Deepening our understanding concerning PhD candidates learning and risk of burnout while doing thesis work Open Access

Studying for a PhD can be stressful which can contribute to some PhD candidates experiencing exhaustion and burnout (Hermann, 2021). There is evidence that PhD candidates are a high risk population who experience substantial stress during their studies (Evans et al., 2018). Thus, investigating candidates’ well-being is important because PhD candidates’ sense of distress, cynicism and inefficacy are also central components of their disengagement (Pyhältö et al., 2023a, 2023b, 2023c; Virtanen et al., 2017). Previous research has focused on exploring the factors that influence PhD candidates’ experiences of exhaustion and burnout and has successfully identified many central factors, such as the research community and supervisory support (Peltonen et al., 2017), lack of interest (Cornér et al., 2021), experiences of inadequate supervision (Cornér et al., 2017; Vekkaila et al., 2013), personal life (more precisely life–work) relations (McAlpine et al., 2020), as well as lack of sense of belonging in scholarly communities (Pyhältö et al., 2024). Moreover, work and organisational context have been found to exert an important effect on PhD candidates’ well-being (Levecque et al., 2017), as has the COVID-19 pandemic and its consequences (Pyhältö et al., 2023b). However, there is still a gap in our knowledge about the relationship of PhD candidates’ learning processes and their experiences of burnout, and to our knowledge, no study has explored this relationship. This is surprising, as among undergraduates, there is growing research indicating that students’ learning processes are clearly related to their experiences of burnout (Asikainen et al., 2022; Kok et al., 2024; Sjöblom et al., 2024). It is evident that a similar relationship can be expected to exist in the PhD context. In addition, although learning is central to the PhD process, only a small number of studies have focused on exploring PhD candidates’ learning processes (Cantwell et al., 2012; Cantwell et al., 2017; Vekkaila and Pyhältö, 2016). Thus, we still lack clear evidence of the PhD candidates’ learning as such and, further, their relations to experiences of burnout (Sverdlik et al., 2018). Therefore, it is important to focus on PhD candidates’ different learning processes and to explore these in a larger sample.

The lack and need for studies focusing on PhD candidates’ learning processes has already been acknowledged several years ago (Vekkaila and Pyhältö, 2016). Previous studies have, for example, focused on PhD candidates’ writing processes which have been widely examined, especially from a sociocultural perspective (Inouye and McAlpine, 2019). Another study by Cantwell et al. (2017) focused on PhD candidates’ metacognition and found that PhD candidates differed in their understanding of the complexity of doctoral studies and how they accepted that complexity. Moreover, they differed in how they assumed ownership of the task by accepting their own responsibility for progress. Cantwell et al. (2017) found that those candidates with an inability to perceive the abstractness and complexity of knowledge struggled with their studies or viewed them as unmanageable. These studies provide an interesting perspective on PhD candidates, but there is still a need to provide a wider perspective on candidates’ learning as PhD studies are a multidimensional learning process.

One way to examine learning processes is the approaches to learning framework (Asikainen and Gijbels, 2017). Approaches to learning theory share the same, basic assumptions as cognitive psychology that students’ prior experiences, expectations and beliefs influence how they approach learning in different situations (Heikkilä and Lonka, 2006). Moreover, they have been detected in several educational and working life contexts (Kirby et al., 2003; Janeiro et al., 2017; Moreira et al., 2021). Thus, it may be expected that similar approaches emerge in PhD contexts.

Many studies exploring learning have detected three different approaches to learning (Biggs, 2001; Entwistle et al., 2006; Entwistle, 1988, 2009; Janeiro et al., 2017). A deep approach to learning refers to the intention to analyse and understand information through comprehending the bigger picture, focusing on underlying meanings and integrating new information into previous knowledge. By contrast, a surface approach is characterised by a focus on memorisation, often resulting in a fragmented knowledge base where information is seen as atomised and unrelated (Entwistle, 2009). Recent research suggests that the core of the surface approach is unreflective study, which consists of the experience of a fragmented knowledge base, difficulty in reflecting on one’s own learning and, instead, a focus on memorisation and repetition of knowledge. Therefore, today, the concept of an unreflective approach is preferred to that of a surface approach (Lindblom-Ylänne et al., 2018). A third approach is organised studying – a term currently used to refer more to students’ everyday practices of managing and organising their studies and time and effort than to their achievement orientation. Consequently, it is also considered more of an approach to studying than an approach to learning (Entwistle, 2009; Entwistle and McCune, 2004), but it is related to learning processes, as the ability to regulate and reflect on one’s own learning is crucial for effective use of learning processes and student achievement (Magno, 2009; Tuononen et al., 2023). Previous research has found that undergraduates with high scores on measures of the deep and organised approach have less burnout (Asikainen et al., 2022), whereas those with an unreflective approach (previously surface approach) are more likely to have study-related burnout (Asikainen et al., 2022). Thus, it could be expected that disorganised and unreflective PhD candidates also experience more study-related burnout.

Approaches to learning have hardly been studied among PhD candidates. Only one small interview study by Vekkaila and Pyhältö (2016) focusing on PhD candidates’ learning patterns suggests that candidates representing the passive producer approach, that is, using surface approach for managing concrete learning outcomes such as measures and analyses, emphasise performance and mastery of different tasks related to research and studies but are unable to affect and change the conditions in which they operate. These PhD candidates also appear to experience problems in regulating and controlling their learning processes (Vekkaila and Pyhältö, 2016). These different learning patterns strongly resemble approaches to learning. For example, the passive producer approach resembles the surface approach to learning detected among undergraduates. Thus, we can expect to find similar kinds of learning patterns among PhD candidates.

Burnout in the context of work has been defined as emotional exhaustion, cynicism and reduced professional efficacy (Maslach et al., 2001). A similar definition of burnout has emerged in school and university contexts, where study-related burnout has been defined through three different components: study-related exhaustion, cynicism and lack of study-related efficacy (Salmela-Aro et al., 2009; Salmela-Aro and Kunttu, 2010; Schaufeli et al., 2001). Study-related exhaustion refers to feelings of being burdened or exhausted resulting from overtaxing work; cynicism refers to lack of interest and a cynical or indifferent attitude toward studying generally and in relation to others; and lack of professional efficacy refers to feelings of incompetence and poor achievement in studying (Salmela-Aro et al., 2009).

All these aspects of study-related burnout have been found to affect students’ engagement with or dedication to their studies (Salmela-Aro and Upadyaya, 2017), also in the PhD context (Vekkaila et al., 2013). Previous review studies have explored the aspects that influence PhD candidates’ well-being (Schmidt and Hansson, 2018; Sverdlik et al., 2018). For example, Schmidt and Hansson’s (2018) review study has found many aspects which affect candidates’ well-being including external factors such as the quality of the supervision, fit between supervisors and supervisees, personal and social lives, financial opportunities as well as internal factors comprising motivation, writing skills and regulatory strategies in writing and academic identity. However, these reviews do not report a single study exploring PhD candidates’ learning when doing thesis work.

Experiences of burnout can be understood through the demand-resources model (Salmela-Aro and Upadyaya, 2017; Schaufeli and Bakker, 2004), which indicates that an imbalance between experiences of demands and resources can lead to burnout if experiences of demands exceed experiences of resources. Previous studies with undergraduate students have shown that students’ scoring high on unreflective approach to learning can experience more negative emotions in studying and have more negative perceptions of the teaching–learning environment than students applying a deep approach to learning (Kuittinen and Meriläinen, 2011; Trigwell et al., 2012). A clear positive relationship between unreflective approach and experiences of burnout has been found among undergraduates (Asikainen et al., 2022; Parpala et al., 2021). However, the relation can also be bidirectional, that is, unreflective may lead to burnout and vice-versa (Kok et al., 2024). Thus, we suggest that there is a relationship between PhD candidates’ learning approaches and their burnout. More precisely, we suggest for PhD candidates applying the unreflective approach to learning can lead to a more demanding experience of the doctoral work and, thus, lead to experiences of burnout. In addition, we suggest that applying a deep approach to learning can lead to a more comprehensive understanding and act as a personal resource and through interest, lower the risk of burnout. However, this relation can also be bidirectional.

The purpose of this study is to examine what kind of approaches to learning can be detected among PhD candidates in the contexts of doing thesis work and to determine how these approaches are related to candidates’ study-related burnout. The specific research questions are the following:

This study applied a mixed methodology by combining survey data with interview data to complement the findings. The study followed the ethical principles of research with human participants and ethical review in the human sciences in Finland by following Finnish National Board on Research Integrity guidelines 2019.

The survey data for this study were collected from four doctoral schools at a large Finnish university. An open invitation to participate in the study was sent to all of the doctoral school candidates via doctoral school coordinators. The study subjects consist of 418 PhD candidates. The sample represents about 10% of all of the doctoral candidates of the university. Informed consent was obtained from all subjects involved in the study. Of the respondents, 293 were female, 122 were male and 3 not specified. Moreover, 246 respondents were from the humanities and social sciences, 62 from natural sciences, 65 from health science and 35 from environmental, food and biological sciences. In all, 12 PhD candidates did not specify their field of study. A total of 9.5% were first-year PhD students, 20% second-year students, 15% were third-year PhD students, 15.7% were fourth year students and 39.8% had started their studies five years ago or more. In the questionnaire, the candidates were offered the possibility of participating in an interview and leaving their contact information after completing the inventory. A total of 93 candidates were contacted and interviewed to gather the interview data for the project.

The data was gathered using the HowULearn questionnaire (Parpala and Lindblom-Ylänne, 2012), which was contextualised to PhD candidates to measure their approaches to learning in thesis work. This was important, as studying in Finland to be a PhD is very different from being an undergraduate student. The Finnish PhD education has a strong emphasis in conducting research and writing their thesis work instead of having courses (Pyhältö et al., 2012). The studies involved in PhD education are mainly methodological courses and international conferences. Moreover, the candidates participate in research seminars. Linking research training and professional development is not very typical for Finnish doctoral education.

In the present study, we used two sections of the questionnaire: the part concerning students’ approaches to learning and the part related to study-related burnout. Concerning approaches to learning, three different approaches were measured: a deep approach, an unreflective approach and organised studying with a five-point Likert type scale (1 = totally agree, 5 = totally disagree). The section measuring learning approaches (Parpala and Lindblom-Ylänne, 2012) has been used and validated in many university contexts for many years (Broseghini et al., 2024; Cheung et al., 2020; Herrmann et al., 2017; Asikainen et al., 2022; Postareff et al., 2018; Ruohoniemi et al., 2017). The section concerning study-related burnout originated from the instrument SBI-9 (Salmela-Aro et al., 2009), developed and validated for measuring study-related burnout in a university context (Salmela-Aro and Read, 2017). The measure consisted of nine Likert-type items (1 = disagree and 5 = agree) measuring three dimensions of study-related burnout: cynicism, inadequacy and exhaustion. Moreover, HowULearn has previously been used to measure the relation between approaches to learning and study-related burnout among undergraduate students (Kok et al., 2024; Sjöblom et al., 2024; Yin et al., 2024).

The contextualisation of the HowULearn questionnaire to the PhD context consisted of multiple different phases. Firstly, the contextualisation of HowULearn began with a team of experts in university pedagogy adapting items to PhD candidates. The most significant change in the new inventory was related to the focus of the items. Instead of concentrating on study processes, PhD candidates were asked to consider their experiences of researching and writing their doctoral thesis, as completing a thesis represents a concrete learning process that is mandatory for PhD candidates in their studies. Secondly, several focus group interviews were organised in different disciplines to validate the questionnaire before data collection. A sample of PhD candidates from each doctoral school was invited to the focus group interview. They completed the questionnaire and marked the questions they did not understand; afterwards, these questions were discussed together with the first author. After this, these items were discussed within a team of university experts and modified according to the comments.

In the next phase, data was collected with the contextualised questionnaire, and an exploratory factor analysis was conducted. As a consequence, two items were removed because they had extremely low communalities (below 0.25) or they loaded on many factors. One item from scale deep approach: “In my PhD thesis, I try to form a coherent whole of its contents” was removed. In addition, the item initially measuring an unreflective approach “Often I have to repeat things to learn them” was removed. After these changes, the scale deep approach focused especially on searching for evidence and analysing and understanding information through comprehending the bigger picture (Entwistle, 2009). Moreover, the scale unreflective approach focused especially on difficulties on relating ideas, therefore resulting in a fragmented knowledge base (Lindblom-Ylänne et al., 2019). These dimensions are crucial elements in both deep and unreflective (previously termed surface) approaches (Entwistle and McCune, 2004; Lindblom-Ylänne et al., 2019), ensuring that the theoretical foundation of the instrument remained solid and consistent with the approaches to learning theory (see  Appendix 1 for the final items).

In the final phase, after the questionnaire data collection, participants were invited to individual interviews to discuss the different dimensions of approaches to learning. The procedure was done as follows: The candidates were asked to explain their answers to the questions on “approaches to learning in thesis work”. The interviewees were shown their own answers to the questionnaire and were asked to reflect retrospectively on those answers and explain why they answered the way they answered. The interviewees did not know the dimension of the items; rather, they focused solely on single items. The interviews lasted from 30 to 90 min. This data provided descriptions of the different dimensions of approaches to learning.

After the interviews, the researchers identified candidates from the survey data who had scored high on a particular approach (e.g. a deep approach). Descriptions of the dimensions were read from students who scored ≥ M = 4.33 for the unreflective approach (n = 16), students’ who scored M = 5.00 for organised studying (n = 19) and for deep approach, from the students’ who scored M = 5.00 (n = 66) to reach a similar proportion of interviews as with the other approaches. Then these candidates’ descriptions of their learning were carefully read through and compared to the survey results. The same procedure was repeated for each approach. The interview data provided support for the dimensions detected in the survey data, and the differences in the descriptions between the three dimensions were evident. For example, participants who scored high on deep approach described the use of such learning processes that aim at deeper understanding of their own research and how it relates to a wider research area. Oppositely, the participants scoring high on unreflective approach described challenges with forming a coherent understanding of their own thesis topic. These various descriptions were in line with candidates’ answers to the items measuring dimensions of the learning processes in the inventory and provided further support for the different dimensions (for excerpts, please see  Appendix 2).

The initial analysis for missing values showed that ten items measuring PhD candidates’ approaches to learning contained one missing value. These were replaced with means. Firstly, confirmatory factor analysis (CFA) was performed to explore the factor structure of items measuring learning approaches and study-related burnout using R 4.2.2 with the package Lavaan 0.6–12. CFA is a statistical technique used to test whether a set of observed variables represent the underlying latent constructs (factors) that they are theoretically supposed to measure. It is commonly used in psychometrics, social sciences and behavioural research to validate measurement models. Confirmatory analysis is considered a good method for testing a theoretical model (Brown, 2014), and thus, it is suitable for this study aiming to test whether approaches to learning can be detected among PhD candidates. After the CFA, sum variables of the different scales were formed, and their interrelations were initially measured with Pearson correlations.

The relationship between approaches to learning in thesis work and different factors of study-related burnout and starting year were then examined by linear regression. Three different regression analyses were performed using approaches to learning in thesis work and starting year as dependent variables and factors of study-related burnout as independent variables.

The CFA performed with the items measuring approaches to learning in thesis work indicated that the three-factor model fitted the data well. The fit indexes showed adequate fit (χ² = 139.9, df = 41, p < 0.001, CFI = 0.913 and RMSEA = 0.076), indicating that the three-factor solution was valid in the data. The first factor consisted of items measuring an unreflective approach to learning (M = 2.46, SD = 0.85, sk = 0.414 and ku = −0.299) and slightly more than 30% of the respondents received a score of at least 3 (on a scale of 1–5). This factor focused on candidates’ difficulties in forming a coherent whole from their doctoral thesis and consisted of four items (e.g. “I have trouble forming a coherent whole from my PhD thesis”). The second factor emphasised candidates’ deep approach to learning (M = 4.11, SD = 0.62, sk = −0.739 and ku = 0.244), such as relating ideas and forming a coherent whole from their own research; this factor consisted of four items (e.g. “While writing my thesis, I try to make use of different viewpoints on the subject matter as much as possible”). The organised studying factor (M = 3.56, SD = 0.80, sk = −0.256 and ku = −0.510) consisted of four items measuring PhD candidates’ time and effort management (e.g. “On the whole, I work on my PhD thesis in a systematic and organised way”). The descriptives can be seen in Table 1.

The CFA concerning study-related burnout was performed with three dimensions measuring different aspects of burnout – Exhaustion (M = 2.67, SD = 1.03, sk = 0.449 and ku = −0.549), cynicism (M = 2.15, SD = 1.06, sk = 0.882 and ku = 0.025) and inadequacy (M = 3.22, SD = 1.16, sk = −0.234 and ku = −0.865) – that have been successfully used by previous research (Salmela-Aro et al., 2009). This three-dimensional model fitted the data adequately (χ² = 123.5, df = 24, CFI = 0.941 and RMSEA = 0.100).

Based on the skewness and kurtosis of the scales measuring approaches to learning and study-related burnout, the normal distribution assumption was met (Kim, 2013), and parametric tests were decided to be used. Next, we explored whether there were differences in participants’ approaches to learning and study-related burnout based on their starting year. Our ANOVA analysis showed that there were no differences between the students’ starting year and their approaches to learning (p = 0.442–0.597 and η² = 0.007–0.009). Thus, we analysed the data as a whole. Concerning study-related burnout, we found that exhaustion did not differ between the starting years, but there was a statistically significant difference in cynicism (p = 0.005 and η² = 0.035) and inadequacy (p = 0.006 and η² = 0.034). Tukey’s test showed that the first-year students differed from the fifth- and further-year students concerning both cynicism and inadequacy.

Firstly, correlations between the PhD candidates’ burnout scales and approaches to learning in thesis work were explored (Table 2). All the correlations between the scales were statistically significant. The deep approach was positively related to the organised approach, and they were both negatively related to the unreflective approach. The different burnout scales – Exhaustion (M = 2.67, SD = 1.03, sk = 0.449 and ku = −0.549), cynicism (M = 2.15, SD = 1.06, sk = 0.882 and ku = 0.025) and inadequacy (M = 3.22, SD = 1.16, sk = −0.234 and ku = −0.865) – were strongly and positively related to unreflective learning and negatively related to the deep approach and organised studying. The analysis revealed a relationship between approaches to learning in thesis work and different dimensions of study-related burnout. The unreflective approach was strongly related to two dimensions of study-related burnout. The opposite pattern was true for the deep approach and organised studying.

Next, we analysed how approaches to learning and starting year explained components of the study-related burnout. Regression analyses were first performed to predict exhaustion based on candidates’ approaches to learning in thesis work and starting year. A statistically significant model was found [F(4, 415) = 13.07 and p < 0.0001] with an R² of 0.10. The results showed that unreflective approach and the starting year were both negatively associated with exhaustion, but there was no statistically significant association with the deep approach and organised studying. In addition, a model where cynicism was predicted by all the three approaches to learning in thesis work and starting year was found statistically significant [F(4, 415) = 47.48 and p < 0.0001] with an R² of 0.31. Unreflective approach and starting year were positively associated, and the deep approach and organised studying were negatively associated with cynicism. Finally, a statistically significant model was found when examining how inadequacy was predicted by the approaches to learning in thesis work [F(4, 415) = 36.11 and p < 0.0001] with an R² of 0.25. Inadequacy was positively and statistically significantly associated with the unreflective approach and starting year and negatively associated with organised studying. Results can be seen in Table 3.

The first aim of the study was to explore the approaches to learning that can be detected among PhD candidates. On the basis of previous research among PhD candidates (Janeiro et al., 2017; Moreira et al., 2021; Vekkaila and Pyhältö, 2016), we expected to find many similarities between undergraduates’ approaches to learning and PhD candidates’ learning patterns. We used CFA to determine whether the data from PhD candidates supports the theory and factor structure of approaches to learning. Importantly, the results of this study show that the dimensions that emerged from the instrument were consistent with the research on undergraduates’ approaches to learning, that is, with the deep, unreflective and organised approaches identified in different contexts (Entwistle, 2009; Lindblom-Ylänne et al., 2018; Asikainen et al., 2022). Moreover, the correlation structure between the scales was also consistent with the previous literature on approaches to learning (Parpala et al., 2013), providing more support for the results. Interestingly, the approaches of PhD candidates did not differ based on their year of study. The participants’ years of study ranged from 1 to over 5 years, suggesting that learning approaches can be observed consistently throughout the various stages of PhD studies, from the initial to the advanced phases. Furthermore, the PhD candidates’ interviews and descriptions of their learning while doing thesis work were also in line with the dimensions and theories behind each scale measuring approaches to learning among undergraduate students (Entwistle, 2009).

The quantitative results showed that PhD candidates scored the highest on the deep approach to learning and second highest on organised studying. The candidates scored lowest on the unreflective approach, which is not surprising, as it measures undesirable learning patterns compared to deep and organised approaches (Richardson, 2012). Still, a particularly interesting finding was that the scale for the unreflective approach, which measures learners’ fragmented knowledge base and inability to form a coherent whole (Lindblom-Ylänne et al., 2018), emerged clearly as a separate and clear dimension. The original scale, used among undergraduates, included item measuring repetition of knowledge, but during the contextualising the instrument, this particular item was removed because of low communality. Therefore, the scale unreflective approach is measuring mainly the PhD candidates inability to form a coherent whole from their thesis work which has shown to be one of the crucial elements of the unreflective (surface) approach to learning (Entwistle and McCune, 2004; Lindblom-Ylänne et al., 2019). PhD candidates’ own descriptions during the interviews provided explanations for their answers to the survey items measuring this approach. These candidates described not only difficulties in forming a coherent whole from their thesis but also unclear aims and an inability to reflect on their choices while researching and writing their PhD theses. By contrast, when PhD candidates explained their answers to the scales measuring the deep approach, they highlighted the importance of argumentation and critical thinking in increasing their in-depth understanding of their own research. Finally, the participants’ explanations for their answers to the items measuring organised studying included descriptions of planning time and effort. For some participants, this occurred in a disorganised manner, where setting clear aims, goals and schedules was experienced as difficult. These descriptions indicated differences in PhD candidates learning processes although the data was just used for questionnaire development purposes. Future research should focus more deeply on the qualitative learning processes of the PhD candidates.

Similar to studies among undergraduate students (Salmela-Aro and Upadyaya, 2017), we were able to detect three different dimensions (exhaustion, cynicism and inadequacy) of PhD candidates’ study-related burnout. Of these three dimensions, candidates in our study exhibited the highest mean level in inadequacy. Previous research on candidates’ well-being has focused, for example, on exhaustion and cynicism (Hunter and Devine, 2016; McAlpine et al., 2020). Consequently, this study highlights the need to examine PhD candidates’ sense of inadequacy when focusing on their well-being. This is partly in line with a recent study which shows that many PhD students can have an impostor syndrome (Nori and Vanttaja, 2023). Moreover, we found more variation in the dimensions of study-related burnout than in the dimensions of learning processes (deep, unreflective and organised). This suggests that, in the future, person-oriented research on PhD candidates’ well-being is important to reveal the individual differences in the dimensions of burnout.

The second aim was to explore the relationship between the approaches to learning in thesis work and the risk of burnout among PhD candidates. Here, again, our correlation results were in the line with findings among undergraduate students (Asikainen et al., 2019): the unreflective approach was positively and significantly related to the dimensions measuring the risk of burnout, more precisely to exhaustion, cynicism and inadequacy. Furthermore, the deep and organised approaches were negatively and statistically significantly related to all the dimensions regarding the risk of burnout. This finding also resembles the results produced by Stubb et al. (2011) that the ability to relate ideas supports the well-being of PhD candidates. Our study implies that also the organised approach, that is, time and effort management, plays an important role in PhD candidates’ well-being. Time and effort management can hinder experiences of burnout, but lower well-being might also cause difficulties in time and effort management. In addition, our regression analyses supported that the unreflective approach exhibited a systematic positive association with all dimensions of study-related burnout. All the approaches to learning in thesis work were significantly associated with cynicism in the multivariable model: the unreflective approach positively and both the deep and organised approach negatively. The results, thus, imply that, among PhD candidates, organised studying and a deep approach to learning (and, conversely, low adherence to the unreflective approach) discourages cynicism in the process of writing the thesis although causal relationship cannot be assumed. Thus, an unreflective approach to learning in thesis work is an unfavourable learning pattern in the light of risk of burnout in PhD process. However, more research on these relationships is required. Furthermore, the regression analyses showed that the starting year was significantly and positively associated with all the aspects of study-related burnout. Thus, it seems that the risk of study-related burnout is smaller in the beginning of studies. Similar results have also been found among undergraduates suggesting that exhaustion increases during studies (Räisänen et al., 2020).

In both the inventory data and interviews, three different learning approaches were detected which resembled the learning dimensions emerging also among undergraduates (Entwistle, 2009). It is particularly important to acknowledge that many PhD candidates may struggle with an unreflective approach to learning. One might assume that inability to perceive the connection between various elements of a PhD thesis is an inevitable aspect of the process of writing a thesis and becoming a researcher. This study shows that this is something which occurs even in every phase of the PhD studies, and thus, PhD candidates need support in constituting a whole of their studies. Supervisors can support this, for example, by asking PhD candidates to keep a learning diary, drawing a mind map or other figures of the core concepts in their studies, discussing with them about their experiences and prior knowledge on their thesis, make them to consider what is the main aim of their study and asking them to speak it out loud and reminding them of those aims and research questions and how their thesis is in line with them. Tools such as Research Canvas could also be used to support forming a coherent whole of the thesis (Ketchen et al., 2019). In addition, it would be beneficial to make PhD candidates acknowledge and monitor their approaches to learning which could be done, for example, by completing an inventory regarding their approaches to learning (Backhaus and Liff, 2007) and discussing the results together with the supervisors. This is important as an unreflective approach to learning is related to feelings of inadequacy and cynicism, as the present study demonstrates. Candidates’ ability to reflect on their learning and critical thinking could be fostered using, for example, literature assignments (Turner, 2013). An interesting approach would be to ask PhD candidates to read or discuss a fictional book or story, as it could help them change their thought patterns, encouraging critical thinking and leading them to reject previously held assumptions (Gouthro and Holloway, 2018). Most importantly, the candidates should be encouraged to notice their approaches to learning in thesis work and understand the effect they have on their well-being. With the supervisors, the candidate may find different ways to organise the thesis learning process and reflect the aims and choices made during the process.

This study had some limitations. Firstly, the data was self-report data and cross-sectional in nature, and thus, causal relationships could not be explored. Thus, in this study, the purpose was to explore students’ experiences about the learning approaches and study-related burnout. Future research should also consider using objective measures of well-being and conducting longitudinal research on the relationship of approaches to learning in thesis work and study-burnout risk. This would also make it possible to examine the possible bi-directional relationship between these measures. Secondly, it should be noted that the unreflective approach was in this study measured only with three items as mentioned in the theory. Thus, the repetition or memorisation in learning was not included in the measurement. It may be possible that the scale, thus, measured the struggling in the learning processes more. In other words, the dimension of a fragmented knowledge base, which is a core element of an unreflective and surface approach to learning, is emphasised in PhD studies (Entwistle and McCune, 2004). However, this has also been the case in higher education contexts more generally (Nieminen et al., 2019).

Thirdly, this study did not take the study programs and their differences into account, even though the participants represented various fields of study. This might affect the generalizability of the results. However, recent research has shown that the dimensions of the approaches to learning emerge similar in different disciplines although there might be differences in how they are emphasised in different learning contexts (Asikainen et al., 2022). Moreover, there could be disciplinary variation in the PhD candidates’ risk of burnout, as previous research implies that PhD candidates’ social support is related both to discipline and burnout (Peltonen et al., 2017).

Finally, the study focused only on the dimensions of approaches and study-related burnout, although there is evidence of the individual variation in how candidates combine the different dimensions and thus suggests a person-oriented focus on approaches to learning to be more fruitful (Asikainen et al., 2019; Asikainen et al., 2022; Kymäläinen et al., 2023; Mendoza et al., 2022; Yin et al., 2023). This is also supported by the low correlations between the deep and unreflective approach in this study which imply that these two approaches are not mutually exclusive. Therefore, more research is needed on the different individual combinations of PhD candidates’ approaches to learning in thesis work in different disciplines. Our study offers a reliable instrument to continue examining PhD learning processes in more detail.

Exploring PhD candidates’ approaches to learning in thesis work offers a promising new perspective to explore factors influencing candidates’ well-being. This study emphasises that PhD candidates’ experiences of high workload and burnout could be partly explained by their approaches to learning in thesis work. Especially difficulties in forming a coherent understanding of own thesis and an inability to reflect one’s own choices appear to be a crucial element and clearly related to all dimensions of burnout. The supervisors have an important role in supporting their PhD candidates in forming a coherent understanding of their thesis work.

• I often have trouble forming a coherent whole of my PhD thesis.

• Many themes related to the contents of my thesis remain disconnected from each other.

• The contents of my PhD feel so complicated that I often have trouble understanding them.

• While doing my PhD thesis, I often contemplate the ideas and their connections from multiple perspectives.

• I carefully look for evidence to reach my own conclusions while doing my PhD thesis.

• While writing my thesis, I try to make use of different viewpoints on the subject matter as much as possible.

• I put a lot of effort into my PhD thesis.

• On the whole, I work on my PhD thesis in a systematic and organised way.

• I organise the time reserved for my PhD thesis carefully to make the best use of it.

• I have made a schedule so that I can complete my PhD thesis as planned.

• I feel overwhelmed by my PhD studies.

• I feel a lack of motivation in my PhD studies and often think of giving up.

• I often have feelings of inadequacy in my PhD studies.

• I often sleep badly because of matters related to my PhD studies.

• I feel that I am losing interest in my PhD studies.

• I constantly wonder whether my PhD studies have any meaning.

• I brood over matters related to my PhD studies a lot during my free time.

• I used to have higher expectations of my PhD studies than I do now.

• The pressure of my PhD studies causes me problems in my close relationships with others.

Source(s): Authors’ own creation

The following interview excerpt describes a typical way in which PhD candidates scoring high on the deep approach described their learning process:

The high scores on organised approach were in line with the PhD candidates’ own descriptions of their time and effort management skills. Candidates who had received high scores on the organised studying dimension described how they attempted to make the best use of their time by planning carefully, setting deadlines and prioritising and using different ways to manage and regulate their own learning. The following excerpts demonstrated good time and effort management skills:

Finally, when candidates were explaining their answers to items measuring the unreflective approach, they described difficulties in forming a coherent whole and reflecting on their learning, which is in line with the theory of an unreflective approach (Lindblom-Ylänne et al., 2018):

Source(s): Authors’ own creation