Principals’ views on the implementation of grade-free middle schools in Norway: justifications, challenges and opportunities

This article reports on four principals’ views on the implementation of so-called grade-free middle schools, i.e. schools that drop all grades on students’ performances except the two required by the national assessment regulation, as part of their work with Assessment for Learning (AfL). More specifically, we were interested in the under-researched area of how principals justify introducing and implementing grade-free schools, and what their experiences are regarding challenges and opportunities that have arisen during and as a result of the implementation. Semi-structured in-depth interviews were used to collect data. Findings suggest that principals rely on research and unsatisfying assessment practices when justi-fying a change to grade-free schools. However, they do not find the involvement of students, nor the information directed at parents about the implementation, to be sufficient. They also mention challenges related to the current assessment system, which they believe underpin a behavioristic understanding of learning. The article calls for more research on trust among various stakeholders and student involvement when implementing grade-free schools. og implementere karakterfrie skoler, og om deres erfaringer når det gjelder utfordringer og muligheter som har oppstått i løpet av og som et resultat av implementeringen. Sem-istrukturerte dybdeintervjuer ble anvendt til å samle inn data. Funnene antyder at rektorene støtter seg på forskning og utilfredsstillende vurderingspraksiser når de legitimerer implementering av karakterfri skole. Imidlertid mener de at elevinvolvering og informasjon rettet mot foresatte om implementering ikke er tilstrekkelig ivaretatt. Rektorene nevner også utfordringer knyttet til dagens vurderingssystem som de mener bygger opp under et behavioristisk læringssyn. Studien impliserer mer forskning på tillit blant ulike aktører, og elevinvolvering i forbindelse med implementering av karakterfrie skoler.


Introduction and context
In 1984, Lysne claimed, partially wrongly, that "it does not seem possible to reduce dramatically or abandon the use of marks in our schools" (p. 164). More and more schools are doing precisely this in Norway, dropping or downplaying grades on students' performances (Ertesvåg & Haegeland, 2018;Mellingsaeter, 2018). To our knowledge, there is little research on grade-free schools' assessment practices. In this study, we asked four principals about their justifications for introducing and implementing grade-free middle schools, what their experiences are when it comes to challenges, and opportunities that have arisen during and as a result of the implementation. We were also curious about success criteria they would recommend to others wanting to implement grade-free schools.
One issue that has caused public debate (Ertesvåg & Haegeland, 2018;Mellingsaeter, 2018) is the extent of grading students' performances -or whether grading should take place at all. Little is known about principals' views on this matter, who indeed have a decisive role in implementing, developing and sustaining Assessment for Learning (AfL) practices (Smith, 2011). There have, however, been several feature articles in the media, where students, teachers and principals express positive experiences as a consequence of dropping or downplaying grades. Less stress, less comparisons with peers, more useful written feedback and more stable effort throughout the school year are some of the responses from students (Mellingsaeter, 2018).
Development work and classroom research into AfL internationally have influenced the Norwegian context of AfL to a large extent over the last 10 years. There have been numerous projects, seminars and workshops aiming at increasing schools' competency in implementing AfL (Birenbaum et al., 2015). Assessment regulations of the Education Act were introduced in 2009 and revised in 2013, which emphasized "good feedback" on continuous assessments, supervision on students' work in order to advance their learning, and student self-assessment. Teachers are not allowed to grade students' work before lower secondary level, i.e. Year Eight. From Year Eight, students have the right to mid-and final term grades. Thus, when we refer to "gradefree schools" who "drop" providing grades on students' work, it means that teachers provide only mid-year-grades and end-of-the-year-grades, in accordance with the national assessment regulations (Regulation of The Education Act, 2013). Since the launch of the national program Vurdering for laering [Assessment for Learning] in 2010 (Ministry of Education and Training, 2010), more than 300 (out of 422) municipalities have participated in various forms of implementing AfL. A national survey (Larsen, Vaagland & Federici, 2017) concluded that school principals believe they have an important role in facilitating competence building in assessment and they are positive toward AfL implementation, but their responses vary across municipalities. With this as a backdrop, the research question for this study is as follows: What are principals' views on implementing grade-free middle schools?
In the following, we elaborate on the theoretical framework in which the implementation of grade-free schools has taken place, before reflecting on the principal's role in leading AfL implementations.

Testing, AfL and the issue of grading
At the beginning of 1990, Gipps strongly advised a shift from a testing culture to an assessment culture in her book Beyond Testing (Gipps, 1994). A testing culture is based on a behavioristic and cognitivist theoretical foundation, whereas an assessment culture is based on sociocultural theory (Gipps, 1999;Pryor & Crossouard, 2008). The process, not the product, is at the center of an assessment culture within a sociocultural theory of learning and development (Gipps, 1999;Lantolf & Thorne, 2006). Assessment and learning are thus intertwined, as elaborated on in theories of dynamic assessment. In dynamic assessment, the emphasis is on understanding the sociocultural interactive relations, and the actual and the potential learning zones (see Poehner, 2008Poehner, , 2011. The social relations in which students and teachers interact are the primary force of development (Poehner, 2008). When principals implement grade-free schools, they have to cater for the social relations that affect student assessments, which in turn facilitate students with competencies required outside of school.
The relations between assessment and learning (Gardner, 2012) were synthesized in a review article in 1998 by Black and Wiliam, in tandem with development work and other research publications by the Assessment Reform Group (2002) which docu mented the positive effects AfL has on students' learning. The widely-used definition of AfL, "all those activities undertaken by teachers, and/or by their students, which provide information to be used as feedback to modify the teaching and learning activities in which they are engaged" (Black & Wiliam, 1998, pp. 7-8), put feedback at the core of AfL, as pointed out by Butler (1987, p. 475): "The normative grades prevalent in schools seem a clear example of information that focuses attention on the self by emphasizing outcome and social comparison, or both, rather than process or task mastery". Grading in secondary education goes as far back as the first half of the 1800s and discussions about grading or not in the political debate in Norway date back a hundred years from now (Lysne, 1984). The controlling function of grading was prominent before the guidance and counseling function became more important, after the 1930s. In her oft-cited experimental study on feedback information and students' motivation, based on 200 students in Years Five and Six, Butler divided students into four groups: one group received individual comments on their product, another group received grades (in the form of a number), a third group received nonspecific praise ("very good") and the last group received no feedback on their products (Butler, 1987). Pre and post tests were conducted in order to establish students' performance on tests and their task and ego-involving motivational orientations. The results confirmed the hypothesis that individually-tailored comments yielded higher performance and higher task-involving motivational orientation compared to all of the other three forms of feedback information or lack of such, and that students were more ego-involving motivated and their performances decreased when receiving grades or non-specific praise. Furthermore, Butler's study not only confirmed other research findings stating that low performing students were the ones suffering the most from ego-involving feedback information, such as non-specific praise or grades, but also showed that high-performing students are affected negatively. In other words, ego-involving feedback information -referred to as "external markers" by Black and Wiliam (1998), i.e. stars, smileys, scores and numbers -are most probably counterproductive to intrinsic motivation which, in turn, is at the core of AfL and students' self-regulation (Black & Wiliam, 1998, 2009Butler, 1987;Hattie & Timperley, 2007). Research on students' self-regulation of learning (see Schunk & Zimmerman, 2011) is in line with Hattie & Timperley's (2007) and Shute's (2008) descriptions of useful feedback in AfL, showing the close connection between learning and AfL.

Principal's role in leading AfL implementation
In this article, we consider school leadership and implementation of AfL within Peter M. Senge's concept of learning organizations (Senge, 1992). Senge (1992) identifies five core disciplines that enable organizations to be learning organizations. 1) Personal mastery; 2) Mental models; 3) Building shared visions; 4) Team learning, and 5) Systems thinking.
Personal mastery relates to continually clarifying what is of importance to us, not only our visions but also our purpose and, at the same time, gaining a clearer understanding of what reality is. Senge describes it as "The ability to focus on ultimate intrinsic desires, not only on secondary goals […] It is a process of continually focusing and refocusing on what one truly wants, on one's visions" (Senge, 1992, pp. 148-149). Moreover, Senge argues that one cannot be forced to develop personal mastery. The development of personal mastery is an ongoing process, and the leadership must communicate and model that personal growth is valued in the organization. Principals can work to develop a climate where staff and students can create visions and dare to challenge the status quo.
Mental models refer to internal images and tacit understandings of how things work. Although our mental models are carried in our minds, they contribute to how we act, and they make our perceptions selective. Furthermore, mental models are simplifications and can serve as barriers in terms of the extent to which we are open to unfamiliar ideas that challenge our mental models which, in turn, can affect our willingness to change. A learning organization should strive to recognize and articulate existing mental models, and balance inquiry and advocacy in order to promote learning. It is reasonable to assume that when introducing and implementing gradefree schools, the initiator could face resistance from colleagues and other stakeholders rooted in mental models.
Shared visions in Senge's vocabulary refer to what the organization wants to create, and are described as a powerful force. Shared visions are what motivates us to learn and are thus vital in a learning organization. Shared visions have their starting point in personal visions, which are an important component of personal mastery. This suggests that shared visions cannot be a top-down project as they emerge from personal visions. Like personal visions, one cannot persuade someone into sharing a vision; they have to choose it themselves. Senge points out that people sharing a vision may be willing to take greater risks if failing the burden is shared among numerous individuals. Moreover, the vision has to be anchored in governing ideas, described by Senge as the what (what do we want to create); why (what is the purpose) and how (how do we want to act when working towards our vision). In the Norwegian school regulations, the purpose or objective of education is stated in paragraph 1-3 of the Education Act for elementary and secondary education. To exemplify, one can say that a principal can have a short-term goal of an average grade of 5 on final exams in mathematics or to eliminate bullying. However, the purpose of practices taking place in school is, among others, to see to it that students "develop knowledge, skills and attitudes so that they can master their lives and can take part in working life and society" (cf. the introduction of the Education Act).
Team learning builds on shared visions and personal mastery, and it refers to "the process of aligning and developing the capacity of a team to create the results its members truly desire" (Senge, 1992, p. 236). As Senge points out, shared visions and personal mastery are not enough; the team must also work together as a team, which is a critical step when building learning organizations.
Lastly, systems thinking refers to the four above-mentioned core disciplines and how they are interconnected.
Regarding the process of implementing AfL changes, the benefits are well documented (Black & Wiliam, 1998, 2009Black et al., 2003;Hattie, 2009). 1 However, the methods and experiences of implementing AfL in classrooms and schools have varied across and within nations, regions and municipalities (Birenbaum et al., 2015;Hopfenbeck, Tolo, Florez & El Masri, 2013). A systematic review of prerequisites needed to be considered in implementation of AfL (Heitink et al., 2016) underlines the principal's important role in facilitating implementation, focusing on schoolwide AfL culture of collaboration, scheduling AfL into practical activities, providing professional development, and making available time for preparing and carrying out AfL practices. Systematic studies of AfL implementation in Norway, including school leaders' views, suggest the following success criteria: teachers collaborating and sharing knowledge in teams, spending time on involvement and participation of the various stakeholders, providing good explanations to the teachers about the reasons behind AfL implementation, clear communication and trust at different levels, focus on the students' voices, deep knowledge of what AfL is, and keeping focus on AfL over time (Hopfenbeck et al., 2013).
Teacher and student perspectives have been quite well represented in the research on implementations of AfL. However, less is known about the principals' views (Smith & Engelsen, 2013). A relevant study from Norway on two principals' views of AfL implementation was conducted by Smith and Engelsen (2013). The principals in their study expressed positive views about AfL implementation and witnessed substantial changes in students' development of assessment literacy by knowing more about AfL and feedback. At the same time, it was a challenge for the teachers to adjust their AfL practices to the extent that was expected by their students. The study mentions some success criteria for the implementation of AfL: that the principals are part of the implementation from the beginning, that they are learners just like the teachers, that they create opportunities for the staff to implement AfL, that they motivate and make an atmosphere for trying out new things, that they manage and organize teachers' time, and that they make sure new teachers are acculturated into the AfL practices (Smith & Engelsen, 2013).

Method
In order to get rich descriptions of the perceived situation, and to gain an understanding of the informant's life world, we conducted semi-structured in-depth interviews.

Sample
We conducted an online Google search for grade-free middle schools. In order to obtain valuable information, we asked the principals if 1) the school had been giving grades at an earlier stage 2) the school is currently a grade-free school 3) the current principal had been responsible for the implementation of the grade-free practice. Four principals from four different schools matched the criteria and accepted the invitation. The schools were all located in urban areas with mainly middle-class residents. School 1 had 590 students and has been grade-free since 2016, School 2 had 220 students and has been grade-free since 2014, School 3 had 400 students and has been grade-free since 2014, and School 4 had 270 students and has been grade-free since 2008.

Data collection
An interview guide was developed (see Appendix 1). It served as a useful frame for the conversation that provided structure and progression in the interviews. The interviews were carried out in line with Kvale and Brinkmann's (2009) twelve aspects of the qualitative research interview. They were carried out and recorded in the principals' offices and lasted 60-90 minutes each.

Data analysis
The interviews were transcribed in NVivo 12 and were analyzed using Strauss and Corbin's constant comparative method of analysis with open and axial coding and categorization (Strauss & Corbin, 2014). The interview transcripts were read in their entirety and openly coded in NVivo. Similarities and differences between the emerging concepts, through constantly comparing them (axial coding), led to the following main categories: -Justifications and opportunities regarding terminating the practice of giving grades and concentrating on AfL, hereunder the principals' understanding of AfL and grading. Also the possibilities grade-free schools provide for students' learning and well-being were included. -Implementation and involvement: the principals' actions and who were involved (stakeholders), including points of improvement as to how this process could have been carried out differently. -Challenges the principals identify during the process of and after the implementation. -Success criteria: the preconditions for successfully implementing and further developing the practice described.
When similarities and differences between the categories were mapped, we once again turned to the informants' original statements in an attempt to gain a better understanding of their life world. This alternation between isolated statements, statements in the context of the whole interview and statements in the context of other similar or different statements in a constant comparative mode revealed new qualities in the data which, in turn, opened up for a more in-depth understanding.

Ethical considerations
The project was carried out in accordance with the Guidelines for Research Ethics in the Social Sciences, Humanities, Law and Theology (The Norwegian National Research Ethics Committee, 2019). As the project involved personal information about the informants, The Norwegian Centre for Research Data (NSD) was notified. The NSD approved of how the processing of personal data in our project was handled and how the principals were informed. All the informants gave their written consent to participate.

Findings
In the following, the four principals' views on grade-free schools will be presented according to the above-mentioned categories.

Justifications and opportunities
All four principals refer to research as one of the main justifications for implementing grade-free schools. Students' distraction when receiving grades, the ineffectiveness of grades for learning and the poor quality of feedback due to grading were issues that were raised by the principals: There was so much focus on that number [grade] and comparisons … Even though we talk and talk about this, we hear parents … and one can understand it, because it [the grade given as a number] is safe, good and very concrete; it's much easier to face your child receiving a 4 in Social Studies than practicing a skill called "discussion", but then we need to be much more concrete in the feedback -something we have struggled with … There is a solid appreciation of psychological stress and many [students] with high shoulders here; 2 that was also part of the background [for implementing grade-free school] (Principal 1).
More focus on learning, less on numbers. We find support in research by Dylan Wiliam [renowned formative assessment researcher], which says that if you assess with grades it often disturbs learning because the student will be so preoccupied with the number -that that will be the most important thing. And he [Wiliam] has conducted research on students where comments only, comments with grades and grades only were provided, which reveals that when assessment is given without grades, the effect of learning is much higher than providing assessment with grades, and providing assessment with grades is almost like providing only grades. It [the grade] is a disturbing factor… We have experienced more focus on learning and the development of students' competencies, that students think, relax more, and suffer less mental stress (Principal 3).
In the same vein, Principal 2 emphasized the teacher perspective, saying that they experienced that the "feedforwards" they provide to the students were not read, that the students only looked at the grades and compared them with peers, and that the teachers felt that grades were an imprecise assessment of students' work. The principal referred to episodes where students had the same test in the same subject and received the same grade but had responded totally differently from each other, and that the teachers did not think this was a good practice. Thus, "they [the teachers] agreed and asked themselves what would happen if the teacher team tried out grade-free assessment practices?" Principals 3 and 4 described another type of justification for implementing gradefree schools. When they commenced their role as principal, teachers welcomed grade-free implementation. One of them had earlier worked at another school where she had been through the same process.

Principal's understanding of AfL and grading
Principal 1 expressed a lack of trust in the current grading system, criticizing it for supporting a behaviorist theory of learning: The concept of assessment for us in Norwegian schools is a lot about receiving a grade in the form of a number, you know, but the guidance that is involved in AfL is about something else. In any case, I don't believe in stimulus-response theories, and our curriculum doesn't support that either, so therefore it's a bit peculiar that we have this grading system. However, she added that there is a need for "a little grading" during the school year when students are in Year Ten of their education (last year of middle school). She said that she builds her knowledge on research, theory and the national assessment regulations, pointing out that "we aren't that fond of the AfL term; I think it's all about good teaching". Similarly, according to Principal 4, "AfL is a mindset, a view of students". Principal 1 was convinced that students' learning output is much more solid if they manage to provide concrete feedback. Three of the principals used sport as a metaphor: athletes practice (Years Eight and Nine) until the time comes for competition where they need to perform (Year Ten): "We separate the school year into a practice period and a test period because there should be room for making mistakes. The teacher has the trainer hat on in the practice period and they [the students] are assessed with feedforward; then there are two times a year [with a test period]" (Principal 2). Principal 3 used a judge and a cook as metaphors as well, stating that it was crucial for the students that "they understood how the teacher became a resource for them in a totally different way than a judge", and "A cook in the kitchen practices making food".
When asked about the possibilities that grade-free schools provide, Principal 1 mentioned the opening in the assessment regulations for only giving grades twice a year and highlighted the focus through feedback on what students achieve and their potential for improvement. In particular, she said, in line with Principal 2, a grade-free school is most beneficial for students performing "average" or "a little below expected level" and students who are at their expected level but who are "super stressed" by grades with "stress on themselves, from home and from the whole world around them" (cf. Bakken, 2018). Grades were criticized for being "wishy-washy" and "old-fashioned", and for not encouraging students to move their learning forward. "It's the cleverest students that prefer grades", stated Principals 2 and 4, which is "a confirmation of how fabulous they are", concluded Principal 3, whereas a 3 (average grade) would stick with that student from Years Eight to Nine, "I'm a '3 student'; typical for a '3 student' . However, dropping giving grades does not mean that students will perform better on tests, claimed Principal 3.
To sum up, the principals refer to research and unfortunate assessment practices as justifications for implementing grade-free schools. Moreover, they are critical to the current assessment system and the assessment language used, which they believe is in the way of AfL.

Implementation and involvement
The leadership, teachers and politicians are considered as important stakeholders when implementing grade-free schools. Principal 1 claimed that "The leadership had such a clear direction for it [the implementation of grade-free school]". However, she added that "Maybe, in a way, it was a bit sudden; at the same time, we feel it was anchored to the staff ". Thus, she believes the implementation of grade-free school was well planned and that uncertainties and questions would have arisen no matter how well she and her co-leaders planned it, succinctly described as follows: "You can anchor to practice but when you haven't really tried it [grade-free school], there will be continuous questions coming up". At School 2, the principal said that "there was a lot of brouhaha over this [implementation of grade-free school] -whether it should be allowed or not, and politicians… we were in the media a lot". The politicians were concerned and challenged the students by asking what grade-free schools would mean to them when they graduated from middle school and started in senior high school. "They [the politicians] thought I had a home-alone party", the principal added. She actively used students' voices in the start-up phase of the implementation, in an "expert panel" with bureaucrats from the municipality, politicians and assessment researchers, where the students expressed their views on the feedback culture at school. The principal contacted one of the students who had graduated from middle school and the student addressed the politicians' concerns by reporting from senior high school: "This is not a problem. It's not that we never receive grades or that we don't know what level we are at; the only thing we lack is the relationship to our teachers, we don't have that here [in senior high school]". Principal 3 had also experienced a lot of activities in relation to the implementation of grade-free schools: "We were asked to go to other schools to talk about our assessment practices. We went together with students and they could relate how excited they were and how they experience a grade-free school". She also referred to "change agents" at her school being part of the project group responsible for the implementation, who were described as the driving force. Her most important role, she said, was to make sure that continuing education in the field of assessment took place. In employing new teachers, Principals 3 and 4 made sure that they possessed knowledge about AfL. In sum, knowledgeable and informed stakeholders seem to be the crucial factors for implementing grade-free schools.

Points of improvement
Principal 1 said that several stakeholders were involved, but "The students could have been involved too". When asked what the next step in the implementation of gradefree school is, she responded "to obtain more cohesion, real collectivity. It's probably a never-ending process". Principal 2 admitted that she should have been better at informing the others about the implementation and not started the implementation by assigning "low, average, high" on students' performances instead of grades: "The students immediately understand what low, average and high are… it was like deceiving yourself ". The only point of improvement mentioned by Principals 3 and 4 was to involve the students to a greater degree. Thus, concerning the implementation of grade-free schools, it seems that involving and informing are the two main points of improvement the principals agree on.

Challenges
Principal 1 mentions the assessment system as challenging because of "a lot of focus on that average grade on exams". She said that students and parents expect grades from Year Eight, and that some teachers feel it is easier to just give a grade: "[dropping grades] faces a little resistance from some teachers too, who think the concreteness of a grade is nice and safe". Dropping grades "requires a lot" since "the system measures us through grades". The same concern was articulated by Principals 3 and 4, stating that students and parents often can be "very conservative when it comes to receiving grades which tell them about the level of the student". Another challenge mentioned by Principal 1 was that she believes "we have been poor at giving information [about the justifications of grade-free schools]. We have experienced before and also now that parents are a little bit frustrated, because we haven't informed them and because the quality of what they receive is poor and variable". Principals 3 and 4 thought the biggest challenge was to involve everyone in the process of implementation and keep up the interest, since "it's about change in basic views of learning, it's about a focus on learning" and "so much else is going on in schools". Finally, Principal 4 maintained that the gaps between teachers' willingness to implement grade-free school, lack of competence and varying feedback practices posed a significant challenge, since they are the ones who stand for the fundamental work of making the changes happen.
Summing up, the principals relate challenges with the implementation of gradefree schools to the school and assessment system, resistance among some teachers, parents and students (mostly parents), lack of competence, and being able to involve and inform everyone.

Success criteria
Regarding the success criteria for implementing grade-free schools, Principal 1 mentioned the importance of the terms that are used, such as "practical test", where the word test automatically gives "some 14-year-old girls high shoulders". Principals 3 and 4 said they do not use words such as "tests" and "correcting" any longer, but rather "assessment situations". Secondly, Principals 2 and 3 underlined the importance of working collectively to improve the quality of feedback through use of cases and modeling. Third, she said you need to be genuinely interested and involved in professional development: "Many claim that the principal or the leadership has to be hands-on, and so you should…" by "being present, knowing what is going on, being out in the classroom, and talking over lunch".
Principal 2 highlighted the importance of changing a middle school to becoming grade-free already from Year Eight, because "once you have started with grading, it's difficult to quit". Furthermore, she mentioned close collaboration with researchers, which "eases people [the staff]". She also said "Even though I never talk about school results, they are there in Skoleporten [public platform]". In other words, results are important, but that does not mean that the principal should talk about that all the time, taking away attention from the continuous work with formative feedback. Another strategy for success mentioned by Principal 2 was to stay in touch with the student representative who went through the implementation in Year Eight (that student is in senior high school now). Similar to Principal 3, Principal 2 underlined the importance of a bottom-up implementation: "If teachers don't want to do this, they might just as well drop doing it", in addition to providing sufficient information to the teachers and being there for them when they experience hardships in the implementation process. Principal 3 also talked about the role of the teacher union representative and her positive attitude to the implementation. Both principals mentioned that their knowledge of assessment research made teachers trust them.
All in all, the everyday language that is used about assessment, collective development, the principal's engagement, involvement and knowledge, early start of gradefree implementation, close collaboration with researchers and stakeholders, and bottom-up processes of change were the success criteria the principals highlighted for implementing grade-free schools.

Discussion
In this study, we set out to explore four principals' views on the implementation of grade-free middle schools and analyzed them according to four main categories: justifications and opportunities, implementation and involvement, challenges, and success criteria. Regarding the first category, justifications and opportunities, all four principals are positive towards the implementation of grade-free schools, which is in line with Smith and Engelsen's study (2013). Furthermore, there seems to be consensus among the principals that students' learning is stimulated and enhanced when the teachers are working in accordance with the principles of AfL (Larsen et al., 2017). Insight into educational research on AfL (Assessment Reform Group, 2002;Black & Wiliam, 2009;Gardner, 2012;Hopfenbeck et al., 2013) and the national assessment regulations that open up for a partly grade-free practice in middle school appear to be the principals' main motivations when deciding to implement and develop gradefree schools. This concurs with research on purposeful staff development (DuFour & Berkey, 1995;Mulford, 2006). All the four principals underlined the need to reduce stress on middle-school students (Bakken, 2018), arguing that a reduction of grading could lead to a reduction of stress and pressure (see Mellingsaeter, 2018). However, more research is needed before we can conclude that grade-free schools contribute to a reduction in stress.
Concerning implementation and involvement the picture appears more nuanced, and the statements regarding initiatives from Principals 2, 3 and 4 differ substantially from the statements of Principal 1. Principals 2, 3 and 4 claimed that the change of practice was a result of an initiative, in fact a demand, from the teachers, which was appreciated by the principals. In these cases, it is likely to assume that the principals and the teachers had some shared values (Fullan et al., 2005;Mulford, 2006;Senge, 1992). Principal 1, on the contrary, claimed that the initiative solely came from the school leadership, and that the teachers perceived the change of practice as unexpected. None of the principals reported that the parents, the school board or the students had been involved or engaged during the decision-making and implementation of a grade-free school. The lack of student involvement was an issue they all regretted not having attended to. Hopfenbeck et al. (2013) also point to this as one of the success criteria in order to implement AfL.
The principals experienced challenges with implementing grade-free schools at both system level and school level. They report that it is challenging to implement a change that involves a significant reduction of grading when the school authorities measure and monitor the students' performances and the school's quality in terms of academic achievement through grades. This can be viewed as tension between a testing regime and an assessment culture rooted in sociocultural theory (Gipps, 1999;Poehner, 2008Poehner, , 2011Pryor & Crossouard, 2008). Besides challenges at the system level, the principals also experience challenges related to actions (or the lack of such) undertaken by the school itself. Insufficient information to staff and parents, along with varying quality regarding teachers' work with AfL (Larsen et al., 2017), seem to be key aspects. Particularly, skepticism among parents makes up the better part of the challenges faced. We believe this skepticism must be addressed against a backdrop of the parents' own school experiences. This can be interpreted as a sign of what Senge (1992) refers to as mental models, and how these can act as barriers to educational change. We assume that the parents' skepticism is closely related to the regulations regarding admission to secondary education, where admission is granted almost exclusively on the basis of grades (Regulation of the Education Act, 2013, Section §6-20). The principals also described challenges that arose as a result of teachers not being convinced that grade-free schools are a step in the right direction (cf. Senge's mental models), and that this, in turn, requires the principals to put a great effort into guiding and instructing their employees to change and implicitly improve their perspectives and opinions regarding AfL (Heitink et al., 2016;Hopfenbeck et al., 2013;Smith & Engelsen, 2013). The teachers need to develop personal mastery in teams and through communication and modeling by the leadership (Senge, 1992). Finally, poor feedback practices and insufficient information to the students about grade-free schools also came up when challenges were discussed, which are factors that hinder successful AfL implementation (Butler, 1987;Hattie & Timperley, 2007;Shute, 2008)). Regarding success criteria for implementing grade-free schools, the principals point to actions directed at the school's staff and actions directed at stakeholders outside of the school. Actions directed at the school's staff included facilitating for collaboration between teachers, providing supervision and one-to-one support, finding time for the teachers to learn and develop practice, involving the union representative, and transforming existing terminology to a more professional and AfL-oriented way of communicating (DuFour & Berkey, 1995;Fullan, 2002;Fullan et al., 2005;Heitink et al., 2016;Mulford, 2003, 2006, Senge, 1992. Moreover, the principals underlined the importance of their role as the school's chief executive in possessing deep and updated knowledge of what AfL is (Gardner, 2012;Hopfenbeck et al., 2013).

Implications for further research, leadership practice and policy
The informants in the present study report that they are working in accordance with knowledge provided by international research over the past couple of decades regarding implementation of change (Fullan et al., 2005;Senge, 1992;Timperley et al., 2007) and, more specifically, implementation of AfL (Birenbaum, 2015;Black et al., 2003). However, the principals did not focus on an in-depth knowledge and critical understanding of the inter-relations between assessment and learning for the students when describing the implementation of grade-free schools, as is required by dynamic assessment (Poehner, 2008(Poehner, , 2011. Their focus is rather on the leadership aspects of implementing grade-free schools. Further research could analyze principals' views of grade-free schools from a sociocultural perspective, thus shedding light on the social relations pertinent to dynamic assessment. Moreover, we argue that our findings also include some new aspects that have not yet been discussed in previous research. Firstly, all the four principals regret not involving the students throughout the process of planning and implementation. Further research is needed to investigate how the school leadership can involve students in implementing grade-free schools. Secondly, two of the principals mention the union representative and how they perceive him/her as having a decisive role during and after the implementation process for establishing and maintaining grade-free practice. While previous research has concentrated on the significance of facilitating professional development and collaboration among teachers (DuFour & Berkey, 1995;Fullan, 2002;Fullan et al., 2005), we argue that, in addition to that, involving the school staff 's union representative seems to be advantageous.
Moreover, the principals' concerns with replacing grades with the terms "low", "average" and "high" was interesting. Although the grade expressed as a number was eliminated from the feedback, "low", "average" and "high" were regarded as being as imprecise and non-informative as the grade ranging from 1 to 6. An implication of this is that other ways of categorizing student performance are not recommended when implementing grade-free schools. All the informants reported establishing a new way of communicating about assessment and providing a common vocabulary when implementing grade-free schools (Hopfenbeck et al., 2013;Smith & Engelsen, 2013).
Two of the principals were concerned about the politicians' mistrust toward the implementation of grade-free schools, cf. "home-alone party". Moreover, one principal said that she had returned to grading in Year Ten, due to teachers' concerns regarding parents' rights and willingness to appeal against the grades given on the final assessment. In conclusion, we suggest that more research is needed in order to explore the issue of trust, both between politicians and principals, and between parents and schools. As one of the main findings in this study was the principals' regrets regarding the lack of student involvement, and given the fact that the students are the main stakeholders who are supposed to benefit from grade-free schools, we argue that more research on successful student involvement is needed.