Between the office and the coffee shop: an examination of spaces used for research degree supervision

The last two decades have seen increased attention given to the role of space within the university campus, with numerous new learning spaces forming part of both the physical and the digital campus. Much of the focus of how these spaces work to create supportive learning environments has been on undergraduate teaching. However, these spaces offer a great opportunity to also enhance the doctoral researcher’s supervision process through the creation of new learning spaces that break away from the traditional office setting. In taking the coffee shop as the antithesis of the office, this paper examines theories around space-making in relation to doctoral research, adding in the experiences of UK doctoral researchers to provoke further thought and discussion about how new spaces within a university and outside the campus might be considered part of the pedagogical approach to supervision. Results suggest that although there is much to be considered, doctoral researchers spend the majority of their time in traditional spaces—where they feel the most comfortable—and become progressively less comfortable the further supervision moves towards public spaces.


Introduction
Throughout the last two decades, higher education has increasingly examined the spatial dimensions of the campus. Billions have been spent on buildings (Else, 2017), and in recent years-accelerated by COVID-19-universities have increasingly invested in virtual spaces. The development of these spaces has been influenced by learning theories that have pushed for more collaborative, student-centred learning spaces. Investment is also Specht Between the office and the coffee shop: an examination of spaces used for research degree supervision Journal of Learning Development in Higher Education, Issue 23: March 2022 2 driven by the changing shape of many workplaces and the university's shifting practical, emotional and symbolic roles (Ellis and Goodyear, 2016). Many decisions around the use and roles of university spaces are informed by undergraduate programmes, and to a lesser degree, taught master's degrees (Deed and Alterator, 2017;Morieson et al., 2018).
However, considerably less attention has been given to how these spaces affect doctoral supervision.
Supervision affects the success of doctoral researchers. Although there are variations between disciplines, typically, this relationship involves a student's work being overseen by one or more senior academic staff members (Garvis and Pendergast, 2012). Despite the fact that universities have undergone many changes in their teaching approaches, the supervision process remains a distinct pedagogical practice in which only the supervisor (or supervisory team) and supervisee might take part, and this practice is resistant to change (Firth, 2020).
The supervision process is becoming increasingly scrutinised, both pedagogically and administratively. There is a growing body of research related to many aspects of the process including how interdisciplinary supervision might work (Kiley and Halliday, 2019), the introduction of co-supervisors and supervisory teams (Olmos-López and Sunderland, 2017), and the developmental quality of supervision (Taylor, 2020). Guides that provide education on supervision are also a large and established part of the academic market (Hemer, 2012;Grant, 2005). However, spaces or locations in which doctoral supervision can take place remain under-researched. To better understand how the location of and the spaces used for supervision meetings might affect the student experience, this paper undertakes a brief exploration of the literature around learning spaces before looking at supervisory spaces and the relationships they might foster. A small-scale study of UK doctoral researchers is then presented, along with a discussion of how the supervision spaces might be reconsidered in order to improve the experiences of doctoral researchers.

Transformed learning spaces
Space is a term that has been widely used in both social and educational research. Its wide usage has also led to a wide range of physical and metaphysical meanings. The relationships between physical spaces and the more abstract, social conceptualisations are 'dialectically interactive [and] inter-dependent' (Soja, 1980, p.210). Lefebvre (1991), in his thesis on the production of space, notes how daily routines give an identity to space.
For example, a room might be called an office because of what happens inside that space.
Conversely, people act as if they are in an office (even when they are not) because of the norms that have become connected to the space that they call the 'office'. Or, to put it another way, social interactions can shape a space, but space shapes the way people interact (Massey, 2008). Ellis and Goodyear (2016) also note how educational spaces, such as libraries, seminar rooms, corridors, and residences, are made meaningful as people use and interact with them.
Of course, spaces for teaching and learning should never be thought of as vacuums that are transformed by use; rather, the design and contents of a learning space create an environment that is conducive to learning and expression (Fassinger, 1995;Ellis and Goodyear, 2016). Indeed, in their 2005 study of what makes a classroom a safe space for exploring new ideas, Holley and Steiner (2005) found that the physical aspects of a space played a significant role, with participants in their research outlining 93 characteristics that indicate an unsafe classroom. The most prominent characteristic was seating arrangement-formal rows versus circles of grouped tables. These studies have helped make university learning spaces more diverse, inclusive, and varied, with informal learning spaces now making up the largest part of most university estate portfolios (Ellis and Goodyear, 2016). These transformations are generally welcomed. There have been important shifts in the ways in which people think about higher education and the learning that takes place on campus, and universities are increasingly moving away from traditional modes in which the lecturer organises, presents and structures learning (Vaatstra and De Vries, 2007). These spaces can perhaps be best understood through Radcliffe et al.'s (2008) 'places for learning spectrum' (See Figure 1, which seeks to map both informal [unstructured] and formal [structured] learning environments within and around the university, along with a simple taxonomy of learning modalities and those who support learning in different spaces). In contrast, some critiques of this framework-such as the conflation of structure with teacher presence-provide a useful jumping-off point for examining how doctoral supervision might fit into the learning spaces provided by a university.

Supervision spaces and relationships
Given the wide spectrum of learning spaces now available to university staff, we might expect to have seen transformations in spatial thinking with regard to doctoral supervision.
Although there have been changes to the process of supervision-such as more emphasis on co-or even team supervision-supervisory pedagogy remains heavily influenced by the Oxbridge tutorial system that was introduced in 1917 (Simpson, 1983). It is characterised by intensive individual relationships. The supervisor is the master, and the student is the disciple (Manathunga, 2005). The student learns through observing the habits of their supervisor (Maor et al., 2016). This symbiotic relationship mirrors those sought through the development of more inclusive and 'safe' learning spaces and takes the idea of nonjudgemental supervision (Latting, 1990) further to consider the power dynamics within the process. Under the master-disciple model, the location of supervisory meetings might move from the formal settings of the office to more casual spaces such as a coffee shops or other alternative spaces. These casual spaces are typically more natural and put both the supervisor and supervisee on an equal playing field, which 'levels' them and erases their distinctions (Oldenburg, 1989). The broad spectrum of supervisory locations frequently used is noted in the author's additional annotation of Figure 1. Here, though, even more important considerations must be given to the production of spacethe kinds of work that might be facilitated through this production, as well as who is producing the space and to what end (Lefebvre, 1991).
Although the seemingly hard boundaries presented by Figure  Much like Holley and Steiner's (2005) work on the characteristics of safe classrooms, in interviews conducted by Hemer (2012) about supervision, space was one of the first topics that came up. Particular attention was paid to cafés, which were seen as very different spaces for supervisions. Furthermore, Hemer's work found that, through having a meeting in a café, the traditional models of power change because the location is not a 'workplace' and instead is seen as a more neutral setting. Additionally, students in Hemer's study suggested that they preferred criticism in a more informal setting. This allowed them to accept and understand evaluation better than in an office, where they felt they were in the supervisor's domain. Not only does changing the location of meetings to less formal spaces begin to reject the master-disciple duality and allow for better collaboration, it also signals to the student that the supervisor has time. Be it true or not, meeting at a café suggests the supervisor has endless time for the student. This can enhance the relationship between the supervisor and the supervisee, lending a sense that the supervisee's work is of enough importance for the supervisor to set time aside.
Garvis and Pendergast (2012)  supervision is perhaps less of a break from traditional supervision than a move towards another imagined tradition. The net effect of demonstrating time for the supervisee remains the same and should be welcomed. Moreover, as doctoral research positions are now increasingly taken up by first-in-family students (although numbers remain low), the generally more familiar norms of coffee shops and cafés may serve other roles. As Goffman (1959, p.79) notes, when an individual moves 'into a new position in society and obtains a new part to perform, he [sic] is not likely to be told in full how to conduct himself'.
A meeting that takes place in a space in which the norms of the space itself are already known-instead of the norms of typical supervisory spaces-might be of great benefit to some. It is also unlikely to have a detrimental effect on other students. Bourdieu (cited in Gaston and Duschinsky, 2020) argues that those with sufficient levels of cultural capital and familiarity with the dominant culture within the university are much more likely to succeed. Moving to a coffee shop is unlikely to detract from this.
In this respect, a supervision meeting over coffee might offer several things to different students: (1) a return to long, philosophical conversations, (2) a move towards safe spaces, or (3) a reduction in the pressures of conducting oneself due to the familiarity of the setting (Gaston and Duschinsky, 2020). Conversely, a short meeting in an office might signal the opposite of these. This is not in any way to suggest that all meetings should move to coffee shops or that one space is preferential over another. Rather, these examples seek to highlight that, through an understanding of the messages that different spaces portray during a supervisory meeting, it might be possible to better strike a personalised balance for each student. Cafés do not remove all power from supervisory relationships, and there are times when they might be either effective or inappropriate (Hemer, 2012). A blend of locations for meetings, carefully considered for the content of the meeting, might well lead to a better navigation of the complex power dynamics between supervisor and student.

Power relations
Spaces alone will not transform the supervisory relationship. The close and tight nature of such work indicates that the interpersonal dynamics are perhaps more important than those occurring in a large class. Power relations can mean that, although a space is familiar, how that space is produced as a learning environment can be heavily influenced by one party, potentially at the expense of the other. Therefore, it is worth giving some attention to the supervisory relationship as it might be performed within these spaces.
There are two generally discussed forms of supervision: the hierarchical approach (Campbell and Campbell, 1997), in which the supervisor does the majority of the teaching and instructing, and mentoring, which is generally taken to be a more horizontal power structure (Terrion and Leonard, 2007). Although many doctoral research relationships will consist of both forms-often moving from hierarchical to more horizontal approaches-the mentor style is becoming the prevailing pedagogical approach, driven by the suggestion that it can provide high levels of confirmation, emotional support, personal feedback and friendship (Robertson, 2017;Al Makhamreh and Stockley, 2019). It does present other challenges in terms of the power dynamics between supervisor and supervisee. However, there are significant pedagogical differences between being a supervisor and being a mentor, and this is borne of the differing balance of power within supervising/mentoring relationships.
Power is a recurring topic in the literature on supervision and is attributed to various factors, including levels of experience, expertise and training (Murphy and Wright, 2005).
This power is often difficult to discern, even when used explicitly, and there is no consensus on whether such power is good or bad within a supervisor-supervisee relationship. Murphy and Wright (2005, p.284) see supervisors as being able to 'use their power productively to enhance the supervisory relationship', suggesting this can be done through an acknowledgement of the power dynamic, which enables students to learn about the power structures of the academy and thus trains them in how to assume their power (Murphy and Wright,2005). Manathunga (2007) also suggests that power is important in a relationship in which the student is not seen as an 'autonomous and rational being' (p.208), and that the student is already subject to institutional powers and responsibilities, and so to attempt to remove the power from the relationship is to pull the rug of the academy from beneath the student's feet.
These pro-power ideas appear to contradict the 'supervisor as mentor' model, which some have suggested helps to remove the hierarchical nature of the relationship, moving away from the supervisor being in the position of master and the student being relegated to the role of devoted apprentice (Manathunga, 2007). The underlying philosophy and rationale behind mentoring are the same as that which underpins supervision as a wholethat the mentor is wiser and more experienced than the mentee (Manathunga, 2007). This view is supported by Eby et al.'s (2008) work, which found that 'academic mentoring typifies the apprentice model of education where a faculty member imparts knowledge, provides support, and offers guidance to a student protégé on academic skills' (p.255). For Redmond (1990), mentors are seen as having two distinctive roles. The first deals with the transfer of marketable and discipline-based skills, behaviours, and attitudes. The second is related to social and emotional interactions that facilitate the transfer of knowledge. Redmond (1990) believes this combination is of particular importance in supporting minority groups. The idea that imparting knowledge is only a component of the mentoring process also reflects the work of Leung and Bush (2003), who suggest the most desirable characteristics of students' mentors are that they are (1) 'understanding and sympathetic', (2) 'accessible to students', (3) able to 'communicate well', and (4) 'enthusiastic' (p.268).
Supervision then is 'not wholly about the exchange of knowledge, skills or advice, but . . .
is often an exploratory process which needs time' (Garvey & Alred, 2000, p.124). The supervisor should be more concerned with helping the supervisee use the method of understanding and exploring.
Yet, despite the noted benefits of blending pastoral and academic care, the concept of moving from supervisor to mentor is not without its opponents. It is, of course, once again an issue of power. 'Supervision as mentoring', asserts Manathunga (2007, p.218) 'is not an innocent, collegial practice'. Although mentoring ideas have emerged from a desire to move away from more overt displays of power, people have only succeeded in submerging these tensions and complexities (Manathunga, 2007). As supervisors becomes more mentor-like, they may begin to transgress boundaries and offer advice on a range of personal issues that may only tangentially relate to research (Manathunga, 2007). dynamics (Manathunga, 2007). It is then the balance between informal and formal that is essential 'to create conditions for cooperation, cohesion and innovation. . . the fine tuning of informality and formality is central to the creation of social trust' (Misztal, 2000cited Hemer, 2012. And although the supervisory relationship directly affects the performance outcomes of supervision (Hemer, 2012), the beginning of this paper describes how different locations affect this relationship. Thus, it is crucial to explore how location affects power.
This leaves a rather tangled set of dynamics in which to place a supervisor. As the importance of social learning spaces on the university campus has grown, supervisors have an increasingly wide range of spaces in which they can conduct meetings. New opportunities have been presented by the reimagining of learning spaces in higher education in addition to the erosion of didactic teaching practices. However, within these new spaces, old modes of power and relationships can still persist. Although changing spaces might change some aspects of the supervisees' need to conform to certain norms, it does not, as the literature suggests, remove the power dynamic-even when supervisors take on a more mentor-like quality. Indeed, these dynamics themselves influence the learning space more than the objects that make the space an office or a coffee shop. Hemer (2012) notes that, in her studies, supervision was undertaken in a wide range of contexts, from offices to meeting rooms, cafés and other third spaces, such as parks or museums. She also acknowledges that the use of different spaces differed for each individual student, and that both supervisors and supervisees sensed that a supervisory meeting over coffee offered a 'different sense of time and space from the more formal and sometimes intimidating spaces of offices' (Hemer, 2012, p.833 Although the length of the survey is a limitation, its brevity allowed the researchers to capture more student responses than they would have been able to capture in a longer survey. The survey and analysis were also conducted pre-COVID-19. Although it is clear that the pandemic has reshaped some thinking in regard to higher education, the research team decided that re-issuing the survey to show the results of online-only supervision would not add much to the pedagogical discourse. Indeed, given the small (often one-toone) nature of supervisory relationships, there is a high likelihood that supervision will return to being in person more quickly than large undergraduate courses. Furthermore, there is extensive research about remote supervision, especially from Australia. Instead, this paper focuses on physical spaces and provides insights that will aid supervisors as

Methodology
campuses re-open their doors.

Results and discussion
The survey yielded 110 respondents from UK higher education providers; disciplines were not recorded for this research. The survey comprised three questions enquiring about the stage of study, the place where supervision meetings occurred and the comfort level of students in each of these spaces.
Regarding the first question, the majority of respondents were doctoral researchers (n=88), with 19 respondents from MPhil programmes leading towards a PhD. Three respondents listed themselves as 'professional', and these were removed when analysing the context elements of the survey.
It was crucial to determine the places where supervision occurred. The second question was designed using a Likert scale where the categories were pre-determined by the researcher and from which the respondents could choose. All respondents (n=110) were included.

Figure 2: Location of supervision meetings (n=110).
Figure 2 shows that most respondents (n=78) had supervision meetings in their supervisors' offices, with online (n=38) and other university locations also being important.
Take note that this was pre-pandemic. This indicates that the transformation to other mediums-in contrast to physical spaces-had started occurring inevitably in the last few years.
Although it is not the focus of this paper, it has been suggested that supervisors might treat supervisees differently because of their gender (Halse, 2011;Conrad, 2017). In the data, no significant difference was presented, although female students responded that they are more likely (p=0.255) to feel comfortable having their meetings in a universitybased café compared with their male counterparts. The gender of the supervisor did not appear to change these results.
Full doctoral researchers were also significantly more likely to have meetings in a café discussions that are perhaps more difficult, or which go a little beyond the parameters of the subject. This kind of availability might be better suited to the later stages of the programme, as they lead to more 'pure relationships' as opposed to office-based 'exchanges' (Hemer, 2012, p.833). In terms of other public spaces, for most respondents, this was either something they had not experienced (n=97) or something that they found more uncomfortable (n=23).

Figure 3. How comfortable do you feel being supervised in each location? (n=110).
These 'Third Spaces', characterised by their informality (Hemer, 2012), are then perhaps not utilised to the extent that might have been imagined when considering the higher education shift towards shared learning spaces in other aspects of teaching. Although the literature notes the issues of power and hierarchy that come with supervising from an office, Paré et al. (2011) are keen to point out that the department itself is a key home base for many students and that, for some students, being supervised in a third space constitutes being further alienated from the academy and the 'invisible curriculum of academic life' (Paré et al., 2011). Students who responded beyond the questions indicated that these third spaces might not always work for them. For example, students who had all or most of their supervision online responded, 'Skype works for me as I'm located in a different country. But-it also increases the feeling of distance and isolation in a way'.
Similarly, some students-specifically females-did not feel comfortable meeting outside the university. Alarmingly, one student responded that their supervisor 'would regularly meet his male students in the bathhouse, and leave his female students waiting Although this study was set up with only three short questions to ensure a higher response rate from time-poor doctoral researchers, the optional longer form answers helped shine more light on the importance of location in supervision. Indeed, many respondents took the time to add longer responses, adding value to the qualitative aspect of the research.
These insights also help us better understand the literature above and the nuances of how space can be formed through supervision.
Many of the open-ended responses echoed the idea that it is the supervisory relationship that most affects the student's performance (Hemer, 2012;Gaston and Duschinsky, 2020): 'My supervisor is honestly such a good person that I'd have no qualms meeting him wherever'; 'I think his availability and prioritisation of our supervisory relationship is much more important than the mode'; 'Location has no bearing, but content and approach does';  Hemer (2012) warned in her research that, although some students prefer to meet in coffee shops when receiving critical feedback, these spaces are also difficult places in which to maintain a more traditional master-disciple relationship, professionalism that Hockey (1995) asserts must be maintained. Some respondents raised concerns over location: 'I wish we had meetings in more of a formal setting such as an office, at least most meetings' or '[it] helps create healthy boundaries if the meetings are in the office'.
Others were specifically against the notion of holding meetings in public spaces: 'I feel less comfortable having work conversations in public as I'm having to be aware of who is around and who is listening'. Some stated that they prefer 'a neutral place-neither his or [sic] mine. But never in public'. These stronger reactions could be related to the sensitive For others, location remained just as important, but they would rather not undertake their meetings in a supervisor's office. One respondent, whose supervisor has a desk in an open-plan office, noted that they 'feel uncomfortable about interrupting the meeting and the possibility that I might be intruding on their workspace. I also feel more conscious of my supervisor's other workload when in their office and more acutely feel I'm taking their time, and so am more likely to end the meeting a little earlier in order to reduce this'.
Others reported that having meetings in the office can have negative effects because 'just a quick catch-up can feel very 'business"'. Perhaps more alarming, though, are those who reported, without further comment, that they 'Would rather door is open rather than shut when meeting in supervisor's office', suggesting that they are extremely uncomfortable in this situation. Missing meetings was also commonly reported: 'He refuses to meet with me.
We have only had a 15-minute call in the last 2.5 years. Everything is in writing only'.
There is a huge range of experiences of supervision, even within this small pool of respondents. An optimistic view might be that the supervision process responds to individuals' needs by tailoring their supervision experience. For some respondents, this appears to be the case. For the majority, though, the location and space within which supervision takes places appear to remain by happenstance and under-considered. The office is the primary location for many, and where third spaces are used, this appears to be due more to constraints of shared offices (or no offices) and other managerial special decisions rather than those linked to pedagogy. The very nature of supervision means that power imbalances are hard to remove (Maxwell and Smyth, 2011), and changing physical locations does not automatically address this. Without careful consideration, 'traditional' spaces can favour those whose habitus matches that of the universitymiddle-and upper-class students (Bourdieu cited Gaston and Duschinsky, 2020 the roles of the supervisory relationship, which also requires sufficient cultural capital to negotiate and decode (Bourdieu cited Gaston and Duschinsky, 2020).
It is not impossible, though, to use space to enhance learning and growth within supervision. Rather than looking for spaces to provide all the signals of learning, spaces for learning can be produced through the supervisor's actions. As Madikizela-Madiya and Atwebembeire (2020) note, these spaces can be developed through 'trust, compassion and care', creating 'favourable space through these values'. However, as they also point out, there is still much to be done in revealing the social aspects of these micro spaces (Madikizela-Madiya and Atwebembeire, 2020). It is clear that the development of researchers is becoming a priority for governments and institutions around the world, but it is also true that how this is implemented, and the range of experiences, is broad at best and that there is still a great deal more to be understood about the pedagogical and institutional circumstances within which supervision takes place (Manathunga, 2005;2009;Firth and Martens, 2008;Madikizela-Madiya and Atwebembeire, 2020).

Conclusions
This paper set out to open up a conversation about how location can affect doctoral supervision-noting that this is an area that remains under researched. The literature noted that there is an important set of questions to be asked concerning how moving from the office to other learning environments and even third spaces might affect the traditional master-disciple supervisory relationship (Hemer, 2012), and how such changes in practice affect power (Murphy and Wright, 2005), isolation (Paré et al., 2011) and pastoral care (Hockey, 1994). The results from the short survey showed that there is still a huge range of different experiences and expectations of students, in terms of where they are being supervised, how comfortable doctoral researchers feel in these spaces and the impact this has on their education and learning. It is clear, though, given the respondents' experiences, that the capability to 'read and analyse emotional signals and psychological processes' needs to be a fundamental part of the training process and pedagogical development of a supervisor (Strandler et al., 2014, p.80). Understanding that a lot of issues around trust, development, questioning and power can be resolved-or at least lessened-by changing the location in which supervision occurs is a great place to start developing this practice and leads to questions of how university estates can be better designed to support doctoral research supervision. Grant (2005) wrote 15 years ago that graduate supervision was a deeply uncertain practice and an important puzzle to solve. It would seem from this brief encounter that although more research is being undertaken around supervision, there is still much to be done in terms of examining supervision pedagogy and the production of space.

Ethics
Ethical clearance was given by the University of Westminster ethics committee. All data was collected anonymously, and there was deemed to be no risk of harm to any participant.