1. Recent trends in feedback studies

1Feedback is part and parcel of teachers’ and students’ lives. Typically, teachers are the providers of feedback and students the recipients. But feedback can also be provided to students by their peers, their relatives or even by automated digital tools; the practice of self-feedback (Raaijmakers et al. 2019) should not be underestimated either, as it constitutes a valuable means towards improvement but also as it enhances feedback received from others. Feedback can be positive (what the learners ‘do’ well), or corrective (areas that could/should potentially be improved). Whilst positive feedback is “as, if not more, important than negative feedback” (Ellis 2013: 6), our focus in the present paper will be on corrective feedback (hereafter abbreviated as CF) in the foreign language classroom. Ellis, Loewen & Erlam (2006: 340) define corrective feedback as feedback that “takes the form of responses to learner utterances that contain an error. The responses can consist of a) an indication that an error has been committed, b) provision of the correct target language form, or c) metalinguistic information about the nature of the error, or any combination of these.”

2Ellis (2013: 6–7) explains that the reasons underlying the intense academic interest in corrective feedback in the second/foreign language classroom are both practical, i.e. related to “the need for empirical evidence to serve as a basis for pedagogical recommendations” and theoretical, i.e. related to the need of testing and validating the claims made by different theories.

3The two main types of CF are oral (OCF) and written CF (WCF). OCF is provided to point mistakes a student has made while speaking. Lyster & Ranta (1997) listed six types of OCF: explicit corrections, recasts, clarification requests, metalinguistic clues, elicitations, and repetitions (see Annex 1 for illustrations). As for WCF, it consists in written comments, usually provided individually to students. WCF can also take various forms (see Ferris 2003, for a lengthy discussion of the various types). The two key types are direct and indirect WCF. Direct (explicit) WCF includes the provision of the correct/target linguistic form or structure by the teacher, the crossing out of unnecessary words/phrases/morphemes, the insertion of missing words/phrases/morphemes, or the provision of the correct form or structure, the inclusion of written meta-linguistic explanation (e.g. provision of a grammar rule). Indirect WCF, in contrast, is found in cases when errors are simply underlined, circled, or highlighted, or when the number of errors is provided in the margin without further comments. In such cases, it is up to the learner to find or understand the error and come up with a correction. Pawlak (2014) sums up the main differences between OCF and WCF in a Table (see Table 1 below):

4As can be seen from Table 1, whilst sharing some commonalities (OCF and WCF can be input-providing or output-inducing, explicit/overt or implicit/covert, conducted by the teachers, learners themselves or peers, possibly contain metalinguistic information and be didactic), there are also important differences between the two types. Some key differences include the fact that as traces of WCF are more easily accessible, it enhances its corrective force and, second, that WCF is usually more extensive as it often includes comments on numerous aspects of written competence. Although WCF is commonly used for written production and OCF for oral production, both CF types can be used interchangeably (e.g. spoken meta-linguistic explanation when commenting on a written assignment or written comments provided after an oral production).

5Another type of feedback – and/or feedback provider - which is not mentioned in Pawlak’s table is technology-mediated corrective feedback taking place through learner–computer interactions. As explained in Heift et al. (2021) this type of digital feedback focusses mainly on spelling, grammar and writing, and pronunciation. It can be provided by the teacher (e.g. when using tools like Markin or MyAnnotator) or provided fully automatically by an automated language correction tool. Whilst the technological tools used have not yet reached 100% accuracy, the authors recognize the significant progress made over the past decades in that area. They however also point to:

• the need for more studies investigating long-term efficacy of such feedback type,

• the need for guidance from language instructors in the appropriate use of such tools and in a careful interpretation of the feedback

• provided (be it written suggestions or more visual types of feedback like waveforms or spectrograms for pronunciation for instance).

6Five key questions have commonly been debated in CF literature (Mackey et al. 2016: 499): should learners’ errors be corrected ? which errors? when? how? and finally, by whom? They will each be briefly commented on here below and will also help us structure our research questions in the methodological section of the paper.

1.1. Effectiveness of O/WCF?

7Whilst authors generally agree on the definition of CF and on its purpose, they have been debating its effectiveness for several decades. Ellis (2009) and Lee (2017) refer to the Truscott (2007) and Ferris (2004) debate on the value of WCF, with – in broad terms – Truscott claiming that WCF is ineffective and Ferris defending that WCF can be profitable to learners depending on the quality of the correction. Bitchener (2008) conducted a pre-test post-test experimental study comparing a control group receiving no CF to other groups receiving various forms of CF: direct corrective feedback; direct corrective feedback plus written meta-linguistic explanation; and direct corrective feedback plus written and oral meta-linguistic explanation. The results indicated that significant progress was achieved by students who received a form of CF and that this progress was retained over a long period of time (2 months). Whereas all groups significantly improved their accuracy compared to the control group, the group that received direct WCF and metalinguistic explanation did not. The author used those results to put an end to the Truscott-Ferris debate. Li (2010), in a meta-analysis of CF studies, concluded that it was effective in assisting acquisition and reported a medium effect on acquisition (d=0.64). Sheen & Ellis (2011) also concluded that the effectiveness debate should be closed as CF has been proved efficient in both oral and written forms. They also indicate that learners want to receive CF in order to progress and that their consciousness of being corrected plays a major role in their acquisition.

8In terms of language skills, Basturkmen & Fu’s (2021) research review suggests that CF is effective in aiding the development of L2 grammar in terms of accuracy and fluency. Kamiya & Nakata (2021), for their part, confirm the potential benefits of WCF on L2 vocabulary, be it in cases of immediate or delayed revisions. Whilst successful revisions do not per se guarantee vocabulary acquisition, the authors also show that the feedback was retained up to two months after the feedback session, hence pointing to vocabulary storage in the long-term memory of learners. When it comes to pronunciation, Saito (2021: 422) shows that CF facilitates the development of segmental and suprasegmental accuracy but that it is subject to individual variability. He states that the effect of CF in that domain can be maximized “(a) when L2 learners have enough phonetic knowledge, conversational experience, and perceptual awareness of target sounds; (b) when CF provides model pronunciation forms (e.g., recasts rather than prompts); and (c) when the target of instruction concerns communicatively important and salient features”. In a review of nine pioneering studies on the impact of CF on pragmatics, Bardovi-Harlig & Yilmaz (2021) point first to the lack of studies in that area and conclude that, at this stage, they can “neither advocate for the implementation of CF in instructional pragmatics nor advocate against it”. Finally, Tarone (2021: 465) in a study of the link between alphabetic print literacy level and noticing of oral corrective feedback, shows that “an absence or limited amount of alphabetic print literacy substantially impairs L2 learners’ ability to notice and take up form-focused corrective feedback”. She also stressed the urgent need to review theoretical claims about the role of CF in the learning and teaching of L2 for learner populations including immigrants, refugees or illiterate and low-literate adult L2 learners to (re-)assess the efficacy of CF types with those populations and support alternative pedagogical practices. Overall, empirical evidence shows that CF is an essential component of the language learning process, but that its effectiveness is modulated by many learner variables and other factors such as the targeted aspects of language competence.

1.2. Which errors should be corrected?

9Two main trends are generally debated in the CF literature, namely: focused/selective/intensive CF where only a limited number of preselected linguistic features are commented on (sometimes even only one), and unfocused/comprehensive/extensive CF where feedback is provided on all types (or a much wider range) of errors. Space prevents us from going into details as to which error types should be corrected for various proficiency levels and/or task types, and/or for which native vs additional language(s) pairs. Nevertheless, Van Beuningen (2021) offers a recent and rather comprehensive discussion on the characteristics of and issues at stake with those two distinct types, including their (un)planned nature and their (un)systematicity, and the factors influencing the effectiveness of focused/unfocused CF (such as feedback provider, feedback intensity, and feedback timing). She also presents the theoretical, pedagogical and methodological arguments in favor of focused and unfocused CF both for their written and oral modes. In her conclusions, the author warns against the danger of oversimplifying the ‘focused vs unfocused’ CF debate as numerous individual and contextual factors mediate CF effectiveness. She therefore advocates for a combination of various CF approaches depending on the teaching/learning objectives (i.e. focused CF might be more effective for improving learners’ accuracy in some language areas recently treated in class). Finally, she encourages (teachers to encourage) learners to engage with the feedback they receive and recommends to use CF that prompts learners to modify/reformulate/revise their initial output.

1.3. When should errors be corrected?

10Arguments are found in favor of delaying correction until learners have completed an activity (in the case of oral interactions for instance) whilst other are in favor of immediate correction (in accuracy work for instance) as there seems to be an optimal ‘window of opportunity’ (Doughty 2001) immediately after an error has been made.

11Links are being made in the literature between timing, correction modes and types of tasks. Henshaw (2011) investigates the effects of timing on written CF for participants using CALL software. Four conditions were compared: immediate CF after each question, delayed CF at the end of the task, 24-hour delayed CF, and absence of CF. She showed that the three CF conditions statistically outperformed the no-CF condition, and that written CF was effective over time, regardless of its timing. Kang & Han’s (2015) meta-analysis shows that delayed written CF facilitates foreign language learning. They also found greater effect sizes for advanced learners (as opposed to intermediate and beginner learners). Shintani & Aubrey (2016) compare the effectiveness of immediate, delayed and no-CF on participants’ written production. Here again CF conditions statistically outperformed the no-CF condition, but only immediate CF was found effective over time. Oral CF being mainly immediate, fewer studies assess the timing effect. In addition, Quinn (2021) lists some studies comparing immediate and delayed oral CF but argues that some of the studies lack clear construct definitions for immediate and delayed CF, sometimes confound CF timing and type, or lack no-CF control groups, thereby making it difficult to interpret the results.

12Quinn (2021: 334) also warns against opting for simplified and opposed dichotomic views as “providing CF immediately and/or delaying CF provision in both written and oral modes can help your students”. He adds that in the case of oral CF “both the empirical and theoretical literature more strongly support the use of immediate than delayed CF”. If delayed oral CF is provided, learners should then ideally reengage in the communicative activity, which might not be easily operationalizable in instructed contexts.

1.4. How should errors be corrected?

13Ellis (2013: 11–12) lists a number of general SLA principles that can guide the implementation of how CF should be provided:

• CF needs to be intensive. A single correction directed at a linguistic feature cannot be expected to have much effect on learning. An advantage of focused CF is that it is intensive.

• Explicitness is important. For CF to be effective, learners need to recognize the corrective force of the CF.

• Teachers need to vary how they correct according to instructional context. In a fluency-oriented instructional context, learners may fail to recognize a recast as corrective as they are primarily focused on meaning. In an accuracy-oriented context, however, they are more likely to treat a teacher’s recast as corrective. This suggests that in a communicative activity, brief explicit forms of correction may be needed. In a grammar exercise, recasts can be effective.

• Input-providing and output-prompting CF strategies can be combined. One way of achieving this is by means of ‘corrective recasts’ where the teacher first prompts learners to self-correct and then, if that fails, provides a recast.

• Encourage uptake with repair. Learning can take place without uptake providing learners notice the correction. But inducing learners to produce the correct form may lead to deeper processing.

14In the same article, Ellis also lists typical recommendations present in English language teaching (ELT) guides, adding that such guides “are wary of recommending the use of any particular strategy” (2013: 5). Variety is here again the key, with a general preference for strategies that require learners to correct their errors and a reminder of taking into account the importance of the affective aspect of correction which should be perceived by the learners as supportive and assisting and not as judgmental.

15Regarding technology-mediated feedback, the same key principles can broadly be adhered to, whilst also considering the advantages and limitations of current technologies. In the case of automatic (also called ‘automated’) corrective feedback, advantages include the fact that computers are never tired of correcting the same mistake repeatedly or that learners may feel less ‘judged’ by computers than by teachers. Limits include the types of errors and tasks that lend themselves to accurate feedback (mostly closed tasks, mostly for spelling, grammar and writing, and pronunciation). As automatic feedback is a vast domain and as it will not be addressed in our practical part, the specificities of that type of feedback will not be further developed here but see Ranalli (2018) or Barrot (2021) for interesting discussions.

16Finally, the language used to provide feedback also plays an important role in SLA. One section of the 2016 Cambridge papers in ELT sums up recent research related to the use of the students’ L1 in the foreign language classroom. Beyond being unavoidable in some specific contexts, providing CF in the students’ L1 can be beneficial to their learning for affective, psycholinguistic, sociolinguistic, and pedagogic reasons. The literature on the topic has evolved from a ‘deficit view’(interference, native-speakerism) to an ‘asset/multilingualism’ view, ranging from measured pleas for judicious and principled use of the learners’ own language (Widdowson 2003) to more radical manifestos (e.g. “teachers who are to produce bilinguals should themselves be bilingual, i.e. be reasonably fluent speakers of both the target language and the language of their pupils” (Butzkamm & Caldwell 2009: 25).

1.5. Who should provide the feedback?

17As explained at the top of section 1, teachers are the typical providers of feedback and students the recipients. Other feedback providers include student peers, relatives, digital tools, and also the learners themselves in cases of self-feedback. Depending on who provides the feedback, on the objectives met when providing the feedback, and on the knowledge potential of the feedback provider, it can be a mix of the following types: an implicit or explicit indication that an error has been made, the possible provision of the correct target language form, additional metalinguistic information about the nature of the error. Here again space limitations prevent us from going into more detail, but we refer the readers to the four chapters of part VII of the recent Cambridge Handbook of Corrective Feedback in Second Language Learning and Teaching (Nassaji & Kartchava 2021) which address learners’ and teachers’ feedback perspectives, perceptions, and preferences.

18Now that we have provided a summary of recent trends and recommendations related to feedback studies in the SLA literature, we will compare what the literature suggests with the current beliefs and practices of teachers in the Fédération Wallonie-Bruxelles. We will also comment on learners’ perspectives on written and oral corrective feedback with a view to spotting similarities and/or differences between teachers’ and learners’ views.

2. An ecological dive in CF beliefs and practices in the FWB: methodological design, data and participants

19To explore the relationship between research on CF and pedagogical practices in our local context, we decided to adopt a multiple case-study design in which several instrumental and bounded cases are examined using multiple data collection methods. The types of data types and participants are provided in Table 2. The various studies each addressed a number of research questions that cannot be extensively listed or described here (format and length issues). Also, our aim here is not to go into the details of each study but rather to highlight general trends in CF practices in a situated context, with a view to promoting – should the need arise – the adoption of new feedback practices that best meet learners’ needs, help teachers optimize the type and amount of feedback they provide to their students. We have thus decided to comment on a selection of results to illustrate some of the key trends presented in section 1 and which, we believe, constitute focal points and issues to address in the future in pre- or in-service teacher training.

20There are 3 main case studies (CS 1, 2 and 3) each with their own participants. Each case study includes sub-case studies (for instance in CS 3, 3a focused on pupils and 3b on teachers).

21We have identified three focal areas that have received little attention in language teaching research:

• Discrepancy between reported practices, beliefs and actual practices regarding focused/selective vs unfocused/comprehensive CF

• Practices in terms of revision post CF

• Teachers’ practices vs learners’ beliefs on OCF

22In what follows, we will report on selected results that substantially add to the understanding of these focal areas. We will also provide comments on the (potential) discrepancies between practices and instructed SLA research recommendations.

2.1. Discrepancies between reported practices, beliefs and actual practices regarding focused/selective vs unfocused/comprehensive WCF

23Our results in this subsection are based on a total of 30 surveyed teachers (CS 1a and 2a) for reported practices and beliefs, and on the qualitative analysis of 6 teachers’ actual correction practices (CS 1b and 2b). Although the (survey) questions were not similar4 in CS1a and 2a, we were able to merge the results of (almost) identical questions for the calculation of overall results.

E.g.: the answers to the following two questions were merged
CS 1, question 5: When grading, which aspects do you correct?
Form only (grammar, spelling, vocabulary…) - Content only (structure, coherence, relevance…) - Both, but mostly form - Both, but mainly content - Both, in an equal way
CS 2, question 24: Lorsque je corrige les textes écrits, je corrige toutes les fautes (Engl. ‘When I correct written texts, I correct all the mistakes’)
Comments found in the open questions also helped us refine our results.

24A vast majority of the teachers (90%) stated they corrected all mistakes in their learners’ written productions. Whilst no teacher reported correcting only mistakes about language features that were being taught in class when the written task was given, 10% said they corrected only what they considered as ‘serious’ mistakes in the productions (and this could include currently taught language features).

25When compared to actual practices, we noticed that all teachers had provided unfocused (comprehensive) WCF. A close analysis of the few mistakes that had not been pointed out to learners showed that those mistakes had most probably been spotted by the teachers (as identical mistakes or very similar mistakes had been signaled in other paragraphs of the student’s production under study).

26We can thus point out some difference between reported and observed practices: 10% of the teachers said that they corrected only ‘serious’ mistakes, but our qualitative analyses found no sample essay in which focused/selective correction has been provided.

27Another difference between reported and observed practices that we found is related to the provision of WCF for formal (grammar, spelling, vocabulary, etc.) and/or ‘content’ aspects (identified in our surveys as being related to structure, coherence, relevance, ideas, etc.). 65% (rounded percentages) of the teachers said they considered and corrected both equally, 30% said they paid attention to both but with a stronger focus on form and 5% reported correcting form only.

28Here again, the qualitative analyses of actual practices showed that WCF on contents were extremely rare. Counting number of instances and comparing them to the survey results would not be an appropriate numerical comparison but suffice it to say that the very few comments on contents that were found in the corrected essays are far from being equivalent to the reported practices listed above. What some teachers did, however, is taking content aspects into account in assessment grids (see screenshot below in Figure 1) but no actual CF was provided for such aspects.

Figure 1. Screenshot from one assessment grid section related to content aspects

29If we now compare actual practices vs instructed SLA research recommendations, a first striking finding is that the observed practices analyzed in our case studies ignore or disregard the need for – and benefits of – focused feedback (see section 1.2). Our results are in line with previous research (see Lee 2009, Ferris 2007 for example in other countries/contexts, and Noiroux 2019 for a study in the FWB context) which also pointed to a mismatch between recommendations in SLA research and actual classroom practices. The claim made by Ferris (2007: 167) that it is was pivotal to urge teachers to stop aiming for perfection and consistently correcting recurrent errors is thus also valid in our context. In a study on reasons put forward by teachers to avoid using focused feedback, Lee (2009) listed teachers’ comments on the fact that the school policy required them to attend to every single student error, a comment also provided in one of the semi-guided interviews in CS 2c – a comment whose veracity was unfortunately impossible for us to check. What is certain, however, is that the official language curricula in the FWB do not demand the provision of unfocused comprehensive feedback. Another teacher’s comment from CS 2c was very similar to what Lee (2013: 110) calls “the more the better” maxim. Some teachers believe that extensive correction is expected from them (by the school policy, from the learners, from their own professional beliefs) and that it “cannot do any harm” (teacher 3; CS 2c).

30The second finding pertains to the prevalence of focus on form to the detriment of content. Our results are in line with Lee’s 2009 study which also showed that teachers tend to focus too much on language form compared to WCF on content. Similarly to what was found in Lee’s study, whilst most teachers concur with the view that it is important to pay attention to the content in written productions, there is a clear gap between what teachers think is necessary and what they actually do in practice. Potential explanations for this gap can be related to time (once teachers have corrected learners’ language errors, they may be lacking time to comment on content) and to their use of assessment grids containing features related to content (see Figure 1) for which they provide marks – but omit to provide feedback.

2.2. Practices in terms of revision post CF

31Our data (CS 1a) shows that rewriting tasks are not commonly done. More than half of the teachers (approx. 54%) said they would sometimes ask students to rewrite their productions once WCF had been provided. Almost 40% of the teachers (approx. 37%) said they never ask for revisions/rewriting. Just above 8% of the teachers said they always asked for rewritings.

32The data at our disposal in CS 1b and 2b could not allow us to find out whether rewriting tasks had actually been proposed to learners, but these were certainly not mentioned – or even alluded to – in the corrected versions of the essays. In the semi-guided interviews carried out in CS 2c, none of the teachers reported asking for revised versions.

33Here again, we see a difference between research recommendations (see sections 1.1. and 1.2) and practices.

2.3. Teachers’ practices vs learners’ beliefs on OCF

34In order to investigate the similarities and/or differences between teachers’ and learners’ views on CF, we compared the data from the questionnaires with the classroom observations made in terms of feedback on speaking skills with the same sample. We first asked students about whether they took into consideration oral or written CF on their speaking skills and whether they think this would have an impact on their learning (Likert scale with five levels). A vast majority of students (92%, rounded percentages here and below) seems to value CF and thinks it helps further develop their skills (90%). This goes in line with Lyster et al.’s (2013) findings, which suggest that learners expect more feedback than what teachers think they should provide.

35We then asked students about their preferences regarding the different types of CF. Surprisingly, 57% of the students preferred written CF, i.e. a note with written comments, instead of oral CF, i.e. oral correction by the teacher. They also stated that written CF can be made accessible in future oral exercises more easily. Teachers, however, exclusively used oral CF when correcting students in their speaking skills. Three quarters (74%) of the students who favored oral CF indicated that they preferred being corrected in French (i.e. in their mother tongue) rather than in English (i.e. their FL), since French would help them better understand their mistakes. The classroom observation revealed, however, that teachers used either the target language or French (e.g. French was often used for comments related to grammar), although one of the teachers corrected far more often in English than in French.

36Furthermore, 65% of the students reported that receiving delayed oral CF, i.e. after their speech, was more supportive than receiving the oral CF immediately, i.e. during their speech. This contrasts with the teaching practice, where teachers predominantly recurred to immediate oral CF to correct the students. In fact, students perceived immediate CF as more disruptive and thus more frustrating and demotivating. However, it is important to note that the oral exercises mostly focused on formal accuracy rather than on fluency or manner (confidence, eye contact), which might explain why teachers favored immediate CF.

37Finally, slightly more students (54%) preferred implicit over explicit oral CF. In fact, some students perceived explicit oral CF as more invasive and as causing more speaking anxiety. The observed teachers’ practice coincides only partially with those learners’ beliefs: whilst one of the teachers used more frequently recasts as a form of implicit oral CF, the others mostly used explicit correction.

3. Conclusions and recommendations

38After providing a brief summary of recent trends and recommendations related to feedback studies in SLA literature (section 1), we analyzed the data collected in the framework of 3 case studies to compare teachers’ reported CF practices/beliefs and actual CF practices in the Fédération Wallonie-Bruxelles.

39We presented results on two key areas, viz. focused/selective vs unfocused/comprehensive CF and practices in terms of revision post CF (sections 2.1 and 2.2). We commented on discrepancies between beliefs, practices and SLA research recommendations. Discrepancies were found between self-reported practices and actual practices, and also between recommendations found in instructed SLA research and teachers’ practices. Even though scholars tend to encourage teachers to make use of focused WCF, very few of them actually use focused WCF despite their recognition that it is an interesting practice (cf. sample analysis in CS 1). This lack of focused feedback is detrimental both for teachers (time wasted correcting all mistakes for poor impact on learners who cannot take up extensive feedback) and learners (who might feel disheartened and/or cognitively overloaded by unfocused feedback). The lack of revisions post feedback is another case in point. Here again a change in practices would be welcome.

40We also commented on learners’ perspectives on written and oral corrective feedback with a view to spotting similarities and/or differences between teachers’ practices and learners’ views (section 2.3). The analysis revealed that teachers’ practice and learners’ beliefs coincide regarding the language to be used in oral CF, but that they differ in that learners prefer delayed CF whereas teachers tend to privilege immediate oral CF. Learners’ beliefs that implicit oral CF would be too disruptive and would cause anxiety should be investigated in further studies measuring both variables with more specific data collection instruments.

41We believe that the promotion of new/renewed feedback practices should be made both at pre- and in-service level to make teachers aware of the benefits (for them and their learners) that such new practices entail. Providing good-quality feedback is a demanding and time-consuming task that teachers perform almost daily in their profession. They would benefit from some guidance as to how best manage their time and to ensure a higher uptake level from learners. We recommend sharing research results in an accessible way by starting from working on teachers’ own beliefs, inviting them to compare those beliefs (or reported practices) with their actual practices and with curricular recommendations. We also suggest providing teachers with instructed SLA research-based guidelines for sustainable and enhanced CF practices as food for thought in the discussion to make sure that teachers’ time and impact on learners’ progress is maximized. Such teacher-researcher/teaching-researching collaboration would also be beneficial for the development of differentiated CF practices according to students’ needs and proficiency levels. One such example of differentiated practice is related to comprehensive and unfocused WCF. Bitchener & Ferris (2012) have shown that teachers should first take the level of the learners into account and adapt their WCF accordingly (Bitchener & Ferris 2012) as learners with a higher level of proficiency may benefit – more than their fellow students with lower proficiency levels – from comprehensive corrective feedback.

42Our final recommendation is thus to promote teacher-researcher collaboration to the service of learners. Ferris’ s (2007: 167) comment that teachers should help learners develop strategies to progress in their learning and that the most important end-product is each student’s progress and increasing awareness of and skill in using various strategies to compose, revise, and edit their own work, is still valid 15 years later. Promoting teacher-researcher collaboration can be one way of meeting this end-product.