Scaffolding writing development : How formative is the feedback ?

This paper contrasts students’ perceptions of feedback practices on written texts and their utilization of the feedback against formative assessment pedagogy. It uses observations of nine lessons of English writing in combination with semi-structured interviews of 39 first year upper secondary students in Norway. The findings show that students in general appreciate the feedback, but have diverging views about the different types and forms of feedback provided. The majority want to have one-to-one discussions with the teacher about their texts in addition to written feedback, and to receive more specific comments on both content and language. Most are content to use feedback in a delayed fashion, as a reference point for future writings. This is a result from the working routine in the classrooms where feedback in most cases comes with a grade – without a strict requirement to follow-up comprehensively, with the exception of correcting language errors. The findings highlight the primacy of form in the feedback provided and, in more than one way, poor quality formative feedback from the students’ perspective. It concludes that in order to maximize the formative potential of feedback there is a need to make it more usable, e.g. by giving it to ungraded texts and asking students to produce improved work based on feedback.


Introduction
One of the main challenges when teaching writing is providing feedback that is read, appreciated and used.Indeed, use is a precondition for moving learning forward (Sadler 1989, 1998, Black & Wiliam 1998, Hattie & Timperley 2007, Shute 2008), and improving writing (e.g.Hyland 1990, Jonsson 2013, Ferris 2010).However, unless teachers provide the time and motivation to do so (Zamel 1985, Jonsson 2013), students will not use and work actively with feedback (Straub 1997).Indeed, the lack of systematic follow-up will make it a disappointing task for the teachers and an unduly onerous task for the students.Further, it is only when feedback information is used to forward learning that we can refer to it as formative feedback.Or, as Sadler (1989), notes, "the information about the gap between actual and reference level is considered as feedback only when it is used to alter the gap" (Sadler 1989:121, italics in original).This is the main principle that underpins formative assessment (FA) and contributes to students' learning.
The first step in this process is to recognize feedback as an important learning tool in the classroom.The next step is to provide good quality feedback, the importance of which is duly acknowledged in FA (e.g.Black & Wiliam 1998, 2009, Sadler 1998, Shute 2008, Hattie & Timperley 2007), alongside other basic principles of this pedagogy (e.g.learner autonomy, success criteria, self-assessment).However, it is problematic that the literature shows "little consensus about what constitutes good quality external feedback" (Nicol & Macfarlane-Dick 2006:208).To compensate, writing experts provide lists of recommended feedback practices (e.g.Hyland & Hyland 2006, Ferris 2003, 2014), but these will not suffice unless the teachers know what kind of feedback the students need and consider useful (Sommers 2006, Underwood & Tregidgo 2006, Hattie & Gan 2011).
Due to the widespread introduction of FA in the late 1990s following the Reform Group project in the UK (e.g.Black & Wiliam 1998), FA was introduced into curricula at almost all educational levels in many countries, Norwaythe context for this study being no exception (KD 2009).It was one of the main school developmental strategies for the 2010-2014 period (UDIR 2010), and was followed up with a number of courses and workshops for teachers (e.g.Burner 2015, KD 2011).This means that teachers are required and expected to use FA extensively, and to be familiar with its main principles.However, implementing FA into everyday teaching takes time and requires support (Lee 2011), and a number of studies show that this is lagging behind in Norway in general (e.g.OECD 2011 review, Smith 2011).The aim of this study is to investigate whether this is also the case, or not, in the first year of upper secondary school English writing instruction.
Internationally, there have been many studies on feedback in a broad range of disciplines and contexts that have a somewhat limited relevance for the domain of second language (L2) writing, the context of this study.In L2 writing, the formative potential of feedback (e.g. Lee 2007, Lee 2008, Lee & Conian 2013) and students' perspectives (e.g.Hyland 1998, Hedgcock & Lefkowitz 1994, Leki 1991, Zacharias 2007) are largely unexplored, and have been suggested to set the agenda for future research (Ferris 2010).Similarly, in Norway, studies of formative feedback have mainly been set in multidisciplinary contexts, focused mostly on the L1, and only a few of them have examined the students' perspectives (e.g.Havnes et al. 2012, Gamlem & Smith 2013).The existing studies of feedback and FA in L2 writing are few (Burner 2016, Horverak 2015, 2016, Saliu-Abdulahi, Hellekjaer & Hertzberg 2017), and only two look at students' perceptions (Burner 2015, Saliu-Abdulahi & Hellekjaer forthcoming).This shows the need for additional research in this area, and the aim of this study is to fill this knowledge gap.

English in the Norwegian context
In Norway, English has been a compulsory subject for all since 1959, from primary school to the first or second year of upper secondary school.Levels of proficiency are high, and English teaching is considered an educational success story (Simensen 2010).by attending to and addressing teachers' comments, and 3) peer and self-assessment emerged as a supplement to teachers' feedback.It also required teachers to ensure that the feedback was helpful and useful (Ferris 2003).This led to new studies that looked at student perceptions of feedback (e.g.Leki 1991, Hedgcock & Lefkowitz 1994, Hyland 2003), and provided teachers with information about what students find helpful and problematic in feedback comments (e.g.Hyland 1998, Sommers 2006).
The importance and power of feedback in the teaching and learning process is acknowledged in FA pedagogy (e.g.Black & Wiliam 1998, Sadler 1989, 1998, Shute 2008).The basic principles of feedback in both pedagogies (i.e. in process oriented writing pedagogy and in FA pedagogy) are very similar, and in both cases, it is the formative feedback that scaffolds learning and accounts for success (e.g.Huot & Perry 2009, Shute 2008).This role of formative feedback differs a lot from summative feedback, where the comments are made to finished texts after the learning process, and to a large extent serve to justify the grade.Instead, formative feedback is to be delivered during the writing process, often in response to multiple drafts (Hyland 1990), and needs to be used for feeding forward (Hattie & Timperley 2007).However, a number of studies indicate that the dominant culture in L2 writing instruction is to give single-draft feedback (i.e.summative feedback) to graded texts after the writing process (e.g. Lee 2007, 2008, Lee & Coniam 2013).This is also the case in Norway (see Saliu-Abdulahi, Hellekjaer & Hertzberg 2017 and Saliu-Abdulahi & Hellekjaer forthcoming).Jonsson (2013) reviewed multidisciplinary studies of writing feedback, including some L2 studies, and examined why students do not use teachers' feedback productively as required in FA.He found that the main obstacle is that students are not given the opportunity to use feedback, despite the widespread agreement on the centrality of revision (Ferris 2010).

Feedback focus and form
A challenging aspect of teaching and responding to writing in the L2 is that teachers need to balance a number of decisions.First, they need to teach writing conventions while also working to develop the target language proficiency (see Hedgcock & Lefkowitz 1994).Next, while keeping in mind the goals for why the L2 writing is taught, teachers need to decide what should be in focus and comment accordingly.However, one of the most debated topics in the field of L2 writing is whether one should treat content and form equally, or if one should prioritize one over the other.
Evidence from studies show mixed results.For example, while some studies show that error feedback is unsuccessful in helping students improve their writing (e.g.Truscott 1996), others show that error feedback helps writing development (e.g.Ferris 2004, Bitchener 2008) if done effectively (Ferris 2010).Furthermore, in a large number of studies (e.g.Hyland 2003, Montgomery & Baker 2007, Lee 2004, 2007, Furneaux, Paran & Fairfax 2007, Evans et al. 2010), L2 teachers are criticized for prioritizing form issues before content.For example, a study by Evans et al. (2010) with teachers from 69 countries shows that this persists to be a global problem.In Norway, only one study of L2 shows a shift from form to content (Horverak 2015).
Informed by new teaching methodologies and pedagogies that advocate studentcentred classrooms, such as FA, teachers are expected to update their classroom instruction as well as feedback practices.One of the main principles of FA is involving the students in the feedback process, by giving and receiving comments (peerassessment), by reflecting on their own work by referring to success criteria (selfassessment), and by discussing feedback with the teacher (one-to-one).These alternative forms of feedback correspond with what the experts in L2 writing recommend as "best feedback" practices (e.g.Ferris 2014, Hyland & Hyland 2006), and contribute to dialogic interaction around feedback (Hyland 1998), and selfregulated learning (Nicol & Macfarlane-Dick 2006) as important aspects of formative comprehension of feedback.

Students' perceptions of feedback
A number of studies have looked at students' perceptions of writing feedback in L2 contexts (e.g.Leki 1991, Hedgcock & Lefkowitz 1994, Hyland 1998, 2003) and in L1 contexts (e.g.Straub 1997, Sommers 2006, Lipnewich & Smith 2009).These have tried to inform both the teachers and researchers about what feedback student writers want and need, and to help teachers adjust their feedback to meet the students' learning needs and preferences.
Regarding student perceptions, Leki's (1991) survey showed a preference for error correction, reflecting that the students aimed for error free writing, in contrast to their teachers who prioritized the development of ideas.Students' preference for comments on form is confirmed in another case study by Hyland (2003).In Ferris (1995), on the other hand, students wanted comments on both form and content.Similarly, a more recent study by Calhoon-Dillahunt & Forrest (2013) shows that students were less concerned with error correction and preferred more holistic and global comments, most probably due to their growing experience as writers.Nevertheless, it remains a challenge that there will always be students with individual preferences.For example, a case study by Hyland (1998) with six students shows that students had quite diverging reactions to written feedback.Therefore, she suggests a face-to-face dialogue between the teacher and student to allow feedback to be better tailored to students' expectations and needs.Yet another important issue that impacts on student perceptions of feedback is grades (Underwood & Tregidgo 2006).A study by Lipnewich & Smith (2009) of students' views of the ideal form of feedback shows a preference for feedback without a grade, because grades are seen as obstacles for the formative function.This has been found to be a problem in Norway as well (see Saliu-Abdulahi, Hellekjaer & Hertzberg 2017, Burner 2015).
To sum up, we can conclude that feedback is important, and that it has been shown to influence student writers.Indeed, as stated by Ferris (2003:117), students "value and appreciate it, attend to it, and utilize it to write revisions and make progress in their writing".Nonetheless, we still know all-too-little about what kind of feedback students get compared to what they want.The present study will address this gap in the Norwegian upper secondary school setting.As a second step, we will consider to what extent current feedback and student follow-up are in line with the main principles of FA theory, as is expected by the Norwegian educational authorities.Consequently, the present study addresses the following research questions: 1. What are students' perceptions of current feedback practices used in writing lessons? 2. To what extent are the current feedback practices in writing utilized in line with FA pedagogy?

Study context and participants
For this qualitative study, a number of upper secondary 'university schools' were contacted and invited to take part.These are schools involved in different collaborative projects and activities with the University of Oslo to improve teaching quality.The six schools that agreed to contribute represent a good mix with regard to profile, students' backgrounds and school admission standards.

Observations
The data in this study was collected through observations combined with focus group (FG) interviews.To get the best possible picture of what is happening in the classroom, and to relate the interview questions to classroom activities, I started by observing writing instruction when teachers had a feedback-related lesson.I observed eight English writing classes in six schools, and one class twice (Class 3), a total of 9 observations (see Table 1 below).In the seven classes that I observed once, the teachers handed out the graded texts with comments, gave some general feedback concerning the whole class and asked students to follow-up on the feedback (e.g.revise sentences with language errors, revise the topic sentence and/or write a reflection).Some teachers (in Class 1, 2 & 6) would hold back the grade until students did so, while others would reveal it immediately.In Class 3, where the teacher gave feedback during the writing process, there was more extensive work with feedback.For example, during the first observed lesson the students got comments and tasks to work on.They had the opportunity to revise the text for the entire lesson/at home and hand it in for grading.During the second observation, the teacher returned the graded text with comments (on global and local issues) drawing on assessment criteria that were already familiar to students, and asked them to make a note of what they need to work on in the future and to write a reflection note on their writing3 .As for the writing tasks, students at this level are usually asked to write argumentative and expository texts.In addition, they may also write texts as a follow-up to literary readings, such as book reviews or other types of reflections on the texts.
For all the observations, I used a semi-structured observation form with predesigned tasks and took field notes.I looked for the following categories: different sources of feedback; focus of feedback; text revision; feedback and grades; amongst others (see Appendix A).

Interviews
After the observations, I carried out focus group (FG) interviews that were first recorded and then transcribed.The students were sixteen-year-olds that participated voluntarily after being selected by the teachers.On my request, the groups comprised of students with varying skills in English, but for ethical reasons I did not know their levels.Each FG consisted of four to seven students (39 in total, 20 boys and 19 girls) with a mixture of boys and girls to maximize heterogeneity.However, later on I realized that having large FGs is a limitation with regard to getting all students to communicate all answers verbally, and makes it difficult to quantify student answers.Table 2 below provides an overview of the student participants.Each interview was conducted in English and lasted about 45-60 minutes.Although the interviews were not conducted in the students' mother tongue, their good levels of language competence and ability to elaborate on or explain each other's comments when needed, prevented potential misunderstandings.I also asked them to bring their written samples as a stimulus for discussions and sometimes we would refer to them for exemplification.I used an interview guide that was developed on the basis of relevant literature, my knowledge from teaching experience, and observations when piloting4 my instruments.I did not follow the guide slavishly and rephrased the questions when necessary (Johnson & Turner 2003).There were questions about feedback importance, feedback type and form, feedback focus, revision opportunities, amongst others (see Appendix B).

Analysis
To analyse the data, I followed the procedure of thematic analysis as described by Braun & Clarke (2006).With a careful and repeated reading of the interview transcriptions, I identified meaningful patterns relevant to the focus of the study.Then, I used structured coding based on the initial categories to simplify the data (Dörnyei 2007).During this process, salient themes were created and grouped with illustrative and representative quotes for each category.Most of the categories were derived deductively from the categories in the interview guide and observation form, and a number of themes emerged inductively (e.g. that some students use feedback as a reference point for future writings).I read observation notes and memos to make reliable inferences of the feedback, and then searched for the repeated patterns of feedback practices in both observations and interviews (see Appendix C & D).There were many repeated patterns that confirmed the validity of the data.An assistant researcher peer-checked the reliability and validity of the categories by testing one third of the material and through a moderation process we agreed on the differences.The data was later sorted into categories (see Figure 1).The analyses resulted in two main categories.The first main category (i) feedback form and type with the sub-categories: (1) feedback source (teacher, student self, peers), (2) mode of delivery (written and oral) and (3) feedback focus (general vs specific), reflects all various forms and types of classroom feedback and sheds light on the first research question of this study, i.e. students' perceptions of current feedback practices.Then, the second main category (ii) feedback utilization consisted of the sub-categories: (1) time of feedback use (e.g. in the process, in the end and/or next assignment) and (2) different ways of acting upon feedback (e.g.reading the comments, correcting errors, revising, ignoring), answering the second research question, i.e. to what extent feedback acts formatively.Last, there was a third category that consisted of (iii) themes, identified as closely related to feedback, namely grading and assessment criteria.In the next section, I will present the findings according to these categories.

Findings
Findings are presented as answers to the two research questions of this study.First, I present the findings about students' perceptions of feedback, and then follow up with the findings of formative utilization of feedback.

Research question 1: What are the students' perceptions of current feedback
practices used in writing lessons?In the answer to this question, findings show that students appreciate feedback as a learning tool, but show diverging attitudes about its particular types and forms.The following is a detailed presentation of various views.

Feedback source and mode of delivery
First, teacher feedback was the regularly used form, and was delivered in two modes: written (hereafter used as teacher written feedback -TWF) or oral (to the whole class and/or to individual students).Students received TWF on all written texts on a regular basis, and considered this form of feedback very important and useful.It was inserted in the text and/or delivered using the LMS5 (Language Management System), depending on its focus (this will be discussed in section 5.1.3).However, despite their positive attitude to written feedback, students expressed a need for verbal clarification of the comments.When asked what they thought would be the best feedback form, the vast majority agreed that a combination of written and oral is to be preferred: Even though on some occasions teachers would go around and approach students randomly (see Appendix C, Class 5) when working with feedback, this was not done in organized and structured manner.However, students said that it is usually possible to ask for a one-to-one discussion on their own initiative (FG2, 4, 5, 8), but that this rarely happened.Normally, they have one-to-one discussions with the teacher about the subject in general (e.g.FG1, 2, 4), but only on special occasions about writing (FG3, 7).In the whole class oral feedback, teachers mostly go through the most common mistakes (e.g. as observed in Class 1, 2, 3, 5) and may show anonymized samples from students' work to illustrate good/bad writing and comment on that (e.g. as reported in FG2).
The observations confirmed that the teacher was the sole source of feedback.Thus, in the interviews I asked students about peer-feedback.The answers indicate that only about a third of the students have done this in one way or another (e.g. with smaller writing tasks, for oral presentations, and in their L1 classes), and they lack systematic experience with it.However, in spite of this limited experience, they expressed uncertainty about the trustworthiness of peer-feedback, saying: - The teacher knows better… I would trust them [peers], but not as much as I trust the teacher.(FG7) -Because I know that some of them are the same level as me, so I really don't think that they are the right person to judge my text and give feedback.(FG6) Their negative attitude is perhaps due to their overall underlying understanding of feedback as part of summative assessment.On the other hand, a couple of students were positive to peer feedback: "I could see that teacher and students see the same mistakes" (FG2), and a few others expressed limited knowledge in dealing with peer response: "I don't have a problem with [peer-feedback], the problem is to find the mistakes" (FG4).
With regard to self-assessment, however, the replies were quite different than those of peer-feedback.With the exception of two groups (FG1&5), self-assessment was practiced regularly and that made them more confident and positive.They said they did this in slightly different ways, for example, mainly self-evaluating the text before handing it in to the teacher (FG3,4,6,7,8) and self-evaluating it before they saw the grade (FG2).They often self-evaluated the text using the assessment criteria the teachers usually provided them with before the writing task.

Assessment criteria
Students were familiar with assessment criteria because they were regularly communicated to them.Students look at them at different occasions, such as: the beginning of the school year (FG6); before a task is written to see what is evaluated (FG2, 6, 8), after the writing to self-correct the text (FG4, 5, 6, 7) and/or before they get the grade (FG2).However, they did not mention doing so during the writing process (except some students in FG2 & 8), and complained about having difficulties in understanding the criteria even when the teacher offered an explanation: -

Feedback focusgeneral vs specific
One form of the dominating TWF was referred to as general feedback.This kind of feedback, the students said, sums up the quality of writing and gives advice on how to improve.It covers different aspects of the text, but mainly content, structure and language, and may occasionally involve comments on other aspects (e.g.cohesion, punctuation, etc.).Student answers show that not all aspects are equally prioritized by allsome get more comments on content and others grammar feedback (e.g.FG8), most likely depending on their level of ability.All students agree that general feedback contains a lot of information and is very important, yet often vague and difficult to understand.The following quotes illustrate this: These quotes show dissatisfaction with the content and specificity of this form, which might be a reason why a minority of students do not pay attention to them (e.g. in FG5&6).However, more than two thirds of the students acknowledged that these comments are very important and help them develop their writing in the future.Nevertheless, those that find general feedback useful wish for more explicit guidance and examples of how to improve, and if they are to use it, students say, it has to be tailored to their specific needs (FG5).Another way of giving TWF was interlinear and/or in the margins of the student texts, and was referred to as specific feedback.This feedback, students explained, covers various language issues, such as grammar, vocabulary, spelling, mechanics and punctuation.The overwhelming majority, however, agreed that it mainly focuses on grammar and most of them are satisfied with that because: grammar comments are more important… because for bad grammar you get bad grades.(FG4) you can't write a good text with many mistakes.(FG2) On the other hand, there were students who disagreed saying: "I think you'll always have some grammar mistakes, the general advice helps you build up the text" (e.g.FG2).The students likewise agree that the lexical comments they get help them build their vocabulary.Sometimes there might be specific comments about the content, students added, in a form of a question or just a simple statement (FG4).Almost all students were positive about these 'in-text' comments because here they get more concrete and specific guidance.The following example illustrates their positivity: -They are more concise and if I see something she has written: "you have to put in or change this one with another", I can see, Ok, this sentence is wrong instead of just saying "you have some structure errors in your sentence".(FG8) To sum up, in general, while students are content with the dominating TWF, the overwhelming majority wish to have it supplemented with a one-to-one discussion for better understanding.As for the two different forms of TWFgeneral and specificboth are perceived as equally important, however, the preferences for each type varied a lot.

Grading
Observations showed that with the exception of one group, students got feedback after the text was graded.I therefore asked the students for their preferences with regard to feedback and grading.The answers varied a lot.While many said that they would prefer to work with feedback first and then get a grade, a minority in FG4&5 expressed the opposite view.These were either good writers who would get good grades and would feel no need to revise their work, or poor writers who were not interested in getting better at writing or getting better grades.The latter said they wanted to know the grade immediately and be done with the writing.Those who wanted the feedback first and the grade later responded that the grade may disrupt them in following up on the feedback, and one student noted, that it "would be good to do tasks without a grade, just to practice" (FG8), and another one said that "I would like to maybe try to have just the task with the corrections and then do it and fix the mistakes and then hand it in and get a grade" (FG4).Only two groups out of eight said that they remember having done that once or twice, either correct language mistakes and get a better grade (FG6), or revise the whole text and then get a grade (FG8).But, not all students in the group agreed, which might be linked to them being unfamiliar with this approach.

Research question 2:
To what extent are the current feedback practices in writing utilized in line with FA pedagogy?
In the following I examine to what extent the current feedback practices reported above are in line with FA principles.The findings point to an absence of systematic opportunities and requirements for comprehensive work with feedback.What follows is a more detailed picture of this pattern.The timing of feedback has a great deal of impact on its formative functioning, which in this study was predominantly on a finished and graded text.Thus, it is not surprising that the late delivery pretty much 'directs' what will happen nextwhether the students will be engaged or disengaged with the feedback.Obviously, getting feedback on a text that is graded limits the follow-up.Therefore, when I asked students about using feedback, the overwhelming majority said they would use the feedback for a "term test", their "next assignment" or "throughout the year, in new projects, but not in this one" (e.g.FG1,2,5,6,7).This answer applied especially to the general feedback: "Maybe go there [LMS or text] before next writing" or "I go through them, but not in depth, more as a referencing point for future" (FG4, 7).The common understanding of feedback as being of importance for future writing assignments can be further exemplified by the following quote: On the other hand, there were some students (in FG1, 4, 5 & 6) who disagreed, saying that comments they received on one text might not be applicable to the next, and they therefore do not pay much attention to the feedback: The quotes clearly show that student follow-up on general feedback is typically reduced to reading only and/or using it in the future.However, the situation was different with interlinear and marginal comments (i.e.specific comments).Students from all groups (except FG3 who did multiple drafts) are normally asked to act upon these comments by correcting a limited number of sentences with language errors, usually 8 to 10. Three groups out of eight (FG1, 2 & 6), those with more demanding teachers, have to do these corrections before they get the grade.Other groups (FG4, 5, 7 & 8) faced no strict requirements, but were assigned some classroom time for sentence corrections.The following quotes reveal quite diverging views about error corrections: -Now [for the end of the year exam] you don't really have to do anything if you don't want to, but if you have to hand it in again then you have to see all the mistakes that you have done and try to change them.(FG5) -I can just look at them and remember, Ok, instead of writing them all over again.(FG6) A very small number of students in FG5 & 8, especially boys, say that reading feedback comments might be enough.This could be seen during observations when students talked to one another instead of doing what they were asked to.They confirmed this in the interviews when I asked about the number of sentences they managed to correct: "Nothing, I was too busy with something…" [student laughs and then all laugh]; "zero, but I looked them through and corrected them in my head sort of, but didn't write them down" (FG8); and some tried to justify their behaviour by saying "But we don't have the opportunity to improve the grade, we never had that" (FG4 & 2).Nevertheless, during observations there were students (mainly girls, presumably the skilled writers) who were actively involved in error correction.Still, only two students (in FG5) out of 39 reported that they revised beyond sentence level correction, like the thesis statement or a whole paragraph.Not surprisingly, no student said that they ever revised the whole text in single draft writing.On some occasions, students are assigned different reflective tasks instead of and/or in addition to error correction (FG2 & 7).
Interestingly, only one group (FG3) out of the eight reported that they get a chance to revise the entire text, and resubmit it for evaluation.They agree that multiple drafting is quite useful and they feel motivated to revise a text before getting a grade.Even in the groups that do not get this opportunity, some students who were particularly interested in developing their writing said: "Sometimes I want to change the whole text and write a new one" (FG6).Yet, there are students, as mentioned earlier, who see the role of feedback mostly as providing justification for the grade.For example: "I really often read the feedback because I need to understand why I didn't get the grade that I hoped" (FG5).This belief is most likely derived from their understanding of feedback as summative since it is invariably delivered along with grades to finished texts.
To sum up, whilst the focus of feedback has broad coverage and students value it, the follow-up opportunities mostly involve correcting language mistakes.As for other comments, for instance regarding content, there is no requirement to act on these, and students report reading them only for possible use in future assignments.Obviously, this incomplete use of feedback does not align with the best feedback practices as recommended by writing experts and FA research.The only exception is class 3 where students were allowed to hand in revised texts for a new evaluation.The need for the utilization of both specific and general comments in the current text, and students' awareness of this fact is illustrated when one student in FG5 says that they need to be involved in more comprehensive follow-up with feedback and argues that they "would have learned a lot more, maybe" if they were asked to revise beyond sentence error correction.

Discussion
This study set out to investigate current feedback practices as perceived and used by Norwegian upper secondary school students.In general, students notice and show appreciation of feedback as a learning tool, but their attitudes vary immensely when it comes to different forms and types.The students' formative utilization of feedback is limited primarily to correcting language mistakes.

Student perceptions of current feedback practices
The study offers some insight into upper secondary students' attitudes and preferences towards the forms of feedback they have experienced, and/or wish to experience in the classroom.The students were mostly positive towards the written feedback they received.The findings about teacher written feedback (TWF) echoed Jonsson's findings about the dominance of this type of feedback on written texts (Jonsson 2013).In addition, there is a clear need, as expressed by the students, to supplement written comments with oral discussion in order to clarify the information given.This is also supported by other international studies (e.g.Duers & Brown 2009, Reid 2010in Jonsson 2013, Hyland 1998, Lee 2007) and other studies in the Norwegian context (Saliu-Abdulahi & Hellekjaer forthcoming, Burner 2016, 2015, Gamlem & Smith 2013).All show a high demand for more detailed oral feedback due to the need to clarify vague comments, and thus extend their formative function (Ferris 2003, Zacharias 2007), and make feedback less authoritative but more productive (Jonsson 2013).
Individual oral feedback together with peer-feedback, are relevant practices in FA pedagogy and play a key role in the interactive conceptualization of feedback (e.g.Hyland 2003, Nicol & Macfarlane-Dick 2006).In general, students' limited involvement in these forms makes for a lack of dialogic interaction, and as such the formative functioning of feedback can be questioned (Gamlem & Smith 2013, Havnes et al. 2012).While the informants were negative towards peer-feedback, one-to-one discussion was highly desired (like in Burner 2015).Hyland (2010) reiterates the importance of conducting studies of written feedback supported with oral feedback.On the other hand, students' negativity towards peer-feedback in this study can be explained as a result of the lack of systematic involvement and lack of trust.In fact, the attitudes in the present study stand in contrast to others where it is perceived as more secure and less threatening (e.g.Zacharias 2007, Ferris 2003).
Students' involvement in self-assessment in this study is also in line with other studies in the Norwegian context (e.g.Gamlem & Smith 2013, Burner 2016) and indicates a commitment to implementing the recent regulations on FA with a particular emphasis on the use of assessment criteria.Indeed, the students, with the exception of those who found self-assessment vague, are in one way or other involved in acquiring the assessment criteria.However, more efforts should be made to explore the full potential of self-assessment as an important practice in facilitating self-regulated learning (Hyland 2010).
A further interesting insight of this study comes from students' attitudes towards specific and more general feedback.Students report that both are relevant (like in Straub 1997, Ferris 1995), but appreciate specific feedback slightly more because it is concise, concrete and straight to the point (like in Leki 1991).The preference for this feedback type reflects the findings from other studies in that specific comments are highly preferred (e.g.Lipnewich & Smith 2009, Sommers 2006, Underwood & Tregidgo 2006).
With regard to grades, almost a third of the interviewed students indicate that they do not follow up on the feedback (i.e.language corrections) after they see the grades, unlike the remainder.This has also been found in other studies (e.g.Underwood & Tregidgo 2006, Lipnewich & Smith 2009, Jonsson 2013, Burner 2016).Instead of looking at feedback as a learning opportunity, it would seem that these students are distracted by the grade, which has the effect of limiting revisions (cf.Huot & Perry 2009).An interesting finding about the grade-feedback relation in this study is that students expressed a readiness to try out new writing practices without a grade for the sake of improving their writing.This has also been shown successful in a Norwegian portfolio intervention study by Burner (2015).

Current feedback practices and their utilization against FA pedagogy
Based on the findings we can see that current feedback practices and students' engagement with them, unfortunately, lag behind FA principles.This is echoed in two other similar studies in Norway in multidisciplinary contexts, one showing "generally weak formative assessment culture" (Havnes et al. 2012:23) and another showing "no formal opportunities to apply the feedback" (Gamlem & Smith 2013:160).This situation reaffirms the need for further professional training for successful implementation of FA in Norway (Smith 2011, OSCD 2011review, Saliu-Abdulahi, Hellekjaer & Hertzberg 2017), in particular in the context of writing, and calls for a follow-up by the school authorities.
Although the students unanimously recognized the value of feedback for improving writing (e.g.Gamlem & Smith 2013, Straub 1997, Zacharias 2007), the utilization of it is reduced to formal aspects of the text.Perhaps this is a reflection of the current school system and its constraints.Furthermore, following up on these comments is what "will pay off in terms of grades" (Jonsson 2013:68) in the next assignments.Students state that for bad grammar you get bad grades, which means they tend to prioritize language in order to be awarded better grades later on.However, content is another important aspect that influences the grade and for unknown reasons students neglect to articulate it.The language primacy is echoed in many feedback studies reviewed by Jonsson, 2013(e.g. Asswell 2000, McDowell 2008, Williams 2004) and with L2 learners (e.g.Evans et al. 2010, Furneaux, Paran & Fairfax 2007, Leki 1991).It seems that students have good reasons to strategically ignore general feedback and move to easier language fixes, especially when there is no requirement to do so and no grading reward for the current text.
As for the general comments on textual issues, the follow-up is left entirely to the students.This reflects another study of FA in writing in Norway (Burner 2016) and other studies of feedback in the Asian upper secondary context (e.g. Lee 2007, 2008, Lee & Conian 2013) where form is prioritized.The most problematic result of this finding is that the engagement with form issues overshadows the importance of global issues.While language accuracy is important for improving writing skills, revisions will be incomplete without textual improvement (Ferris 2014, Hyland 2003).As such, an 'unconscious' message of language primacy is conveyed to the students, and may create the wrong understanding of writing development (Montgomery & Baker 2007) that does not concur with the curriculum aims for writing at this level.This is the opposite of the situation of developing writers in Calhoon-Dillahunt & Forrest (2013) where students valued feedback for text improvement.
Another worrying finding comes from the handful of students, mainly boys, who claimed that they do not need to do any work with feedback because looking at the comments and trying to remember is sufficient.This corresponds with Jonsson's passive users of feedback, who as a result of a lack of strategies prefer to make only mental notes of feedback instead of using it productively (Jonsson 2013).That being said, there should be more emphasis on the crucial dialogic interaction (e.g.Nicol & Macfarlane-Dick 2006, Sommers 2006, Sadler 1998) in order to encourage the students to change their habits from passive to active users, and build on the follow-up strategies.This should be one of the main implications of this study.On the other hand, there are some who say they would bring the feedback to the final exams and use it as a reference point.However, not all agree on the applicability of feedback to other texts.For instance, as shown by Carless (2006) and Saliu-Abdulahi & Hellekjaer (forthcoming), if the comments are text specific, they cannot be applied to other assignments.On the other hand, if teachers write more generic comments that can create a bridge to future writing assignments (Sommers 2006), it will be at the expense of specific comments that are highly valued by students (see Saliu-Abdulahi & Hellekjaer forthcoming).In other words, this remains a complex issue.
Finally, I would argue that the combination of classroom observations with student group interviews has enhanced the validity of this study by allowing data triangulation from two qualitative sources that reflect the reality of writing classroom.What I saw during the observations was in accordance with what the students said during the interviews.In addition, they reflected what the teachers said in the interviews (see Saliu-Abdulahi, Hellekjaer & Hertzberg 2017) and what other studies in L2 context in Norway (e.g.Burner 2016) and internationally have shown (e.g. Lee 2007, 2008, Lee & Conian 2013).Last, it can also be mentioned that these results were confirmed by the findings of a follow-up survey of 14 classes from seven schools in the same area (see Saliu-Abdulahi & Hellekjaer forthcoming).To sum up, I would argue that although this is a small, qualitative study, there are good reasons to contend that the findings are relevant to similar contexts in Norway.

Conclusions
The aim of this study was to provide insights about how students experience and envision feedback on written texts in a context expecting adherence to FA pedagogy.The study builds on the premise that formative feedback has great potential to move learning forward (i.e.help writing development), when and if used appropriately.However, students' voices in this study reveal that current feedback practices diverge from central FA principles and recommended feedback practices from experts.Likewise, they do not fully meet the curriculum aims for this level.In fact, they are overshadowed by summative assessment purposes.
This study confirms previous observations that there is no 'one-size-fits-all' approach to feedback.Students reveal diverging views about what the best feedback form is for them.Given this complexity of perceptions, one-to-one discussion should become the main provision of feedback in this context.This, together with other student-centred feedback strategies, will enable students' active involvement in the feedback process as aligned with the FA foundational principles.For this to happen, teachers need to create a context where these forms of feedback can be used on a regular basis.Another important point with regard to realizing the formative potential of feedback is that teachers need to set aside time to work with feedback before the text is graded.One way of doing this is by delivering feedback during the writing process and by engaging the students to respond to both form and content feedback in due time.This also entails ensuring that teachers can devote more time to working with texts in progress.
Returning to Ferris' quote that students "value and appreciate [feedback], attend to it, and utilize it to write revisions and make progress in their writing" (Ferris 2003:117), this is unfortunately only partly true for the students of this study.First, although the students show appreciation of feedback, this pertains to those limited forms that they have experienced.Second, the lack of follow-up opportunities these students get, do not allow for revisions beyond sentence level.And, finally, it is unclear if they make progress as a result of the feedbackunless we count the students' own judgments about their making progress.This calls for other studies to investigate students' written work and their progress over time.In fact, this is one of the limitations of this study, that what is reported is based on the students' voices and classroom observations.However, I still contend that it provides interesting insights into students' perspectives that can lead to more productive use of formative feedback in writing instruction, where my findings certainly show that there is great room for improvement.

-
On the term test I tend to bring 3-4 reviewed papers so I can have an overview of what is good and what is bad and work from there instead of working on the task.(FG4)

Table 1 .
Observed classes and times of observations

Table 2 .
Profile of the focus groups in the study

-
Sometimes you get comments like "work on your vocabulary" or something like that, and that's the only comment you get, then is like "what else do you need to work with" and then you just don't do anything.It's easier when it says: "work on the verbs" or something similar, but when it's like this [unspecified] then it's quite wide range.(FG1)

Feedback according to the subjects involved (i.e. source of feedback):
II. Revisionstudents are asked to revise: _________ in the classroom, as a second draft/final _________ at home as a second/final draft _________ next class as a second/final draft _________ students are not asked to revise the text at all Other aspects: Interviewer: everybody agrees on that that sometimes you have to revise that?And what do you think about that, is it really useful to go through the same text and make the changes, mistakes you had…?Student 1 girl: sometimes I want to change the whole text… Interviewer: yes that's ok, but do you do that?Student 1 girl: not really.I just correct the grammar mistakes when she has commented about… but, I… sometimes, I really want to change the whole text and write a new one… Interviewer: if you are not given the opportunity to give it to the teacher, which probably would change your grade, then you don't see any point in doing that, is that right or?