The role of individual differences in the development and transfer of writing strategies between foreign and first language classrooms

Abstract While the importance of considering the wide variation among language learners has been brought to the forefront in recent years, the impact of such individual differences on the process of second or foreign language writing has been largely neglected. This paper aims to explore the ways in which individual students develop and transfer strategies within and between foreign language (FL) and first language (L1) writing. A two-phase intervention of strategy-based instruction was conducted primarily in the FL German classroom, and later also in the L1 English classroom of a Year 9 (age 13–14) class in a secondary school in England. This paper draws on in-depth qualitative data from writing tasks and stimulated recall interviews. A range of students’ trajectories through the intervention were evaluated and four distinct writer ‘profiles’ were identified: the strategic writer, the experimenter, the struggling writer and the multilingual writer. Both the development and transfer of strategies for these students were shown to be influenced by a complex and dynamic range of factors such as the learner’s proficiency level, their level of metacognitive engagement with the task, their attitude towards writing and their strategic use of other languages in their repertoire.


Introduction
While there is evidence to suggest that strategy use can be of great benefit to language learners (e.g. Macaro 2001;Cohen 2011;Plonsky 2011;Oxford 2017), it is also important to consider the role of learners' individual differences on their development and cross-linguistic transfer of strategies. Much research in the fields of second language acquisition and foreign language education has focused on investigating general principles and developing widely applicable pedagogical theories; as such, there has been a tendency to think in terms of the collective rather than the individual (Dörnyei 2005). While the importance of considering the wide variation among learners has been brought to the forefront in the field of second language acquisition more generally by Dörnyei and Skehan (2003) and Dörnyei and Ryan (2015), the impact of such individual differences on the process of second language writing KEYWORDS Learning strategies; individual differences; transfer; writing has been largely neglected (Kormos 2012). The aim of this paper is therefore to critically evaluate a range of different students' trajectories through an intervention of strategy-based instruction in order to explore the role of key factors, such as their level of proficiency, their attitude towards writing and their knowledge of other languages, in negotiating their process of strategy development and their ability to transfer strategies from one language context to another.

Development of writing strategies
This study is situated within the wider theoretical framework of language learning strategies. Language learning strategies are generally considered as a means of ensuring that language is stored, retained and able to be produced when necessary; that is, they affect learning directly. They are 'optional' (Bialystok 1978, 69), 'consciously selected by the learners' (Cohen 1998, 4) and the aim of learning strategies according to O'Malley and Chamot (1990) is to 'enhance comprehension, learning or retention of new information' (1). Early studies into learning strategies within the second language context sought to identify and classify strategies used in the classroom according to their function or skill area (Naimen et al. 1975;Rubin 1975;O'Malley and Chamot 1990;Oxford 1990).
The current study focuses particularly on developing strategies with a metacognitive function, which 'involve thinking about the learning process, planning for learning, monitoring of comprehension or production while it is taking place, and self-evaluation after the learning activity has been completed' (O'Malley and Chamot 1990). This focus is due to evidence of a correlation between learners' success and the use of metacognitive strategies (Cohen 2011;Griffiths 2013), and also the importance of metacognitive strategies in maintaining strategy use over time and transferring strategies to new tasks (O'Malley and Chamot 1990).
Yet, it is important to recognise that the study of writing strategies in particular is also situated 'within a wider research movement known as 'process writing' , which emerged in the field of native language composition research with the aim of gaining insights into the mental actions writers engage in while composing' (Manchón, de Larios, and Murphy 2007, 229). Process writing emerged in the 1980s and was pioneered by Hayes and Flower (1980) with their Cognitive Process Model, where composing is viewed as a writer-centric, goal-oriented, problem-solving task. However, although Hayes and Flower set the precedent for a more strategic approach to writing instruction, they have been criticised for attempting to describe features common to all writers in a single model (Grabe and Kaplan 1996;Hyland 2002). Building on their work within the context of first language (L1) writing, Bereiter and Scardamalia (1987) later posited the need for not one, but two process models in order to account for differences between 'immature' and 'mature' writers. They identify these as the knowledge-telling and the knowledge-transforming models, respectively. Such models mark the beginning of differences in proficiency among learners being taken into account when investigating strategy use in writing.
Such a trend was also reflected in second language (L2) studies, which similarly focused on exploring differences in proficiency by identifying the writing strategies used by skilled and less skilled writers (Zamel 1983;Raimes 1987;Cumming 1989). Such studies found that skilled L2 writers spent more time planning, revising and editing at the discourse level, while less skilled writers tended to view writing as a series of words and sentences rather than as a global text. However, when comparing the same writers using both L1 and L2, when writing in the L2 the learners tended to plan less (Albrechtsen 1997), generate fewer goals (Skibniewski 1988) and spend more time dealing with formulation problems (de Larios, Manchón, and Murphy 2006).
However, from the mid-1990s, this approach received criticism for neglecting the sociocultural context and more emic perspectives of writing. It was therefore acknowledged that writing should be considered as 'a sociocognitive activity which involves skills in planning and drafting as well as knowledge of language, contexts, and audiences' (Hyland 2002, 23). This sociocognitive perspective on writing in turn influenced pedagogy and research in both L1 and L2 contexts. It resulted in increasing attention being paid to the role of individual differences in writing and strategy use beyond proficiency, such as the influence of learning styles, learner beliefs, motivation and language(s) spoken. Relationships were found, for example, between the use of strategies and learners' self-efficacy beliefs (Yang 1999), and it was recognised that L2 writers also have the unique opportunity to use the 'resources of both first and second languages together for strategic purposes while composing' (Cumming 2001, 6). A methodological shift also occurred, moving from a focus on a writer's individual cognitive processes by means of verbal protocols in artificial conditions, to a consideration of writing in an authentic social environment such as a classroom, which is reflected in the current study.
As outlined above, most of the research that has been conducted in relation to writing strategies has been primarily concerned with the identification, description and comparison of strategies used by L1 and L2 writers, and in particular the difference in strategies used by 'good' and 'poor' learners. However, as Grenfell and Harris (1999) state, 'describing the end product is not the same as prescribing the means to get there' (38), and it is therefore surprising that the influence of instruction on strategy use is a comparatively under-researched area (Chamot 2005). Some studies have been conducted into the influence of an intervention of strategy-based instruction (SBI) in either an L1 context (e.g. De La Paz and Graham 2002;Graham andHarris 2003), or an L2 context (e.g. Sengupta 2000;Sasaki 2000;Macaro 2001;De Silva and Graham 2015), and on the whole have shown positive effects on the performance of those who received writing strategy instruction. However, while some considered the effect of proficiency level among groups of learners (Sasaki 2000;De Silva and Graham 2015) and found that gains were greater among the lower attaining groups, on the whole the effect of other individual differences in negotiating the process of strategy development has been largely neglected.

Cross-linguistic transfer of strategies
As outlined above, the vast majority of research into the development of writing strategies has taken place within a single context of either L1 or L2 education and has had a tendency to disregard the potential interactions between the two. In addition, it is almost taken for granted within the literature that any reference to transfer between these two contexts implies the one-way transfer of pre-existing skills and strategies from the L1 to the foreign language (FL). It has also been suggested that it may not be necessary to teach strategies at all in the FL classroom, as learners will already have developed strategies from their learning of their L1, and can simply transfer these to the FL (Kellerman 1991). However, as explored above, many writers may be considered less skilled or 'novice' writers even in their native language (Bereiter and Scardamalia 1987) and therefore may not have developed effective strategies they can transfer to a new language.
The focus of the current study therefore emerged from a hypothesis that the reverse may be equally valid; if writing strategies are explicitly developed within the FL classroom, then it seems logical that this knowledge could not only benefit FL writing tasks, but may also positively affect L1 writing. One of the few studies on the influence of learning a FL (English, French or Russian) on the L1 (Hungarian) conducted by Kecskes and Papp (2000) provides evidence of such bidirectionality of transfer, but suggests that a threshold of proficiency in the FL must be reached in order for the FL to affect the L1. However, there is a lack of research into how individual learners transfer strategies between different language contexts and the complex factors which may hinder or facilitate such transfer. This in turn will have implications for approaches to pedagogy in both first and foreign language classrooms. The current study therefore explores the following research question: • How do students' individual differences influence the way in which they develop and transfer strategies within and between foreign language and first language writing?

Methodology
A sociocognitive perspective was adopted in relation to the design of the current quasi-experimental study which was conducted within an authentic classroom context. A two-phase intervention of strategy-based instruction was implemented primarily in the FL German classroom, and later also in the English classroom of a Year 9 (age 13-14) mixed ability class of 22 students in a secondary school in England. The majority of the students were in their third year of learning German and fourth or fifth year of learning French at the time the study was conducted. While the broader study from which this paper is drawn also considered trends at a whole-class level and compared the outcomes of the experimental group with a control group using a mixed methods approach (see Forbes 2016), this paper will focus on the qualitative findings from a smaller group of students from the experimental group. This section will firstly outline the overall research design of the study and then describe the process through which the four writer profiles discussed in this paper were identified.

Research design
At the beginning of the study, the students were asked to complete a narrative-style writing task on a specially designed writing task sheet in English, German and French in order to explore their general approach to writing in each subject. Topics included, for example, a piece of travel writing in English and an introductory email to an exchange partner in German. This information was used to develop an intervention of strategy-based instruction to be carried out in the German classroom of the experimental group over the course of four months (Phase A). The SBI was designed to be integrated into the existing scheme of work and involved the introduction of activities to aid the planning, monitoring and evaluation of written work and to encourage students to reflect on and assess their personal learning strategies and approaches.
Examples of activities include: the introduction of guided planning sheets which encouraged students to set goals and to consider the main content ideas and language features they would include in their text; the creation of co-constructed checklists for evaluating work and setting individual targets for improvement; and error correction activities to help students to independently identify errors and self-correct their writing. The strategy-based instruction was carried out by the class teacher with resources developed in collaboration with the author. The teacher started by raising awareness of the strategies learners were already using and then presented and modelled each of the new strategies. The teacher then provided opportunities for students to practise and evaluate them in different contexts and gradually removed scaffolding and explicit instruction over time.
Next, a second set of tasks were set in order to investigate whether or not this explicit focus on metacognitive strategy use in German lessons affected students' strategy use in German, and also whether any of the students transferred these strategies to their French or English writing tasks without any explicit encouragement to do so. During Phase B of the intervention, which also took place over the course of four months, the SBI continued in the German classroom, however was also explicitly reinforced in the English classroom of the experimental group. Similar activities and resources were used by both teachers in order to encourage learners to transfer the skills and strategies they developed in one curriculum area to another. A third and final set of tasks were set in order to explore any further changes which took place as a result of this additional explicit strategy instruction in English, as shown in Figure 1.

Data collection methods
This paper focuses on the qualitative data which were collected from the writing strategy task sheets and stimulated recall interviews. This allowed for a more in-depth exploration of the role of individual differences in the development and transfer of students' strategy use between the different language contexts.
The writing strategy task sheets, based on a method used by Macaro (2001), were designed to capture students' pre-task planning, the use of resources and problem-solving strategies while writing (by asking students to use underlining and the notes section of the margin), and included some post-task questions about the evaluation process, such as whether they checked over their work and what they checked for. As the task sheets required students to report their strategy use relating to a specific and contemporaneous task, it was felt that this would provide a more accurate reflection of their behaviour than a more general questionnaire. The completed task sheets then acted as a stimulus for the stimulated recall interviews. They provided a concrete experience for students to reflect on and further explain their rationale behind the strategies they used at the time. Stimulated recall interviews were also considered to be less disruptive and cognitively demanding than think aloud protocols, a particularly important consideration when working with younger participants. In order to minimise any 'contamination of memory' (Gass and Mackey 2000) and increase the reliability of data, the stimulated recall interviews were conducted by the author within several hours of completion of the writing task.

Analysis
Following the data collection process, all of the writing tasks were analysed in order to explore whether the students shared any similar characteristics with regard to their strategy use and proficiency in writing. These were typed verbatim into word processing software and imported into NVivo software for analysis. The coding scheme for the writing tasks emerged from the data itself and was organised into five categories in order to capture a range of strategies used by students: • Planning -this included 20 sub-codes such as the inclusion of content items, language features, goal-setting and the language of planning. • Monitoring -this included 11 sub-codes such as the use of a dictionary, asking the teacher or a peer for help and reformulation of a sentence. • Errors and error correction -this included 36 sub-codes to identify particular errors in spelling, grammar, vocabulary and punctuation which had been left uncorrected and also those which had been self-corrected. • Evaluation -this included 24 sub-codes relating to what the student indicated that they checked and/or changed during the evaluation process. • Post-task thoughts -this included 16 sub-codes relating to what the students reported finding easy or difficult about the task and their self-assessment of the outcome.
Following this, an individual summary of the strategies coded per task was created for each student in order to gain an overview of their patterns of strategy use in each language over time.
The stimulated recall interviews were then analysed using a thematic coding approach (Robson and McCartan 2016). An initial set of codes was generated in line with the writing task coding scheme above and incorporated references to strategies within the key areas of planning, monitoring and evaluation relating to the particular writing task in question.
Other key codes were then added where students had made more general comments about their strategy use. This included comparisons between their approaches to writing and attitudes towards writing in different languages, comments which related to strategy use and achievement and, crucially for this study, comments relating to a change in their approach over time and the transfer of strategies from one context to another. The codes were created in such a way as to ensure that links could easily be made between the writing task data and the interview data.

Emergence of the four writer profiles
Following analysis of the above data at a whole-class level, it became evident that some of the students shared similar characteristics in terms of their strategy development and use and a range of four writer 'profiles' emerged; I refer to these as 'the strategic writer' , 'the experimenter' , 'the struggling writer' and 'the multilingual writer' . Table 1 provides an overview of the general characteristics of each profile along with the number of students associated with each. It is acknowledged that not all students neatly fitted into one of these four categories and that some may have shifted over time. However, based on the evidence from the writing tasks and interviews, Table 1 provides an indication of the general characteristics arising from the data which were broadly shared by each group. The following section explores each of these profiles in turn with in-depth reference to one case study student as a typical example for each.

The strategic writer (Carissa)
The first type of writer which will be examined is referred to as the 'strategic' writer. Such writers emerged as being relatively effective strategy users from the beginning. A typical example of one such writer is Carissa, a native English-speaking student. She was a conscientious and consistently high achieving student across all of her subjects in school and she was on the school's list of 'able, gifted and talented' students in languages. In the writing tasks she improved consistently over time in both FLs and maintained her high scores in English throughout. Carissa commented that she quite enjoyed writing in all three languages, although reported feeling more confident with writing in English. The following key characteristics emerged:

Used a range of strategies effectively from the beginning
Carissa was among the minority of students who engaged in planning in all three languages from the beginning. She took a relatively similar approach to each by focusing on the main content items she would include. In addition, the number of errors she made in each task was well below the class average, and also decreased over the year, suggesting that her monitoring and evaluation strategies were becoming more effective. Such close monitoring of her work was evident in her frequent reformulation and rephrasing of sentences. For example, in the first English task, she changed the phrase 'the wondrous display' to 'the stunning display' because she thought it sounded better. She also demonstrated evidence of being an independent learner, and was much more likely to look a word up in the dictionary or check her notes than to ask for help, even at the beginning.

Evidence of the ability to transfer strategies independently from one language context to another, both from L1-FL and FL-L1
At the beginning of the year, Carissa was very capable of writing in all subjects; however, she seemed to conceptualise L1 and FL writing in distinct ways: I think I see them as quite different, cause when I'm writing in, yeah, like when I'm writing in French and German like, I just feel like I'm being marked on like, like, spelling and like words Table 1. General characteristics of the four writer profiles.

Writer profile
Number of students General characteristics arising from the data the strategic writer 7 • consistently achieved well in both english and foreign language writing tasks; • used a range of writing strategies effectively; • demonstrated a self-awareness of their strategy use. the experimenter 7 • Generally average achieving students in both english and foreign language writing tasks; • showed evidence of experimenting with their strategy use over the course of the year and a willingness to try different approaches.
the struggling writer 6 • consistently lower achieving students in both english and foreign language writing tasks; • did not particularly enjoy writing and generally found it difficult; • Made use of a range of strategies, however not necessarily effectively.
the multilingual writer 2 • spoke a first language other than english; • showed evidence of using their mother tongue strategically when writing in a variety of ways.
and stuff, but with English it is more about what you're writing, so I do think they're quite different and I treat them differently.
Yet at the end of the year, although she still thought about them differently, she acknowledged that: 'I think that the way that I approach them is probably more similar, like, I do more planning with French and German and stuff ' . When asked further about the development of her writing, Carissa highlighted the influence of her L1 strategies on her FL writing, suggesting that: 'things that I've done in English have helped in French and German' . This was particularly evident in her approach to the first set of tasks, where she used similar strategies across all three subjects, especially in relation to her approach to planning. However, she also mentioned in the final interview that she felt the transfer worked 'both ways' and that some elements practiced more in the FL classroom also helped her in English. This was particularly evident in her approaches to evaluation and error correction.

Demonstrated a self-awareness and self-evaluation of strategy use
It is worth noting that Carissa's perceptions of her strategy use throughout seemed to align very closely with what she actually did; for example, not only did she say she checked over particular aspects of her work more thoroughly, but it is evident from the reduction in errors and from the type of corrections she made that she was in fact doing so. Such an astute sense of awareness regarding her strategy use is another characteristic feature of the 'strategic' writer.

Continuous development and refinement of strategy use over time
There was no evidence that the particular strategies she used changed significantly over the course of the year, but from the decrease in errors made over time and the increase in marks in the FL tasks, it seems as though she was gradually using these strategies in a more focused, reflective, and effective way. For example, in relation to her English planning Carissa commented in the first interview that she had recently changed her approach. While for Task 1 she planned key content items and structure in bullet point form, she said she previously used to complete full drafts of her writing 'and then like, go through it and change it'; these changes generally consisted of 'little things' like spelling, grammar and rephrasing sentences to make them sound better. However, she felt that this approach was very time consuming and therefore decided to change to be 'more organised' . Such a change in approach was entirely self-initiated and occurred before the start of the study, suggesting not only that she was aware of the strategies she used, but also that she was constantly evaluating their effectiveness and making deliberate changes when necessary.

The experimenter (Chris)
The second type of writer to emerge is referred to as the 'experimenter' . These writers demonstrated a willingness to try new strategies introduced during the intervention in order to improve their performance. An example of such a writer is Chris, a native English-speaking student. He generally scored around the class average, or just below average, in tests and exams in German, French and English. Over the course of the year, his marks improved consistently in both German and English. He enjoyed creative writing tasks in English, but experienced frustration when writing in German and French because 'you don't know how to say everything' . As a result, his approach to writing in both contexts was very different at the beginning of the study. The following key characteristics emerged:

Displayed a willingness to experiment with and try out new strategies and approaches
While at the beginning Chris used only a narrow range of strategies, particularly in German and French, over time he experimented with different approaches. For example, while in the first set of tasks he did not engage in any written planning at all for German or French, in the second tasks he drafted his ideas in the L1 for the French task and for German he used the target language to draft some sentences and identify four key content items to include. Similarly, in English, he progressed from only planning some bullet point content items in the first task, to also considering structure and language features (such as the use of the past tense for a diary entry) in subsequent tasks. He also became increasingly willing to try a range of strategies to solve a problem himself, and in the final tasks he used a dictionary or his notes for help instead of asking the teacher. He was willing to try the majority of strategies presented during the intervention in order to help him to 'get a better mark at the end' .

Strategy use developed most following the periods of explicit SBI with the aid of scaffolding and opportunities to practise
The most notable improvements in Chris's writing took place in German following Phase A of the intervention and in English following Phase B. Yet, in comparison to Carissa, he seemed to rely more heavily on the explicit instruction and scaffolding, and commented that he found the resources used in class helpful 'to kind of prompt me' . This is also evident by the fact that he experimented considerably less in his French writing, perhaps due to the absence of explicit SBI in this context.

Some evidence of an ability to transfer strategies
However, there was some evidence that he was beginning to transfer some strategies between subjects without being deliberately encouraged to do so, particularly in relation to the quality of his planning and evaluation strategies. In the interview following the second set of tasks, Chris hinted at this implicit transfer by stating that: Well, I didn't used to do much planning but I do some now, and I didn't used to check it over in French and German but now I've started checking it over and it's started to work a lot more cause I'm getting better marks, so if I've done something and my marks have improved then I' d start to do that more.
[…] Then I was trying some of the things out in English as well, sort of reading through afterwards to check, make sure my spellings are OK and stuff like that.

Often judged the success of his strategy use according to marks given by the teacher rather than independently
The above quotation also highlights that Chris perhaps had not yet developed the ability to self-evaluate his strategy use, and while he was willing to experiment and try different things, he relied on the marks given by his teacher to decide whether or not the strategies he had used were successful, and by extension, whether or not he would continue to use them or try something else. This was also evident in his approach to checking over his work. In the second German task, for example, he checked over his work more carefully as he went along, rather than waiting until the end to check for superficial errors. This corresponded to a considerable reduction in errors and similarly led him to attach increasing importance to the evaluation process and to claim that 'once I've checked it over then usually I get a better mark than when I haven't checked it over' . This reiterates the suggestion that he largely measured the success of his strategy use by the marks given by his teacher.

The struggling writer (Zoe)
The next type of writer to emerge was the 'struggling' writer. Such writers tended to be lower performing students, however, that is not to say that these students did not engage in strategy use; they often made use of a range of writing strategies, yet unlike the first two types of writer discussed above they did not necessarily do so effectively. A typical example of such a writer is Zoe, a native English-speaking student. She consistently scored below average in class tests and exams in German, French and English, but her results were particularly low in German. She described herself as 'not very good at languages' and her results in the writing tasks remained low throughout. She felt that writing across all three subjects had 'got harder' since the beginning of secondary school and did not particularly enjoy it. Such views remained consistent throughout the year. The following key characteristics emerged:

Made use of some strategies, however not necessarily effectively, particularly in the FLs
For example, in the final German task, Zoe engaged in some planning of content items (four items written in English) and wrote down 12 translations of vocabulary items, however fewer than half of these words were integrated into the final text, and a third were either incorrect or misspelt. In addition, in the first set of tasks, she made no attempts to use a dictionary or to solve problems independently, and simply asked the teacher for help; however, in later tasks she tried to use a wider range of problem solving strategies. In the second German task, for example, she did not know how to say 'I walk my dog' , so decided to use a dictionary. Though she then commented that she did not know which of the translations to use and gave up. This example shows that although Zoe was attempting to use resources to solve problems more independently in German, she did not always fully understand how to do so. It seems as though she lacked the confidence, and perhaps also the competence, to engage in strategy use effectively, and that further time and practice were necessary to develop her skills.

Effective strategy use was slow to develop and required considerable scaffolding, guidance and practice
In spite of her ineffective strategy use in the FLs outlined above, Zoe did show some evidence of developing her planning in English in the final task by creating a 'checklist' of key points and integrating them all into the final text. She also succeeded in self-correcting a higher proportion of errors in this task with the help of prompt sheets, however this only emerged in English at the very end of the year and did not yet result in an improvement in marks.

Less willing to experiment and try out new strategies and approaches
Although there were some parallels between Zoe and Chris at the beginning, especially in terms of their lack of FL planning, unlike Chris, Zoe did not seem to be willing to experiment with strategies or to deliberately change her approach to writing. She used a narrower range of strategies throughout than the majority of her peers and there was less variation in her approach. Following the final English task when she was asked whether she had tried out any of the other strategies introduced in class, she responded that she hadn't and didn't think that they would be useful, 'because I've never done it before and I don't, I just don't think it would help me' .

Lack of awareness of strategy use, which often led to a discrepancy between what she thought she did and what she actually did
It is interesting to note that, even though Zoe did take on board some new strategies, as suggested above, she often did not seem to be aware that she was doing anything differently. In the second English task, for example, her planning consisted of a draft of around half of the final text. She said she took this approach because she had 'always done it like that … I do it rough and then change a few bits and do it neat' . Yet, it was evident from her previous task and class work that this was not the case. Similarly, the justification that she had 'always done it like that' reiterates her lack of willingness to change her approach. In addition, in the final interview following the third German task, she commented that she just did 'the same sort of thing' as always, and lacked the awareness and deliberateness displayed by Carissa throughout and by Chris towards the end.

The multilingual writer (Mei)
The final type of writer which emerged is the multilingual writer. These were students who spoke English as an additional language (EAL) and therefore were often approaching writing in English from a different perspective to their native speaker peers. Obviously these students were not solely defined by their EAL status and shared some of the same characteristics of the other profiles discussed above, however this section aims to specifically consider how having an additional language at their disposal affected their interaction with and negotiation of strategies. One example of such a student is Mei, a native Mandarin speaker who came to the UK from China with her mother six years prior to the beginning of the study. Although she was fluent in English at the time of the study, Mandarin was still the primary language spoken at home. She was considered to be a middle-low performing student and generally scored around the class average in English tasks and below average in German and French tasks. The following key characteristics emerged:

Her overall conceptualisations of writing in English, German and French were more similar than many of her native English-speaking peers
Mei seemed to position herself primarily as a Chinese speaker learning English, German and French, which led her to conceptualise the nature of writing in these three languages as being 'quite similar' , unlike the majority of her peers who viewed writing in English and the FLs as being very different.

Strategic use of the mother tongue as an additional resource in language learning
For Mei, her native language was omnipresent and largely constituted her primary language of thought when writing in any subject. As a result, she became very adept at translating from and through Chinese; when writing in German she said that her thought processes went from 'Chinese to English to German... cause I think in Chinese before I go into English, I have to start there, cause it's, Chinese is my starting point for a language' . One of her FL planning strategies, for example, was to consider first of all what she knew in the target language, then to 'simplify that in English to make it easier to put in German' . She also frequently used backtranslating as an evaluation strategy by translating her FL text back into English in her head to make sure she hadn't missed anything. Interestingly, when asked if she found it easier to think in German via either Chinese or English, she commented that 'it's kind of the same, because my German skill level is the same' . It seems therefore that her high level of proficiency in both Chinese and English allowed her to use both as a resource for accessing and learning a new language.

The key problems and concerns encountered while writing tended to be more crosslinguistic in nature
Mei reported experiencing some of the same problems across all three languages, and had a prevailing concern that her writing wouldn't 'make sense' due to her continuous translation via Chinese. She engaged a lot in monitoring her work while writing, to a greater extent than the majority of her peers. This, at least partly, seems to have been driven by her prevailing concern with 'making sense' in her writing, especially given that she was operating through multiple languages. As she wrote, she frequently rephrased sentences or substituted words for synonyms, not because what she had written initially was incorrect, but because she felt the alternative would make more sense. For example, in the final English task she changed 'they agree' to 'they go along with it' as she felt it sounded better. In addition to reformulating her work as she was writing, Mei also engaged in problem-solving using a range of strategies, such as using a dictionary or reading her work aloud. Interestingly, the most common type of problems she encountered seemed to be cross-linguistic in nature and were often related to either vocabulary or punctuation. She also seemed to take a similar approach to solving these problems across all of the subjects. When asked about whether she often made connections or comparisons between the different languages, she replied: 'I think about it sometimes, but then sometimes I just do it naturally, without even knowing it' , suggesting perhaps that being a multilingual writer has helped her to more easily transfer strategies across and between various languages.

Discussion
The previous section explored the varied trajectories through the intervention of SBI of four students in the experimental group, who each represented one of the distinct writer 'profiles' which emerged from this study. However, both the development and cross-linguistic transfer of strategy use were shown to be influenced by a complex and dynamic range of factors relating to the role of the learners' individual differences, such as the learner's proficiency level, their level of metacognitive engagement with the task, their attitude towards writing and their strategic use of other languages.

The symbiotic relationship between proficiency and strategy use
The relationship between strategy use and achievement is a complex and mutually interdependent one: just as the development of strategies can positively influence achievement, learners' proficiency level can also impact the extent to which they are able to develop and transfer strategies in different contexts. As such, it is important to acknowledge the wide range in individual variation in proficiency among the learners in the current study. This is a key factor which has been taken into consideration in a number of existing studies into writing processes and strategies (e.g. Sasaki 2000; De Silva and Graham 2015). There was evidence that Carissa, who was identified as a high achieving writer across all three languages, used strategies more effectively and deliberately than many of her peers, even from the beginning of the study. She shared many characteristics with Bereiter and Scardamalia's (1987) knowledge-transforming model and with skilled writers from other studies, such as engaging in more planning (Raimes 1987;Cumming 1989;Sasaki 2000), using more complex strategies (Bereiter and Scardamalia 1987) and engaging in more editing and revision (Skibniewski 1990;Myhill 2009). Her approach to writing did not change dramatically over the course of the SBI; however, she was engaged in continuous refinement of her strategy use and was aware of the occurrence of transfer between subjects. She succeeded in maintaining her high level of attainment in English, while improving slightly in both German and French over the course of the year. This provides evidence in support of the correlation between strategy use and achievement which has similarly been acknowledged by O'Malley and Chamot (1990), Macaro (2001) and Cohen (2011). Interestingly, however, out of the four students examined above, it was Chris, rather than Carissa, who developed his strategy use the most and also improved his attainment in all three subjects over time, particularly in German and French following Phase A of the SBI. For Chris, the link between strategy use and achievement was very explicit, as he had a tendency to evaluate the success of the strategies he used according to the marks he received from the teacher.
Zoe, on the other hand, provides an example of a lower achieving writer who seemed to benefit relatively little from the intervention of SBI in terms of either strategy development or attainment. Although she employed a range of strategies, she did not necessarily use them effectively and as a result, many characteristics of her strategy use are reminiscent of Bereiter and Scardamalia's (1987) knowledge-telling model and of less skilled or struggling writers in other studies. Throughout the SBI, she engaged in minimal planning, spent little time revising and any revision which did take place focused on superficial errors (Bereiter and Scardamalia 1987), and she tended to view writing as a series of words and sentences rather than as a global text (Zamel 1983;Raimes 1987). There were no perceptible improvements in terms of her achievement in any of the three subjects over time and any changes in her strategy use were much more limited than in the case of Chris and seemed to take longer to develop.
Given Zoe's limited proficiency and lack of confidence in German from the beginning, it is important to acknowledge that this may in turn have acted as a barrier for the development of strategy use in this context during Phase A of the study, particularly given that the few changes which did occur took place following Phase B. While a low level of FL proficiency in comparison to the L1 may not in itself prevent strategy development or transfer, as this was evident for both Carissa and Chris, the case of Zoe provides some evidence to suggest that, at an intra language level, a certain level of underlying proficiency in the FL does help to facilitate the successful development and transfer of strategies in a particular context. As such, it would seem that while beginner or low proficiency FL learners can be taught in such a way as to encourage strategy development and transfer to the L1, some learners will require more scaffolding and practice than others.
The above discussion raises complex questions surrounding the issue of cause and effect relating to level of proficiency and the development and transfer of writing strategies: Does strategy development increase proficiency? Does a higher level of proficiency increase the potential for strategy development? Or, in what Griffiths (2013) refers to as the 'Tornado effect' , is the relationship a spiral one, 'with one factor augmenting the other? ' (93). Evidence from the current study would suggest that the two are inextricably intertwined; however, it would also seem that the learners' level of metacognitive engagement with the task is equally, if not more, influential a factor in the development and transfer of writing strategies.

The importance of metacognitive engagement
Strategies in themselves are not inherently good or bad, but can be applied successfully or unsuccessfully (Grenfell and Harris 1999;Cohen 2011) and the findings of this study highlight the importance of the learners' ability to engage metacognitively with the writing task in determining the success of their development, application and transfer of strategies.
The relationship between metacognition and strategy development has been underlined in a number of studies (e.g. Graham 2006;Grenfell and Macaro 2007;Griffiths 2013) and was similarly evident in this study. Carissa, for example, emerged as a strategic writer from the beginning, and as such, she independently and consistently engaged in a high level of self-evaluation in order to continuously develop her strategy use over time. Zoe, at the other end of the spectrum, was less successful in developing her strategy use, and one of the key inhibitory factors was her lack of metacognitive awareness and engagement with the task. This was manifest primarily in the discrepancies which emerged between what she thought she did and what she actually did, and is in line with the findings of a case study by Vann and Abraham (1990) into two unsuccessful adult ESL learners. The authors similarly identified that these learners 'lacked certain necessary higher-order processes … which would enable them to assess the task and to bring to bear the necessary strategies for its completion ' (191). This suggests that it is not the particular strategies in themselves which make a difference, but the learner's explicit awareness, self-evaluation and ultimately metacognitive engagement with these strategies which leads to successful strategy use.
However, it is important to recognise that such profiles of struggling and strategic writers are not static and as such, helping learners develop the ability to reflect metacognitively is key to enabling them to progress from the former to the latter. As stated by Zhang and Zhang (2013), 'metacognition should be treated as dynamic systems, and it should be construed as something embedded in language learners, which is intertwined with many modifiable variables, both cognitive and sociocultural' (114). Such dynamism is effectively captured in the case of Chris; although not a particularly high achieving or strategic writer at the beginning, Chris actively developed an increasing awareness of his strategy use over time. However, while Carissa engaged in such self-evaluation relatively independently, Chris had a tendency to rely more heavily on summative marks and teacher feedback in assessing the effectiveness of his strategy use. This highlights the importance of building in opportunities for students to engage in guided practice of a range of strategies, but also the importance of including time for feedback and reflection.
Just as enabling learners to engage metacognitively with their writing is a key factor in the development of successful strategy use, it is similarly fundamental to facilitating the transfer of such strategies from one language context to another (James 2006). There was evidence in the current study that the intervention of SBI in the German classroom during Phase A helped to develop the students' awareness and use of writing strategies in both the FL and L1 contexts. By extension, I would argue that the FL classroom is an environment which is highly conducive to the cultivation of language-related metacognitive engagement more generally. In this context, learners are more conscious of their thought processes, unlike the L1 classroom where such processes are more likely to have become automatic and proceduralised (see Forbes In press).

Attitudes towards writing
The development and transfer of strategies in this study were also shown to be affected by learners' attitudes towards writing. Although there was no evidence that the intervention directly affected students' enjoyment of writing in any subject, there was some evidence to suggest that the reverse may be true; that the extent to which students enjoyed writing in a particular subject may have influenced their willingness to develop their strategy use.
Zoe, for example, commented that she would 'try really hard' if she enjoyed the particular topic she was writing about, particularly in English, the implication being that she would invest more time and effort in such tasks and perhaps be more likely to seek strategies to help her to achieve well (however, as discussed above, these strategies may not be used effectively). Yet she disliked writing in German, which is also likely to have hindered her from transferring strategies both to and from this FL context. This is in line with findings from Kobayashi and Rinnert (2008), who similarly suggested that 'students' attitudes toward English writing could prevent their knowledge from being transferred to L1 writing ' (19-20).
Although not examined directly as part of the current study, motivation has also been shown to affect the development of language learning strategies (Yang 1999;Griffiths 2013), in that those learners who are more motivated to achieve are therefore more likely to view the use of strategies as a means of helping them to do so. This can be observed in the case of Chris, who was motivated to improve his marks in writing across all subjects and was willing to experiment with strategies in order to achieve this.

Strategic multilingualism
Another important factor influencing the development and transfer of strategies which emerged from this study was the strategic use of other languages throughout the writing process. As emphasised by Tullock and Fernández-Villanueva (2013), 'perhaps one of the most important resources multilingual writers possess is their ability to refer to their full linguistic repertoire while composing' (421). Although this could also apply to learners such as Carissa who, for example, used English strategically when planning for her German writing task, it was most strikingly evident among the bilingual EAL learners such as Mei, who were able to draw on their native speaker competencies in another language, in addition to English, when learning German and French.
It is often posited that multilinguals have increased metalinguistic awareness and increased language learning awareness (Cenoz 2003;Psaltou-Joycey and Kantaridou 2009), and this was reflected to some extent in the current study among the EAL learners by the omnipresence of their mother tongue during all stages of the writing process across all languages. Although Mei had been living in England for many years and spoke fluent English, Mandarin still constituted her primary language of thought and she therefore engaged in a lot of translation back and forwards between languages. She reported experiencing some of the same problems across all subjects, such as her prevailing concern with 'making sense' , using the right vocabulary and correct use of punctuation. Consequently, she often approached these problems in a similar way across the various subjects. This provides some evidence in support of the hypothesis that bilinguals are 'capable of transferring skills from the two languages they know to a third language' (Cenoz 2003, 77). It is also in line with findings from recent studies into the learning strategies of bilingual adolescents learning an additional foreign language by Grenfell and Harris (2015) and Mitits and Gavriilidou (2016).
Yet, although having an additional language at her disposal was certainly a strategic asset for Mei, it did not necessarily make her a more proficient FL learner, which seems to go against the commonly held view that the more languages a learner speaks, the more proficient they are at acquiring additional languages. For Mei, therefore, it would seem to hold true that the strategic advantage of bilingual learners 'is to be found at the metalinguistic level, which includes both communicative and learning strategies' (Bono and Stratilaki 2009, 211), rather than at the level of academic attainment.

Conclusion
The aim of this paper was to explore the role of individual differences in the development and cross-linguistic transfer of strategies between foreign and first language writing. A range of students' trajectories through an intervention of strategy-based instruction were evaluated and four distinct writer 'profiles' emerged: the strategic writer, the experimenter, the struggling writer and the multilingual writer. Both the development and transfer of strategies for these students were shown to be influenced by a complex and dynamic range of factors such as the learner's proficiency level, their level of metacognitive engagement with the task, their attitude towards writing and their strategic use of other languages in their repertoire. It is important to acknowledge the limitations of presenting just four students; however, one of the key aims of this paper was to conduct an in-depth analysis of individuals, which would not have been feasible at a whole-class level. It is also important to note that it is by no means expected that all students would fit neatly into one of these four categories, they may share characteristics of more than one 'type' of writer, and likewise may shift over time from one type of writer to another. The teacher therefore also plays an important role in facilitating strategy development through effective modelling, scaffolding and feedback in order to help students to move towards being more strategic and proficient writers.

Notes on contributor
Karen Forbes is a lecturer in Second Language Education at the Faculty of Education, University of Cambridge. Her current research interests include the development and transfer of language learning strategies, the influence of multilingual identity on foreign language learning, and the language development of EAL students.