Studies in Second Language Learning and Teaching

This paper reviews and reflects on developments in classroom-based research on second or foreign language (L2) writing from 2001 to 2020, based on schol-arship drawn from the Journal of Second Language Writing , the flagship journal of the field. The review covers a total of 75 classroom-based studies and examines the major research themes and key findings under three research strands: (1) students and student learning of writing; (2) teachers and teaching of writing; and (3) classroom assessment and feedback, as well as the key theories and research methodologies adopted in extant classroom-based studies on L2 writing. The article ends with a discussion of the practical implications arising from the review, as well as potential research gaps that inform future directions for L2 writing classroom-based research. By providing a state-of-the-art review of developments in classroom-based research on L2 writing, this article contributes to a nuanced understanding of salient issues about learning, teaching and assessment of writing that take place in naturalistic classroom contexts, with relevant implications for both L2 writing practitioners and researchers.


Introduction
In the field of education causal inference research that takes place in controlled conditions, which produces generalizable findings, is often seen to provide the gold standard.To make sense of complex, real-world phenomena in education, however, classroom-based studies that occur in naturalistic classroom conditions are valuable as they can bring nuanced understandings of classroom learning and teaching through in-depth investigation of salient issues in authentic classroom contexts (Allwright & Bailey, 1991).Although not intended for the purpose of generalization, classroom-based research produces knowledge that is context-specific, with potentially powerful, practical impact that can enhance teaching and learning (Kostoulas, 2015).
In second or foreign language (L2) writing, research that has taken place in naturalistic classroom contexts has been in the minority.For example, in the Journal of Second Language Writing (JSLW), the flagship journal in the field of L2 writing, out of a total of 349 studies (full-length articles and short reports) published from 2001 to 2020, only 75 (21.5%)are classroom-based.Given the insight classroom-based research can produce to guide pedagogical practice in real-world contexts, there is a need to understand what classroom-based research has been conducted on L2 writing, how it has informed classroom teaching and learning, and what further issues merit research attention.The purpose of this paper, therefore, is to provide a review of developments in classroombased research on L2 writing with a view to yielding a nuanced understanding of key issues about writing in naturalistic classroom contexts, with relevant implications for both L2 writing practitioners and researchers.
In this paper, classroom-based research refers to investigations undertaken by teachers themselves (e.g., action research) or by external researchers (e.g., case study research) in naturalistic classroom contexts (Allwright & Bailey, 1991;Kostoulas, 2015).The term classroom is used in a broad sense, referring to a learning space in which teaching and learning take place in naturalistic conditions.Such learning space can be physical or virtual and hence both pen-and-paper and digital writing is included.The learning space is not confined to the traditional classroom where a teacher is teaching a class of students through lectures and/or other pedagogical activities, but it also includes other learning spaces where teaching and learning occur in situ, such as tutorials in writing centers (Nakamaru, 2010), conferences between supervisors and supervisees (Merkel, 2020), doctoral writing workshops (Thurlow et al., 2019), students' group writing conferences, and dissertation support groups (e.g., Russell-Pinson & Harris, 2019).The learners of classroom-based research are situated in educational contexts ranging from pre-school, elementary, secondary to college/university contexts that include undergraduate and graduate students as well as teacher learners in pre-service or in-service settings.

The review: Scope, method, and objectives
This review covers the 20-year period (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020) of classroom-based studies published in JSLW.Although it does not include relevant papers from other language journals and hence cannot qualify as a comprehensive review of classroom-based research on L2 writing during the period, in light of the status of JSLW as the flagship journal of L2 writing, the review can provide rich and useful information about the major research foci, theories and methodologies that have informed classroom-based research on L2 writing in the last 20 years.
After collecting the 75 articles published in JSLW from 2001 to 2020, I had a careful read through the abstracts to get an overall view of the research covered in the articles, which quite unsurprisingly turned out to be wide-ranging.Previous researchers have focused on five areas in classifying topics in L2 writing, namely writers' characteristics, writing process, writing feedback, writing instruction, and writers' texts (see Zhang, 2008).My preliminary analysis of the 75 abstracts showed that while writing instruction and writing feedback were pertinent, there did not appear to be a prominent emphasis on writers' characteristics, writing process, and writers' texts (probably because research on these areas does not typically take place in naturalistic classroom contexts).Instead, quite a number of studies addressed different aspects of student learning of L2 writing.Initially, I had writing instruction and writing feedback as two possible areas, plus learning of writing, but then realized that it was necessary to broaden feedback to assessment, with feedback as part of assessment.As a result, teaching, learning, and assessment of L2 writing were the three broad areas that provided the point of departure for my analysis.Such a broad classification coincides with that adopted by Javadi-Safa (2018) in his overview of key issues in L2 writing and research.
I then re-read the 75 studies, categorizing them into one of the three main areas, that is, teaching, learning, and assessment.Within the assessment area, a few studies have focused on feedback specifically; hence I renamed it as the classroom assessment and feedback strand.As for the teaching area, several studies have addressed teachers themselves; therefore, the second area was refined as the teachers and teaching of writing strand.Similarly, in the third area on learning, some studies have focused on students themselves such as student beliefs and motivation; as a result, the third area was refined as the students and student learning of writing strand.The review reveals that the large majority (N = 51) (68%) of classroom-based studies have addressed the students and student learning aspect of L2 writing, with 22.7% (N = 17) focusing on teachers and teaching, and 9.3% (N = 7) dealing with classroom assessment and feedback.Within each strand, I further identified the major themes of the studies drawing on titles, keywords, and major research foci; the themes and key findings are reported in the following section.
It is noteworthy that of the 75 classroom-based studies, 15 (20%) have been conducted in digital environments.Unsurprisingly, the large majority of these 15 studies (13 -86.7%) have been published in recent years, that is, from 2016 to 2020, addressing topics like multimodal composing (Jiang, 2018), wiki collaborative writing (Li & Kim, 2016), and electronic feedback (Ene & Upton, 2018).The review was guided by the following questions: RQ1: What are the major research themes and key findings of classroombased research on L2 writing under each of the three strands?RQ2: What key theories and methodologies are adopted in extant classroom-based research on L2 writing?RQ3: What pedagogical implications can be drawn from the review?RQ4: What recommendations for further research arise from the findings of the review?

Major research themes and key findings
This section provides the findings for the review study, that is, the major research themes and findings of classroom-based research on L2 writing under each of the three research strands.

Research strand 1: Students and student learning of writing (N = 51)
The bulk of classroom-based studies on L2 writing have focused on students and their learning of writing, under which four themes are identified: development of academic writing/literacy (N = 20); cognitive and affective dimensions of student learning/writing (N = 12); student writing in the digital space (N = 11); and student learning from interactions with tutors/supervisors (N = 8).

Development of academic writing/literacy (N = 20)
Classroom-based research on students' development of academic literacy has mainly taken place in undergraduate and graduate contexts.In this particular research theme, how students make use of source texts and engage in citation practices is crucial to their success in academic writing (Friedman, 2019;Li & Casanave, 2012;Wette, 2010Wette, , 2017a)).While use of source texts continues to present a challenge to L2 students (Li & Casanave, 2012), socialization into citation as an academic practice is a complicated matter (Friedman, 2019), since students tend to consider citation a formal requirement of academic writing rather than engage with source texts to project their own academic voice.Although instructional intervention can help cope with the use of source texts in academic writing, students have difficulties grappling with complexities in texts and incorporating citations with their own voice (Wette, 2010).On the whole, L2 writers need support to better understand rhetorical and disciplinary purposes of source text use and to engage in effective citation practices (Wette, 2017a).Development of genre knowledge and awareness plays a significant role in L2 students' development of academic literacy, whether it is research article writing for doctoral students (Negretti & McGrath, 2018), writing book reviews and literature reviews for graduate students (Wette, 2017b), or email and summary writing for undergraduate students (Yasuda, 2011(Yasuda, , 2015)).The development of genre knowledge entails the development of metacognition (Negretti & McGrath, 2018).Mind maps (Wette, 2017b), genre-based tasks (Yasuda, 2011) and SFL-informed genre analysis tasks (Yasuda, 2015) are found to have a useful role to play in developing students' genre knowledge (Wette, 2017b), genre awareness, linguistic knowledge, and writing competence (Yasuda, 2011), as well as genre awareness and meaning-making choices (Yasuda, 2015).
Within this research theme on the development of academic writing/literacy, collaboration among students is another topic of investigation.Such collaboration includes collaborative prewriting discussion (Neumann & McDonough, 2015), which fosters student reflection with potentially positive impact on the quality of their written products.Prewriting collaboration in the form of online text chat before writing, compared with face-to-face discussion, can better promote linguistic accuracy in post-chat individual writing, and is more beneficial to less skilled L2 writers (Liao, 2018).Apart from prewriting collaboration, student collaboration can take place during writing, that is, collaborative writing.When compared with individual writing, collaborative writing can improve student texts in terms of task fulfillment, written accuracy, and complexity (Storch, 2005).Overall, the effectiveness of collaborative writing may hinge upon a number of factors, such as students' first language (L1) as well as judicious use of this resource, L2 proficiency, education experiences, and group rules (such as task division, keeping group harmony, and taking leadership) (Yang, 2014).In addition to collaborative writing, students can collaborate in group writing conferences, during which they engage in negotiation process regarding the ideal forms of genre, for example, thesis genre for doctoral students in the gradate context (Mochizuki, 2019).
Two studies have addressed the challenges younger L2 students face in acquiring academic literacy, namely, adolescent English learners (Spycher, 2007) and community college students (Finn, 2018) in the US.To help students succeed in academic writing, teachers can provide explicit instruction that addresses particular features of academic language (e.g., authoritative stance, conjunction, and reference) (Spycher, 2007), as well as opportunities for students to reflect on and articulate their struggles (Finn, 2018).
Finally, the research theme on students' development of academic writing/literacy covers a few under-researched groups, namely, multilingual doctoral candidates (Thurlow et al., 2019), late arrival immigrant students in US schools (Fránquiz & Salinas, 2011), and language minority adolescent students in the US (Kibler, 2010).A novel area of inquiry is the role of creativity in doctoral writing and its potential in helping doctoral candidate take risks and discover their voice in doctoral writing (Thurlow et al., 2019) -creativity as personal, product/outcome, practices/processes, and subject to social, cultural, political, and environmental influences.Another interesting topic concerns EFL students' development of academic literacy as they exercise their agency to negotiate with an ESL writing pedagogy that is imported from the USA (Liu, 2008), and relatedly, the role of racial and language ideology in EFL students' learning of academic writing (Liu & Tannacito, 2013).These classroom-based studies highlight issues of writers' identities, white prestige and native speaker ideologies, and unequal power relations, which are intertwined with students' literacy practices.

Cognitive and affective dimensions of student learning/writing (N = 12)
The second research theme under the students and student learning strand includes several topics, one of which concerns student beliefs, views, perceptions, or attitudes, that is, students' perceptions of non-native English speaking tutors in writing centers (Okuda, 2019), their perceptions of L1 and L2 texts (Abasi, 2012;Liu & Du, 2018), beliefs and goals regarding writing tasks (i.e., mental models) (Nicolás-Conesa et al., 2014), views of content area writing (Kibler, 2011), reactions to teacher feedback (Lee, 2008), and attitudes towards corpus use (Yoon & Hirvela, 2004).These studies have enriched our understanding of the student perspectives on various important issues in L2 writing.For example, trust of non-native Englishspeaking tutors is found to hinge on tutees' preferences for specific tutoring strategies, confidence in tutors' writing, and alignment between tutor and tutee motives (Okuda, 2019).Students' L1 has been found to play an important role in influencing their perceptions of writing in L2, for example, L2 students' interpretations of evidence use in argumentative writing of the target language are influenced by their L1 rhetorical knowledge (Liu & Du, 2018).L2 students' beliefs and goals regarding the writing task (i.e., their mental models of writing), on the other hand, are related to their motivation, self-regulation, and writing performance.In particular, more sophisticated mental models (e.g., representing writing as a problem-solving process) are likely to result in more active student engagement and better writing products (Nicolás-Conesa et al., 2014).Regarding feedback, although students welcome explicit error feedback from teachers, low-proficiency students react less positively to teacher error feedback since it is probably too overwhelming for them (Lee, 2008).Overall, these studies suggest that student cognitive and affective variables can have a crucial role to play in their writing.
Specifically, a few classroom-based studies have addressed student engagement, motivation, and anxiety/stress.The two studies that address student engagement both adopt the tripartite cognitive, behavioral, and affective engagement framework, exploring how students engage with peer feedback (Fan & Xu, 2020) and teacher written corrective feedback (WCF; Han & Hyland, 2015).These studies have enhanced our understanding of this under-explored aspect of feedback, suggesting that L2 students tend to engage with form-focused peer feedback more extensively in the cognitive and behavioral sense than content-focused feedback (Fan & Xu, 2020), and that student engagement with teacher WCF is a complex phenomenon mediated by individual and contextual factors (Han & Hyland, 2015).To enhance student engagement and motivation in the writing classroom, primary L2 learners benefit from a writing program that provides interesting and relevant writing topics as well as genuine audiences (Lo & Hyland, 2007).That said, student motivation may change over time during a writing class, students may experience an increased level of writing anxiety even though they display control of the target genre, and such anxiety can coexist with adequate levels of self-regulation and self-efficacy (Han & Hiver, 2018).The sources of L2 writer anxiety or stress can be attributed to different reasons, and for doctoral students, these could relate to not only challenges arising from text production, but also interpersonal difficulties and intrapersonal conflicts (e.g., lack of perseverance and poor time management skills) (Russell-Pinson & Harris, 2019).

Student writing in the digital space (N = 11)
The ubiquity of technologies has, in recent years, witnessed an increase in classroom-based studies that investigate writing in the digital space.Several studies have investigated online collaborative writing with specific attention to multimodal interactional skills (Abe, 2020), interaction patterns (Li & Kim, 2016;Li & Zhu, 2017), as well as factors that mediate online collaborative writing (Cho, 2017).L2 students are able to adapt their multimodal interactional skills to manage negotiations with peers as they engage in computer-mediated collaborative writing (Abe, 2020).Students demonstrate changes in interaction patterns during wiki-based collaborative writing within the group and across different tasks (Li & Kim, 2016), showing the dynamic nature of small group interactions and the fluid nature of scaffolding across wiki-based collaborative writing tasks.Additionally, group interaction patterns are directly linked to the quality of writing products, with a collective pattern yielding the greatest potential, followed by an expert/novice pattern, in producing high-quality collaborative writing products (Li & Zhu, 2017).Factors that mediate synchronous web-based collaborative writing include communication modes (e.g., text-chat and voice-chat), task representations (i.e., students' understanding of the nature of writing tasks), alignment/misalignment between students' self-perceived roles and other-perceived roles, and students' perceptions of peer feedback (Cho, 2017).Two studies have looked into student collaboration in cross-national digital spaces via email (Liu, 2011) and an online forum (Zhu, 2020).They demonstrated how Taiwanese undergraduates use their textual identities and writing styles to negotiate power differentials with their American peers (in interactions that are characterized by balance, endurance, and resistance) (Liu, 2011), as well as how EFL students use mediational resources (e.g., Microsoft Word, corpus, and Wikipedia) in their digital literacy practices to communicate with American undergraduate peers in a Sino-US telecollaboration project (Zhu, 2020), respectively.
A number of recent classroom-based studies have specifically focused on digital multimodal composing (Jiang, 2018;Shin et al., 2020, Smith et al., 2017).As students engage in multimodal composing, they enhance their awareness of semiotic systems of linguistic and visual resources and develop metalanguage mainly in mode choices and intermodal resources, which facilitate their writing development (Shin et al., 2020).As a relatively new kind of writing, digital multimodal composing has different impacts on students' investment, influenced by factors such as students' different identity positions and sociocultural norms (e.g., value placed on high-stakes testing and mainstream curriculum) (Jiang, 2018).In addition to the use of digital tools and modes, multilingual students can engage in multimodal code-meshing, using their heritage languages in the digital composing process to benefit their writing (Smith et al., 2017).
Writing in the digital environment can benefit both younger and older learners, such as high school and graduate students.An ICT-rich environment can benefit younger, high school students, who learn to appropriate the writing process in an ESL course and transfer their use of the writing process to assignments for other school subjects (Parks et al., 2005).For older learners such as graduate students, through email interactions with the instructor, they acquire rhetorical strategies that enable them to shuttle between formal and informal language according to the rhetorical context (Bloch, 2002).

Student learning from interacting with teachers/tutors/supervisors (N = 8)
A number of studies have addressed learning that takes place in writing center tutorials, advancing our understanding of what actually happens in the learning space afforded by writing centers.Compared with US-educated students, international students devote more time to the discussion of lexical issues with tutors in writing center sessions (Nakamaru, 2010).Affective support from tutors proves to be particularly helpful for international students attending writing center tutorials (Weigle & Nelson, 2004).Tutees benefit most (e.g., engage in sentence-level revisions) when tutors give direct suggestions, mark critical features in student texts, emphasize goals of the session, and when tutees themselves actively participate in the conversation and put their plans down in writing during the tutorial (Williams, 2004).In some cases, the benefits of writing center tutorials may not be immediately evident but only become noticeable after a period of time (Weigle & Nelson, 2004).
Within this research theme, several studies have focused on the graduate context, demonstrating the potential of tutor-tutee interactions in facilitating L2 students' writing development.Tutor-tutee interactions, which provide an optimal site for supporting L2 students' writing development, are shaped by interpersonal, contextual, and institutional factors, suggesting that tutors should use tutoring strategies that are sensitive to tutees' emerging abilities and responsive to their individual needs (Yu, 2020).Dialogic interactions, informed by Bakhtin's dialogism, can help tutees clarify their ideas through verbalizing their thoughts and transferring them to writing; tutees' knowledge can be an important resource for tutors and empower tutees as they work on improving their writing (Merkel, 2020).Drawing on activity theory, multi-directional learning between tutors and tutees allows tutees to learn from tutors, and at the same time enables tutees to bring change to tutors' teaching practices (Fujioka, 2014).These findings are particularly relevant to the graduate context when students may be asked to write occluded genres.Gender may play a role in tutor-tutee interactions in the graduate context too, as in the case of tutoring involving two transnational women, which facilitates the development of authorial voice, expertise, and disciplinary identity in the tutee (Sánchez-Martín & Seloni, 2019).
While one-on-one tutoring (or supervision) is prevalent in the graduate context, which explains why studies about tutor-tutee interactions are mostly conducted with graduate students, child-teacher interaction also plays a significant role in facilitating the literacy development of 5-and 6-year-olds; specifically, synchronicity -dynamic oneness characterized by a caring child-teacher relationship -can account for the bountiful L2 literacy development in children (Blanton, 2002).

Research strand 2: Teachers and teaching of writing (N = 17)
Within the second research strand on teaching and teachers, three themes are identified: pedagogical focuses or approaches in L2 writing (N = 12); teacher learning (N = 3); and teacher beliefs and practices (N = 2).

Pedagogical focuses or approaches in L2 writing (N = 12)
The first research theme has addressed a range of pedagogical focuses in L2 writing classrooms, that is, teaching of coherence (Lee, 2002), teacher conferences (Ewert, 2009), peer review instruction (Min, 2006), literacy narrative instruction (Harman, 2013), genre-based instruction (de Oliveira & Lan, 2014; Worden, 2018), summary-based instruction (McDonough et al., 2014), metaphor-oriented intervention (Wan, 2014), dialogical pedagogy for helping students construct voice (Canagarajah, 2015), use of models in academic writing instruction (Macbeth, 2010), critical reflection and critical language awareness pedagogies (Britton & Leonard, 2020), and multimodal authoring pedagogy (Unsworth & Mills, 2020).Such pedagogical approaches have broadened the field of investigation to address not only how teaching can help students write better, such as through a focus on textual coherence (Lee, 2002) and genres (Harman, 2013;de Oliveira & Lan, 2014;Worden, 2018), but also how explicit instruction can empower learners to take greater responsibility for learning, for example, through conducting effective peer review (Min, 2006).

Teacher learning (N = 3)
The second research theme concerns pre-and in-service writing teachers.A genre-based approach has been found to be productive in helping pre-service teachers enhance their metacognitive awareness, which in turn facilitates their academic writing (Negretti & Kuteeva, 2011); it can also help L2 writing teachers develop their pedagogical content knowledge of unfamiliar genres, which at the same time enhances their sympathy and empathy for the writing challenges their own students face in their academic literacy development (Worden, 2019).Pre-service teachers' learning can be an extended process even when it involves the teaching of simple yet unfamiliar concepts (such as parallelism), which is influenced by various factors including their interactions with course materials and teacher educators, as well as their own values and beliefs regarding the content of teaching (Worden, 2015).

Teacher beliefs and practices (N = 2)
The final research theme has addressed L2 writing teachers' beliefs and practices.Instead of investigating the beliefs and practices of a large number of teachers through questionnaire surveys, the two classroom-based studies under this theme are situated within the teachers' own classrooms -one novice ESL university teacher in the US (Junqeira & Payant, 2015) and two novice secondary teachers in Macao (Yu et al., 2020).In line with the findings of previous research (e.g., Lee, 2009), novice teachers' beliefs regarding feedback are misaligned with their practices; for example, belief in the importance of global issues in written feedback is not reflected in teachers' actual practice dominated by feedback on local issues (Junqeira & Payant, 2015).Instead of viewing the relationship between beliefs and practices simplistically in terms of matches or mismatches, there are complex interactions among beliefs, practices, and contexts (e.g., contextual constraints arising from the curriculum and schools), suggesting that novice teachers' beliefs and practices are best studied in situated contexts (Yu et al., 2020).As such, complexity theory has great potential as an analytical lens to illuminate the complex interactions among teacher beliefs, practices, and contexts (Yu et al., 2020).

Research strand 3: Classroom assessment and feedback (N = 7)
Three themes are identified for the classroom assessment and feedback strand: different aspects of feedback (i.e., teacher, peer, electronic feedback, and error correction) (N = 5); assessment for learning (AfL) (N = 1); and multimodal composing (N = 1).

Different aspects of feedback (N = 5)
With a substantial amount of feedback research conducted in controlled classroom environments, several feedback studies (Ene & Upton, 2018;Ferris et al., 2013;Lee & Schallert, 2008;Yang et al., 2006) have shed important light on what happens when feedback is provided to students in naturalistic classroom contexts.Focused written corrective feedback (WCF) is perceived as useful by students, especially when it is followed by opportunities for revision and one-onone discussion about errors (Ferris et al., 2013).In classrooms where teachers build a trusting relationship with students, they tend to take teacher feedback seriously in their revision (Lee & Schallert, 2008).While research has found that a lot of L2 teachers focus on micro issues in their feedback (Lee, 2008), electronic teacher feedback tends to be content dominated; specifically, the combined use of asynchronous feedback (via Word comments) and synchronous text-based chats allows teachers to put more emphasis on higher-order issues in their feedback (Ene & Upton, 2018).In addition to electronic feedback, peer feedback is likely to engender student revisions at the meaning level (Yang et al., 2006).Although students tend to prefer teacher feedback to peer feedback (Yang et al., 2006), peer feedback still has a significant role to play as it is "associated with a greater degree of student autonomy" (Yang et al., 2006, p. 179).Also, self-feedback generated from data-driven learning (through the use of corpora) is found to assist students' self-correction of errors, with certain error types (e.g., omission, noun number, and agreement errors) being more frequently and accurately corrected with corpus use than others (e.g., verb voice and verb form errors) (Satake, 2020).

Assessment for learning (N = 1)
Assessment for learning (AfL) involves not only peer and self-feedback but also explicit sharing of learning goals and strengthened pre-writing instruction (Lee & Coniam, 2013).Although these strategies sound uncomplicated, the implementation of AfL in L2 writing classrooms is not a straightforward matter since teacher, student, school and systemic issues especially in examination-oriented school contexts in Hong Kong) are found to be restraining factors (Lee & Coniam, 2013), posing impediments to teachers' AfL practices.A case in point is the need for teachers to correct all errors in student writing, which is against the spirit of AfL.

Multimodal composing (N = 1)
An important AfL strategy is to clarify the learning goals of writing, which directly inform the design of assessment criteria for classroom writing.Little, however, is known about how AfL can be designed to reflect learning goals in digital environments, such as the development of multiple-mode competencies in digital multimodal composing.Recent classroom-based research has found that through adopting a process-based model for assessing digital multimodal composing, both teachers and peers can provide formative feedback on rhetorical organization and language use in the drafts of student scripts (Hafner & Ho, 2020), suggesting that the assessment design of digital multimodal composing should draw on both formative and summative strategies, and engage teachers, peers and self at different points in the design process (Hafner & Ho, 2020).
In terms of methodology, 20 (26.7%) out of 75 studies are designed as teacher research, self-study (Abasi, 2012;Britton & Leonard, 2020;Worden, 2019) or action research (Han & Hiver, 2018;Wette, 2010), with the investigations conducted by the teacher-researchers.Other common methodologies are case study research (Spycher, 2007;Zhu, 2020) and ethnography (Harman, 2013;Kibler, 2011).Myriads of data collection methods are used (primarily qualitative), notably student texts/textual data, interviews/focus groups, stimulated recall, think-aloud protocols, journals, written reflections, questionnaires, classroom observations, group discussions, and recording of writing sessions/interactions.To enhance the trustworthiness of findings, most classroom-based studies rely on several data sources with data triangulation performed.

Pedagogical implications
A number of pedagogical implications can be drawn from the review of the 75 studies published in JSLW from 2001-2020, which are examined according to the three research strands that guide the review.

Students and student learning of writing
To facilitate L2 students' development of academic literacy, it is important that teachers help students cope with source texts to facilitate effective citation practices (Li & Casanave, 2012;Wette, 2017a).Teachers can also make use of mind maps (Wette, 2017b) and genre-based/analysis tasks (Yasuda, 2011(Yasuda, , 2015) ) to foster students' genre knowledge and genre awareness.Student collaboration can be exploited at different stages of the writing process; for instance, prewriting collaboration through face-to-face discussion or online text chat (Liao, 2018;Neumann & McDonough, 2015), collaborative writing (during writing) (Storch, 2005;Yang, 2014), and group writing conferences (after writing) (Mochizuki, 2019).Further, teachers can pay greater attention to issues of student voice (Thurlow et al., 2019), identities, agency (Liu, 2008;Liu & Tannacito, 2013), as well as the obstacles students may face in their academic literacy development (Finn, 2018;Spycher, 2007) so that they can better design writing instruction to empower students and cater to their needs.
Moreover, it is crucial that L2 writing teachers bring to their teaching a good understanding of their students ' beliefs, views, attitudes, perceptions, motivations, anxieties, etc. (Abasi, 2012;Lee, 2008;Okuda, 2019;Yoon & Hirvela, 2004), since these cognitive and affective dimensions have direct bearing on student learning of writing.With a better understanding of different student variables, teachers are more likely to provide motivational writing instruction and assessment/feedback that enhance student engagement and self-efficacy (Han & Hiver, 2018;Han & Hyland, 2015).
The development of digital literacy has taken on new importance in the 21st century, as shown in the increase of published research on digital writing in the past few years.Drawing from the findings of this review, teachers could provide opportunities for digital writing, encourage online collaboration (such as pre-writing online chats and online collaborative writing), and equip students with multimodal resources to develop their digital literacy.They can provide explicit instruction in digital interactional skills to maximize learning outcomes, and explore opportunities for students to engage in cross-national or telecollaboration projects that involve them in writing collaboratively with peers from other regions in digital spaces.
Research on tutor-tutee interactions has important implications not only for instructors based in writing centers, but also for teachers interested in using one-on-one conferences to support student learning of writing.In addition to affective support of the tutors and the use of learner-centered tutoring strategies, as demonstrated in this review, it is important to engage students in dialogic interaction, and empower them to take responsibility for learning in one-one-one interactions by, for example, making them set goals and monitor them, and for tutors to use strategies that cater to individual student needs and abilities.

Teachers and teaching of writing
As revealed in the findings, L2 writing teachers can explore a host of instructional or pedagogical approaches to promote L2 students' writing development, such as coherence instruction (Lee, 2002), summary writing instruction (McDonough et al., 2014), and metaphor-oriented instruction (Wan, 2014).Specifically, genre-based pedagogy (de Oliveira & Lan, 2014;Worden, 2015) has considerable potential for learners at different educational levels and contexts, including pre-and in-service teachers.For pre-and in-service teachers, it is important that they engage in ongoing professional development to enhance their own knowledge of writing (e.g., genre knowledge) and develop ability to critically evaluate and reflect on teaching content so as to enhance their professional competence as teachers of writing (Worden, 2015(Worden, , 2019)).Additionally, teachers can be made aware of their own beliefs about writing and the teaching of writing, how their beliefs may diverge from their practices, and what they can do to foster greater alignment between beliefs and practices.

Classroom assessment and feedback
Feedback is a central component of classroom writing assessment.Aside from conventional written teacher feedback, teachers can explore alternative modes of feedback to maximize student learning, such as synchronous and asynchronous electronic feedback and other kinds of technology supported feedback, which is found to generate more feedback on content and higher-order issues (Ene & Upton, 2018).Given that peer feedback can facilitate learner autonomy, it should be used in conjunction with teacher feedback (Yang et al., 2006).Teachers can incorporate data-driven learning to empower students to play a more active role in self-editing (Satake, 2020).Above all, AfL has a pivotal role to play in classroom assessment (Lee & Coniam, 2013); despite the challenges, teachers should explore ways to integrate it into writing classrooms, including classrooms that practice digital multimodal composing (Hafner & Ho, 2020).

Recommendations for future research
Future classroom-based research on L2 writing can continue to address the unanswered questions that emanate from previous research and explore new frontiers to push forward classroom-based writing research.This concluding section examines research directions by examining some possible topics/areas for further investigation, as well as methodological and theoretical perspectives that can be considered for future research.

Research areas/topics for further research
The majority of studies reviewed in this paper have focused on students and their learning of writing.It is envisaged that students' development of academic writing and academic literacy is a topic that will continue to attract considerable research attention.While the bulk of such research has focused on older learners in higher education contexts, future studies can address underrepresented groups like young and adult L2 learners with emergent literacy.
Classroom-based research on different pedagogical focuses or strategies adopted by L2 writing teachers will continue to be of interest to researchers.Apart from conventional focuses on form, structure, and language features as shown in the review study, it is useful to approach this research area from the perspective of the 21st century skills that L2 writers need, such as critical thinking and digital literacy skills, to ascertain how these 21st century skills could be fostered in L2 writing classrooms.
This review has found a number of classroom-based studies that address L2 writing teachers, both pre-service and in-service; however, attention to L2 writing teacher educators is non-existent.Recent L2 teacher education literature has expanded to include a focus on language teacher educators, such as who they are (issue of identity), what they do (their teacher education practice), and how they develop professionally (Barkhuizen, 2021;Yuan & Lee, 2021).In the realm of L2 writing, a lot more classroom-based research can be conducted to better understand the identity, work and lives of L2 writing teacher educators.
Feedback as a component of classroom writing assessment is of perennial interest to researchers.In this review, there is only one classroom-based study focusing on WCF.As suggested by Atkinson and Tardy (2018), research on WCF has to be conducted in situ, with more attention paid to WCF as one component of feedback and the broader pedagogical context of the classroom (see Lee et al., 2021).We are entering the post-COVID era.How technology can be further leveraged in classroom writing assessment and feedback is a promising area for further classroom-based research, for example, data driven learning, electronic feedback, and technology enhanced feedback in various forms (such as screencast feedback).With a rising number of studies on digital writing (13 in total from 2016 to 2020) in recent years, it is envisaged that digital multimodal writing will continue to gain currency.More classroom-based research could investigate how assessment for and as learning, and formative feedback strategies can be applied to writing in digital spaces to promote student learning.

Methodological and theoretical perspectives in future research
In terms of research methodology, less than 30% of the studies reviewed in this paper have been conducted by teachers themselves, with the majority undertaken by external researchers.The emic perspective can generate rich insight that benefits teaching, learning and assessment in the writing classroom.As such, L2 writing teachers can be encouraged and empowered to research their own practice by undertaking action research, teacher research, self-study or autoethnography.L2 writing teacher educators' self-study can also provide an interesting methodological approach that generates new knowledge about the work and lives of writing teacher educators.This review shows that mixed-methods research has been under-utilized; future classroom-based research could use mixed methods to investigate teaching, learning, and assessment in the L2 writing classroom.
Given that classroom-based research captures what happens in the classroom, theoretical perspectives that draw on sociocultural theory, ecological systems theory, and complexity theory have great potential in L2 writing contexts.For example, they could be employed for exploring the interaction between student and context variables in the teaching and learning of L2 writing, and uncovering the complexity of feedback practices in situ -so as to yield research findings of greater pedagogical relevance.The utility of these theoretical perspectives remains to be more thoroughly investigated in future classroom-based research on L2 writing.

Conclusion
To conclude, this state-of-the-art review of developments in classroom-based research on L2 writing, despite its limitation (covering articles in JSLW only), contributes to a nuanced understanding of salient issues about learning, teaching, and assessment of L2 writing in naturalistic classroom contexts.It is hoped that the pedagogical implications and recommendations for future research can provide useful information for L2 writing practitioners and researchers, respectively, to guide their practice.