Doctoral supervision is a fundamental component of doctoral students’ success. Research indicates that doctoral outcomes, including thesis defence, publication productivity, and even future academic trajectories, are highly dependent on the quality of supervision (Bao et al., 2018; Bagaka’s et al., 2015; Li & Fernandez, 2024; McAlpine & McKinnon, 2013; Sverdlik et al., 2018). While many countries have gradually transitioned to distributed supervision models, such as co-supervision, the traditional mentoring model remains prevalent in Russia (Olmos-López & Sunderland, 2017; Terentev & Kuznetsov, 2024; Zhuchkova et al., 2023). This concentration of authority and responsibility in a single supervisor necessitates the exploration of different approaches to interpersonal relationships. As a result, much of the research in this area focuses on various styles and functions of supervision and their impact on doctoral students’ achievements (Salinas-Perez et al., 2019).

Less frequently do studies examine doctoral supervision from an initial-stage perspective, specifically focusing on how the relationships between doctoral candidates and supervisors begin. Yet, interaction with a supervisor before the start of doctoral studies can play a crucial role in future achievements, as it helps set expectations and allows both parties to understand each other’s strengths and weaknesses (Nelson et al., 2006; Zhuchkova & Bekova, 2023). The decision-making process of selecting a supervisor is also given attention in the literature, as making an informed and rational choice can help prevent potential mismatches in the supervisor-student relationship (Moxham et al., 2013). 

However, the selection process from the perspective of doctoral candidates has been researched, though not as extensively as other aspects of the supervisor-student relationship. Candidates may prioritise various factors, such as academic expertise in a specific research topic, previous supervision experience, positive personal relationships, or demographic characteristics (Onen, 2016; Shafiq & Jan, 2017; Zhao et al., 2007). Although the effects of these different selection criteria are not well understood, a small body of literature does address the topic (Shafiq et al., 2020). In contrast, little is known about how supervisors make decisions to take doctoral students on and, more importantly, how these factors relate to student satisfaction and successful thesis defence. The literature that does exist is descriptive in nature and identifies potential factors linked to supervisors’ decisions, such as students’ motivation, previous interactions, and alignment of research interests, but not the effects of these decisions (Friedrich-Nel & Mac Kinnon, 2019). 

This empirical study addresses a gap in previous research by examining how different reasons for supervising relate to satisfaction levels of both doctoral students and supervisors, as well as thesis defence outcomes. To explore this, we utilize a dyadic survey of doctoral students and their supervisors conducted across six Russian universities. We apply three logistic regression models to investigate the link between the dependent variables and the combination of reasons to supervise. 

Drawing on a unique dataset of paired survey responses from doctoral students and their supervisors (N=313) across six Russian universities, this research addresses critical gaps in understanding the initial stages of the supervisor-student relationship. To identify the various reasons to supervise, we used principal component analysis (PCA) on the tetrachoric correlation matrix (psych package in R) with promax rotation (Revelle, 2017), and then we obtained factor scores using the thurstone method. We choose this approach because of expected correlation between different options and the binary nature of the variables (presented as 1 if respondents agreed with the statement in the question and 0 otherwise). Three primary binary outcomes are examined in this study. The first reflects whether a doctoral student encountered difficulties with their supervisor (faced difficulties = 1). The second represents whether the supervisor expressed satisfaction with the doctoral student’s performance (satisfied = 1). The third outcome addresses whether the doctoral student successfully defended their dissertation within the normative period (defended = 1). To better describe the differences in the average values of the components and gain a deeper understanding of the various reasons for supervising, we conducted T-tests for two-sample comparisons and one-way ANOVA-tests for comparisons involving more than two groups. These statistical tests were performed using SciPy library in Python. Next, we examined how the different reasons for supervising were associated with three outcome variables described in the previous section. Since each of these outcome variables is binary, we constructed three logistic regression models using Python’s statsmodels library and presented results with odds ratio, confidence intervals, and standard errors. Our data has a hierarchical structure, as one supervisor can mentor multiple students. This creates potential dependencies within groups, as a supervisor may exhibit a consistent behavioural style and other shared characteristics across their mentees. To account for possible correlation between observations within these groups, we applied the method of clustering standard errors by using their unique ID in the survey. This approach adjusts standard errors for in-group correlations without introducing additional variables into the model, thereby enhancing the reliability and interpretability of the results (Huang et al., 2023).

First, we found that across all our models, the decision to supervise a doctoral student based on positive prior interactions was significantly associated with a reduced likelihood of encountering difficulties in the advisor-student relationship, increased supervisor satisfaction, and a higher probability of dissertation defence in the normative period. The second finding shows that even strong motivation from doctoral students cannot compensate for the lack of prior working experience with a supervisor. We interpret the results from the perspective that supervisors should begin working with doctoral candidates as early as possible, such as during their master’s programs. The 'new route' doctorate (Bao et al., 2018, p. 9), which combines a master’s and doctoral degree, is becoming more popular in European countries and the USA, but remains less common in Russia (Bao et al., 2018; Zhuchkova & Pavliuk, 2024). This extended period allows supervisors and students to have more time to develop their relationship, which has been shown to positively impact outcomes (Seeber & Horta, 2021). Doctoral students may have more time to develop hard skills and gain deeper expertise in their dissertation topic, both of which are positively associated with successful outcomes (Zhuchkova & Bekova, 2023). It also gives supervisors the opportunity to better understand the strengths and weaknesses of their doctoral students, allowing them to tailor an individualised trajectory. This is particularly important in the traditional mentoring model, where distributed supervision is less common. Additionally, this system could be especially beneficial in Russia, where dissertation defence demands considerable time and investment due to the high publication requirements, which are often difficult to meet within the typical 3-4 year timeframe (Bekova et al., 2017).

Bao, Y., Kehm, B. M., & Ma, Y. (2018). From product to process. The reform of doctoral education in Europe and China. Studies in Higher Education, 43(3), 524–541. https://doi.org/10.1080/03075079.2016.1182481 Barnes, B. J., & Austin, A. E. (2009). The Role of Doctoral Advisors: A Look at Advising from the Advisor’s Perspective. Innovative Higher Education, 33(5), 297–315. https://doi.org/10.1007/s10755-008-9084-x Bekova, S. (2021). Does employment during doctoral training reduce the PhD completion rate? Studies in Higher Education, 46(6), 1068–1080. https://doi.org/10.1080/03075079.2019.1672648 Gisemba Bagaka’s, J., Bransteter, I., Rispinto, S., & Badillo, N. (2015). Exploring Student Success in a Doctoral Program: The Power of Mentorship and Research Engagement. International Journal of Doctoral Studies, 10, 323–342. https://doi.org/10.28945/2291 Gruzdev, I., Terentev, E., & Dzhafarova, Z. (2020). Superhero or hands-off supervisor? An empirical categorization of PhD supervision styles and student satisfaction in Russian universities. Higher Education, 79(5), 773–789. https://doi.org/10.1007/s10734-019-00437-w Huang, F. L., Zhang, B., & Li, X. (2023). Using Robust Standard Errors for the Analysis of Binary Outcomes with a Small Number of Clusters. Journal of Research on Educational Effectiveness, 16(2), 213–245. https://doi.org/10.1080/19345747.2022.2100301 Li, L., & Fernandez, F. (2024). Mentoring matters: Examining the relationship between adviser interactions and doctoral student publications. Higher Education. https://doi.org/10.1007/s10734-024-01302-1 Seeber, M., & Horta, H. (2021). No road is long with good company. What factors affect Ph.D. student’s satisfaction with their supervisor? Higher Education Evaluation and Development, 15(1), 2–18. https://doi.org/10.1108/HEED-10-2020-0044 Shafiq, A., & Jan, A. (2017). Factors Influencing Gen-Y Undergraduates’ Choice of Research Supervisor: A Case Study of a Malaysian Private University. Educational Process: International Journal, 6(4), 20–34. https://doi.org/10.22521/edupij.2017.64.2 Shafiq, A., Pahlevan Sharif, S., & Jan, A. (2020). Psychometric Analysis of a Proposed Model to Determine Factors Influencing Selection of a Research Supervisor. International Journal of Doctoral Studies, 15, 285–304. https://doi.org/10.28945/4567 Weidman, J. C., & Stein, E. L. (2003). Socialization of Doctoral Students to Academic Norms. Research in Higher Education, 44(6), 641–656. https://doi.org/10.1023/A:1026123508335 Zhuchkova, S., & Bekova, S. (2023). Building a strong foundation: How pre-doctorate experience shapes doctoral student outcomes. PLOS ONE, 18(9), e0291448. https://doi.org/10.1371/journal.pone.0291448 Zhuchkova, S., Terentev, E., Saniyazova, A., & Bekova, S. (2023). Departmental academic support for doctoral students in Russia: Categorisation and effects. Higher Education Quarterly, 77(2), 215–231. https://doi.org/10.1111/hequ.12389 Zhuchkova, S. V., & Pavliuk, D. M. (2024). Is Doctoral Education A Priority? Doctoral Education Improvement as an Element of 'Priority-2030' Universities’ Development Programs. University Management: Practice and Analysis, 28(1), 21–33. https://doi.org/10.15826/umpa.2024.01.002