2.1 Amount of WCF

Traditionally, writing teachers give comprehensive WCF to students, responding to every single error as it occurs. Such WCF is unfocused. There is recent WCF research that points to the benefits of a selective (i.e. focused) approach, in which teachers target a selected number of error types in their feedback. Ferris (Reference Ferris2002) suggests that responding to recurrent patterns of errors in a focused manner, especially rule-governed items (e.g. verb tense and form, articles, subject-verb agreement), may be more beneficial than responding to all errors in an unfocused manner. Drawing upon his study on WCF, Bitchener (Reference Bitchener2008: 108) suggests that writing teachers should respond to ‘one or only a few error categories’ at a time. While the studies by Bitchener (Reference Bitchener2008) and others (e.g. Ellis et al. Reference Ellis, Sheen, Murakami and Takashima2008; Sheen, Wright & Moldawa Reference Sheen, Wright and Moldawa2009) focus on only one error type, some researchers feel that selecting only one error type for feedback is not practical because in writing, students do need to focus on different error types at the same time (Van Beuningen Reference Van Beuningen2010). Thus, a middle ground is proposed, where teachers select several error categories, instead of one, for feedback (Ferris Reference Ferris2010; Storch Reference Storch2010). Compared with comprehensive WCF, selective WCF is more manageable (Evans et al. Reference Evans, Hartshorn, McCollum and Wolfersberger2010) for both teachers and students. From the affective perspective, selective and focused WCF is also better for students, as their papers are no longer inundated with red ink, which is likely to hurt their ego and damage their confidence in writing, and may in turn affect the uptake of feedback (Storch & Wigglesworth Reference Storch and Wigglesworth2010). Although there is research that demonstrates the efficacy of comprehensive and unfocused WCF, studies that indicate its benefits are as yet few and far between (Truscott & Hsu Reference Truscott and Hsu2008; Van Beuningen, De Jong & Kuiken Reference Van Beuningen, De Jong and Kuiken2008). Overall, the thrust in current WCF research is that selective and focused WCF (especially involving rule-based error categories) is beneficial, particularly to students of lower language proficiency, and is also more manageable and motivating. More recently, Bitchener & Ferris (Reference Bitchener and Ferris2012) have added that comprehensive and unfocused WCF may be useful to advanced learners with a high level of accuracy. They suggest that, for some situations, a combination of focused and unfocused WCF may be desirable, but teachers have to factor student's needs and proficiency levels into their decision-making.

Despite WCF research that casts doubt on the usefulness of the traditional practice of comprehensive, unfocused WCF for less advanced L2 learners in general, in many EFL writing classrooms (including Hong Kong and similar contexts) such a practice persists at almost all levels, especially primary and secondary. Teachers still find themselves marking student errors in great detail, giving meticulous attention to written errors in an unfocused manner (Lee Reference Lee2004, Reference Lee2008, Reference Lee2010). The WCF finding about the care teachers should take about giving unfocused WCF has failed to reach classroom teachers. Even in places where WCF research findings have made an impact on curriculum guidelines, such as Hong Kong, where the local curriculum guide recommends that teachers mark errors selectively (rather than comprehensively) (Curriculum Development Council 1999), teachers still continue to mark errors comprehensively, focusing primarily on written accuracy (Lee Reference Lee2008). This is partly because of their lack of training, as well as the apprenticeship of observation (Lortie Reference Lortie1975) that makes them adhere to the tried and trusted method of detailed WCF. Also, teachers feel under pressure to demonstrate, through comprehensive WCF, that they are hardworking teachers, who have to satisfy the expectations of students and parents. More importantly, in Hong Kong, teachers mark errors in detail because the internal appraisal system in schools requires them to do so: they are evaluated positively, at least on that score, if they mark errors meticulously, but negatively if they mark errors only selectively (Lee Reference Lee2008). In similar EFL contexts, it is likely that writing teachers are influenced by the ‘more is better’ maxim, thinking that the more errors they respond to, the more responsible teachers they are.

In EFL contexts where learning to write is a protracted process, responding to errors in an unfocused manner can have harmful effects on learners, especially the weaker ones. It is not only discouraging but also confusing to receive feedback on a large number of errors at the same time, without knowing which error categories require more urgent attention, and what learners can do to bridge the gaps in their interlanguage system. The recent assessment for learning literature has pointed out the importance of having one target at a time during assessment, and that formative feedback is selective and focused rather than comprehensive and unfocused (Chappuis Reference Chappuis2009). Although some researchers are beginning to draw our attention to the benefits of and need for comprehensive WCF in L2 writing classrooms (as it reflects the reality of the classroom and addresses students’ need to produce grammatically accurate writing in the long run) (Van Beuningen et al. Reference Van Beuningen, De Jong and Kuiken2008; Van Beuningen Reference Van Beuningen2010), in EFL contexts teachers have to consider how best to help students meet the long-term goal of written accuracy development, when and how (if at all) comprehensive WCF should be provided, and particularly how it can be combined with focused WCF to maximize the overall benefits of WCF. Evans et al. (Reference Evans, Hartshorn, McCollum and Wolfersberger2010), for instance, suggest that comprehensive WCF can be given to shorter texts or paragraphs rather than extended pieces of writing. Alternatively, as suggested by Bitchener & Ferris (Reference Bitchener and Ferris2012), teachers can lead students through detailed, comprehensive self-editing in the penultimate or final draft (with focused WCF given to interim drafts), helping them catch as many errors as they can so as to produce a relatively error-free text. Given the current research findings about the benefits of selective WCF for lower proficiency learners and the need for more research regarding the role of comprehensive WCF, EFL teachers should reflect critically on their practice and consider the relative weight they should give to selective and comprehensive WCF.

2.2 Choice of WCF strategies

WCF research has yielded conflicting findings about the efficacy of different WCF strategies, such as direct and indirect feedback. While direct WCF involves providing correct answers for students, indirect WCF allows teachers to provide hints (e.g. underlines, circles and symbols) and let students come up with their own correct answers. Although there is recent research that shows the efficacy of direct over indirect WCF for longer-term acquisition of target items, there is evidence to demonstrate the short-term benefits of both direct and indirect WCF strategies (Bitchener & Ferris Reference Bitchener and Ferris2012). In spite of the inconclusive findings, a general principle emerging from research is that the choice of strategies and the explicitness of WCF (i.e. whether codes should be used) should depend on a range of factors, including the type of error, the nature of the writing task, and the students’ proficiency level (Storch Reference Storch2010; Van Beuningen Reference Van Beuningen2010). In general, indirect WCF is able to allow greater cognitive engagement (Ellis Reference Ellis2010), and it may be best to give indirect WCF to treatable errors (i.e. errors that students are able to self-correct because they are rule-governed), and direct feedback to untreatable errors (errors that are not amenable to self-correction, such as word choice and word order) (Ferris Reference Ferris, Flowerdew and Peacock2001). In addition, students of lower language proficiency are likely to benefit from direct WCF (Ferris & Roberts Reference Ferris and Roberts2001). As for the explicitness of indirect feedback, research has found that coded feedback can be as effective as uncoded feedback (Robb, Ross & Shortreed Reference Robb, Ross and Shortreed1986; Ferris Reference Ferris, Flowerdew and Peacock2001; Ferris & Roberts Reference Ferris and Roberts2001). Specifically, error codes may not be helpful if teachers use the codes for grammar items that students have not yet mastered (Lee Reference Lee1997; Ferris Reference Ferris2002). Overall, it may be best to use a combination of direct and indirect WCF to suit different learners, writing tasks and error types.

Research findings about the choice of WCF strategies have not been well applied in EFL contexts, as there is little evidence that EFL teachers choose their WCF strategies according to the abovementioned factors. My experience working with Hong Kong teachers shows that their WCF practice is rather haphazard, not well informed by research or theoretical principles. There is a tendency to emphasize coded over uncoded feedback. This may be because the local curriculum guide has a recommendation about coded WCF, suggesting that teachers can ‘indicate mistakes so that learners can correct them . . . using a correction code to alert learners to the kinds of mistakes they have made’ (Curriculum Development Council 1999: 96). This piece of advice, which is given more or less out of context, may have resulted in many schools in Hong Kong compiling a long list of error codes for students, encouraging teachers to provide coded feedback. However, in so doing, teachers are not paying sufficient attention to whether the errors in question are treatable (rule-governed) or untreatable (also less amenable to self-correction) (Ferris Reference Ferris1999, Reference Ferris, Flowerdew and Peacock2001). In some error coding lists I have seen, the list contains as many as 20 to 26 error types, which makes marking difficult and cumbersome for teachers. When students receive papers filled with a large range of error codes, they can be easily put off. In fact, it is common to find students simply copying error codes such as ‘art’ (for ‘article’) and ‘prep’ (for ‘preposition’) when they do their corrections (a common practice in Hong Kong schools after students receive teacher feedback), instead of writing out the correct answers. This shows that when error codes are used for errors that students are unable to fix, real learning cannot take place.

Apart from the popularity of error codes, which tend to be over-used in some EFL classrooms, a number of teachers go to the other extreme and provide direct WCF to the majority of errors (Lee Reference Lee2004), without really taking into consideration the error type and students’ language proficiency. Some local teachers I have worked with tell me that giving direct WCF is convenient, as they can make sure that when students do their corrections they will not make any mistakes. This will also save them the trouble of having to answer students’ queries in a large class and having to correct errors again when students submit their corrections. On the whole, whether WCF is direct or indirect (coded or uncoded), teachers do not appear to be aware that the choice of WCF strategies should be, as Bitchener & Ferris (Reference Bitchener and Ferris2012) have shown, based on a principled approach.

In the EFL writing classroom, therefore, I see a great need to address the under-application of the judicious choice of WCF strategies. Without heightened awareness of what WCF strategies to use for different error types or different student needs, teachers will simply repeat their existing practice without bringing any major benefits to student learning.

2.3 Which errors to correct

Related to the choice of WCF strategies is the choice of errors: that is, which errors should teachers correct? Research on this has yielded mixed findings. There has been a call to make a distinction between ‘errors’ and ‘mistakes’: the former are systematic inaccuracies that indicate gaps in learners' interlanguage system, whereas the latter are unsystematic inaccuracies such as slips and memory failures (Corder Reference Corder1967). Bitchener & Ferris (Reference Bitchener and Ferris2012) think that teachers should correct errors but not mistakes, which can easily be self-corrected by students. Another way to distinguish errors is to find out whether they are local or global. Local errors are those that do not impede communication (e.g. morphological errors), while global errors result in communication breakdown (e.g. syntax and lexical errors). Burt (Reference Burt1975) believes that only global errors need to be corrected. Ferris (Reference Ferris1999, Reference Ferris2002), on the other hand, distinguishes between treatable (rule-governed) and untreatable errors (non-rule-governed), claiming that WCF is most effective when directed at treatable errors (Van Beuningen Reference Van Beuningen2010). Van Beuningen, De Jong & Kuiken (Reference Van Beuningen, De Jong and Kuiken2012) have differentiated grammatical (e.g. morphological and syntactic errors) from non-grammatical errors (e.g. spelling and punctuation errors). Their study shows that direct WCF is more effective for grammatical errors, and indirect WCF works better for non-grammatical errors. In addition, Bitchener & Ferris (Reference Bitchener and Ferris2012) suggest that teachers respond to high-frequency and stigmatizing errors: that is, errors that are likely to label students as less proficient writers.

Although the research literature has not produced conclusive evidence about the errors that teachers should target for WCF, what is clear is that the choice of errors is important. Storch & Wigglesworth (Reference Storch and Wigglesworth2010) remind us that students’ uptake of WCF depends on the depth of their processing of WCF, and that uptake is possible only when students notice and understand the reasons for the WCF provided by their teacher. Such noticing is substantive rather than perfunctory, and is thus more likely to result in students’ accurate revision in the following drafts. Given this, the choice of errors for WCF is absolutely crucial, since WCF given to certain errors may facilitate deeper processing and uptake than WCF given to others.

Despite the lack of definitive findings about which errors teachers should correct, a clear research finding for EFL teachers is that they need to choose their errors carefully, making sure that students can notice the gap and understand why the errors were made. This finding, however, is not applied in classrooms, where teachers are marking all errors, without making a distinction between mistakes and errors (Corder Reference Corder1967), local and global errors (Burt Reference Burt1975), treatable and untreatable errors (Ferris Reference Ferris2002) and grammatical and non-grammatical (Van Beuningen et al. Reference Van Beuningen, De Jong and Kuiken2012). The outcome is that WCF is provided perfunctorily, without taking account of the significant issue of student uptake. I believe that teachers should steer away from giving detailed, comprehensive WCF reflexively, but instead reflect on fundamental issues such as which errors they should correct, why, and what WCF strategies they should use for different types of error. In EFL school contexts, for example, it may be best to focus on treatable rather than untreatable errors in WCF, using the indirect strategy to enable students to self-correct. The last thing we want to see is teachers consuming their time and energy without making a real difference to students’ written accuracy development, which is what seems to be happening in many EFL classrooms.

2.4 Student uptake

Research has clearly demonstrated that WCF is one thing, but student uptake can be quite another. This is because WCF can be thwarted, for example, by teachers’ inconsistent or inaccurate error correction (Cohen & Robbins Reference Cohen and Robbins1976; Zamel Reference Zamel1985; Ferris Reference Ferris2002), illegible teacher comments (Lee Reference Lee2008) and confusing symbols. Another factor that can influence uptake is the affective factor. As Hyland (Reference Hyland2010: 177) rightly points out, WCF is useful only when learners are ‘willing and motivated to engage with it’. Also, student uptake is directly related to the issue of learnability; students can acquire grammatical features only when they are developmentally ready (Shaughnessy Reference Shaughnessy1977), irrespective of WCF (also see Pienemann Reference Pienemann1989 for the ‘Teachability or Learnability Hypothesis’). When learners are not ready to acquire certain grammatical items, WCF related to these items may not be useful.

The threats to WCF uptake from both the teacher's and student's perspective should have significant ramifications for classroom practice. Unfortunately, EFL teachers do not seem to be fully aware of the possible threats to their own WCF practice. They continue to spend a huge amount of time correcting written errors, risking giving inconsistent and inaccurate correction because of fatigue and burnout. They continue to do what seems efficient for themselves but demotivating for students, scrawling detailed written comments on student texts (which may not be legible), filling it with circles, underlines and error codes, without seeming to care that such feedback is likely to confuse, frustrate and put off the recipients. In giving WCF in an unfocused manner, EFL teachers also tend to ignore the issue of learnability. Their practice appears to be guided by the underlying premise that students are able to correct/learn all the errors pointed out by the teacher. Such an approach, however, does not take account of the developmental stage of individual students and is unlikely to be effective because for some errors, students may simply lack the readiness to correct them.

The issue of uptake is pivotal because without student uptake teachers’ efforts in WCF are futile. I believe many EFL teachers know that their WCF practice is not cost-effective, but still appear to disregard the issue of student uptake and continue with the same old approach. They seem to be much less concerned about student uptake than getting the job done. The reasons are complicated, but I think teachers are greatly influenced by the values of their institutions and lack the autonomy to change their conventional WCF practice. To enhance student uptake, awareness-raising through teacher training and ongoing professional development in the area of WCF is essential. What needs to be impressed on teachers is that they should consider using strategies that enhance student engagement with feedback and facilitate uptake, such as targeting a few error types, sharing responsibility with learners by training them to conduct peer- and self-editing, and experimenting with alternative forms of corrective feedback such as electronic feedback (see Burstein, Chodorow & Leacock Reference Burstein, Chodorow and Leacock2004; Milton Reference Milton, Hyland and Hyland2006; El Ebyary & Windeatt Reference El Ebyary and Windeatt2010). Bitchener & Ferris (Reference Bitchener and Ferris2012) report an innovative approach adopted by Brown (Reference Brown2010), who color-coded a few selective error patterns in student texts and found that students enhanced their awareness of their recurring error patterns as a result. Other options for corrective feedback include feedback delivered through the word processor and face-to-face feedback. No single approach is perfect, and it is important to bear in mind that WCF could be combined with other forms of feedback to facilitate student uptake. Whatever forms and strategies of WCF teachers use, student uptake should be the key guiding principle.

3.1 Need to act upon WCF

Research has suggested that WCF is most beneficial when it is given to intermediate rather than terminal drafts (Lalande Reference Lalande1982; Frantzen & Rissell Reference Frantzen, Rissell, VanPatten, Dvorak and Lee1987; Ferris Reference Ferris1995, Reference Ferris1997; Chandler Reference Chandler2000), since this can allow student engagement in the correction process. Although there is yet no compelling evidence to show that revision does make a difference in students’ long-term written accuracy (Ferris Reference Ferris, Flowerdew and Peacock2001), the majority of WCF studies are conducted in process-writing classrooms where students are required to act on the errors made.

In EFL contexts, although a product-oriented approach is dominant in the large majority of writing classrooms, it is common for teachers to require students to correct errors after receiving teachers’ WCF. In Hong Kong schools, for instance, teachers make it compulsory for students to write out their corrections. Practice varies, with some teachers asking students to re-write the whole piece of writing and correct the errors made, while others ask students to pick out sentences that contain errors and re-write those sentences only. On the whole, the research finding about opportunities to revise writing after receiving teachers’ WCF, as far as written errors are concerned, has been reasonably well applied, though it is noteworthy that in many EFL contexts like Hong Kong, the revision is restricted to written errors and the approach is primarily product-based. For this reason, in Hong Kong, teachers refer to such practice as ‘correction’ rather than ‘revision’.

3.2 Timely WCF

Effective, consequential WCF is timely (Evans et al. Reference Evans, Hartshorn, McCollum and Wolfersberger2010), that is, feedback that is given at the right time. Although no research has defined exactly what ‘timely’ feedback means, or has specified the optimum time for feedback delivery, there should not be ‘an unproductive time gap between when the error was made and when it is corrected’ (Evans et al. Reference Evans, Hartshorn, McCollum and Wolfersberger2010: 456). In other words, the less delay in the provision of WCF, the better it should be for students. In the experimental study by Bitchener & Knoch (Reference Bitchener and Knoch2009), WCF was given to students one week after the writing test, and in the study by Evans et al. (Reference Evans, Hartshorn, McCollum and Wolfersberger2010), WCF was provided ‘by the next class period’ (p. 455) (there were four sessions per week in the study). It is important to note that Evans et al. experimented with a special kind of comprehensive WCF, which was given to 10-minute paragraph writing, thus shortening the time teachers had to spend on WCF.

While timely WCF is beneficial for students, it may be less so for teachers because it requires hard work and gives them extra pressure, and this is particularly true for EFL teachers faced with material challenges like large class sizes and heavy workloads. In real classroom situations, therefore, giving WCF several days or one week after the writing is not always feasible. Nonetheless, the need for timely WCF has become a conventional wisdom, as teachers see the provision of timely feedback as a goal, trying to make sure that WCF is returned to students without a long delay. Viewed in this light, I think the research finding about the importance of timely WCF has been reasonably well applied.

In EFL contexts such as Hong Kong, being able to give timely WCF is important. This is, in part, because of the regular writing students are required to do in schools, and hence the need for teachers to deliver timely feedback to clear their backlog of work before the next pile of compositions is collected. Simply put, if teachers are to survive the system, they need to provide timely WCF to students, ‘timely’ as defined, in their context, by returning WCF to students before the next regular writing task is assigned. Although it is hard to gather evidence to show that EFL teachers do give timely WCF, overall, I think the importance of such feedback is well recognized and many teachers see this as a sensible goal to be achieved.

4.1 WCF as a strategy to aid L2 students’ language acquisition

SLA research argues for the use of error correction as an intervention strategy to facilitate language acquisition (Van Beuningen Reference Van Beuningen2010), helping learners notice the gap between what they can currently produce and what they aim to produce (Schmidt & Frota Reference Schmidt, Frota and Day1986). In writing, compared with speaking, learners have plenty of time (and hence cognitive resources) to make a comparison between their output and the WCF provided, and this can make it more likely for them to notice the gaps in their writing (Esteban & de Larios Reference Esteban and de Larios2010; Sheen Reference Sheen2010; Van Beuningen Reference Van Beuningen2010). Although there has been research (Ellis et al. Reference Ellis, Sheen, Murakami and Takashima2008) that suggests WCF can help learners develop greater control over certain grammatical forms (mainly the rule-governed ones), the extent to which WCF is helpful in treating other linguistic forms (such as non-rule-based and more complex structures) remains unknown. Additionally, there is as yet no compelling evidence to show that WCF can aid students’ long-term writing development.

The research finding about WCF as a means to enable ‘noticing’ and facilitate language acquisition has been over-applied. Teachers who spend an inordinate amount of time correcting written errors seem to be affected by the ‘information transfer fallacy’ (Truscott Reference Truscott2010: 629), operating on the assumption that learning is simply a matter of information transfer from teacher to learner. In EFL contexts, such a preoccupation has made teachers respond almost exclusively to the language form in student writing. This, however, ignores the fact that writing competence is a multi-faceted concept, and that the ability to write requires attention to other dimensions of writing, such as content, organization and genre.

I think the over-application of the value of WCF as an intervention strategy to facilitate language acquisition has an undesirable effect on students as well. Teachers’ form-focused feedback practice may have led students to believe that writing is primarily a vehicle for language practice and reinforcement. As a result, students are less ready to experiment with creative ideas and take risks in writing, because in doing so they are likely to commit more errors. Over-emphasis on WCF as a means to enhance students’ written accuracy may, therefore, short-circuit students’ writing development.

It is imperative, therefore, that EFL teachers re-examine the needs of students in their writing classroom and clarify the goals of written feedback in their specific contexts. Do students write to learn the language, or do they learn to write (see Murphy & de Larios Reference Murphy and de Larios2010)? Although learning to write requires students to learn to use the language accurately, students cannot possibly perfect their writing in a classroom that over-emphasizes written accuracy. While WCF can be used as an intervention strategy to assist students in their written accuracy development, ways have to be found to balance WCF with feedback on other areas of writing so that students learn to develop competence in writing.

4.2 WCF and oral metalinguistic explanation

Recent WCF research has shown that oral/written metalinguistic explanation combined with WCF can benefit students’ written accuracy development, especially where the more treatable and rule-governed errors are involved (Bitchener, Young & Cameron Reference Bitchener, Young and Cameron2005). A more recent study by Bitchener & Knoch (Reference Bitchener and Knoch2010) has yielded slightly conflicting findings, suggesting that as long as WCF is provided it does not really matter whether it is given alone or in conjunction with written or oral metalinguistic explanation.

In spite of the inconclusive evidence on the effectiveness of oral metalinguistic explanation alongside WCF, in EFL school contexts it is very common to see teachers giving mini-lessons on student errors, providing oral metalinguistic explanation in addition to WCF. Given the existing indefinite findings, I think this attention to oral metalinguistic explanation is somewhat overdone, especially when such in-class oral metalinguistic explanation is given irrespective of error types and students’ motivation and metalinguistic knowledge. While the studies by Bitchener & Knoch (Reference Bitchener and Knoch2010) focus on metalinguistic explanation of one or two selected error types, EFL writing teachers tend to provide oral metalinguistic explanation for a large number of error categories in an unfocused manner, based on the common errors made by their students in the writing under consideration. I have no objection to teachers providing oral metalinguistic explanation, as in grammar lessons and strategy training sessions, but there is a need for teachers to rethink how such explanation is best provided, together with WCF, to benefit student learning and facilitate uptake. For example, if teachers provide focused WCF on a few selected error types, oral metalinguistic explanation given to the same error types can reinforce student learning. However, oral metalinguistic explanation not geared towards specific error types but provided haphazardly, unsystematically and regardless of students’ metalinguistic knowledge as well as willingness to listen and to engage with the oral explanation, is unlikely to benefit students. This, unfortunately, is what is happening in many EFL classrooms I have observed. Aside from oral metalinguistic explanation, teachers should think of what alternative post-feedback strategies can be combined with WCF to maximize student learning and to make sure that their time and energy is well spent. For instance, after students receive teachers’ WCF, they can work in pairs and help each other with editing by referring to teachers’ oral and/or written metalinguistic explanation.