Self-regulating writers’ uses and non-uses of peer feedback

Abstract This study contributes to knowledge about peer review by identifying how different components motivate self-regulating learners’ adaptations in later writing. The study followed the changes to writing approach that first-year undergraduates deployed before, during and after peer review workshops. The workshops consisted of: holistically reviewing and discussing examples, developing criteria, writing a practice essay, reviewing three peer essays and receiving peer comments. Case study analysis drew on self-regulated learning theory and triangulated self-report with textual and/or think-aloud evidence. Eight of the ten case writers made adaptations to their writing which could be traced back to workshop learning. Analysing examples sparked students’ initial adaptations to their writing approach in the workshop essay. They then refined, adjusted and added new adaptations in the post-workshops essay. Students made varied use of peer comments, which were not often new information as students had already reflected on writing. Only two writers demonstrated direct uptake of peer feedback information, implementing suggestions in their post-workshop essay. Yet comments also verified the success of strategies, boosted self-efficacy and confirmed priorities for next steps. Behind some apparent non-uses there were interim steps for development. The study contributes to understanding of what it means to use peer feedback.


Introduction
Research on feedback has increasingly focussed on the agentic role of learners (Sadler 2010;Boud and Molloy 2013;Van der Kleij, Adie, and Cumming 2019), to the point that it has recently been posited that we need to re-frame the whole concept of feedback around learners' metacognitive internal feedback as only learners themselves can generate change in their work (Nicol 2020).One key way in which learners have been afforded a more agentic role in feedback is through peer review.Research has identified a range of learning that students can take from peer review, such as an impact on performance (Li et al. 2020), enhanced capability to make evaluative judgements on work (Tai et al. 2018), and induction into professional practices and communities (Reddy et al. 2021).The process of peer review can involve multiple components: training on exemplars, drawing on criteria, giving feedback, receiving feedback, and then applying the results of learning to a redraft or new piece.The components of peer review often vary between studies, and there is still relatively little research separating out how students learn from different aspects (To, Panadero, and Carless 2022).
Broadly, two main strands of research have looked at how students learn from peer review.One strand has analysed the extent to which students implement peer feedback information.In open-ended written work, uptake of even good quality comments tends to be relatively low, with a third (e.g.van den Berg, Admiraal, and Pilot 2006;Patchan, Schunn, and Correnti 2016) or half of comments implemented in re-drafts (Walker 2015).There could be a multitude of reasons why students may not implement feedback, ranging from the specific features of comments (Patchan, Schunn, and Correnti 2016) to issues of power and trust (Panadero 2016).Other reasons are generalisable to all sources of comments, such as not knowing how to implement suggestions, struggling with agency, or viewing skills such as writing as fixed (Winstone et al. 2017).Much of the peer feedback research focussing on uptake of comments could be characterised as using a feedback-as-communication lens (Van der Kleij, Adie, and Cumming 2019) and so overlooks what the learner might have learned from other activities within the peer review intervention.
A further strand of research has identified a trend around students perceiving that they have learned more from giving feedback than from receiving comments (Nicol, Thomson, and Breslin 2014;McConlogue 2015;Gaynor 2020).The research into perceptions is complemented by experimental research showing that reviewing examples can lead to improvements in learners' own writing (e.g.Cho and Cho 2011;Cho and MacArthur 2011;To, Panadero, and Carless 2022).However, there can be much variation within a cohort as to preference -or indeed ambivalence -towards either giving or receiving feedback (Ion, Sánchez, and Agud 2019), and there are similar performance gains when students solely give or receive feedback (Huisman et al. 2018).The question may not be whether giving or receiving is better, but could be reframed as: what do learners take from different parts of the peer feedback process?
There is surprisingly little research that addresses the specifics of learning from different components of peer review.Recent studies have indicated that peer feedback can enhance self-assessment processes (Yan and Brown 2017;To and Panadero 2019).Nicol and McCallum (2022), isolating comparison as the mechanism behind internal feedback that generates future changes, asked students to engage in three cycles of composing feedback on peer work then self-reviewing their own essay, before receiving feedback.The self-reviews were comprehensive, such that three-quarters of the students had anticipated teacher comments in their self-reviews, and peer feedback had little to add.This almost suggests that received comments could be redundant as students' internal feedback was so rich.Yet students in this study still attributed learning to peer feedback, suggesting that the role of feedback may be something more than just providing information.
As the field continues to move towards process and learner-centred views of feedback (Van der Kleij, Adie, and Cumming 2019; Winstone et al. 2022), the present research seeks to add to knowledge of what it means to use peer feedback to improve writing in future work.The study followed writers over four months, tracing the changes they made in essay approach following a series of peer review workshops.Self-regulated learning (SRL) was used to examine students' process and adaptations for three main reasons.Firstly, SRL positions students as agentic decision makers, who weigh up feedback, 'filtering' it through their own beliefs about learning, their knowledge, and indeed the conclusions of their own internal feedback (Butler and Winne 1995, 264).Secondly, SRL focuses on learning as dynamic and developing.The model used in this study, Winne and Hadwin (1998), describes four recursive and overlapping phases: task perception, plans and goals, enactment and large-scale adaptation.The first phase, task perception, is crucial as fulfilling the assignment brief is essential for success and often easiest understood through examples rather than explanation as prior genre knowledge relating to abstract criteria can over-transfer (Sadler 2010;Reiff and Bawarshi 2011).The final reason for drawing on SRL is that the large-scale adaptation phase draws attention to the decisions learners make about which strategies to keep and which to discard next time they approach similar work, and which components of peer review play into these decisions.

Research questions
This study addressed the following research questions: • How do students make adaptations to writing during and after peer review workshops?• (How) do peer feedback comments contribute to students' self-regulation of writing?design-based research (dBR) methods were used to develop and evaluate a series of workshops for first-year undergraduates on writing timed essays.Case study analysis was used to address the research questions.Table 1 outlines the research timeline.

Workshop design
The workshops and study were developed through design-based research principles.A signature feature of dBR is that as well as addressing a common educational challenge, interventions are tailored to the requirements of the specific teaching context (Anderson and Shattuck 2012).The setting for the study was the first year of a BA(Hons) English Language and Literature programme at an élite university in England.In non-pandemic circumstances, summative assessment for the first-year modules is one take-home portfolio and three three-hour unseen, closed-book, essay-based handwritten examinations.The workshops focussed on two examinations with the same assessment criteria: Literature in English 1830-1900 and 1900-present day.The programme is writing-intensive with frequent essay assignments, but to prepare for examinations, students have to make a transition in genre and process.The system has been characterised as a 'constructive misalignment' between teaching and assessment (Horn 2013), offering academics and students the freedom to tailor the course to their interests, and pursue learning without the potential pressures of continuous assessment.Nonetheless, students find it challenging and stressful to be assessed in a different format to which they learn, raising concerns around inclusivity.
The workshops were designed to support students in the transition to writing examinations by offering the opportunity to analyse and discuss authentic examples of essays written in timed conditions, and to exchange anonymous peer feedback on a practice essay.Assessment was holistic, with no pre-specified criteria, following Sadler (2010), to prompt close reading, a key practice in the discipline, and to uncover how criteria are inter-related and overlap.Example essays were selected to represent a range of quality, so students could contextualise performance and see criteria in practice (Sadler 2010).Second and third year students shared with informed consent timed essays written for formative assessment.Students were paid £5 in recognition of their time and contribution.To select exemplars representative of common strengths and challenges, I interviewed four English tutors, using the same process as in the workshops to discuss a shortlist of essays and develop teaching notes to ensure consistent coverage in workshop discussions.

The workshops
Table 1 gives the timeline for the workshops and outlines the procedure.The features of note are: • For each example, students individually read the essay and wrote a response, focussing on what seemed to them to be most important to quality holistically.We then discussed the essay as a group, and after reviewing three essays, developed a list of potential criteria, which was updated in the final workshop.Coding of students' criteria and comments showed a high degree of overlap with programme criteria.• After analysing example essays and developing criteria, students completed reflections, connecting what they had noticed in the examples to previous essays they had written.• Next, students handwrote a timed practice essay: an immediate opportunity to apply learning.
• Students gave written comments on three randomly allocated peer essays, using the same holistic process, with the criteria in their own words to hand.To promote psychological safety, feedback was double-blind and comment-only.As training is important to the quality of peer comments (van Zundert, Sluijsmans, and van Merrienboer 2010), we also discussed guidelines and ground rules for feedback.We examined the differences between a personal and critical response arising from a holistic reading versus reader-oriented feedback.Students were encouraged to channel their holistic response into helpful, selective and actionable feedback comments, and to aim for a constructive tone.• After writing feedback, students used prompts to reflect on what they had noticed about the peer essays.They then re-read their own essay and reflected again.• In the final workshop, students received written comments from three peers.Students reviewed one further example and discussed possible strategies for writing development.
• At the end of each workshop, students completed open-ended 'reflections' questions to make connections with their own writing, and develop metacognitive awareness.

Participants and data generation
Participants were 14 women and 7 men who volunteered to take part in the workshops in addition to usual teaching.Two participants were international students and all students reported that English was their first language.Students were invited to participate through a purposive sample of tutors, who taught the relevant modules to tutor groups of 2-10 students.The 7 tutors had themselves all completed a programme on teaching and learning and were supportive of educational research.10 of the students (8 women and 2 men) were case study participants, a convenience sample of students from the first three tutor groups to recruit to the study in the term before the workshops.As outlined in Table 1, case study participants took part in 1:1 pre-and post-workshop research meetings, involving writing a one-hour practice essay, thinking aloud for the planning stage -followed by an interview both on the process of the essay just written, and writing more generally.The research meetings lasted up to two hours, and participants were paid £20 in recognition of their time and contribution.In addition, I gathered copies of the material all 21 participants produced during the workshops: annotations and responses to example essays, reflections, practice essays, feedback and responses to feedback, as well as a final questionnaire.The dataset totalled 251,583 words.All participants gave their informed consent and the study was approved by the University's Central Research Ethics Approval Committee.All names are pseudonyms, some self-chosen.

Data analysis
To examine how participants self-regulated their writing over time, I used case study analysis.data were transcribed and then coded deductively in NVivo 11 with a guide based on the phases and aspects of SRL in the Winne and Hadwin (1998) model (Miles, Huberman, and Saldana 2013).As coding fragments the data, I then hand-mapped and visualised each case study participant's three writing episodes, arranging coding output to make it easier to trace progress and connections across pre-, during-and post-workshop writing.Throughout the process, I used analytical memos to draw out within and cross-case themes, and to interrogate my own reflexivity.
The analysis reported here takes as a starting point students' post-workshop interviews, cross-referenced with workshop data to corroborate perceived influence of workshop learning (annotations on example and peer essays, reflections on own writing, feedback to peers, responses to feedback).Interview data were also cross-referenced with essays and think-aloud data to identify evidence of changes.Starting with interview data honours the intentionality of participants as self-regulating learners, while triangulation serves to check memory effects and build a rich picture of how students used workshop learning, but without assuming that this is unproblematic, or complete (Hammersley 2008).An example of the case study analysis process can be found in the supplemental online material.

Changes to writing after workshops
The first research question asked: How do students make adaptations to writing during and after peer review workshops?In their post-workshops essay, eight of the ten case study participants made adaptations to their writing approach which they attributed to workshop learning and could be evidenced through case study analysis.The remaining two students had started to make changes beforehand, and the workshops served to confirm that they were on the right track.
The eight who adapted writing each reported 2-3 changes in approach to their post-workshops essay which could be corroborated with textual and/or think-aloud evidence.In order of frequency, changes included: an increased focus on the question, narrowing scope to ensure they finished in time, signposting, modifying introduction moves and experimenting with planning.In interviews, students attributed the changes to be motivated by what they had noticed about successful examination essay writing in the workshops.To check whether and how workshop activities had raised students' awareness in the areas in which they then made changes, I cross-referenced with annotations on example and peer essays, additions made in different coloured pen during discussions, reflections written at the end of workshops and peer feedback.For further details of the adaptations students made, together with an example of the analytical process, see the supplemental online material.
All eight students started adapting their writing in the workshop essay, which was written after analysing four examples and reflecting on writing they had done prior to the workshops.Analysing and discussing the exemplars was the initial catalyst for making adaptations.These adaptations were then often adjusted and refined in the post-workshops essay, as writers drew on what they had learned from the process of writing the workshops essay, reviewing peer essays, receiving feedback and continuing to reflect on work.
The second research question asked: (How) do peer feedback comments contribute to students' self-regulation of writing?Beginning with the feedback information itself, in reflections and interviews, students overall reported comments to be helpful.Sometimes different reviewers gave different perspectives, but having multiple reviewers meant they could spot outlying views.Suggestions made by multiple reviewers were compelling.Comments were extensive: the mean length, from the three reviews combined, was 496 words (Sd = 90.97).Coding showed that feedback was rich in solutions and well aligned with the criteria students identified in example essays.
Yet only two students made changes which were direct implementations of suggestions made by reviewers.Chloe and Megan both reported using feedback to increase their use of signposting in post-workshops essays.They were aware of the issues which the reviewers suggested signposting would resolve, but signposting itself was a new strategy.Other participants attributed changes to peer feedback, but similarly to Nicol and McCallum (2022), they had actually self-suggested these actions in reflections before they received feedback.Feedback confirmed rather than generated actions.
Feedback, then, didn't provide completely novel information.Students came into peer feedback having already made judgements on their own essay on two occasions (immediately post-writing and re-reading their essay after providing feedback).All eight had started making adaptations in the workshop essay that were linked to comments and discussion on the example essays, and reflections on their own writing.Much of what peers suggested confirmed students' own judgements or further raised awareness.Heather explained in the post-workshops interview: I thought [peer feedback] was really good because they all made really similar points and it was both things I understood about my writing and I was like, "Oh, yeah, you're right about that.You only read one of my essays, but you're right" -and also things that I didn't really realise that, like, slightly deconstructed… what I did.And so I was sort of like, "Oh, so that's why I do it" or "That's how I can not do it", so it was like new ideas about things that I was already aware of.
As Heather describes it, feedback complemented and enriched students' metacognitive awareness of writing.Feedback itself didn't necessarily provide brand new information, but students did draw on it as they decided how to adapt next.The next sections show how students 'amalgamated' , as Ben put it, their own judgements with insights from peer feedback.Students used feedback to confirm and clarify their priorities for future work.Feedback was also helpful for highlighting the strengths of work -what students should continue doing.For some, this was new information, and served to boost their self-efficacy.The non-uses relate to how, as self-regulating learners, students integrated feedback information with their own beliefs and goals to make decisions about what to do next.For some students, suggestions were not yet realistic, or misaligned with how they wanted to write.Only in one instance was some feedback outright unhelpful.

Use 1: Feedback 'verifying' success of writing strategies
For two writers, peer comments confirmed that new strategies they had tried in their workshop essay had worked, and they should continue with them.The introduction is a key site of differences between the tutorial and examination essay genres, and both Heather and Tessa introduced new moves into their peer assessed essays.They both then received feedback that highlighted the introduction as a strength, and so continued with these changes in their post-workshops essay.
More widely, participants valued comments which offered 'verification' , to use Phoenix's term, about which aspects of writing worked, and should be continued.As Tessa explained in the post-workshops interview: It was quite nice as well, where it sort of showed you things that you did well, because then you could sort of see 'OK, this is obviously working, I'll try and keep that' .
It is easy to assume that feedback should always provide next steps, but taking stock and deciding what doesn't need to be adapted is also important for self-regulating learning, too, particularly under time pressure.

Use 2: Feedback 'boosting' self-efficacy
For some students, feedback about the strengths of their work was new information and provided a 'boost' , as Nicole described it in reflections.In contrast to her reviewers, Nicole 'focused mostly on the negatives in my own review' , which reflected the general tendency among the participants in this study.For instance, it was difficult to elicit answers to the interview question that asked about the strengths of the essay they'd just written; participants found it much easier to talk about what they would do differently.
In interviews, some students expressed how the 'boost' feedback provided was to their beliefs about their capability to write a successful examination essay: their self-efficacy (Bandura 1997).Felicia found her feedback was 'quite encouraging' , showing her that 'actually, I can do this' .This last phrase is echoed by Chloe, who was 'so surprised' that her feedback highlighted 'strengths' in her work: Chloe: The strengths of my writing, I wasn't aware that there were strengths [laughter] so that was nice.Umm but it was probably just on this particular essay, but yeah.So that was [trails off ] Researcher: So it was nice to have someone point out what's good about it?
Chloe: Yeah, it was like 'you can do this' , sort of like 'oh really?' [laughs] Here, Chloe doesn't appear to be entirely persuaded by her reviewers -she immediately downplays the presence of strengths, and her 'oh really?' conveys her surprise at the reviewers' message that 'you can do this' .In self-efficacy theory, 'social persuasions' such as positive feedback are a relatively weak influence on self-efficacy (Bandura 1997), compared with stronger evidence of capability, such as a mastery experience, like having written a successful essay previously.Chloe did not seem to have mastery experiences to draw on, stating, 'I'm clearly not used to getting good feedback' , which hints that previous feedback from tutors may have been quite robust.However, the positive feedback becomes more credible as Chloe had read example and peer essays.As research on exemplars has found (e.g.Hawe, Lightfoot, and dixon 2019), discussing others' work served to calibrate her expectations of what is achievable in timed conditions.When asked in the interview what she had learned in the workshops, Chloe's first point was that 'I learned that I'm really critical, both of my own essays and other people's' .This awareness allowed her to contextualise feedback and realise that her own judgements were overly critical.Thus, feedback can have a more powerful impact, and the message that 'you can do this' becomes more credible.

Use 3: Feedback confirming what to do next/judgements about priorities
For three students, feedback confirmed their judgements about what they should work on next.They then made these changes in the post-workshops essay.For example, Nicole, who had struggled with time, decided to use shorter paragraphs in order to move on with her argument more quickly.In the interview, she attributed this strategy to feedback.However, analysis showed she had already considered using shorter paragraphs in reflections written after giving peer feedback and re-reading her own essay.Her decision to adapt her paragraphing arose from reviewing others' work, but the attribution suggests that peer feedback cemented it.

Non-use 1: Suggestions as yet too ambitious
In some cases, writers chose not to act on feedback despite agreeing with it.The first reason for non-use was that suggestions seemed too ambitious.All three reviewers of Phoenix's essay commented on the need to develop the conclusion, and he agreed -as shown in the underlining and annotations in Figure 1.Phoenix was already somewhat aware of this issue, commenting in post-writing reflections: 'I felt it was a little rushed at the end' .The feedback served to underline the importance of the conclusion, and it is the first thing he mentioned in the interview a month later: 'completely true about the need for a conclusion, and I think that is a big thing' .However, he didn't manage one in the post-workshops essay, and tellingly noted that 'I don't think I've ever written an essay where I've been able to write a nice conclusion at the end' .
Nonetheless, Phoenix did make a related adaptation: he 'narrowed… down' his material, jettisoning 'some of my lovely points' .This was a prerequisite to making space and time for a conclusion.In the interview, hypothesising what he would do with an extra ten minutes, Phoenix voiced conflicting priorities.While he first stated that he would have 'written a conclusion' and edited the introduction, he also acknowledged that: But saying that, I know me and I also want to try and fit in as much as I can, so I might be stupid and try and put in another point.But I hope by now I realise that that's not a good idea.
Phoenix's comment highlights how students can experience conflicting standards, and particularly the strong pull of those rooted in enjoyment in the subject.Such standards may be long-held and anchored in old writing habits.Here, he is torn between writing about as many of 'my lovely points' as possible and writing a conclusion to meet his readers' expectations.With metacognitive knowledge of his writerly tendencies -'I know me' -he can choose which to follow.'Narrowing down' material is an achievable first adaptation towards writing a conclusion, and so meeting readers' expectations.

Non-use 2: Suggestions 'compromise' writing style
Phoenix, then, although amenable to feedback, retained strong ownership of his writing.This was also the case for Holly, one of the two participants who didn't make adaptations linked to the workshops.In her interview, Holly thought her feedback was 'insightful' and appreciated how her reviewers 'analysed it in depth' .All three converged on the need for further signposting in the introduction.In post-feedback reflections, she wrote that 'this is definitely worth noticing because it is something an examiner might pick out' .However, she decided not to act as it was beyond what she was comfortable with.Prior to the workshops, she had already begun using some signposting in her introduction -signalling her argument in a sentence beginning 'this essay will explore' .This move was not solely for the reader: it worked as a 'pointer' to clarify the argument for her, too, but still felt 'a little bit awkward' even post-workshops.So, Holly did not feel willing to add the move her reviewers requested to signpost her texts: 'that's still further away from my writing style than I'd like it to be' .She explained that making too many changes was risky: I think also if you stray too far away from your writing style, it can kind of compromise the rest of your essay.It's better to try and stick with yourself, and it's not necessarily hundred percent what other people want, but then it's never going to be.
Holly voices a limit as to how much she would adapt her writing style, recognising that it may not be possible to please completely any reader, even an examiner.Both Holly and Phoenix's cases, then, demonstrate that changes which may seem straightforward to reviewers -include a conclusion, try more signposting -can often seem bigger to students, particularly when filtered through personal beliefs and standards (Butler and Winne 1995).Change is often incremental and slow, and there may be intermediate steps that students need to take in order to reach that ultimate change -if they are willing to 'compromise' their style to do so.

Non-use 3: 'blunt' and cutting' comments without actionable suggestions
Only one writer described feedback as outright unhelpful.Ben received 'awful' , 'cutting' comments from one reviewer around clarity of argument.Ben was aware in post-writing reflections that he didn't have a clear thesis; this was confirmed in his reflections after reviewing peer essays and corroborated by other reviewers.However, he was 'affected' by the tone of the unhelpful feedback: It's just really blunt.I -it's really odd that it was the tone that affected me more than the comments, because the comments were right, it's just the… the way something is articulated makes such difference.Umm [pause, reading].Yeah, so 'several points were not properly explained and therefore risk subverting your argument' … I was like, OK, sorry!Here the symmetrical peer-to-peer relationship is disrupted, with the 'blunt' comments forcing Ben into a position of apology: 'OK, sorry!' Solutions are not given as suggestions but as imperatives, mirroring findings in the wider literature that a harsher, more directive tone is used for weaker work (e.g.Wingate 2010).This work is ungraded -it cannot be explicitly positioned as weak -and the reviewer is not a teacher.Yet through tone, the reviewer constructs themselves as an authority, passing judgement.The writer is positioned as unknowing, ignorant of their own writing, when Ben was already aware of many points.
Such feedback that invokes final, absolute judgements is emotive and risks undermining the agency of the recipient (Boud 1995).In this instance, as Ben had received multiple reviews, he was able to contextualise the 'blunt' feedback as one reading: others were more 'complimentary' or at least 'a lot easier to deal with' .He also explained how over his first year, he had learned to extract what was useful from critical feedback: Although… it's cutting on a personal level, it feels like it's cutting on a personal level, you've got to realise that it's about the text, it's about the essay.And you've got to take the points out of that in the same way that you've got to take, you've got to remove yourself from all the feeling of how good you are from other people's responses and find the bad things and work out how to improve this.
Managing affect by reappraising feedback to be 'about the essay' is an important component of feedback literacy (Carless and Boud 2018).It was still up to Ben, though, to 'work out how to improve' as while the feedback-giver took an authoritative stance, they did not actually suggest any adaptations to make in the next essay.The post-workshops think-aloud protocol showed Ben focussing on formulating a thesis statement and re-ordering his material before beginning to write.This adaptation drew on what he had learned from reviewing example and peer writing, and reflecting on his own, and was confirmed by others' feedback.As such, Ben was able to move forward not just because of his feedback literacy, but because he had other sources of information to draw on.

Discussion
The current study examined how students made adaptations to writing during and after peer review workshops, and their uses of peer feedback comments.Eight of the ten case study writers made 2-3 self-reported adaptations in post-workshops writing, which were confirmed with textual and/or think-aloud evidence.Beyond direct uptake, six further uses and non-uses of feedback were identified.
The first contribution of the study is that it adds to knowledge about the interplay between two typical components of peer review: reviewing work and receiving comments.Students began adapting their approach in the workshops essay, and only two students made post-workshops adaptations which originated solely from peer feedback.Information was not often new to students.The anticipation of peer comments in this study chimes with Nicol and McCallum (2022), where students iteratively reviewed work and then reflected on their own, receiving peer comments after three rounds.For only around a quarter did peer comments add something they hadn't noticed.As Nicol and McCallum (2022) argue, this suggests the internal feedback students generated from comparing their work and others was comprehensive.does the high quality of internal feedback make peer feedback redundant, then?Not necessarily.In both studies, students reported that they had learned from peer comments.As Nicol and McCallum (2022) suggest, it may be that students over-state the importance of feedback due to memory effects.But in this study, with so many things to notice and potentially improve about writing, peer feedback cemented what was worth noticing and helped students to prioritise, extending findings that peer review can help students understand quality (To and Panadero 2019).Writers did put feedback to a range of uses, even if these were not direct uptake.
The second contribution of the study, then, is that it adds to knowledge about what it means to use feedback when seen through a learner-centred, process-oriented lens (Van der Kleij, Adie, and Cumming 2019; Winstone et al. 2022).Not so long ago, the finding that only two writers directly implemented peer comments might have been seen as a failure of feedback.But this study has shown how, by participating in peer review processes, students' own judgements became the primary drivers of changes to writing approach, with peer comments put to complementary uses.A specific benefit for this group was that peers noticed the strengths of their work, which brought two implications.Firstly, while in feedback research the focus is often on change, strategic large-scale adaptation of SRL can be about what not to change, too.With just a month between the workshops and examinations, feedback that 'verified' (Phoenix) what worked enabled students to decide what to keep, and so prioritise.It is striking that the writers converged on 2-3 adaptations; alongside subject knowledge revision, this seems to be what was realistic in the time.Secondly, peer comments appeared to give an important 'boost' (Nicole) to self-efficacy for some, helping them realise that 'actually, I can do this' (Felicia).It may be that highlighting strengths was particularly important for this group: first-year students taking high-stakes examinations at an élite university.Research into exemplars has found similar self-efficacy benefits in making standards seem more achievable (Hawe, Lightfoot, and dixon 2019).Future peer review research should incorporate dimensions of learning such as self-efficacy to continue to build a richer picture of how different components of peer review contribute to self-regulated learning.
A third contribution of the study is that it sheds light on the complexities around non-uses of feedback.Contrary to other research, non-uses of feedback comments were not due to lack of agency or volition (Winstone et al. 2017).In one case, some feedback was 'blunt' and unhelpful, but the writer nevertheless made adaptations, drawing on feedback literacy to mitigate the emotional impact (Carless and Boud 2018), but also other sources of feedback, including, importantly, his own judgements.Peer review can sometimes seem a risky, difficult to manage activity for educators.An implication of this finding is that feedback that goes wrong may not be a barrier to adaptation if peer review is set up to empower students to understand quality and reflect on their work.Other non-uses related to writers agreeing with but discarding recommendations, either because they needed to make an interim developmental step, or it was too distant from their preferred style.In both cases, students had a strong sense of ownership of writing.It seems important that success isn't too tightly defined, lest peer review becomes too conformative, and enjoyment and empowerment disappear (Torrance 2012).
A further practical implication of this research is that it can inform educators' decisions about which components of peer review to deploy when.For methodological reasons, the workshops designed for this study ran over a single intensive period, which could be challenging to integrate.Yet different components of peer review can be embedded into programmes at different stages depending on students' needs (Boud et al. 2018).Since the research, a team have scaled up the design, offering all students workshops at three transition points across the English Language and Literature programme, informed by the finding that participants were able to begin to adapt their approach in the workshop essay.Within time and resource constraints, the workshops prioritise holistic analysis of examples as students have later writing and feedback opportunities.In other situations, peer feedback can offer practice, exposure to peer work, and peer comments without significantly adding to teacher workload (e.g.Nicol and McCallum 2022).Future research should evaluate which components of peer review are best deployed at which stages of programmes.
Turning to the methodological limitations of the study, as a first stage of dBR, the design deployed a single, intensive intervention with a small sample of students who volunteered, and the study period was limited to four months.These limitations arise from the fine-grained focus of the study, used to trace the adaptations students made to their approach.Future peer review research could beneficially work with larger cohorts over a phase or whole programme, following a purposive sample of case study learners in some detail while addressing wider questions such as whether such interventions promote equitable outcomes.
A further methodological point of note is that the study started from self-report, and then linked participants' adaptations and attributions back to what they had noticed, commented on and produced during the study activities.However, as Nicol (2020) has argued, comparisons with one's own work can arise from many different sources in the learning environment beyond planned activities such as peer review.Thus, a methodological limitation is that the study design did not capture students' learning outside the workshops.Future research should try to capture both planned and unplanned influences on students' development of understandings about quality, for example through a learning journal, so as to identify how different inputs to internal feedback interact.

Conclusion
This article has investigated how students adapted their writing during and after peer review workshops, drawing on multiple sources of evidence.Students began adapting their writing after analysing examples and reflecting on their own work.Peer feedback was not often novel, and students 'amalgamated' peer comments with their own judgements, using comments to verify the success of strategies, boost self-efficacy and confirm priorities for next steps.Non-uses were more nuanced than outright rejection of feedback.The study contributes to knowledge about what students take from different parts of the peer review process, and what it means to use peer feedback.

Figure 1 .
Figure 1.Feedback to Phoenix on his conclusion, plus his annotations.