1. Background and Aim
As researchers and university teachers, we are used to the challenges related to combining the provision of high-quality academic training with conducting productive and innovative research. Facing increasing enrolment numbers, it sometimes appears that we need to use much of our resources for academic training, limiting our degrees of freedom regarding our own research activities. Instead of perceiving these two aims as competing, we work on the premise that training students by letting them participate in research projects can release synergies and is advantageous for both interests. In response to the influx of new master’s students at our institute, we decided to train them in groups working on related research projects. This would make it likely that the training could improve through more effective and manifold communication. We also assumed that such a group-based approach would provide more learning and a greater sense of meaning to individual candidates.
Graduate thesis supervision is pivotal at the master’s level because it shapes students’ research capabilities, completion trajectories and scholarly identities. Contemporary syntheses emphasise that effective supervision is not a single supervisory ‘style’, but an interplay of structures, roles, and feedback practices that support students’ developing autonomy and competence (Grohnert et al., 2024; Lee, 2008). In response to rising enrolments and diversified student needs, programmes increasingly experiment with group‑ and cohort‑based supervision to widen access to feedback, publicise exemplars, and normalise dialogue around research writing. A long Norwegian line of work is particularly instructive: Dysthe (Dysthe et al., 2006) shows how combining group sessions, peer interaction, and individual meetings creates a multivoiced arena for learning, while Samara (Samara, 2006) details how group supervision can iteratively improve both supervisory skill and text quality. Related scholarship on writing groups likewise demonstrates the value of collective critique alongside one‑to‑one meetings (Aitchison, 2009).
Established learning theories help explain why collective supervision sometimes excels and sometimes fails. From a sociocultural perspective, especially in communities of practice, group formats legitimise participation and make disciplinary norms public through shared practice (Lave & Wenger, 1991; Wenger, 1998). Psychological safety accounts for when students are willing to “learn in public,” ask questions, share partial work, or admit confusion without fear of embarrassment (Edmondson, 1999). Motivation research grounded in self‑determination theory predicts that designs that balance freedom and structure, supporting autonomy, competence, and relatedness, foster deeper engagement and ownership of scholarly work (Ryan & Deci, 2000). Complementary work on feedback for self‑regulated learning and student feedback literacy clarifies how timely, dialogic feedback cycles enable learners to take control of their progress, mechanisms that collective formats can amplify when well staged (Carless & Boud, 2018).
Against this backdrop, we evaluate a novel group-based didactic supervision model implemented in a Master’s programme in health education. Our objectives are twofold: (1) to assess the feasibility of the model, its acceptability, motivational pull, and practical workability—and (2) to understand how students experience and make sense of their learning processes within this design. We position the study within traditions of collective supervision that deliberately combine group arenas, peer interaction, and individual guidance (Dysthe et al., 2006; Samara, 2006), and we use sociocultural/Communities-of-Practice, psychological safety, and Self-Determination Theory as sensitising lenses for interpreting participation, voice, motivation, and emerging scholarly identity (Lave & Wenger, 1991; Wenger, 1998; Edmondson, 1999; Ryan & Deci, 2000).
Research questions
How acceptable and practicable is the group-based supervision model for students (e.g., motivation, perceived usefulness, and organisational fit)?
How do students describe their learning processes and sense-making within the model?
Which mechanisms appear to enable or hinder progress (e.g., cohort heterogeneity and pacing, multi-supervisor dynamics, scaffolding/timing of information, and peer processes), and how do these align with the above theoretical lenses?
2. Methods
2.1. Design
In this study, we conducted a test application of the group supervision approach, which we evaluated exploratively by collecting qualitative data from students’ feedback in a focus group interview and our observations on the practicability of the new approach. We also conducted a questionnaire survey investigating learning experiences.
Figure 1. Direction of communication in master’s supervision.
The figure illustrates the directions communication in the master’s supervision is supposed to take. The arrows mark links and directions.
2.2. Intervention
The intervention comprised a group approach to the supervision of master’s projects. To facilitate a group-based learning process, it is advisable to focus on very similar research questions or sub-topics of the same overarching research question. The starting point in our case was an ongoing joint research project to which the individual master’s theses contributed (Kasper et al., 2025). The research project, designed as a multiple cross-sectional study to map the quality of health information, offered the opportunity to conduct virtually any number of sub-studies.
The main objective of this approach is to facilitate and improve the learning processes of individual participants. This would also reduce, or at least redistribute, the overall efforts required to supervise the theses. In particular, the group approach should stimulate motivation to acquire and practise core scientific skills, such as systematic and thorough information searching, independence in data collection, justification of the choice of methods, and interpretation of results in social discourse. As a result of these improvements over the traditional 1:1 supervision setting, we expected strong performance in the master’s assignments but also sustained identification with and a sense of purpose in the research process and with regard to the overarching project.
All rules and measures for implementing the new approach were aimed at catalysing communication (Figure 1). This was attempted by maximising the number and diversity of communication channels, settings, and directions and through lowering barriers for any of these channels. All three supervisors were approachable and tried to contribute to the learning process to a large extent as facilitators moderating the modes and opportunities of information exchange. For specific evaluation steps, such as reading parts of the thesis, a rotation system was developed to ensure that each supervisor was familiar with all tasks and each candidate received the review of all supervisors. Consultations took place in group settings (total or part groups) and individually. Recipients of individual counselling were encouraged to pass on important results and findings during peer counselling sessions. This was facilitated by the operation of a shared Teams platform. For example, corrections to individual work were made available to everyone there in order to avoid duplication of work for the supervisors. On the Teams platform, we also shared learning resources, literature, tools for data collection (Zacher et al., 2025), and methods discussions.
The group supervision approach was realized openly over a longer timeline between autumn 2022 and spring 2024. The students were employed in the same research with individually varying deadlines. This further pronounced the effect that participants were working at different times. Most students worked in pairs. Particular themes / domains were distributed following individual interests and considering the expected amount of work. Some students conducted one single survey; others did several surveys of smaller size.
2.3. Sample and Recruitment
The participants in the group programme were 21 master’s students from various health professional trainings at OsloMet, but mostly (18) from the master’s programme in health sciences. The scientific project in which the students participated was also based there (Kasper et al., 2025). Three supervisors were involved as the leaders. The research fellow employed in the project contributed with administrative tasks and in the data collection. The recruitment of students followed the usual procedure for assigning topics for master’s theses. Students volunteered, for example, in the context of health communication classes, or selected the project from a range of options offered at the so-called master’s market. All students in the supervision group agreed to participate in the scientific evaluation of the new didactic approach.
2.4. Evaluation
2.4.1. Focus Group Interview (Including Feedback Mails)
Students were invited to a focus group interview to evaluate their learning experience. The research fellow and one of the supervisors (VTH) took notes, while another teacher (JK) moderated the interview. By showing up, the students agreed to the use of the collected data anonymously for a scientific publication describing our didactic approach to supervising master’s theses. The interview guide consisted of just one question that was given several times and varied in the way it was asked. “How was your learning experience in your master’s projects shaped / affected by the way, we have worked together in a group and contributed to the same overarching project?”
2.4.2. Supervisors’ Reflective Notes
In our roles as both researchers and teachers / supervisors with varying previous experience in using this type of group approach to managing the conduct of a research project while simultaneously providing scientific training to a larger number of students on the master’s level, we were curious throughout the entire project. Our curiosity was mainly focused on the quality of the academic discourse with and between the students as an indicator for achievement of the master’s level. We also wanted to learn how employing them in small research projects, which all contribute to the same overarching research question and are therefore structurally quite similar, would affect students’ motivation and learning processes. Moreover, we were interested in how we would be able to manage the immense workload in the given time frame. These observations were made unsystematically and documented post hoc.
2.4.3. Questionnaire
A questionnaire evaluating the learning experience was sent out via the Teams platform, and all students were invited to answer. The question was an adaptation of a previous self-developed evaluation of another master’s course (Tables 1-3). Three blocks of questions ask for 1) general ratings of learning experience and outcome, 2) weighting didactic components regarding their importance for the learning, and 3) weights on different learning outcomes. The students were instructed to weigh the presented items using a number between 0 and 10 (Table 1 and Table 2). In addition, the CATS questionnaire, (Odéen et al., 2013) a set of seven items measuring general control beliefs on a four-point Likert scale, was used as a potentially relevant moderator variable. Answering the questionnaire was feasible and took about seven minutes. However, to keep the answers anonymous, students were asked to print the questionnaire and deliver it physically to the university. By submitting the completed questionnaire, students gave informed consent to use of their data in a publication.
2.5. Analyses
We analysed the focus-group interviews using reflexive thematic analysis (TA) following Braun and Clarke’s six-phase approach (Braun & Clarke, 2006; Braun & Clarke, 2021): familiarisation, generating initial codes, constructing candidate themes, reviewing themes, defining and naming themes, and producing the report. Our analysis was primarily inductive and semantic, staying close to participants’ talk while acknowledging researcher subjectivity as an analytic resource consistent with reflexive TA (we did not calculate inter-rater reliability or seek coding consensus). Throughout, the lead analyst (YR) kept reflexive memos and an audit trail, while a second researcher (JK) acted as a critical friend during the theme review to enhance clarity and coherence. The observations by the involved supervisors were analysed narratively. Data collected with the questionnaire were transferred to SPSS and descriptively analysed regarding mean values and variation. The CATS items were aggregated using the mean as an estimate of students general control beliefs. Pearson’s correlation coefficients were generated between the control beliefs score and each of the three general judgements of the learning process.
3. Results
3.1. Description of the Sample
All 21 students (one from the master’s programme on empowerment, two from the master’s programme in clinical nursing, and 18 from the master’s programme in health sciences) submitted their master’s theses as monographies on time. In total they conducted 64 surveys evaluating 1948 websites independently three times by applying the MAPPinfo (Kasper et al., 2025). The results on the quality of health information for users in a well-defined segment of the Norwegian health information landscape were presented in a consistent manner and critically discussed. All candidates passed their exams with good to very good grades. Together, we prepared a manuscript for publication (Kasper et al., 2025).
3.2. Focus Group Interview
The invitation for the focus group was followed by 16 students. All students who did not participate gave understandable reasons for not showing up. The analysis of the notes made during the discussion resulted in three themes and in addition a couple of tips for practical refinement. Here are the themes:
3.2.1. Lack of Directive Guidance (Causing Uncertainty vs. Autonomy)
Many of the students’ comments referred to the experience of although having many resources accessible not getting clear and unequivocal orientation. This experience was based on having three different supervisors instead of one, working together in a group with students of uneven pace (students have varying schedules, starting points, and deadlines). Some students explained that they felt that the information varied slightly between the supervisors and came either too early or too late for them personally. The familiar one-to-one setting, providing individually targeted instructions, was experienced as a scarcity in this context, and the more often used group settings or information resources forced them to filter information regarding their own needs. This situation affected students in different ways, with some complaining of feeling uncertain and others discovering the opportunity to define their own way of doing it.
“We’ve had to be independent and make many decisions. It’s been challenging. There’s an expectation that we should consult one another. The problem is that people are at different points in the process (stages) ...”
“Since everyone has been at different stages, I’ve just gone my own way. Maybe that was the point?”
“Examiners differ, you differ, everyone differs. We have to figure out how I want to do my thesis. It took a while to feel confident about that.”
3.2.2. Peer Support Available (Latent Potential vs. High Thresholds)
We received a lot of feedback on how peer support worked for the students. The feedback showed that the idea of using each other was fully understood. The students also reported on a couple of attempts to self-organise peer to peer communication. A Facebook group and a separate Teams room were established. In addition, some students worked in pairs on the same topic. Clearly, the information source of peers was considered a potential resource. However, from several comments, we also learned that peer coaching did not work optimally. Drafts were not always shared, and questions were not answered.
“We haven’t managed to create a sense of group cohesion. Different expectations from the start. Maybe the supervisors should have held our hand. Maybe it was our job to bring ourselves together, but we didn’t manage.”
“Not everyone has shared their thesis.”
“We can learn from each other by sharing.”
“It’s been great having access to so much literature; we’ve essentially done that work for one another, especially on the theory.”
“It’s been good to be a source of support; in that sense, the group has worked.”
3.2.3. Similarity in Content and Structure (Being Part of Something Bigger/Efforts to Achieve Originality and Impact)
A discourse during the focus group was related to the process and reasons for the single theses becoming individual unique studies despite the high extent of similarity regarding theory, aim, measurement methods, and the structure of the monographies. Nobody seemed concerned about just having delivered a copy of another study, leaving little space for individual development. In contrast, we received many descriptions of how the project was experienced as a process over time, leading to the development of ownership. Some experienced individual supervision by individual supervisors as the reason, while others considered the health domains they had chosen or the opportunity to make their own choice as the origin of ownership. It appears that the similarity in the content and structure of the projects may have encouraged efforts to achieve originality.
“We are all different. We have different experiences, different ways of working, and different ways of producing text. That stems from us being different individuals, hence different theses.”
“Individual supervision has been important, especially for understanding writing style. We ended up with different theses because we had different supervisors.” (students, not participating in our group and conducting individual projects)
“We had to do it many times before we understood. We gained an overview, a different focus. It feels different when we go to those websites today; there’s a sense of mastery.”
“This is something you can take with you. The others have eight informants. This is a clear answer. I, for one, will take this with me, that’s why I’m grateful I got to be part of it.”
“We’ve been through the same thing. It’s hard to talk with someone who chooses an individual thesis. You feel a sense of community. We understand, I know what she/they are going through. There’s a psychological effect of belonging.”
“It strengthens the method and the fact that we arrived at the same findings. We’ll bring this into our professional practice.”
“If we manage to bring about change at the system level, that’s the benefit of working as a group.”
3.3. Summary of Supervisors’ Reflective Notes
The workload of supervising 21 students more or less simultaneously was quite high compared to supervision of four individual master’s theses, despite of being a team of three supervisors and having much opportunity to rationalise supervision communication. This was due to three reasons. First, individual supervision was still desired by the students. Therefore, we had to meet this need. Second, in addition to the usual supervision of research methods, we were also involved in the quality appraisal of nearly 2000 websites, as we functioned as expert coders. Besides caring for the high quality of the research, this role as a coworker in the individual projects was meant to provide supervision. And it was appreciated by the students as such. Third, the development of generic resources, such as lists of comments, administration of the Teams platform, and communication with the entire group, were also effortful endeavours. However, supervision of as many individual theses using an individual supervision approach would not have been possible at all.
Our expectations of using peer support to improve supervision were partly but not fully met. The idea of not needing to correct the same mistake several times did not fully work; however, we experienced that the quality of the first drafts improved over time as a result of students making use of the resources.
Our initial concerns about students feeling too little space for their own development, as research questions were generated like slices of the same salami, were not among the feedback from the students. On the contrary, we experienced good academic discussion with the students within each of the surveys, indicating that the students were highly identified with their individual tasks, and developed curiosity and self-efficacy in their role as responsible researchers in their particular field.
Beyond our initial goals of conducting the research project and ensuring appropriate supervision, we unexpectedly learned about the power of being a research group and the synergy effects of working based on a shared mindset with the same methods and similar research questions. After catalysing critical thinking at the onset of data collection, we experienced a flow of motivation from the students’ side. It was then very easy for us to give them a feeling of achievement, both based on their individual work and their contribution to the entire project. This dynamic facilitated meetings between students and supervisors on equal terms. We also recognised multiple expressions of pride of having been part of a meaningful project and group cohesion, which seemed quite sustainable. However, group cohesion within the MAPPinfo project was accompanied by an in-group/out-group dynamic. Both directly and indirectly, we received unambiguous signals of scepticism, envy, and disapproval from our colleagues looking at our project from the outside. This was reflected in discussions during the regular master seminars for the entire year group. Even not involved students reacted sensitive on how students from our group were met by our colleagues and reported in towards the elected representant. We experienced similar remarks during team meetings and related to the evaluation of the master’s theses from our group. However, both the institute leader and the dean of our faculty became interested in and appreciated our activities. We also noticed that we did not receive any positive feedback from our close colleagues, for example, for taking responsibility for the majority of the master’s projects at that time.
We observed multiple challenges and received feedback addressing practical issues with potentially high impact on the learning process. For example, we understood that we had not been sufficiently attentive in catching up on the need to continue meetings of the entire group. We are still not sure whether this meeting setting would have met the needs of most of the students later in the course of writing their theses. However, we did not offer these meetings for a longer period. In response to requests and our own observations, we made many adaptations. Many of them refer to the timing of feedback, clarity in the communication of plans for work packages and deadlines, and solutions for handling the distribution of supervision resources between students with an uneven pace of the individual projects.
3.4. Questionnaire on Learning Process
Thirteen students completed the questionnaire. The general judgements of the learning process were positive. The highest score was observed for the perceived usefulness of the thesis itself, the result of the learning process (mean = 7.9, SD = 1.8, theoretical range 0 - 10) (Table 1).
Table 1. General judgement of the master’s process.
|
Mean (range 0 - 10) |
SD |
min / max |
Score your learning experience |
7 |
1.2 |
5 / 9 |
Recommend the group approach to another student |
6.5 |
1.2 |
4 / 8 |
Score usefulness of your thesis |
7.9 |
1.8 |
4 / 10 |
Table 1 provides results on the students’ rating of the quality of the learning process in general. Results are based on N = 13 valid responses.
The weights indicating the importance of the 10 didactic components provided as didactic measures for the individual learning process were rather evenly distributed. Lowes impact is rated for the common seminar meetings together with all programmes of the umbrella master and seminar meetings within the health science master class. All specific components that were part of the MAPPinfo group’s approach to supervision received higher scores. The highest impact (mean = 8.2, SD = 1.8, theoretical range: 0 - 10) was perceived with regard to individual supervision provided when small groups worked with data collection and resolved discrepancies of ratings between the students and us as expert raters. However, all components appear to be highly important contributions to learning, including those referring to self-studies.
Table 2. Contribution of components to learning.
|
Mean (range 0 - 10) |
SD |
Methods course all programmes together |
3.9 |
2.6 |
Supervision with all public health nurses |
4.0 |
1.7 |
Group meetings in the MAPPinfo research group |
5.8 |
1.7 |
Individual supervision on writing the thesis |
7.2 |
1.2 |
Individual supervision on data collection |
8.2 |
1.8 |
Peer support |
5.9 |
2.0 |
Learning resources (Teams) |
7.7 |
1.3 |
Self-studies literature |
6.0 |
2.1 |
Data collection |
6.9 |
2.1 |
Writing thesis |
7.2 |
1.6 |
Table 2 provides results on the students’ appraisal of different didactic components’ contribution to the learning process. Results are based on N = 13 valid responses.
Regarding the differentiation of learning outcomes (listed in Table 3), we see high scores for learning outcomes throughout. The lowest score (mean = 6, SD = 2.6, theoretical range: 0 - 10) was given for learning “how to cooperate in groups.” The highest scores were given to learning regarding critical thinking and evidence-based practice. The students also consider their learning experiences highly important with regard to the management of their own project (means = 7.9/7.8, SD = 1/1.2) and perceive that they have learned a lot about how to make a master’s project relevant for practical work.
Table 3. Weight given to different learning outcomes.
|
Mean (range 0 - 10) |
SD |
How to fulfil a big project |
7.9 |
1.0 |
How to master a big task |
7.8 |
1.2 |
What public health nursing is about |
6.9 |
1.9 |
How to analyse and critically apply information sources |
8.6 |
1.4 |
How to cooperate in groups |
6.0 |
2.6 |
How to apply and to justify research methods |
6.7 |
1.3 |
About basic principles in EBP |
8.2 |
1.1 |
Application of research ethics |
6.9 |
2.0 |
About relevance of the master’s thesis for the job |
7.7 |
2.0 |
Table 3 provides results on the students’ weighing of their learning achievements on four different learning outcomes. Results are based on N = 13 valid responses.
Spearman’s correlation coefficients between the students’ general judgements of their learning experience and their general control beliefs were unobtrusive in two cases (“Score your learning experience judgement”, “Recommend the group approach”) and highly significant in one case (“Score the usefulness of your thesis”, r = 0.74, p = 0.004). This correlation indicates that students with generally high control beliefs tended to consider the impact of their own project high.
4. Discussion
Our study describes a novel group-based approach to the supervision of master’s projects. We sought to generate practice‑relevant insights for programmes that aim to purposefully blend group, peer, and individual supervision at the master’s level. Our preliminary evaluation used reflexive thematic analysis of a focus group interview, a questionnaire on the learning experience, and unsystematic observations by the supervisors.
4.1. Important Results
The group-based supervision model was both feasible and motivating for the cohort of students. Students consistently described the multivoiced character of the format, seeing others’ drafts, hearing alternative framings, and comparing analytic choices, as energising and instructive, even if it sometimes created uncertainty about one’s own next step. Despite that ambivalence, the model supported steady progress, timely submission, and strong overall performance, alongside substantial shared outputs from the overarching project.
Survey responses reinforced this picture. Students regarded individual supervision embedded in small-group, hands-on data work as the most impactful element of the design, whereas large cross-programme seminars were seen as less directly helpful for moving a thesis forward. Learning outcomes were rated positively overall, particularly for critical appraisal and evidence use, while collaborative know-how was viewed as less developed, mirroring reports of relatively high thresholds for sharing work in larger or more formal settings.
Focus-group accounts pointed to a productive balance between belonging and authorship. The students valued contributing to a collective endeavor while still cultivating a clear, individual thesis voice, even when projects shared a common structure and method. Peer support was acknowledged as a meaningful resource but did not always self-activate; differences in pacing meant some preferred to “go their own way” at times. Notably, students who felt greater personal agency tended to describe their thesis as more useful, suggesting that perceptions of control may shape how a collective supervision design translates into impact. Amongst others this is in view of the fact that even uninvolved students complained about unfair treatment of our students by our colleagues. Although not obtained from systematic evaluation, we consider our findings related to “ingroup-outgroup” dynamics as significant. This is in particular due to the fact that critical and unsettling comments by colleagues addressing our students were observed and reported by not involved students.
4.2. Limitations
The students in our supervision group represented a self-selected cohort. Therefore, and because we have not used a controlled design, we do not claim that our study can tell about the new methods efficacy. In addition, our role as researchers involved into the data collection and analyses can be questioned as we were not neutral. However, our focus was predominantly on exploring potential mediating mechanisms.
The response rate of 62% of the student group delivering the questionnaire is a weakness as this selection can entail any bias. Although we attribute this rate to organisational challenges that could have been avoided, there remains uncertainty regarding the quantitative results.
In our evaluation, the didactic approach is confounded by the nature of the particular research project. We can say that the infrastructure of the research project, MAPPinfo, was suitable for applying the group-based approach. However, it will need further trials to learn more about which types of research projects harmonise with the method and how individual research topics should be related to each other to make this method work.
4.3. Critical Discussion in Context of the Literature
The students’ ambivalent response to non-directiveness, at once liberating and disorienting, maps closely onto Self-Determination Theory: autonomy-supportive designs can heighten ownership and motivation, but only when they are supported by clear pathways to competence and relatedness (Ryan & Deci, 2000). In our model, multiple supervisors and mixed-pace group meetings diversified perspectives but also produced moments of role ambiguity and timing misfit (e.g., “information came too early or too late”), which the supervision literature predicts when supervisory functions are not explicitly coordinated (Grohnert et al., 2024; Lee, 2008). From a communities-of-practice lens, the large, heterogeneous cohort offered legitimate peripheral participation but insufficient graduated participation opportunities; students effectively asked for stage-aligned subgroups and “just-in-time” inputs to convert autonomy into mastery (Lave & Wenger, 1991; Wenger, 1998). Read through feedback literacy and self-regulated learning (Carless & Boud, 2018; Nicol & Macfarlane, 2006), the data suggest that the non-directive stance worked best when paired with fixed milestones and decision logs to help students reconcile divergent advice, honing judgement rather than chasing consensus.
Peer processes showed latent capacity but high thresholds for public sharing—especially in digitally mediated spaces, underscoring the centrality of psychological safety for learning “in public” (Edmondson, 1999). Where small, steady subgroups formed, students reported reciprocal help and growing confidence; where they did not, the promise of collective supervision was only partially realised. This pattern is consistent with Norwegian multivoiced supervision models that combine group, peer, and individual arenas to stage dialogue and make exemplars visible (Dysthe et al., 2006; Samara, 2006) and with evidence on structured writing groups that lower participation barriers (Aitchison, 2009).
Notably, students articulated strong ownership despite shared templates and closely related projects, a productive “standardization-individualisation” paradox that Communities of Practice theory would explain as identity work through shared practice and differentiated contribution (Lave & Wenger, 1991; Wenger, 1998). The cohort’s forward-looking stance (transfer to work; system-level ambitions) further suggests that when safety and staging are in place, collective formats can scaffold both individual thesis authorship and participation in a broader research enterprise.
4.4. Practice Implications
We consider our development and preliminary evaluation to be very encouraging and a potentially important contribution to resolving challenges related to the increasing number of master’s programmes and enrolments in these programmes. We are motivated to replicate this approach with minor adjustments. We plan to spend more emphasis on managing varying pace and progress being more flexible in defining group settings. We also consider investing in training provision of constructive feedback and engaging in academic discourse early in the master’s course. Moreover, we will distribute responsibilities in form of a peer pair structure, to care for sufficient support and confidence. As an important requirement for implementation, the approach needs acceptance and support by all colleagues of the department. Therefore, potential concerns need to be explored. In the meantime, three other Universities have replicated our approach, and amongst others also used our learning resources: Master programme evidence-based practise at Høgskolen i Vestlandet, Bergen, Norway, Master’s programme in pharmacy at OsloMet and Master in clinical nursing at University Halle Wittenberg, Germany.
5. Conclusion
Supervising Master’s theses through a group-based model is both feasible and educationally valuable. Beyond obvious organisational efficiencies, the approach leverages the inherent social nature of research by making analytic choices, argumentation, and writing practices visible to peers. In our context, it supported multilevel communication (group, peer, and one-to-one) and cultivated a non-directive academic discourse in which students developed judgement by weighing alternative perspectives while retaining individual authorship and ownership.
At the same time, the model surfaces design tensions, especially cohort heterogeneity in pace and progress, occasional ambiguity across multiple supervisors, and high thresholds for sharing in larger or more formal arenas. These are tractable issues. Practical refinements include combining group forums with stage-aligned micro-cohorts, clarifying supervisory roles, anchoring work to clear milestones and “just-in-time” inputs, and deliberately scaffolding peer interaction to strengthen psychological safety. Taken together, the findings suggest that a thoughtfully staged group-based supervision model can deliver high-quality academic training while preparing students for collaborative research cultures. Future work should test variations in cohort size and staging, examine longer-term transfer to professional practice, and explore how the model performs across disciplines and institutional contexts.