Problems for teachers

For teachers, CWCF takes up an inordinate and disproportionate amount of time, distracting them from providing feedback on other important issues in writing, such as content, organization and genre. It also relegates them to grammar teachers, language editors, or in Hairston's words, ‘error hunters’ (Hairston, Reference Hairston and Bridges1986, p. 122) and ‘composition slaves’ (p. 117). My previous research has shown that this can have an adverse effect on teachers’ identity development, making writing teachers feel that they are simply ‘marking machines’ (Lee, Reference Lee2010, p. 148) that repeat the same tedious job mechanically, day in day out. Hyland refers to this behaviour as ‘reflexive’ (Reference Hyland2010, p. 173) – i.e. teachers responding to errors in a habitual and unthinking manner. It can be imagined that having to give feedback on all written errors is also exhausting and psychologically and emotionally draining, especially when teachers have piles of compositions to mark with tight deadlines to meet. What's worse, when teachers are faced with enormous pressure from teaching several writing classes with large class sizes, having to return timely feedback to students may cause them to perform the WCF task hastily, risking producing illegible WCF and inaccurate error corrections. When students cannot even read the teacher WCF clearly, they fail to understand and act on it to improve their writing. And when the teacher WCF is inaccurate, it is unhelpful, misleading and even harmful (Ferris, Reference Ferris1999; Lee, Reference Lee2004).

For teachers, another problem arising from CWCF is that they are sending students some unintended messages that are not desirable and conducive to student learning. The first message is that one of the top priorities, if not the top, for students is to aim at producing grammatically perfect writing that is free of errors in every attempt. However, it is unrealistic to expect L2 student writers to produce error-free writing, not least from the second language acquisition (SLA) perspective (Ferris, Reference Ferris2002). Another message sent by CWCF is that students are incapable of identifying any of their own errors and so teachers have to work very hard to underline or circle them all for students. In fact, one of the goals of teacher WCF is to help students become independent writers capable of self-editing. This goal, however, is not in sync with CWCF, which is a teacher-dominated practice that reduces students to passive recipients of teacher WCF.

Problems for students

For students, receiving papers flooded with the red ink is overwhelming, confusing and discouraging (Sheen, Wright, & Moldawa, Reference Sheen, Wright and Moldawa2009) – an awful experience for developing L2 student writers, particularly when such experience is repeated over time. The prejudicial effect on students is echoed by Hairston (Reference Hairston and Bridges1986, p. 128):

Although CWCF is prevalent in many L2 contexts, especially EFL contexts (Furneaux et al., Reference Furneaux, Paran and Fairfax2007), there is research evidence that shows the negative impact of CWCF on student writing motivation (Lee, Yu, & Liu, Reference Lee, Yu and Liu2018). As students accumulate negative experiences from CWCF, they are bound to lose interest in writing. Since the affective component can ‘moderate learners’ attention and processing’ (Bitchener & Storch, Reference Bitchener and Storch2016, p. 28) of teacher WCF, CWCF may adversely affect students’ engagement, which impacts negatively on their learning from teacher WCF. In Zheng and Yu's (Reference Zheng and Yu2018) recent study, the low-proficiency EFL students were found to be affectively engaged with teacher WCF that was focused, suggesting that FWCF could facilitate students’ affective engagement with WCF. From the SLA perspective, acquisition is less likely to take place when WCF targets all errors/error types (Schmidt, Reference Schmidt1994; Ellis, Reference Ellis2005; Ellis, Sheen, Murakami, & Takashima, Reference Ellis, Sheen, Murakami and Takashima2008). This is because when students receive their texts filled with underlines, circles, corrections and/or error codes, they are overloaded with information that lacks a clear focus, and such cognitive and information overload (Bitchener, Reference Bitchener2008) leaves students with little attentional capacity to cope with teacher WCF (Bitchener, Reference Bitchener2008; Sheen et al., Reference Sheen, Wright and Moldawa2009). Drawing on Pienemann's (Reference Pienemann1989, Reference Pienemann1998) processability theory and teachability hypothesis, CWCF that targets all grammar items is unhelpful because some of the grammar items covered in teacher WCF are likely to fall beyond students’ readiness. When L2 learners are still developing their grammatical competence in writing, having to understand and/or fix all types of error in a single piece of writing is likely to be beyond their ability. Practically speaking, CWCF is simply not manageable for students.

Finally, with a focus on CWCF, teachers tend to give much less feedback on other issues. Students are mainly helped to develop grammatical competence in writing rather than overall writing competence. Without a ‘balanced diet’ from teacher WCF, students fail to receive appropriate assistance that benefits their long-term writing development.

Overall, research has failed to produce conclusive evidence about the effectiveness of CWCF (Semke, Reference Semke1984; Robb, Ross, & Shortreed, Reference Robb, Ross and Shortreed1986; Kepner, Reference Kepner1991; Sheppard, Reference Sheppard1992), except more recent studies by Hartshorn and Evans (Reference Hartshorn and Evans2015), Van Beuningen, De Jong, and Kuiken (Reference Van Beuningen, De Jong and Kuiken2008, Reference Van Beuningen, De Jong and Kuiken2012), Bonilla López, Van Steendam, and Buyse (Reference Bonilla López, Van Steendam and Buyse2017), and Bonilla López, Van Steendam, Speelman, and Buyse (Reference Bonilla López, Van Steendam, Speelman and Buyse2018), which showed that CWCF could improve written accuracy. Worthy of note is that these studies investigated CWCF without comparing it with FWCF and were conducted in experimental classrooms with limited relevance for authentic classrooms. In Hartshorn and Evans (Reference Hartshorn and Evans2015), for instance, FWCF was delivered to ten-minute paragraph writing rather than full-length essays, while in Van Beuningen et al.’s studies (Reference Van Beuningen, De Jong and Kuiken2008, Reference Van Beuningen, De Jong and Kuiken2012), CWCF was found to be more beneficial than self-editing without WCF, and writing practice without WCF; the findings did not show the benefits of CWCF over FWCF. The studies by Bonilla López et al. (Reference Bonilla López, Van Steendam and Buyse2017 and Reference Bonilla López, Van Steendam, Speelman and Buyse2018) compared direct CWCF and metalinguistic CWCF, focusing on the form of CWCF rather than comparing it with FWCF.

Benefits for teachers

When teachers do not have to respond to all written errors, their time, energy and effort can be re-distributed so that they can comment on content, organization and other important issues in student writing. Teachers then re-adjust the main purpose of their written feedback, using it to help students focus on the overall message and meaning conveyed by their writing, with language use as a means rather than an end in itself. Another advantage of FWCF for teachers is a practical one. With fewer errors to mark, teachers are more likely to return feedback to students in a timely fashion.

As teachers focus less on written errors, they also send forth a positive message that helps build a more supportive learning environment in the classroom, encouraging a more relaxed attitude to errors as well as willingness to take risks in writing, through which a stronger rapport can be built with students too. Rather than turning a blind eye to errors and ignoring them, FWCF can allow teachers to help students address their gaps in written accuracy in a focused fashion so that students are helped to develop their written accuracy in a gradual and incremental manner.

Through FWCF, and with selected grammar focuses for each writing task, teachers can easily align assessment and feedback with pre-writing instruction and post-writing reinforcement, fostering a very strong inter-relationship between assessment/feedback, teaching and learning. For instance, when the target error types for a writing task are verbs and subject-verb agreement, teachers can focus on the teaching of the target items before students start writing, and based on students’ writing performance with regard to the target error types, teachers can design post-feedback/writing reinforcement activities to consolidate student learning. Feedback, teaching and learning thus become purposeful, focused and meaningful, with supplemental grammar instruction used alongside WCF to help students improve written accuracy (Ferris, Reference Ferris2003).

An extra benefit of FWCF, particularly in contexts where teachers engage in collaborative professional development, is that it entails teachers’ collective planning and learning whereby they work out target error types for each piece of writing, and discuss how these can tie in with the instructional focuses, and how the various error types should be recycled and sequenced at different stages of learning or schooling. Such horizontal (involving teachers teaching the students in different classes/sections of the same level) and vertical (involving teachers teaching students of different levels) teacher planning is essential to the successful implementation of FWCF in the long run. As teachers embrace ‘less is more’ in their WCF practice, they also take active steps towards their own continuing professional development and collective professional learning in their work context (McLaughlin & Talbert, Reference McLaughlin and Talbert2006). In conventional CWCF, however, teachers basically work in isolation, with each and every single teacher responding to errors comprehensively in every writing task, with little need to engage in collaborative planning. As teachers implement FWCF, they have to reflect on a host of related issues to make FWCF work, including how to select target errors, the number of target error types, how to align WCF with instruction, and how to carry FWCF through different stages of student learning/schooling. Such significant issues provide a springboard for school-based/collaborative professional learning, which helps teachers engage in reflective practice and is beneficial to their continuing professional development.

To sum up, for many L2 teachers, ‘less is more’ in WCF represents out-of-the-box thinking. It enables them to revisit and problematize an entrenched practice that is both ineffective and exhausting, leading to growth in their own assessment/feedback literacy.

Benefits for students

With a reduced focus on written errors and more balanced teacher feedback on different dimensions of writing, students can make use of teacher feedback to develop their writing competence rather than focus mainly on grammatical accuracy in writing. With fewer underlines, circles, error codes, etc., WCF is less confusing and intimidating, and students are more likely to take risks (e.g. experimenting with new language to express meaning rather than use grammatically safe expressions that do not necessarily convey their real intended meaning) and as a result build fluency and confidence in writing.

The benefits of FWCF for students can also be examined from the SLA perspective. To elaborate, drawing on the work of Gass (Reference Gass1997), the cognitive processing of WCF goes through several stages (also see Bitchener & Storch, Reference Bitchener and Storch2016). In order for learning to take place, students must first attend to the WCF provided, notice that there is a gap between their own output and the target language input through the WCF, and then comprehend the input (i.e. WCF) – before intake (i.e. matching WCF with existing knowledge – through hypothesis testing), integration, and accurate output can take place. Since FWCF draws attention to several error types, students’ attention to their written errors becomes more focused, which can benefit their learning and L2 development. First, they are more likely to attend to the teacher WCF, which is conducive to learning because learning tends to occur for things that students attend to rather than not attend to (Schmidt, Reference Schmidt and Robinson2001). With FWCF, students are also more likely to notice the WCF/input with awareness and understanding (Bitchener & Storch, Reference Bitchener and Storch2016), since a smaller number of grammar items are selected for students’ attention. And when the target grammar items are chosen because they have been taught or are within students’ existing L2 knowledge system, students are more likely to comprehend the WCF, which in turn facilitates the internalization of the input, and subsequently intake and integration (through hypothesis testing). When post-feedback grammar reinforcement activities are used to consolidate student learning of the specific target grammar items, students are more likely to remember and avoid making the same errors in their future writing (i.e. accurate output). Thus, the cognitive processing perspective can help us understand why FWCF is beneficial to students and able to facilitate their L2 development.

Another obvious benefit of FWCF for students is practical. With fewer errors to attend to, WCF becomes much more manageable. According to Evans, Hartshorn, McCollum, and Wolfersberger (Reference Evans, Hartshorn, McCollum and Wolfersberger2010, p. 451), WCF is ‘consequential when it is manageable’. For students, learning is rendered more focused and goal-oriented as they know exactly what error types to focus on each time they write, and this can facilitate goal setting, self-monitoring and self-evaluation too. Conversely, in a writing classroom that emphasizes CWCF, students have to attend to all grammatical items each time they write, which is unrealistically demanding particularly for younger or weaker learners who are still developing their grammatical proficiency – and not very manageable either.

From a sociocultural perspective, WCF can be considered an affordance (van Lier, Reference van Lier and Lantolf2000), a learning opportunity, but only when students perceive it as such and act on teacher WCF. Therefore, the specific features of the teacher WCF and the forms of assistance provided (Bitchener & Storch, Reference Bitchener and Storch2016) are important as they can impact directly on learning. For developing L2 student writers, CWCF can be overwhelming and confusing (Sheen et al., Reference Sheen, Wright and Moldawa2009), and WCF delivered in this form may fail to provide assistance. Instead, FWCF, being focused and hence more manageable for students, can provide an appropriate form of assistance and is more likely to provide affordances for students. (Different WCF strategies – e.g. direct or indirect WCF can also represent different forms of assistance; however, they are beyond the scope of this position paper and hence not examined here.)

CWCF can help students avoid making the same errors again

According to behaviourist learning theory, error is more or less like ‘sin’ (Brooks, Reference Brooks1960, p. 58), which has to be eradicated. This is because when bad habits are formed, they are likely to negatively influence the learning of new habits (Brooks, Reference Brooks1960). This belief is foundational to CWCF. However, telling students all the errors they have made by giving CWCF is one thing, but students learning from it and avoiding error recurrence is quite another. According to the constructivist learning theory (Vygotsky, Reference Vygotsky1978), learning is not a passive process where students learn from being told, but rather a contextualized process of active knowledge construction influenced by cultural and contextual factors, students’ personal beliefs and prior experiences, as well as their own interpretations of, and responses to, the knowledge presented (i.e. teacher WCF). Viewed in this sense, even though teachers point out all errors (or even correct them), students may not learn and avoid making the same errors again in their writing, since learning is not a simple matter of acquiring knowledge from the WCF provided but instead is mediated by a host of factors, including students’ understanding of teacher WCF. As mentioned in the preceding section, due to cognitive overload through CWCF, students may fail to notice and understand the input/WCF. However, noticing-with-understanding is necessary for uptake and L2 learning (Bitchener & Storch, Reference Bitchener and Storch2016); hence, even though CWCF is provided, students are likely to continue to make the same errors. Also, when CWCF is adopted, it is very likely for teachers to provide WCF on grammar items that students are not yet developmentally ready for (Pienemann, Reference Pienemann1989, Reference Pienemann1998), which explains why students may make the same errors again (Truscott, Reference Truscott1996).

There is some research evidence that shows that FWCF can facilitate cognitive processing and is effective in improving students’ accuracy. For example, the studies by Bitchener (Reference Bitchener2008, Reference Bitchener2012), Bitchener and Knoch (Reference Bitchener and Knoch2008, Reference Bitchener and Knoch2009a, Reference Bitchener and Knoch2009b) and Sheen (Reference Sheen2007), which involved feedback on two functional uses of the article system, showed the effectiveness of focused WCF in improving students’ written accuracy. Ellis et al. (Reference Ellis, Sheen, Murakami and Takashima2008) and Sheen et al. (Reference Sheen, Wright and Moldawa2009) compared the effectiveness of FWCF delivered to the English article with unfocused WCF defined as WCF on a variety of linguistic errors (four in total). While Ellis et al. (Reference Ellis, Sheen, Murakami and Takashima2008) showed that focused and unfocused WCF could be equally effective in improving students’ use of articles, Sheen et al. (Reference Sheen, Wright and Moldawa2009) found that FWCF was more effective than unfocused WCF in improving students’ written accuracy. It can therefore be concluded that WCF may be more effective when it is highly focused (e.g. addressing one linguistic item) than mid-focused (e.g. addressing four linguistic items). However, since these studies were conducted in tightly controlled experimental classrooms involving a limited number of grammar items, much more classroom-based research is needed to shed light on the effects of FWCF in real classroom contexts.

The purpose of teacher WCF is to improve written accuracy in general

Teachers give WCF to help students improve written accuracy in general; however, in L2 contexts this is a long-term rather than a short-term goal. Given this long-term goal, the key issue at stake is whether CWCF or FWCF is a better means to the end. As we compare CWFC and FWCF each as a pedagogical/assessment strategy, CWFC deals with all errors each time students write, whereas FWCF targets a range of errors/error types in each piece of writing. While CWFC is unfocused and random, FWCF is focused and purposeful, with the choice of target errors/error types based on informed decisions – e.g. aligned with grammar/writing instruction and/or based on recurrent error patterns in student writing. While CWFC is overwhelming, FWCF is manageable for both teachers and students. If teachers’ ultimate goal is to help students develop written accuracy in general, prioritizing the target grammar items, focusing on error patterns that are recurrent, and delivering FWCF in a systematic manner, is a fruitful way to help students reduce errors and develop written accuracy in the long run since ‘people master skills and solve large problems by tackling small, manageable tasks a few at a time’ (Hairston, Reference Hairston and Bridges1986, p. 128). From the sociocultural perspective, FWCF better serves students’ learning needs as it is able to help them progress within their zone of proximal development in a gradual manner (ZPD) (Vygotsky, Reference Vygotsky1978), which takes into consideration students’ current and potential level of development. Although feedback provides students with a form of assistance, not all forms of assistance are beneficial to students’ L2 development. In fact, too much assistance (e.g. through CWFC) can be counterproductive and stymie development (Lantolf & Thorne, Reference Lantolf and Thorne2006). In FWCF, even though only a few error types are targeted for WCF in each piece of writing, it does not mean that other grammar items receive no attention at all because teachers can still provide input on these items through grammar instruction. Moreover, over time all the major grammar items will be covered in teacher WCF, and the goal is to help students achieve written accuracy in writing in a gradual and incremental manner. Thus, FWCF, being strategic and dynamic, is a better way to help students achieve written accuracy in the long run than CWFC, which is random, purposeless and lacks a clear focus.

It is noteworthy that highly focused FWCF (i.e. delivered to one or a very small number of grammar items as in recent FWCF research reviewed above) is criticized as unauthentic in real classroom contexts (Hartshorn et al., Reference Hartshorn, Evans, Merrill, Sudweeks, Strong-Krause and Anderson2010; Van Beuningen et al., Reference Van Beuningen, De Jong and Kuiken2012) since real life tasks require ‘accurate production of all aspects of writing, simultaneously’ (Hartshorn et al., Reference Hartshorn, Evans, Merrill, Sudweeks, Strong-Krause and Anderson2010, p. 89). Although I argue for the need to go for less teacher WCF, targeting only one (or two) grammar items in FWCF may not suit most L2 contexts. In this position paper, my objection is mainly to a rigid adherence to CWCF that encourages an unthinking and mechanical approach to WCF. While FWCF is overall more desirable than CWCF, the two approaches are not mutually exclusive. Taking into account the students’ stage of learning and specific needs, teachers can consider adopting a ‘middle’ position (Ferris, Reference Ferris2010) that combines FWCF and CWCF, which will be examined in the ‘What teachers can do’ subsection in the following section.

It is confusing for students if some of the errors are corrected while others are not

One major argument against FWCF is that it is confusing for students if some errors are pointed out while others are not. The potential ‘confusion’, however, can be dissipated if teachers tell students expressly right at the outset of the writing class that FWCF is to be practised, explain the rationale and the procedure involved (e.g. the target error types and how they are selected), and inform students of the other kinds of feedback they should expect to get (e.g. feedback on content and organization) as well as the actions they are required to take (i.e. revising their writing by paying heed not only to language but also other aspects of writing). Once FWCF becomes a routine, confusion, if any, will disappear and instead be replaced by focused and purposeful learning. The initial potential confusion caused by FWCF, if any, is definitely not as damaging in the long term than the confusion, frustration and discouragement caused by CWCF over time, since the red ink on student texts, as well as the dense and at times illegible teacher feedback (Lee, Reference Lee2008b) is enough to confuse students and shatter their confidence. What's worse, confusion can arise from CWCF that contains inaccuracies, which are likely to occur through fatigue or incomplete grammatical knowledge on the part of the teacher. In Lee's study (Reference Lee2004) where 58 Hong Kong secondary English teachers were asked to correct a student essay, analysis of the teacher error corrections showed that only slightly half of them were accurate – the ‘inaccurate’ part comprising omissions (errors not identified), unnecessary WCF (e.g. due to teachers’ stylistic preference), and inaccurate grammatical judgement (e.g. treating ‘damage’ in the student text as wrong by changing it to ‘damages’ when in fact the word should be used as an uncountable noun in the text). When teachers’ WCF itself is inaccurate, it is highly confusing and student learning is likely to suffer. On balance, if FWCF and CWCF may cause some confusion, the confusion triggered by FWCF is resolvable and much less damaging in the long run, whereas the confusion caused by CWCF is cumulative, extremely discouraging, and can be deleterious to students’ writing development especially when teacher WCF itself is not entirely accurate (Lee, Reference Lee2004).

The more WCF, the better the teacher

Conventional wisdom is that teachers who respond to written errors copiously and meticulously are competent writing teachers (Hairston, Reference Hairston and Bridges1986) – and hence the more WCF, the better the teacher. This conventional belief will naturally create problems for teachers who adopt FWCF. In Lee (Reference Lee2008a), the participating teachers said that if they did not adopt CWFC, they would be considered lazy or irresponsible. However, we have to revisit the real reasons why teachers give WCF on student writing. It is not because teachers are paid to catch as many written errors for students as possible, nor to prove that they are responsible, competent and hardworking. The more important reasons are to use feedback to promote students’ learning and turn them into autonomous writers. The long-term goal of WCF is to help students become independent in self-editing so that even without the teacher, they are able to edit and improve their written accuracy and produce grammatically accurate English in the long run. Given these sounder reasons for teachers to give WCF, the ‘better’ teacher should be the one who provides quality WCF, rather than more WCF. Quality WCF is focused and purposeful; it enhances students’ motivation and confidence and improves their performance in written accuracy. When teachers spend less time responding to all written errors, they work smart on WCF but work hard on other areas – e.g. giving feedback on non-language issues and devising pedagogical materials to strengthen pre-writing instruction. With less but more quality WCF, more can be done on training students to conduct peer and self-editing, which can contribute to students’ long-term written accuracy development. Thus, with FWCF, teachers can produce quality WCF (and possibly quality feedback in general), and are definitely no less competent or hardworking than those who give CWCF.

Key stakeholders expect teachers to mark errors comprehensively

When students’ writing has been subjected to CWCF throughout their educational experiences, it is natural for them to expect teachers to continue with the practice as it may be regarded as a tried-and-trusted method. Although there is research evidence that suggests L2 students have a preference for CWCF (Leki, Reference Leki1991; Oladejo, Reference Oladejo1993; Lee, Reference Lee2005), students at lower proficiency level can be inhibited and discouraged by teacher CWCF (Lee, Reference Lee2008b) Hence it is not true that CWCF is necessarily a feedback practice that all L2 students look forward to. When students have a chance to experience FWCF and see for themselves its benefits, their traditional mindsets will be changed. For other key stakeholders like parents and school administrators, especially in contexts heavily influenced by the Confucian heritage culture (Biggs, Reference Biggs1994; Watkins & Biggs Reference Watkins and Biggs1996), such as Hong Kong, China, Singapore and Taiwan, they may hold ingrained beliefs about ‘models’ in learning (and hence CWCF and error-free writing) and the teacher as ‘authority’ – and hence the importance of CWCF and error-free writing, as well as the preponderant role of the teacher in providing CWCF to students.

However, if CWCF had been effective and productive, there would not have been so much research conducted and literature written about the topic – most notably, Truscott's (Reference Truscott1996) vehement, polemical argument for the complete abandonment of grammar correction in writing that triggered the subsequent research on WCF. The problem with the key stakeholders’ expectation of CWCF is that it does not emanate from their expertise in English language education or writing but rather their lack of it; it also signifies the conservatism that characterizes mainstream educational practice in many L2 contexts. Without innovation, however, educational practice will ‘stagnate and produce mediocre outcomes’ (Serdyukov, Reference Serdyukov2017, p. 17), which is happening in the area of WCF. FWCF has strong theoretical support from the research literature (e.g. SLA and sociocultural perspectives), and hopefully with positive outcomes from FWCF based on teachers’ concerted efforts at classroom-based research, even key stakeholders with recalcitrant beliefs can be persuaded, especially when they understand the philosophy of FWCF and witness positive impact on students’ attitudes to and performance in writing. This takes time, and therefore systematic actions are needed – which I turn to in the next section.

What teachers should do

At the individual level, for teachers adopting CWCF, they should start reflecting on their WCF practice, gather evidence about its impact on student learning and writing through critical inquiry or teacher/classroom research, and examine how CWCF may impinge on the quality of their feedback on other important aspects of student writing. Collectively, teachers should get together and engage in professional sharing and exchange, and examine the benefits and problems of CWCF for both students and themselves. Through experience sharing and critical reflection, they can identify core issues that need to be addressed as a team, and discuss how best to plan the delivery of WCF to maximize student learning. At the school level, teachers should garner the support of school administrators through presenting a collaborative proposal that includes the rationale of FWCF and the grammar items targeted for FWCF in different writing tasks and at different stages/levels. In L2 contexts where parents play a significant role in their children's education, it is also important to enlist the support of parents, and this can be done through teachers’ open sharing of the philosophy and rationale of FWCF in newsletters and/or parent forums.

In order not to confuse students who used to receive CWCF, teachers should explain the rationale of the new FWCF practice and make students understand that though there is less feedback on errors, more focus is given to other aspects of writing. And when students revise their writing they should also pay attention not only to language but also content, organization, etc. It is important to remind students that for non-target grammar items not addressed in teacher WCF, they are not necessarily correctly used, and students should engage in self-/peer editing to fix as many errors as they can at different stages of the writing process. So ‘less teacher WCF’ becomes more because teachers can increasingly share responsibility with students, providing them with opportunities to engage in self-/peer editing on a regular basis. Students can also be encouraged to make use of computer-mediated resources to improve the accuracy of their writing – e.g. using Microsoft Word, Thesaurus, concordancing, or web-based grammar check (e.g. Grammarly) to monitor their written accuracy development.

Depending on students’ proficiency level, instead of targeting only one or two grammar items, teachers should provide WCF on a range of error types based on the genre/writing task as well as students’ needs, perhaps gradually increasing the number of target grammar items as their students’ progress in written accuracy. When students reach an advanced level in writing, they are likely to make very few errors; by then the choice between CWCF and FWCF is not going to be of much relevance (Bitchener & Ferris, Reference Bitchener and Ferris2012). In fact, while using FWCF as the staple approach, teachers may combine it with CWCF at different stages of student learning, as they see fit. For example, teachers can give CWCF to paragraph writing but FWCF to essay writing (Evans et al., Reference Evans, Hartshorn, McCollum and Wolfersberger2010; Hartshorn & Evans, Reference Hartshorn and Evans2015), CWCF to the first few paragraphs in an essay and FWCF to the rest of the writing, or CWCF to shorter essays but FWCF to more extended essays. They can give FWCF to intermediate drafts, and engage students in peer and/or self-editing in the final draft, helping students fix as many errors as they can so as to produce a relatively accurate text in the final stage of writing (Bitchener & Ferris, Reference Bitchener and Ferris2012).

To implement an evidence-based approach to the implementation of FWCF, teachers should collect baseline data from students’ writing to track students’ written accuracy development. To do this, teachers combine FWCF and CWCF for diagnostic purposes, where they give FWCF to the majority of the writing tasks but CWCF to two pieces of writing administered at the beginning and end of the writing class to allow comparison of students’ written accuracy performance over time. Through the CWCF provided to the first piece of writing, teachers can analyse all the error types in students’ writing, work out the error ratio and gravity for each error category, and use such information to inform individual students of their strengths and weaknesses. To calculate the error ratio for each error type (see Ferris, Reference Ferris2002), teachers can first count the number of errors in each error category and divide it by the total number of errors in the entire essay. For instance, if there are 30 errors in total and three errors in the use of ‘subject-verb agreement’, the error ratio for this error category is 0.1. When the ratio for each error category is worked out, teachers can order the error ratios and find out the most serious error type based on the largest ratio. Such diagnostic information will be useful for individual students as it can help them understand their own error patterns, and set personal learning goals to monitor their own written accuracy development. The error analysis data for each student can then be compiled to generate an error analysis sheet for the whole class, showing the most serious error types in students’ writing for that class, which can guide teachers’ FWCF (e.g. their selection of error types) as well as grammar instruction. At the end of the academic year/writing class, the same or a similar writing task (e.g. the same genre and a topic of similar difficulty level) can be administered (in the same condition). Using the same error analysis procedure described in the above, teachers can generate useful data that provides evidence about the progress, if any, in students’ written accuracy at the end of the writing class. Hopefully, the data can also shed light on students’ mastery of the grammar items targeted for teacher WCF throughout the year. One caveat, however, is that since the errors students make are likely to vary with the text type, the diagnostic writing task as described above may be limited and fail to provide a complete picture of students’ written accuracy for all text types.

What teacher educators should do

To help teachers realize and understand that less teacher WCF is more, awareness-raising is crucial. Teacher education classes that focus on L2 writing, assessment and feedback should provide teacher learners with plenty of opportunities for critical reflection. I have used a mini-debate on WCF with my teacher learners, and each time my students leave the classroom with renewed perspectives on the issue of WCF. Basically, the teachers form two groups, one in favour of CWFC, and the other against it. They have to take turns to present their arguments and counter-arguments and provide rebuttals. The mini-debate gives teacher learners, pre-service or in-service, a great opportunity to examine the issue of WCF from multiple perspectives and to think outside the box. It is not surprising to hear from my students that after the mini-debate, their faith in CWCF becomes shaken and they begin to realize the problems associated with CWCF.

Indeed, teachers’ feedback literacy development is not a simple matter of acquiring broad WCF principles from the literature; it entails not only an understanding of knowledge of effective feedback principles but also comprises knowledge that is ‘contextualized in the realities of teachers’ contexts of practice’ (Scarino, Reference Scarino2013, p. 316). Since WCF does not take place in a social vacuum (Goldstein, Reference Goldstein, Silva and Matsuda2001; Ortega, Reference Ortega2012), it is not useful for teacher educators to simply promote the best WCF practices (e.g. less is more) without taking account of teachers’ concerns and constraints within their situated contexts. Teacher educators have to be aware of the potential obstacles to the implementation of FWCF in contexts traditionally dominated by CWCF, and provide suggestions to help teachers navigate the contextual challenges (e.g. how to cope with the conservatism among colleagues with entrenched ideas about CWCF). They can play the role of critical friends and/or external experts by offering assistance to teachers, helping them with school-based professional learning or classroom-based research so as to help enhance their WCF practices.

What researchers should do

Although I have argued for the need of FWCF in L2 writing classrooms, I am aware of the lack of a strong research base to support some of the views expressed in this article. Much of previous WCF research has been conducted in experimental or quasi-experimental classrooms that are not authentic for teaching and learning purposes. There is, therefore, an urgent need to carry out WCF research in real classroom contexts – namely classroom-based research with greater ecological validity. Such classroom-based research should address the long-term effects of FWCF (as compared to CWCF), with specific research focuses on students and teachers. In particular, research should investigate the implementation of FWCF on a longitudinal basis, focusing on how teachers implement it in their own situated contexts (e.g. amidst the constraints they are faced with), the effects of FWCF on students’ motivation and written accuracy, and the factors that influence the efficacy of FWCF. For students, as the learner factor can influence their response to and uptake of teacher WCF, students of different backgrounds (e.g. different proficiency and motivation levels) can be surveyed about their attitudes to WCF – e.g. whether and how they improve their motivation and engagement in classrooms that adopt FWCF. Students’ written accuracy development can also be tracked on a longitudinal basis – e.g., through the use of error analysis that focuses on specific error types targeted for FWCF, as well as writing measures that investigate students’ overall written accuracy. For teachers, research can investigate their beliefs and practice about WCF, as well as how pedagogical conceptual change (i.e. beliefs about new ideas being translated into effective practice) happens (Guskey, Reference Guskey2002; Lee, Cawthon, & Dawson, Reference Lee, Cawthon and Dawson2013) – e.g. how teachers implement and engage in WCF innovation, and in particular, how their critical reflection and professional learning facilitate their feedback literacy development.

Sociocultural theory will have promising potential in future WCF research as it provides alternative perspectives to cognitive theories of SLA, as well as more robust explanations for issues of WCF that have remained unanswered in current WCF research – e.g. the factors (both individual and contextual) that impact on WCF and influence students’ engagement with teacher WCF and uptake. Using activity theory (Lee, Reference Lee2014; Storch Reference Storch2018), research can address the tensions and contradictions arising from the conventional WCF activity system (characterized by CWCF) and investigate how innovation through FWCF may impact on the teachers’ provision of WCF and students’ response, engagement and uptake (also see Storch, Reference Storch2018 for research agenda from sociocultural perspectives).

Insights from the above proposed WCF research areas can have direct ramifications for teacher education, informing teacher educators of how best teacher education can be designed and delivered to help frontline teachers cope with WCF. Hopefully, this position paper can stimulate further ideas for researchers and practitioners alike so that teacher WCF can be rendered a positive and productive experience for both teachers and students in the long run.