The more the merrier? PhD supervisors’ perspectives in engaging in co-supervision

ABSTRACT

Over the past decade, co-supervision of PhD candidates has become increasingly common across the disciplines. However, co-supervision has been explored previously primarily from the candidates’ viewpoint and in small-scale qualitative studies. We have extended the research by examining co-supervisory experience from the supervisor’s perspective, and by exploring the influences of co-supervision for supervisory competencies and the occupational wellbeing with a large-scale quantitative survey. 561 supervisors participated in the study. The data were analysed with statistical methods. The results showed that having a supervisory team has become the dominant from of supervision across the disciplines. Furthermore, the frequency of engaging in co-supervision was related to supervisor’s occupational wellbeing, but the association was not straightforward. Having experience in co-supervision seemed to contribute to feeling more confident with one’s supervisory competencies. The findings imply that the development of well-functioning co-supervision practices can enhance supervisors’ competencies and occupational wellbeing alongside promoting PhD candidates’ studying.

KEYWORDS:

• Co-supervision
• doctoral supervisors
• occupational wellbeing
• supervisory experience
• competencies

Introduction

Over the past decade, co-supervision of PhD students, that is, collaboration between two or more supervisors, each of them providing substantial and distinct contributions to a supervisee’s work, has become increasingly common across the disciplines (see review by Kálmán et al., 2022). Complementary drivers for the development have been proposed, including the increase in interdisciplinary and inter-sectorial research, specialisation of methods and research knowledge and quality assurance (Grossman & Crowther, 2015; Guerin et al., 2015; Kálmán et al., 2022). Benefits of engaging in co-supervision include a reduction in supervisor absence/abandonment (Olmos-López & Sunderland, 2016), a decrease in supervisory workload (Buttery et al., 2005), enhancement of supervisory development of early career supervisors (Maritz & Prinsloo, 2015), an increase in PhD student satisfaction (Cornér et al., 2017) and transparency of supervisory practices (Ukwoma & Ngulube, 2020). In turn, potential negative influences of co-supervision are said to include disagreement between the supervisors (Chatterjee Padmanabhan & Rossetto, 2017; Kobayashi et al., 2015), a lack of commitment emerging from the inequitable recognition of the second supervisor’s work (Grossman & Crowther, 2015; Olmos-López & Sunderland, 2016), and problems with time management and organising joint meetings due to conflicting timetables (Olmos-López & Sunderland, 2016). As most of the studies on co-supervision have been small scale qualitative case studies (Angus et al., 2022; Johansen et al., 2019; Robertson, 2017), large-scale quantitative studies on co-supervision have yet to be undertaken. Moreover, thus far, most studies have focused on examining influences of co-supervision on PhD students (see Kálmán et al., 2022); less is known about the influences of engaging in co-supervision for the PhD supervisors. We extended the research on co-supervision in two ways: 1) by examining co-supervisory experience from supervisor’s perspective, and 2) by exploring the influences of co-supervision for supervisory competencies and their experienced occupational wellbeing with large scale quantitative survey design.

Co-supervision

We use term co-supervision to refer to supervisory arrangements in which two or more supervisors collaborate in guiding and supporting a PhD student in their doctoral research (see Kálmán et al., 2022). Characteristic for co-supervision is that each of the supervisory team members offers a substantial and distinct contribution to the supervisee’s work by sharing the responsibility (Paul et al., 2014). Co-supervision can be orchestrated in a less or a more formal manner, and within or across sectors and disciplines. Accordingly, supervisors beyond academiafor example, from industry (Tavares et al., 2020) or from other disciplines may engage in co-supervision (Kobayashi et al., 2015). In prior research on co-supervision, some variation in how typically co-supervision is applied has been detected both across the disciplines and in different countries. For example, the sciences have a much longer tradition of co-supervision compared to the humanities and some of the social sciences (Manathunga, 2012). In UK, Australia, Canada, and Europe, many universities have adopted a co-supervision model (Bitusikova, 2009; Manathunga, 2012; Robertson, 2017).

Studies on co-supervision have identified both vertical and horizontal structures in co-supervision (Guerin et al., 2011; Johansen et al., 2019; Kálmán et al., 2022). In vertical collaboration, responsibilities and power are hierarchically distributed between the supervisors, typically meaning that one supervisor is the principal or leading supervisor, while second or third supervisors may be more junior or may have fewer supervisory responsibilities or their responsibilities are more specific regarding the supervision. In horizontal collaboration, power and responsibilities between the supervisors are more evenly shared. Vertical co-supervision with more hierarchical collaborative structures is more typical in co-supervision than horizontal collaboration (Kálmán et al., 2022). Yet, there is tentative evidence that the arrangements are rarely explicit, though they are mutually accepted (Johansen et al., 2019).

Prior research has identified some of the potential pros and cons of co-supervision. Advantages of co-supervision for the PhD candidate are said to include accommodation of supervisory absences (Grossman & Crowther, 2015), improvement in the quality of learning experiences (Ngulube, 2019), higher odds of receiving frequent support and wider range of well-fitted feedback (Kálmán et al., 2022; Paul et al., 2014), as well as reduced risk of the student dropping out (Cornér et al., 2017). In turn, the potential disadvantages of co-supervision involve getting conflicting advice leading to student confusion, or students getting stuck with their research (Chatterjee Padmanabhan & Rossetto, 2017; Ukwoma & Ngulube, 2020; van Biljon & de Villiers, 2013), dispersion of supervisory responsibility leading to ignoring the PhD student’s problems (Paul et al., 2014; Wald et al., 2023), lack of coherent overall academic view on the doctoral research project (Paul et al., 2014), and problems with orchestrating joint meetings (Olmos-López & Sunderland, 2016). Novice co-supervisors are also reported to experience discomfort if they are left alone with the responsibility or handling students’ severe mental health issues or when they accept tasks outside their responsibilities out of fear of compromising their relationship with the more senior co-supervisor (Almlöv & Grubbström, 2023)

The benefits of engaging in co-supervision for supervisors include reduction in workload (Buttery et al., 2005) and enhancement of supervisory development. Particularly for novice supervisors, engaging in co-supervision with more experienced colleagues has shown to be beneficial their supervisory development (Olmos-López & Sunderland, 2017). Accordingly, engagement in co-supervision may have a positive influence on supervisory development and potentially for supervisor’s occupational wellbeing via reduced workload. Yet, empirical evidence of the influences of co-supervision on the supervisors is scarce. Moreover, supervisors’ interest in engaging in co-supervision regardless of the career stage cannot be taken as granted. In a recent study on research supervisors in medicine, none of the novice supervisors wanted to co-supervise and few sought increased involvement of expert advisors (8%), whereas moderately and highly experienced supervisors were considerably more open to the option (16–20%) (Hart et al., 2022).

Aims

We wanted to understand better how supervisory arrangements and disciplinary variation in them including engaging in co-supervision influences on supervisory experience, and whether engagement in co-supervision is related to experiences of supervisory workload, perceptions of supervisory competencies, involvement in supervisory development activities and their occupational wellbeing. The following research questions were addressed:

1. What variations can be detected in a) supervisory model typically applied in the context, b) engagement in co-supervision, c) typical size of the supervisory team, and d) supervisory workload in terms of number of supervisees?

2. Are there disciplinary differences in a) supervisory models typically applied in the context, b) engagement in co-supervision, c) typical size of supervisor team, and d) supervisory workload in terms of number of supervisees?

3. Is engagement in co-supervision related to a) supervisory workload in terms of the number of supervisees, b) supervisory competencies, c) involvement in supervisory development activities, and d) experiences of burnout symptoms and work engagement?

Methods

PhD education in a Finnish case university

The case university, ranked among the top 100 in the world, is one of the 13 Finnish research-intensive universities. About 4900 PhD candidates study at the case university and about 500 PhDs graduate from it each year. All the PhD candidates belong to one of four doctoral schools: a) Doctoral School in Natural Sciences, b) Doctoral School in Health Sciences, c) Doctoral School in Humanities and Social Sciences, and d) Doctoral School in Doctoral School in Environmental, Food and Biological Sciences. Eligibility criteria for applying to study for a PhD include having a Finnish second-cycle master’s degree or comparable foreign degree and demonstrating academic language proficiency in either Finnish, Swedish or English. In addition, a research proposal and PhD study plan must be included in the application.¹ Research is launched at the start of the PhD, with limited complementary course work (typically 40 ECTs) based on the individual study plan. The doctoral dissertation can be written as either an article compilation (typically three published articles in peer-reviewed journals and a summary) or a monograph. Expected graduation time for full time studying is four years, although once receiving a right to study, it is valid for life. There are no tuition fees, but PhD candidates need to cover their own living expenses. Supervisors² are free to choose who they are willing to supervise and on what topic. While the main supervisor is usually a full professor or research director (career stage 4),³ academic rank can vary from full professor to post-doctoral researcher (career stage 2). Typically, at least one of the prospective supervisors is a full professor (permanent position).

Participants

The survey was completed by 561 doctoral supervisors from the case university. The response rate was 16%. The sample represented the whole supervisory population at the case university well regarding gender and disciplinary distribution. The supervisors were from all 11 faculties of the university. The participants represented four doctoral schools and various supervisor career stages. Most supervisors were working as full professors or research directors (44%). Half of the supervisors (51%) had been supervising doctoral students for more than 12 years (M = 14.7 years; SD = 9.7 years). At the time of the data collection, the average number of doctoral degrees supervised to completion was seven, and the number of PhD candidates under supervision was four. Their supervisees typically wrote article-based dissertations (82%) and were studying full-time (68%).

Measurements

The data were collected online during 2021 by using a modified version of a supervisory experience survey (Pyhältö et al., 2022). The survey had been validated in previous studies (Pyhältö et al., 2015). A link to the survey was sent to all supervisors at the case institution via the Doctoral Schools’ supervisor mailing lists. Participation in the study was voluntary, and no identifying information was collected, nor incentives used. For the present study, the following measures were used: frequency of engaging in co-supervision (never/rarely/occasionally/frequently/always), supervisory competence scale (six items, Cronbach alpha = .728) (adapted from the UK Research Supervision survey, UK Council for Graduate Education, 2021), involvement in supervisory development activities (six items with yes/no options) (adapted from UK Research Supervision survey, UK Council for Graduate Education, 2021), burnout scale consisting of three factors: a) exhaustion (five items), b) cynicism (three items), and sense of inadequacy (three items) (Pyhältö et al., 2018 drawing on Maslach Burnout Inventory: Maslach & Jackson, 1981), work engagement (nine items) (Schaufeli et al., 2006). The burnout, work engagement and supervisory competencies scales were measured with a seven-point Likert scales (1 = strongly disagree …7=fully agree). The supervisory workload in terms of number of supervisees was measured with a multiple-choice question: In my opinion, the number of students under my supervision is: (suitable/too high/too low). In addition, disciplinary clusters based on the doctoral schools were constructed (Environmental, food, and biological sciences, Health sciences, Humanities and social sciences, and Natural sciences).

Analysis

A cross-tabulation with a chi-square test was conducted to analyse the disciplinary differences in typical supervisory models and the frequency of engaging in co-supervision. Also, differences in supervisory workload based on the frequency of engaging in co-supervision were analysed with cross-tabulations and chi-square tests. One-way analysis of variance (ANOVA) was used to analyse the differences in supervisory competencies, burnout and work engagement, and involvement in supervisory development activities. The Bonferroni test was used in post-hoc comparisons for differences in supervisory competencies and Games-Howell for differences in burnout symptoms.

Results

Variation in supervisory arrangements and workload

The most typical way (68%) to structure supervision within their field according to the supervisors was few-to-one, i.e. each PhD candidate had a team of two or more supervisors. Also, the one-to-one supervision model (19%,) i.e. each candidate had a single supervisor, was frequently applied. In contrast, some models were rarely used: the many-to-many model (i.e. a cohort/group of candidates are supervised by a group of supervisors) (5%); the many-to-one model (i.e. a single supervisor advising a group of candidates) (4%); or other ways of structuring supervision (4%).

Most supervisors had experience in engaging in co-supervision. They reported being engaging in co-supervision either frequently (34%) or occasionally (24%). A minority of supervisors reported that they had never (17%) or had rarely (15%) been engaged in co-supervision. Also, the number of those who exclusively engaged in co-supervision was low (10%). According to the supervisors, the size of the supervisory team was typically two (52%) to three (42%) supervisors. Larger supervisory teams were less frequently reported (5%). The supervisors most typically perceived the supervisory workload in terms of number of supervisees as being suitable (79%), while a minority reported that the number of supervisees was either too high (9%) or too low (12%). No statistically significant association between the perceived supervisory workload (number of candidates under supervision being high/suitable/low) and the frequency of engaging in co-supervision were detected (χ² = 5.97, df = 6, p =.43).

Disciplinary variations in supervisory arrangements

Differences between the disciplines in the ways the supervision was organised were detected (χ² = 61.48, df = 12, p < .001). More precisely, one-to-one supervision was more usual in natural sciences than in other fields (see Table 1). In addition, few-to-one supervision was more common in the humanities and social sciences, followed by health sciences, whereas in natural sciences and environmental and biological sciences, the model was less frequently applied.

Table 1. Supervisory arrangements and workload by disciplines.

	Environmental, Food and Biological Sciences	Health Sciences	Humanities and Social Sciences	Natural Sciences
Supervisory model
One-to-one	21%	16%	10%	43%
Few-to-one	62%	65%	84%	42%
Many-to-one	7%	6%	2%	6%
Many-to-many	5%	7%	2%	5%
Other	5%	6%	2%	5%
Frequency of engaging in co-supervision
Always or frequently	64%	43%	39%	38%
Occasionally	22%	16%	27%	31%
Rarely	5%	16%	18%	12%
Never	9%	26%	16%	20%
Typical size of a supervisory team
Two supervisors	41%	52%	54%	63%
Three supervisors	52%	45%	42%	43%
Four or more	7%	3%	4%	5%
Supervisory workload
Suitable	80%	77%	82%	81%
Too high	1%	10%	12%	6%
Too low	18%	7%	7%	13%

Disciplinary differences were also detected in the frequency of engaging in co-supervision (χ² = 30.05, df = 9, p < .001). The supervisors in health sciences reported more often that they did not have experience in co-supervision than did the supervisors in environmental, food and biological sciences, whereas the supervisors in humanities and social sciences reported being engaged in co-supervision less often compared to their colleagues in environmental, food and biological sciences. Accordingly, the supervisors in environmental, food and biological sciences were more commonly engaged in co-supervision frequently or always, compared to the supervisors in other fields. There were no disciplinary differences in the size of a typical supervisory team (χ² = 8.33, df = 6, p = .215).

Disciplinary variations also existed in supervisors’ experiences of supervisory workload (χ² = 14.51, df = 6, p =.024). The supervisors in health sciences (10%) and humanities and social sciences (12%) perceived the number of supervisees as being too high more often than did supervisors in environmental, food and biological sciences (1%).

Interrelation between the engagement in co-supervision and supervisory competencies, involvement in supervisory development and occupational wellbeing

The supervisors’ engagement in co-supervision was associated with their perceptions of supervisory competencies (F(3, 548) = 4.16, p = .006). The supervisors who reported that they never engage in co-supervision were less confident with their supervisory competencies (m = 5.10, sd = .88) than supervisors who reported engaging in co-supervision occasionally (m = 5.44, sd = .72, p = .010) or frequently and always (m = 5.41, sd = .77, p = .008). However, no differences existed in supervisor’s involvement in supervisory development activities based on the frequency of engaging in co-supervision (F(3, 553) = 1.93, p = .123).

An interrelationship between co-supervision experiences and supervisor’s wellbeing was detected. Supervisors’ co-supervision experiences were related to their experiences of exhaustion (F(3, 539) = 3.12, p = .026). Supervisors who reported being engaged in co-supervision rarely reported lower levels of exhaustion (m = 2.96, sd = 1.25) than supervisors who reported engaging in co-supervision either never (m = 3.54, sd = 1.30, p = .018) or always or frequently (m = 3.41, sd = 1.41, p = .044). Differences in supervisors’ experiences of cynicism (F(3, 539) = 3.12, p = .026) or inadequacy (F(3, 539) = 3.03, p = .029) were also detected in the whole sample, but the between-group differences were so small that none of them were statistically significant in post-hoc comparisons. No differences were detected in work engagement based on the frequency of being engaged in co-supervision (F(3, 545) = 1.47, p = .221).

Discussion

Methodological limitations and ethical reflections

This study has some methodological strengths and limitations that should be considered. We applied a cross-sectional design in studying supervisory arrangements, field-related variation, and supervisors’ occupational wellbeing, which does not allow us to draw causal conclusions between the study variables. Furthermore, as the study was conducted in only one case university, one needs to be cautious in generalising the results into other socio-cultural contexts, particularly outside Finland. It is also important to note that regardless of the good representativeness of the sample in terms of gender and disciplinary distribution, the response rate was rather low. The data were collected during the COVID-19 pandemic, which might have affected the results and the response rate. Accordingly, further longitudinal, and cross-cultural studies are needed so we can form a better understanding on the impact of co-supervision to the supervisors’ work and wellbeing.

This study followed the principles of research integrity in Finland (Finnish Advisory Board on Research Integrity, 2012). As participation was voluntary, the participants were adults, and the study itself did not pose risks or involve intervention in the physical integrity of the participants, we did not require a formal ethics review (Finnish Advisory Board on Research Integrity, 2009).

The findings in the light of the previous literature

In this study, we explored the supervisory arrangements and disciplinary variation in them. In addition, we aimed to understand the co-supervision influences on supervisory experience by exploring whether engagement in co-supervision was related to supervisory workload, supervisory competencies, involvement in supervisory development activities and occupational wellbeing.

The results indicated that having a supervisory team has become the dominant from of supervision across the disciplines within the case university. Most supervisors reported having experience of co-supervision, though individual and disciplinary variation in the frequency of engagement occurred. The findings are in line with those from the Kálmán et al. (2022) literature review, implying that co-supervision is becoming more and more popular across the globe. Our results further showed that the supervisors in environmental, food and biological sciences reported that they engaged most frequently in co-supervision compared to their colleagues from other fields. A reason for this might be that those in environmental, food and biological sciences often work in laboratories or conduct field work in teams that also provide well-established structures for co-supervision. We did not detect cross-disciplinary variation in the size of supervisory team. However, we did identify cross-disciplinary differences in experienced supervisory workload: the supervisors in environmental, food and biological sciences reported less often the supervisory workload to be too high compared to those in health sciences and in humanities and social sciences. This is in line with prior research suggesting that engaging in co-supervision may reduce a supervisor’s workload (Buttery et al., 2005). On the other hand, the result may reflect the fact that approximately four out of five PhD candidates at the case university study either in the doctoral school of Health Sciences or in the doctoral school of Humanities and Social Sciences, implying significantly higher supervisory workload for an individual supervisor than in the other schools. However, further studies on cross-disciplinary differences across the varied socio-cultural contexts are needed to confirm/overrule our findings.

Further investigation showed that engagement in co-supervision was related to perceived competencies. The supervisors who engaged in co-supervision were more confident with their supervisory competencies compared to those supervisors who did not have experience in co-supervision. This result is in line with recent findings that the most typical professional development activities the supervisors engaged in were the discussions with colleagues about their supervisory practices (Pyhältö et al., 2022). Yet, further studies are needed on the relationship. The tentative finding implies that co-supervision provides an important informal arena for enhancing supervisory development. Systematically harnessing this potential might be particularly useful in cultivating supervisory development among early career supervisors (Olmos-López & Sunderland, 2016). However, this should be considered in pairing senior and junior supervisors and providing structures for supporting meaningful supervisory development.

Our findings concerning involvement in co-supervision and supervisors’ occupational wellbeing were less consistent, yet they give food for thought for further studies. The results showed that supervisors who were either never or frequently or always engaging in co-supervision reported higher levels of exhaustion compared to those supervisors who rarely engaged in co-supervision. On one hand, a reason for these seemingly mixed results might be that while the supervisors who did not have co-supervisory experience had also missed out on learning opportunities as well as opportunities to share workload with another supervisor that come with co-supervision, which might have contributed to feeling of being less competent as supervisors, leading to increased levels of exhaustion. On the other hand, those who tended to engage in co-supervision frequently or always may have a higher risk of experiencing exhaustion because they constantly need to negotiate across the roles and the timetable of other supervisors (Chatterjee Padmanabhan & Rossetto, 2017; Kobayashi et al., 2015). A source of stress in co-supervision, particularly for novice co-supervisors may also stem from dependency with more experienced supervisors: they may for example, have been invited by seniors as co-supervisors, and feel a debt of gratitude, and hence unable to set healthy boundaries (Almlöv & Grubbström, 2023). Accordingly, the risk of experiencing exhaustion for these groups may derive from different sources. In turn, those who rarely engage in co-supervision may have an opportunity to be more selective with whom they co-supervise, or they may have well-established co-supervisory practices. In general, these findings imply that the interrelation between a supervisor’s occupational wellbeing and involvement in co-supervision is not straightforward. Hence, further studies are needed to understand the complexity of the interrelation better.

To conclude, as prior empirical evidence of the pros and cons of engaging in co-supervision for supervisors has been limited, our study contributes to the research on supervisory experience by providing new insights into the potential influences of engaging in co-supervision for experienced supervisory workload, their occupational wellbeing, supervisory competencies, and supervisory development.

Practical implications

The findings have several implications in developing doctoral education. It can be expected that the need for interdisciplinary and inter-sectorial doctoral programs along with the number of PhD candidates will grow in the future. This implies that the demand for multidisciplinary and cross-sectorial co-supervision will grow as well. Considering this, our results imply that investing in cultivating co-supervision may provide a means for enhancing supervisory development as response to the growing need. However, engagement in co-supervision does not automatically result high-quality supervision. Institutional investment, including resources, support, and structures, is also needed to harness the full potential provided by co-supervision for the benefit of doctoral education. Moreover, at its best, investing in developing co-supervision may promote a supervisor’s occupational wellbeing as well.

However, our results imply that although engagement in co-supervision may reduce supervisory workload and decrease the risk of suffering from exhaustion, it can also increase it. This suggests that ways in which co-supervision is conducted matters. Hence, the quantity and the quality of co-supervision matters not only for PhD candidate’s perspective, but also in terms of supervisory development and occupational wellbeing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

K. Pyhältö

Kirsi Pyhältö, PhD, is a professor of higher education in the Faculty of Educational Sciences, University of Helsinki and an extraordinary professor at the Department of Curriculum Studies, University of Stellenbosch, South Africa. She is an expert in the field of research education and careers.

L. Tikkanen

Lotta Tikkanen, PhD, is a university lecturer in the Centre for University Teaching and Learning (HYPE), Faculty of Educational Sciences, University of Helsinki, Finland. Her research interests include doctoral supervision, and doctoral candidates’ and supervisors’ wellbeing.

H. Anttila

Henrika Anttila, PhD, is a senior lecturer in university pedagogy in the Centre for University Teaching and Learning (HYPE), Faculty of Educational Sciences, University of Helsinki, Finland. Her research interests include doctoral candidates’ academic emotions, learning and wellbeing.

Notes

1 The prospective supervisor(s) reviews the research plan and the application prior to submission and supervisory commitment letter is attached to the application.

2 The primary supervisor must have the title of docent, requiring a corpus of scholarly work equal to at least two dissertations as well as good teaching skills. The nationally recognized rank is granted upon an application based on external review.

3 The career stages applied are aligned with European Research Career Framework see more detailed https://euraxess.ec.europa.eu/europe/career-development/training-researchers/research-profiles-descriptors