advanced EFL learners, it especially might be a good idea for teachers to offer more feedback on the content and structure of the students’ work.

The findings reveal some of the limitation of current WCF practices in Mainland China, especially unidirectional feedback (see Nassaji 2011). In this respect, the majority of the students in the study wanted their instructors to provide more detailed comments on their errors. particularly those related to content and organization They also wanted more follow-up (i.e. in the form of bidirectional communication). This finding reveals a constraint regarding WCF practices in this teaching setting and can be attributed to the institutional context of the survey. English teaching in China, like many other EFL contexts, over-emphasizes form and accuracy (e.g., Hu 2003; Liu 2007). Due to various constraints such as class length and class size, Teaching grammar often presents itself as the only viable option because most students in such contexts believe that grammar can help them pass their English language proficiency test. As a result of too much emphasis on grammar, many students have also been turned off explicit grammar instruction. These impressions remain as they move into the tertiary level, as the survey data shows.. Language instructors ought to adjust their WCF strategies to guard against “grammar fatigue.” Judging by the participants’ comments, more attention could be given to the communicative aspects of English writing.

Finally, the survey also highlighted the importance of considering contextual variables in future WCF research. In response to Goldstein’s (2001) call for WCF research to pay more attention to contextual variables, this study sought to illustrate how the contextual complexity of language classrooms in Mainland China adds difficulties to WCF activities and in many ways the survey results captured the participants’ various stances toward WCF activities. In addition the study showed that this factor did not interact significantly with learners’ English proficiency level. It seemed that learners at all levels disliked emphasis on grammar instruction and too much focus on accuracy Although such an attitude towards grammar correction can be also found in ESL contexts, this negative tendency seemed to be very pronounced in the present study, which indicates how teaching dynamics can differ between ESL and EFL contexts. The EFL context addressed in our study cannot be reprehensive of all language classrooms in Mainland China. Still, the findings are relevant given the large number of schools in China similar to the one we studied. This, however, does not indicate that the results are entirely generalizable and hence more research of this kind in other institutions and places in China should be carried out. An important takeaway from the current study is that more attention should be paid to contextual factors in WCF research. This is not limited to instructional setting variables only but also other contextual variables such as those related to the nature of the curriculum, class sizes, frequency of class meetings, and teacher variables. These factors may all play an important role in the effectiveness of WCF and also how learners perceive it and hence deserves more attention in the future.

This study explored the role of WCF in EFL writing classes in Mainland China. The research was conducted in an instructional context that has not been examined before - -- university level students in Mainland China. The findings show that students tend to hold a positive view on WCF. However,, their views vis-à-vis WCF are tempered by by what they consider to be an over-emphasis on grammar explicit.. Thus,the use of WCF should carefully consider a balance between grammar and content-oriented feedback.

The study contains some important limitations. First, the findings were based on data from a small group of students. Thus the generalizability of the findings needed to be further corroborated by additional research. For instance, research is needed in other instructional contexts in China to empirically test the generalizability of our findings. Although EFL learners from? Mainland China represent a large proportion of the EFL population studying in English-speaking countries, there are students from many other non-English speaking countries whose views on WCF need to be examined in more detail. Thus, there is a need for more research to examine how different learners from different backgrounds or different instructional contexts perceive the importance of error correction and also what factors may also mediate their perception. Such studies are extremely necessary to advance our understanding of the role of corrective feedback in such contexts. Finally, this study investigated only EFL learners’ perceptions of WCF in China. Similar surveys can also be conducted with language instructors in this context.

1. 1.

   Written corrective feedback (WCF) is also commonly referred as grammar correction or written error correction in SLA research (Ferris 2012). In this study, however, we adopt a broad definition of “written corrective feedback” by using the term to refer to instructors’ corrective activities toward both grammatical and content aspects of students’ written assignments. This decision is based upon two considerations: first, the central research question of the reported study is ELF learners’ perceptions of the relation between WCF and the overall effectiveness of their writing; second, instructors from disciplines such as communication, sociology, and political science also conduct extensive WCF activities, yet these activities’ focuses often go beyond grammatical errors.

2. 2.

   See Ferris (2012) for a detailed timeline of previous research on written corrective feedback.

We would like to thank Kam Razavi at School of Communication, Simon Fraser University and the editors and reviewers from Asian-Pacific Journal of Second and Foreign Language Education for their insightful comments on earlier drafts of this article.

Affiliations

1. School of Communication, Simon Fraser University, K9671- 8888 University Drive, Burnaby, BC, V5A 1S6, Canada

   • Sibo Chen

2. Department of Linguistics, University of Victoria, Victoria, Canada

   • Hossein Nassaji

3. School of Foreign Languages and Literatures, Beijing Normal University, Beijing, China

   • Qian Liu

Corresponding author

Correspondence to Sibo Chen.

Competing interest

The authors declare that they have no competing interests.

Authors’ contribution

SC and HN designed the questionnaire of the study. QL conducted the survey. SC and HN drafted the manuscript. All authors read and approved the final manuscript.

Reprints and Permissions