Skip to main content

Understanding the mental health of doctoral researchers: a mixed methods systematic review with meta-analysis and meta-synthesis

Abstract

Background

Data from studies with undergraduate and postgraduate taught students suggest that they are at an increased risk of having mental health problems, compared to the general population. By contrast, the literature on doctoral researchers (DRs) is far more disparate and unclear. There is a need to bring together current findings and identify what questions still need to be answered.

Methods

We conducted a mixed methods systematic review to summarise the research on doctoral researchers’ (DRs) mental health. Our search revealed 52 articles that were included in this review.

Results

The results of our meta-analysis found that DRs reported significantly higher stress levels compared with population norm data. Using meta-analyses and meta-synthesis techniques, we found the risk factors with the strongest evidence base were isolation and identifying as female. Social support, viewing the PhD as a process, a positive student-supervisor relationship and engaging in self-care were the most well-established protective factors.

Conclusions

We have identified a critical need for researchers to better coordinate data collection to aid future reviews and allow for clinically meaningful conclusions to be drawn.

Systematic review registration

PROSPERO registration CRD42018092867

Peer Review reports

Background

Student mental health has become a regular feature across media outlets in the United Kingdom (UK), with frequent warnings in the media that the sector is facing a ‘mental health crisis’ [1]. These claims are largely based on the work of regulatory authorities and ‘grey’ literature. Such sources corroborate an increase in the prevalence of mental health difficulties amongst students. In 2013, 1 in 5 students reported having a mental health problem [2]. Only 3 years later, however, this figure increased to 1 in 4 [3]. In real terms, this equates to 21,435 students disclosing mental health problems in 2013 rising to 49,265 in 2017 [4]. Data from the Higher Education Statistics Agency (HESA) demonstrates a 210% increase in the number of students terminating their studies reportedly due to poor mental health [5], while the number of students dying by suicide has consistently increased in the past decade [6].

This issue is not isolated to the UK. In the United States (US), the prevalence of student mental health problems and use of counselling services has steadily risen over the past 6 years [7]. A large international survey of more than 14,000 students across 8 countries (Australia, Belgium, Germany, Mexico, Northern Ireland, South Africa, Spain and the United States) found that 35% of students met the diagnostic criteria for at least one common mental health condition, with highest rates found in Australia and Germany [8].

The above figures all pertain to undergraduate students. Finding equivalent information for postgraduate students is more difficult, and where available tends to combine data for postgraduate taught students and doctoral researchers (DRs; also known as PhD students or postgraduate researchers) (e.g. [4]). The latest trend analysis based on data from 36 countries suggests that approximately 2.3% of people will enrol in a PhD programme during their lifetime [9]. The countries with the highest number of DRs are the US, Germany and the UK [10]. At present, there are more than 281,360 DRs currently registered across these three countries alone [11, 12], making them a significant part of the university population. The aim of this systematic review is to bring attention specifically to the mental health of DRs by summarising the available evidence on this issue.

Using a mixed methods approach, including meta-analysis and meta-synthesis, this review seeks to answer three research questions: (1) What is the prevalence of mental health difficulties amongst DRs? (2) What are the risk factors associated with poor mental health in DRs? And (3) what are the protective factors associated with good mental health in DRs?

Methods

Literature search

We conducted a search of the titles and abstracts of all article types within the following databases: AMED, BNI, CINAHL, Embase, HBE, HMIC, Medline, PsycInfo, PubMed, Scopus and Web of Science. The same search terms were used within all of the databases, and the search was completed on the 13th April 2018. Our search terms were selected to capture the variable terms used to describe DRs, as well as the terms used to describe mental health, mental health problems and related constructs. We also reviewed the reference lists of all the papers included in this review. Full details of the search strategy are provided in the supplementary material.

Inclusion criteria

Articles meeting the following criteria were considered eligible for inclusion: (1) the full text was available in English; (2) the article presented empirical data; (3) all study participants, or a clearly delineated sub-set, were studying at the doctoral level for a research degree (DRs or equivalent); and (4) the data collected related to mental health constructs. The last of these criteria was operationalised (a) for quantitative studies as having at least one mental health-related outcome measure, and (b) for qualitative studies as having a discussion guide that included questions related to mental health. We included university-published theses and dissertations as these are subjected to a minimum level of peer-review by examiners.

Exclusion criteria

In order to reduce heterogeneity and focus the review on doctoral research as opposed to practice-based training, we excluded articles where participants were studying at the doctoral level, but their training did not focus on research (e.g. PsyD doctorate in Clinical Psychology).

Screening articles

Papers were screened by one of the present authors at the level of title, then abstract, and finally at full text (Fig. 1). Duplicates were removed after screening at abstract. At each level of screening, a random 20% sub-set of articles were double screened by another author, and levels of agreement were calculated (Cohen’s kappa [13]). Where disagreements occurred between authors, a third author was consulted to decide whether the paper should or should not be included. All kappa values evidence at least moderate agreement between authors [14]—see Fig. 1 for exact kappa values.

Fig. 1
figure1

PRISMA diagram of literature review process

Data extraction

This review reports on both quantitative and qualitative findings, and separate extraction methods were used for each. Data extraction was performed by authors CH, CB, SV and LC.

Quantitative data extraction

The articles in this review used varying methods and measures. To accommodate this heterogeneity, multiple approaches were used to extract quantitative data. Where available, we extracted (a) descriptive statistics, (b) correlations and (c) a list of key findings. For all mental health outcome measures, we extracted the means and standard deviations for the DR participants, and where available for the control group (descriptive statistics). For studies utilising a within-subjects study design, we extracted data where a mental health outcome measure was correlated with another construct (correlations). Finally, to ensure that we did not lose important findings that did not use descriptive statistics or correlations, we extracted the key findings from the results sections of each paper (list of key findings). Key findings were identified as any type of statistical analysis that included at least one mental health outcome.

Qualitative data extraction

In line with the meta-ethnographic method [15] and our interest in the empirical data as well as the authors’ interpretations thereof, i.e. the findings of each article [16], the data extracted from the articles comprised both results/findings and discussion/conclusion sections. For articles reporting qualitative findings, we extracted the results and discussion sections from articles verbatim. Where articles used mixed methods, only the qualitative section of the results was extracted. Methodological and setting details from each article were also extracted and provided (see Appendix A) in order to contextualise the studies.

Data analysis

Quantitative data analysis

Descriptive statistics

We present frequencies and percentages of the constructs measured, the tools used and whether basic descriptive statistics (M and SD) were reported. The full data file is available from the first author upon request.

Effect sizes

Where studies had a control group, we calculated a between-group effect size (Cohen’s d) using the formula reported by Wilson [17], and interpreted using the standard criteria [13]. For all other studies, we sought to compare results with normative data where the following criteria were satisfied: (a) at least three studies reported data using the same mental health assessment tool; (b) empirical normative data were available; and (c) the scale mean/total had been calculated following original authors’ instructions. Only the Perceived Stress Scale (PSS) 10- [18] and 14-item versions [19] met these criteria. Normative data were available from a sample of adults living in the United States: collected in 2009 for the 10-item version (n = 2000; M = 15.21; SD = 7.28) [20] and in 1983 for the 14-item version (n = 2355; M = 19.62; SD = 7.49) [18].

The meta-analysis of PSS data was conducted using MedCalc [21], and based on a random effects model, as recommended by [22]. The between-group effect sizes (DRs versus US norms) were calculated comparing PSS means and standard deviations in the respective groups. The effect sizes were weighted using the variable variances [23].

Correlations

Where at least three studies reported data reflecting a bivariate association between a mental health and another variable, we summarised this data into a meta-analysis using the reported r coefficients and sample sizes. Again, we used MedCalc [21] to conduct the analysis using a random effects model, based on the procedure outlined by Borenstein, Hedges, Higgins and Rothstein [24]. This analysis approach involves converting correlation coefficients into Fisher’s z values [25], calculating the summary of Fisher’s z, and then converting this to a summary correlation coefficient (r). The effect sizes were weighted in line with the Hedges and Ollkin [23] method. Heterogeneity was assessed using the Q statistic, and I2 value—both were interpreted according to the GRADE criteria [26]. Where correlations could not be summarised within a meta-analysis, we have reported these descriptively.

Narration

Due to the heterogenous nature of the studies, the above methods could not capture all of the quantitative data. Therefore, additional data (e.g. frequencies, statistical tests) reported in the identified articles was collated into a single document, coded as relating to prevalence, risk or protective factors and reported as a narrative review.

Qualitative data analysis

Coding

We used thematic analytic methods to analyse the qualitative data. We followed the thematic synthesis method [16, 27] and were informed by a thematic analysis approach [28, 29]. We took a critical realist epistemological stance [30, 31] and aimed to bring together an analysis reflecting meaningful patterns amongst the data [29] or demi-regularities, and identifying potential social mechanisms that might influence the experience of such phenomena [31]. The focus of the meta-synthesis is interpretative rather than aggregative [32].

Coding was line by line, open and complete. Following line-by-line coding of all articles, a thematic map was created. Codes were entered on an article-by-article basis and then grouped and re-grouped into meaningful patterns. Comparisons were made across studies to attempt to identify demi-regularities or patterns and contradictions or points of departure. The thematic map was reviewed in consultation with other authors to inductively create and refine themes. Thematic summaries were created and brought together into a first draft of the thematic structure. At this point, each theme was compared against the line-by-line codes and the original articles in order to check its fit and to populate the written account with illustrative quotations.

Research rigour

The qualitative analysis was informed by independent coding by authors CB and SV, and analytic discussions with CH, SV and LC. Our objective was not to capture or achieve inter-rater reliability, rather the analysis was strengthened through involvement of authors from diverse backgrounds including past and recent PhD completion, experiences of mental health problems during PhD completion, PhD supervision experience, experience as employees in a UK university doctoral school and different nationalities. In order to enhance reflexivity, CB used a journal throughout the analytic process to help notice and bracket personal reflections on the data and the ways in which these personal reflections might impact on the interpretation [29, 33]. The ENTREQ checklist [34] was consulted in the preparation of this report to improve the quality of reporting.

Quality assessment

Quantitative data

The quality of the quantitative papers was assessed using the STROBE combined checklist [35]. A random 20% sub-sample of these studies were double-coded and inter-rater agreement was 0.70, indicating ‘substantial’ agreement [14]. The maximum possible quality score was 23, with a higher score indicating greater quality, with the mean average of 15.97, and a range from 0 to 22. The most frequently low-scoring criteria were incomplete reporting regarding the management of missing data, and lack of reported efforts to address potential causes of bias.

Qualitative data

There appeared to be no discernible pattern in the perceived quality of studies; the highest [36,37,38,39,40] and lowest scoring [41,42,43,44,45,46] studies reflected both theses and journal publications, a variety of locations and settings and different methodologies. The most frequent low-scoring criteria were relating to the authors’ positions and reflections thereof (i.e. ‘Qualitative approach and research paradigm’, ‘Researcher characteristics and reflexivity’, ‘Techniques to enhance trustworthiness’, ‘Limitations’, ‘Conflict of interest and Funding’). Discussions of ethical issues and approval processes was also frequently absent. We identified that we foregrounded higher quality studies in our synthesis in that these studies appeared to have greater contributions reflected in the shape and content of the themes developed and were more likely to be the sources of the selected illustrative quotes.

Mixed methods approach

The goal of this review is to answer the review questions by synthesising the findings from both quantitative and/or qualitative studies. To achieve our goal, we adopted an integrated approach [47], whereby we used both quantitative and qualitative methods to answer the same review question, and draw a synthesised conclusion. Different analysis approaches were used for the quantitative and qualitative data and are therefore initially reported separately within the methods. A separate synthesised summary of the findings is then provided.

Results

Overview of literature

Of the 52 papers included in this review (Table 1), 7 were qualitative, 29 were quantitative and 16 mixed methods. Most articles (35) were peer-reviewed papers, and the minority were theses (17). Only four of the articles included a control group; in three instances comprising students (but not DRs) and in the other drawn from the general population.

Table 1 List of studies included in this review

Quantitative results

Descriptive statistics

Thirty-five papers reported quantitative data, providing 52 reported sets of mental health related data (an average of 1.49 measures per study): 24 (68.57%) measured stress, 10 (28.57%) anxiety, 9 (25.71%) general wellbeing, 5 (14.29%) social support, 3 (8.57%) depression and 1 (2.86%) self-esteem. Five studies (9.62%) used an unvalidated scale created for the purposes of the study. Fifteen studies (28.85%) did not report descriptive statistics.

Effect sizes

Of the four studies that included a control group, only two of these reported descriptive statistics for both groups on a mental health outcome [66, 69]. There is a small (Cohen’s d = 0.27) and large between-group effect (Cohen’s d = 1.15) when DRs were compared to undergraduate and postgraduate clinical psychology students respectively in terms of self-reported stress.

The meta-analysis of DR scores on the PSS (both 10- and 14-item versions) compared to population normative data produced a large and significant between-group effect size (d = 1.12, 95% CI [0.52, 1.73]) in favour of DRs scoring higher on the PSS than the general population (Fig. 2), suggesting DRs experience significantly elevated stress. However, these findings should be interpreted in light of the significant between-study heterogeneity that can be classified as ‘considerable’ [26].

Fig. 2
figure2

A meta-analysis of between-group effect sizes (Cohen’s d) comparing PSS scores (both 10- and 14-item versions) from DRs and normative population data. *Studies using the 14 item version of the PSS; a positive effect size indicates DRs had a higher score on the PSS; a negative effect size indicates that the normative data produced a higher score on the PSS; black diamond = total effect size (based on random effects model); d = Cohen’s d; Q = heterogeneity; Z = z score; I2 = proportion of variance due to between-study heterogeneity; p = exact p value

To explore this heterogeneity, we re-ran the meta-analysis separately for the 10- and 14-item versions. The effect size remained large and significant when looking only at the studies using the 14-item version (k = 6; d = 1.41, 95% CI [0.63, 2.19]), but was reduced and no longer significant when looking at the 10-item version only (k = 3; d = 0.57, 95% CI [− 0.51, 1.64]). However, both effect sizes were still marred by significant heterogeneity between studies (10-item: Q = 232.02, p < .001; 14-item: Q = 356.76, p < .001).

Correlations

Studies reported sufficient correlations for two separate meta-analyses; the first assessing the relationship between stress (PSS [18, 19]) and perceived support, and the second between stress (PSS) and academic performance.

Stress x support

We included all measures related to support irrespective of whom that support came from (e.g. partner support, peer support, mentor support). The overall effect size suggests a small and significant negative correlation between stress and support (r = − .24, 95% CI [− 0.34, − 0.13]) (see Fig. 3), meaning that low support is associated with greater perceived stress. However, the results should be interpreted in light of the significant heterogeneity between studies. The I2 value quantifies this heterogeneity as almost 90% of the variance being explained by between-study heterogeneity, which is classified as ‘substantial’ (26).

Fig. 3
figure3

Forest plot and meta-analysis of correlation coefficients testing the relationship between stress and perceived support. Black diamond = total effect size (based on random effects model); r = Pearson’s r; Q = heterogeneity; Z = z score; I2 = proportion of variance due to between-study heterogeneity; p = exact p value

Stress x performance

The overall effect size suggests that there is no relationship between stress and performance in their studies (r = − .07, 95% CI [− 0.19, 0.05]) (see Fig. 4), meaning that DRs perception of their progress was not associated with their perceived stress This finding suggests that the amount of progress that DRs were making during their studies was not associated with stress levels.

Fig. 4
figure4

Forest plot and meta-analysis of correlation coefficients testing the relationship between stress and performance. Black diamond = total effect size (based on random effects model); r = Pearson’s r; Q = heterogeneity; Z = z score; I2 = proportion of variance due to between-study heterogeneity; p = exact p value

Other correlations

Correlations reported in less than three studies are summarised in Fig. 5. Again, stress was the most commonly tested mental health variable. Self-care and positive feelings towards the thesis were consistently found to negatively correlate with mental health constructs. Negative writing habits (e.g. perfectionism, blocks and procrastination) were consistently found to positively correlate with mental health constructs. The strongest correlations were found between stress, and health related quality of life (r = − .62) or neuroticism (r = .59), meaning that lower stress was associated with greater quality of life and reduced neuroticism. The weakest relationships (r < .10) were found between mental health outcomes and: faculty concern, writing as knowledge transformation, innate writing ability (stress and anxiety), years married, locus of control, number of children and openness (stress only).

Fig. 5
figure5

Correlation coefficients testing the relationship between a mental health outcome and other construct. Correlation coefficients are given in brackets (r); *p < .05; each correlation coefficient reflects the results from a single study

Narration

Prevalence

Several studies reported DR mental health problem prevalence and this ranged from 36.30% [54] to 55.9% [67]. Using clinical cut-offs, 32% were experiencing a common psychiatric disorder [64]; with another study finding that 53.7% met the questionnaire cut-off criteria for depression, and 41.9% for anxiety [67]. One study compared prevalence amongst DRs and the general population, employees and other higher education students; in all instances, DRs had higher levels of psychological distress (non-clinical), and met criteria for a clinical psychiatric disorder more frequently [64].

Risk factors

Demographics Two studies reported no significant difference between males and females in terms of reported stress [57, 73], but the majority suggested female DRs report greater clinical [80], and non-clinical problems with their mental health [37, 64, 79, 83, 89].

Several studies explored how mental health difficulties differed in relation to demographic variables other than gender, suggesting that being single or not having children was associated with poorer mental health [64] as was a lower socioeconomic status [71]. One study found that mental health difficulties did not differ depending on DRs’ ethnicity [51], but another found that Black students attending ‘historically Black universities’ were significantly more anxious [87]. The majority of the studies were conducted in the US, but only one study tested for cross-cultural differences: reporting that DRs in France were more psychologically distressed than those studying in the UK [67].

Work-life balance Year of study did not appear to be associated with greater subjective stress in a study involving clinical psychology DRs (Platt and Schaefer [75]), although other studies suggested greater stress reported by those in the latter part of their studies [89], who viewed their studies as a burden [81], or had external contracts, i.e. not employed by their university [85]. Regression analyses revealed that a common predictor of poor mental health was uncertainty in DR studies; whether in relation to uncertain funding [64] or uncertain progress [80]. More than two-thirds of DRs reported general academic pressure as a cause of stress, and a lack of time as preventing them from looking after themselves [58]. Being isolated was also a strong predictor of stress [84].

Protective factors

DRs who more strongly endorsed all of the five-factor personality traits (openness, conscientiousness, extraversion, agreeableness and neuroticism) [66], self-reported higher academic achievement [40] and viewed their studies as a learning process (rather than a means to an end) [82] reported fewer mental health problems. DRs were able to mitigate poor mental health by engaging in self-care [72], having a supervisor with an inspirational leadership style [64] and building coping strategies [56]. The most frequently reported coping strategy was seeking support from other people [37, 58].

Qualitative results

Meta-synthesis

Four higher-order themes were identified: (1) Always alone in the struggle, (2) Death of personhood, (3) The system is sick and (4) Seeing, being and becoming. The first two themes reflect individual risk/vulnerability factors and the processes implicated in the experience of mental distress, the third represents systemic risk and vulnerability factors and the final theme reflects individual and systemic protective mechanisms and transformative influences. See Table 2 for details of the full thematic structure with illustrative quotes.

Table 2 Thematic structure with illustrative quotes

Always alone in the struggle

‘Always alone in the struggle’ reflects the isolated nature of the PhD experience. Two subthemes reflect different aspects of being alone; ‘Invisible, isolated and abandoned’ represents DRs’ sense of physical and psychological separation from others and ‘It’s not you, it’s me’ represents DRs’ sense of being solely responsible for their PhD process and experience.

Invisible, isolated and abandoned

Feeling invisible and isolated both within and outside of the academic environment appears a core DR experience [39, 43, 81]. Isolation from academic peers seemed especially salient for DRs with less of a physical presence on campus, e.g. part-time and distance students, those engaging in extensive fieldwork, outside employment and those with no peer research or lab group [36, 52, 68]. Where DRs reported relationships with DR peers, these were characterised as low quality or ‘not proper friendships’ and this appeared linked to a sense of essential and obvious competition amongst DRs with respect to current and future resources, support and opportunities [39], in which a minority of individuals were seen to receive the majority share [36, 74]. Intimate sharing with peers thus appeared to feel unsafe. This reflected the competitive environment but also a sense of peer relationships being predicated on too shared an experience [39].

In addition to poor peer relations, a mismatch between the expected and observed depth of supervisor interest, engagement and was evident [40, 81]. This mismatch was clearly associated with disappointment and anger, and a sense of abandonment, which appeared to impact negatively on DR mental health and wellbeing [42] (p. 182). Moreover, DRs perceived academic departments as complicit in their isolation; failing to offer adequate opportunities for academic and social belonging and connections [42, 81] and including PGRs only in a fleeting or ‘hollow’ sense [37]. DRs identified this isolation as sending a broader message about academia as a solitary and unsupported pursuit; a message that could lead some DRs to self-select out of planning for future in academia [37, 42]. DRs appeared to make sense of their lack of belonging in their department as related to their sense of being different, and that this difference might suggest they did not ‘fit in’ with academia more broadly [74]. In the short-term, DRs might expend more effort to try and achieve a social and/or professional connection and equitable access to support, opportunities and resources [74]. However, over the longer-term, the continuing perception of being professionally ‘other’ also seemed to undermine DRs’ sense of meaning and purpose [81] and could lead to opting out of an academic career [62, 74].

Isolation within the PhD was compounded by isolation from one’s personal relationships. This personal isolation was first physical, in which the laborious nature of the PhD acted as a catalyst for the breakdown of pre-existing relationships [76]. Moreover, DRs also experienced a sense of psychological detachment [45, 74]. Thus, the experience of isolation appeared to be extremely pervasive, with DRs feeling excluded and isolated physically and psychologically and across both their professional and personal lives.

It’s not you, it’s me

‘It’s not you, it’s me’ reflects DRs’ perfectionism as a central challenge of their PhD experience and a contributor to their sense of psychological isolation from other people. DRs’ perfectionism manifested in four key ways; firstly, in the overwhelming sense of responsibility experienced by DRs; secondly, in the tendency to position themselves as inadequate and inferior; thirdly, in cycles of perfectionist paralysis; and finally, in the tendency to find evidence which confirms their assumed inferiority.

DRs positioned themselves as solely responsible for their PhD and for the creation of a positive relationship with their supervisor [36, 52, 81]. DRs expressed a perceived need to capture their supervisors’ interest and attention [36, 52, 74], feeling that they needed to identify and sell to their supervisors some shared characteristic or interest in order to scaffold a meaningful relationship. DRs appeared to feel it necessary to assume sole responsibility for their personal lives and to prohibit any intrusion of the personal in to the professional, even in incredibly distressing circumstances [42].

DRs appeared to compare themselves against an ideal or archetypal DR and this comparison was typically unfavourable [37], with DRs contrasting the expected ideal self with their actual imperfect and fallible self [37, 42, 52]. DRs’ sense of inadequacy appeared acutely and frequently reflected back to them by supervisors in the form of negative or seemingly disdainful feedback and interactions [41, 76]. DRs framed negative supervisor responses as a cue to work harder, meaning they were continually striving, but never reaching, the DR ideal [76]. This ideal-actual self-discrepancy was associated with a tendency towards punitive self-talk with clear negative valence [38].

DRs appear to commonly use self-castigation as a necessary (albeit insufficient) means to motivate themselves to improve their performance in line with perfectionistic standards [38, 41]. The oscillation between expectation and actuality ultimately resulted in increased stress and anxiety and reduced enjoyment and motivation. Low motivation and enjoyment appeared to cause procrastination and avoidance, which lead to a greater discrepancy between the ideal and actual self; in turn, this caused more stress and anxiety and further reduced enjoyment and motivation leading to a sense of stuckness [76].

The internalisation of perceived failure was such that DRs appeared to make sense of their place, progress and possible futures through a lens of inferiority, for example, positioning themselves as less talented and successful compared to their peers [37]. Thus, instances such as not being offered a job, not receiving funding, not feeling connected to supervisors, feeling excluded by academics and peers were all made sense of in relation to DRs’ perceived relative inadequacy [36].

Death of personhood

The higher-order theme ‘Death of personhood’ reflects DRs’ identity conflict during the PhD process; a sense that DRs’ engage in a ‘Sacrifice of personal identity’ in which they feel they must give up their pre-existing self-identity, begin to conceive of themselves as purely ‘takers’ personally and professionally, thus experiencing the ‘Self as parasitic’, and ultimately experience a ‘Death of self-agency’ in relation to the thesis, the supervisor and other life roles and activities.

A sacrifice of personal identity

The sacrifice of personal identity first manifests as an enmeshment with the PhD and consequent diminishment of other roles, relationships and activities that once were integral to the DRs’ sense of self [59, 76]. DRs tended to prioritise PhD activities to the extent that they engaged in behaviours that were potentially damaging to their personal relationships [76]. DRs reported a sense of never being truly free; almost physically burdened by the weight of their PhD and carrying with them a constant ambient guilt [37, 38, 44, 76]. Time spent on non-PhD activities was positioned as selfish or indulgent, even very basic activities of living [76].

The seeming incompatibility of aspects of prior personal identity and the PhD appears to result in a sense of internal conflict or identity ‘collision’ [59]. Friends and relatives often provided an uncomfortable reflection of the DR’s changing identity, leaving DRs feeling hyper-visible and carrying the burden of intellect or trailblazer status [74]; providing further evidence for the incompatibility of their personal and current and future professional identities. Some DRs more purposefully pruned their relationships and social activities; to avoid identity dissonance, to conserve precious time and energy for their PhD work, or as an acceptance of total enmeshment with academic work as necessary (although not necessarily sufficient) for successful continuation in academia [40, 52, 77]. Nevertheless, the diminishment of the personal identity did not appear balanced by the development of a positive professional identity. The professional DR identity was perceived as unclear and confusing, and the adoption of an academic identity appeared to require DRs to have a greater degree of self-assurance or self-belief than was often the case [37, 81].

Self as parasitic

Another change in identity manifested as DRs beginning to conceive of themselves as parasitic. DRs spoke of becoming ‘takers’, feeling that they were unable to provide or give anything to anyone. For some DRs, being ‘parasitic’ reflected them being on the bottom rung of the professional ladder or the ‘bottom of the pile’; thus, professionally only able to receive support and assistance rather than to provide for others. Other DRs reported more purposefully withdrawing from activities in which they were a ‘giver’, for example voluntary work, as providing or caring for others required time or energy that they no longer had [38, 44]. Furthermore, DRs appeared to conceive of themselves as also causing difficulty or harm to others [81], as problems in relation to their PhD could lead them to unwillingly punishing close others, for example, through reducing the duration or quality of time spent together [38].

Feeling that close others were offering support appeared to heighten the awareness of the toll of the PhD on the individual and their close relationships, emphasising the huge undertaking and the often seemingly slow progress, and actually contributing to the sense of ambient guilt, shame, anger and failure [38]. Moreover, DRs spoke of feeling extreme guilt in perceiving that they had possibly sacrificed their own, and possibly family members’, current wellbeing and future financial security [49].

Death of self-agency

In addition to their sense of having to sacrifice their personal identity, DRs also expressed a loss of their sense of themselves as agentic beings. DRs expressed feeling powerless in various domains of their lives. First, DRs positioned the thesis as a powerful force able to overwhelm or swallow them [46, 52, 59]. Secondly, DRs expressed a sense of futility in trying to retain any sense of personal power in the climate of academia. An acute feeling of powerlessness especially in relation to supervisors was evident, with many examples provided of being treated as means to an end, as opposed to ends in themselves [39, 42, 62]. Supervisors did not interact with DRs in a holistic way that recognised their personhood and instead were perceived as prioritising their own will, or the will of other academics, above that of the DR [39, 62].

Furthermore, DRs reported feeling as if they were used as a means for research production or furthering their supervisors’ reputations or careers [62]. DRs perceived that holding on to a sense of personal agency sometimes felt incompatible with having a positive supervisor relationship [42]. Thus whilst emotional distress, anger, disappointment, sadness, jealousy and resentment were clearly evident in relation to feeling excluded, used or over-powered by supervisors [37, 42, 52, 62], DRs usually felt unable to change supervisor irrespective of how seriously this relationship had degraded [37, 62]. Instead, DRs appeared to take on a position of resignation or defensive pessimism, in which they perceived their supervisors as thwarting their personhood, personal goals and preferences, but typically felt compelled to accept this as the status quo and focus on finishing their PhDs [42]. DRs resignation was such that they internalised this culture of silence and silenced themselves; tending to share litanies of problems with supervisors whilst prefacing or ending the statements with some contradictory or undermining phrase such as ‘but that’s okay’ [42, 52].

The apparent lack of self-agency extended outward from the PhD into DRs not feeling able to curate positive life circumstances more generally [76]. A lack of time was perhaps the key struggle across both personal and professional domains, yet DRs paradoxically reported spending a lot of time procrastinating and rarely (if ever) mentioned time management as a necessary or desired coping strategy for the problem of having too little time [46]. The lack of self-agency was not only current but also felt in reference to a bleak and uncertain future; DRs lack of surety in a future in academia and the resultant sense of futility further undermined their motivation to engage currently with PhD tasks [38, 40].

The system is sick

The higher-order theme ‘The system is sick’ represents systemic influences on DR mental health. First, ‘Most everyone’s mad here’ reflects the perceived ubiquity mental health problems amongst DRs. ‘Emperor’s new clothes’ reflects the DR experience of engaging in a performative piece in which they attempt to live in accordance with systemic rather than personal values. Finally, ‘Beware the invisible and visible walls’ reflects concerns with being caught between ephemeral but very real institutional divides.

Most everyone’s mad here

No studies focused explicitly on experiences of DRs who had been given diagnoses of mental health problems. Some study participants self-disclosed mental health problems and emphasised their pervasive impact [50]. Further lived experiences of mental distress in the absence of explicit disclosure were also clearly identifiable. The ‘typical’ presentation of DRs with respect to mental health appeared characterised as almost unanimous [39] accounts of chronic stress, anxiety and depression, emotional distress including frustration, anger and irritability, lack of mental and physical energy, somatic problems including appetite problems, headaches, physical pain, nausea and problems with drug and alcohol abuse [39, 46, 59, 76]. Health anxiety, concerns regarding perceived new and unusual bodily sensations and perceived risks of developing stress-related illnesses were also common [46, 59, 76]. A PhD-specific numbness and hypervigilance was also reported, in which DRs might be less responsive to personal life stressors but develop an extreme sensitivity and reactivity to PhD-relevant stimuli [39].

An interplay of trait and state factors were suggested to underlie the perceived ubiquity of mental health problems amongst DRs. Etiological factors associated with undertaking a PhD specifically included the high workload, high academic standards, competing personal and professional demands, social isolation, poor resources in the university, poor living conditions and poverty, future and career uncertainty [36, 41, 43, 46, 49, 76]. The ‘nexus’ of these factors was such that the PhD itself acted as a crucible; a process of such intensity that developing mental health problems was perhaps inevitable [39].

The perceived inevitability of mental health problems was such that DRs described people who did not experience mental health problems during a PhD as ‘lucky’ [39]. Supervisors and the wider academic system were seen to promote an expectation of suffering, for example, with academics reportedly normalising drug and alcohol problems and encouraging unhealthy working practices [39]. Furthermore, DRs felt that academics were uncaring with respect to the mental challenge of doing a PhD [39]. Nevertheless, academics were suggested to deny any culpability or accountability for mental health problems amongst DRs [39, 59, 74]. The cycle of indigenousness was further maintained by a lack of mental health literacy and issues with awareness, availability and access to help-seeking and treatment options amongst DRs and academics more widely [39]. Thus, DRs appeared to feel they were being let down by a system that was almost set up to cause mental distress, but within which there was a widespread denial of the size and scope of the problem and little effort put into identifying and providing solutions [39, 59]. DRs ultimately felt that the systemic encouragement of unhealthy lifestyles in pursuit of academic success was tantamount to abuse [62].

A performance of optimum suffering

Against a backdrop of expected mental distress, DRs expressed their PhD as a performative piece. DRs first had to show just the right amount of struggle and difficulty; feeling that if they did not exhibit enough stress, distress and ill-health, their supervisors or the wider department might not believe they were taking their PhD seriously enough [40]. At the same time, DRs felt that their ‘researcher mettle’ was constantly being tested and they must rise to this challenge. This included first guarding against presenting oneself as intellectually inferior [36]. Yet it also seemed imperative not to show vulnerability more broadly [74]. Disclosing mental or physical health problems might lead not only to changed perceptions of the DR but to material disadvantage [74]. The poor response to mental health disclosures suggested to some DRs that universities might be purposefully trying to dissuade or discourage DRs with mental health problems or learning disabilities from continuing [74]. The performative piece is thus multi-layered, in that DRs must experience extreme internal psychological struggles, exhibit some lower-level signs of stress and fatigue for peer and faculty observance, yet avoid expressing any real academic or interpersonal weakness or the disclosure of any diagnosable disability or disease.

Emperor’s new clothes

DRs described feeling beholden to the prevailing culture in which it was expected to prioritise above all else developing into a competitive, self-promoting researcher in a high-performing research-active institution [39, 42]. Supervisors often appeared the conduit for transmission of this academic ideal [74]. DRs felt reticent to act in any way which suggested that they did not personally value the pursuit of a leading research career above all else. For example, DRs felt that valuing teaching was non-conformist and could endanger their continuing success within their current institution [55]. Many DRs thus exhibited a sense of dissonance as their personal values often did not align with the institutional values they identified [74]. Yet DRs expressed a sense of powerlessness and a feeling of being ‘caught up’ in the values of the institution even when such values were personally incongruent [74]. The psychological toll of this sense of inauthenticity seemed high [55]. Where DRs acted in ways which ostensibly suggested values other than prioritising a research career, for example becoming pregnant, they sensed disapproval [76]. DRs also felt unable to challenge other ‘institutional myths’ for example, the perceived institutional denial of the duration of and financial struggle involved in completing a PhD [49]. There was a perceived tendency of academics to locate problems within DRs as opposed to acknowledging institutional or systemic inequalities [49]. DRs expressed strongly a sense in which there is inequity in support, resources and opportunities, yet universities were perceived as ignoring such inequity or labelling such divisions as based on meritocracy [36, 74].

Beware the invisible and visible walls

DRs described the reality of working in academia as needing to negotiate a maze of invisible and visible walls. In the former case, ‘invisible walls’ reflect ephemeral norms and rules that govern academia. DRs felt that a big part of their continuing success rested upon being able to negotiate such rules [39]. Where rules were violated and explicit or implicit conflicts occurred, DRs were seen to be vulnerable to being caught in the ‘crossfire’ [36]. DRs identified academic groups and departments as being poor in explicitly identifying, discussing and resolving conflicts [37]. The intangibility of the ‘invisible walls’ gave rise to a sense of ambient anxiety about inadvertently transgressing norms and divides, such that some DRs reported behaving in ways that surprised even themselves [37].

Gendered and racial micropolitics of academic institutions were seen to manifest as more visible walls between people, with institutions privileging those with ‘insider’ status [36]. Women and people of colour typically felt excluded or disadvantaged in a myriad of observable and unobservable ways, with individuals able to experience both insider and outsider statuses simultaneously [36, 37], for example when a male person of colour [36]. Female DRs suggested that not only must women prove themselves to a greater extent than men to receive equal access to resources, opportunities and acclaim but also are typically under additional pressure in both their professional and personal lives [37, 52, 76]. Women also felt that they had to take on more additional roles and responsibilities and encountered more conflicts in their personal lives compared to men [52]. Examples of professionally successful women in DRs’ departments were described as those who had crossed the divide and adopted a more traditionally male role [40]. Thus, being female or non-White were considered visible characteristics that would disadvantage people in the competitive academic environment and could give rise to a feeling of increased stress, pressure, role conflicts, and a feeling of being unsafe.

Seeing, being and becoming

The higher-order theme of ‘Seeing, being and becoming’ reflects protective and transformative influences on DR mental health. ‘De-programming’ refers to the DRs disentangling their personal beliefs and values from systemic values and also from their own tendency towards perfectionism. ‘The power of being seen’ reflects the positive impact on DR mental health afforded by feeling visible to personal and professional others. ‘Finding hope, meaning and authenticity’ refers to processes by which DRs can find or re-locate their own self-agency, purpose and re/establish a sense of living in accordance with their values. ‘The importance of multiple goals, roles and groups’ represents the beneficial aspects of accruing and sustaining multiple aspects to one’s identity and connections with others and activities outside the PhD. Finally, ‘The PhD as a process of transcendence’ reflects how the struggles involved in completing a PhD can be transformative and self-actualising.

De-programming

DRs reported being able to protect their mental health by ‘de-programming’ and disentangling their attitudes and practices from social and systemic values and norms. This disentangling helped negate DRs’ adopting unhealthy working practices and offered some protection against experiencing inauthenticity and dissonance between personal and systemic values.

First, DRs spoke of rejecting the belief that they should sacrifice or neglect personal relationships, outside interests and their self-identity in pursuit of academic achievement. DRs could opt-out entirely by choosing a ‘user-friendly’ programme [44] which encouraged balance between personal and professional goals, or else could psychologically reject the prevailing institutional discourse [40]. Rather than halting success, de-programming from the prioritisation of academia above all else was seen to be associated not only with reduced stress but greater confidence, career commitment and motivation [40, 50]. It was also suggested possible to ‘de-programme’ in the sense of choosing not to be preoccupied by the ‘invisible walls’ of academia and psychologically ‘opt out’ of being concerned by potential conflicts, norms and rules governing academic workplace conduct [36]. Interaction with people outside of academia was seen to scaffold de-programming, by helping DRs to stay ‘grounded’ and offering a model what ‘normal’ life looks like. People outside of academia could also help DRs to see the truth by providing unbiased opinions regarding systemic practices [39].

A further way in which de-programming manifested was in DRs challenging their perfectionist beliefs. This include re-framing the goal as not trying to be the archetype of a perfect DR, and accepting that multiple demands placed on one individual invariably requires compromise [40, 76]. DRs spoke of the need to conceptualise the PhD as a process, rather than just a product [46, 82]. The process orientation facilitated framing of the PhD as just one-step in the broader process of becoming an academic as opposed to providing discrete evidence of worth [82]. Within this perspective, uncertainty itself could be conceived as a privilege [81]. The PhD was then seen as an opportunity rather than a test [37, 46]. Moreover, the process orientation facilitated viewing the PhD as a means of growing into a contributing member of the research community, as opposed to needing to prove oneself to be accepted [82]. Remembering the temporary nature of the PhD was advised [45] as was holding on to a sense that not completing the PhD was also a viable life choice [76]. DRs also expressed, implicitly or explicitly, a decision to change their conceptualisation of themselves and their progress; choosing not to perceive themselves as stuck, but planning, learning and progressing [38, 39, 81, 82]. This new perspective appeared to be helpful in reducing mental distress.

The power of being seen

DRs described powerful benefits to feeling seen by other people, including a sense of belonging and mattering, increased self-confidence and a sense of positive progress [37]. Being seen by others seems to provoke the genesis of an academic identity; it brings DRs into existence as academics. Being seen within the academic institution also supports mental health and can buffer emotional exhaustion [37, 52, 55, 81]. DRs expressed a need to feel that supervisors, academics and peers were interested in them as people, their values, goals, struggles and successes; yet they also needed to feel that they and their research mattered and made a difference within and outside of the institution [42, 52, 81]. It was clear that DRs could find in their disciplinary communities the sense of belonging that often eluded them within their immediate departments [42]. Feeling a sense of belonging to the academic community seemed to buffer disengagement and amotivation during the PhD [81]. Positive engagement with the broader community was scaffolded by a sense of trust in the supervisor [81]. DRs often felt seen and supported by postdocs, especially where supervisors appeared absent or unsupportive [50].

Spending time with peers could be beneficial when there was a sense of shared experience and walking alongside each other [39]. Friendship was seen to buffer stress and protect against mental health problems through the provision of social and emotional support and help in identifying struggles [39, 43]. In addition to relational aspects, the provision of designated physical spaces on campus or in university buildings also seemed important to being seen [37]. Peers in the university could provide DRs with further physical embodiments of being seen, for example, gift-giving in response to their birthdays or returning from leave [37, 50]. Outside of the academic institution, DRs described how being seen by close others could support DRs to be their authentic selves, providing an antidote to the invisible walls of academia [50]. Good quality friendships within or outside academia could be life-changing, providing a visceral sense of connection, belonging and authenticity that can scaffold positive mental health outcomes during the PhD [39]. Pets could also serve to help DRs feel seen but without needing to extend too much energy into maintaining social relationships [50].

Finally, DRs also needed to see themselves, i.e. to begin to see themselves as burgeoning academics as opposed to ‘just students’ [81]. Feeling that the supervisor and broader academic community were supportive, developing one’s own network of process collaborators and successfully obtaining grant funding seemed tangible markers that helped DRs to see themselves as academics [37, 81]. Seeing their own work published was also helpful in providing a boost in confidence and being a joyful experience [42]. Moreover, with sufficient self-agency, DRs can not only see themselves but render themselves visible to other people [37].

Multiple goals, roles and groups

In antidote to the diminished personal identity and enmeshment with the PhD, DRs benefitted from accruing and sustaining multiple goals, roles, occupations, activities and social group memberships. Although ‘costly’ in terms of increased stress and role conflicts, sustaining multiple roles and activities appeared essential for protecting against mental health problems [50, 68].

Leisure activities appeared to support mental health through promoting physical health, buffering stress, providing an uplift to DRs’ mood and through the provision of another identity other than as an academic [44, 50, 76]. Furthermore, engagement in activities helped DRs to find a sense of freedom, allowing them to carve up leisure and work time and psychologically detach from their PhD [68, 76]. Competing roles, especially family, forced DRs to distance themselves from the PhD physically which reinforced psychological separation [50, 59]. Engaging in self-care and enjoyable activities provided a sense of balance and normalcy [39, 44, 68]. This normalcy was a needed antidote to abnormal pressure [59]. Even in the absence of fiercely competing roles and priorities, DRs still appeared to benefit from treating their PhD as if it is only one aspect of life [59]. Additional roles and activities reduced enmeshment with the PhD to the extent that considering not completing the PhD was less averse [40]. This position appeared to help DRs to be less overwhelmed and less sensitive to perceived and anticipated failures.

Finding hope, meaning and authenticity

Finding hopefulness and meaning within the PhD can scaffold a sense of living a purposeful, enjoyable, important and authentic life. Hopefulness is predicated on the ability to identify a goal, i.e. to visualise and focus on the desired outcome and to experience both self-agency and potential pathways towards the goal. Hopefulness was enhanced by the ability to break down tasks into smaller goals and progress in to ‘baby steps’ [38, 59]. In addition, DRs benefitted from finding explicit milestones against which they can compare their progress [59], as this appeared to feed back into the cycle of hopeful thinking and spur further self-agency and goal pursuit.

The experience of meaning manifested in two main ways; first as the more immediate lived experience of passion in action [76]. Secondly, DRs found meaning in feeling that in their PhD and lives more broadly they were living in accordance with their values, for example, experiencing their own commitment in action through continuing to work on their PhD even when it was difficult to do so [76]. DRs who were able to locate their PhD within a broader sense of purpose appeared to derive wellbeing benefits. There was a need to ensure that values were in alignment, for example, finding homeostasis between emotional, intellectual, social and spiritual parts of the self [46, 59, 90].

The processes of finding hopefulness and meaning appear to be largely relational. Frequent contact with supervisors in person and social and academic contact with other DRs were basic scaffolds for hope and meaning [52]. DRs spoke of how a sense that their supervisors believed in them inspired their self-agency and motivation [42, 62, 76]. Partners, friends and family could also inspire motivation for continuing in PhD tasks [44, 76]. Other people also could help instil a sense of motivation to progress and complete the PhD; a sense of being seen is to be beholden to finish [39]. Meaning appeared to be scaffolded by a sense of contribution, belonging and mattering [81] and could arise from the perception of putting something into the collective pot, inspiring hopefulness and helping others [39, 42]. Moreover, hopefulness, meaning and authenticity also appeared mutually reinforcing [81]. Finding meaning and working on a project which is in accordance with personal values, preferences and interests is also helpful in completing the PhD and provides a feedback loop into hope, motivation and agentic thinking [39, 81]. Furthermore, DRs could use agentic action to source a community of people who share their values, enabling them to engage in collective authenticity [39].

The PhD as a process of transcendence

The immense challenge of the PhD could be a catalyst for growth, change and self-actualisation, involving empowerment through knowledge, self-discovery, and developing increased confidence, maturity, capacity for self-direction and use of one’s own autonomy [44, 82]. The PhD acted as a forge in which DRs were tested and became remoulded into something greater than they had been before [44, 82, 90]. The struggles endured during the PhD caused DRs to reconsider their sense of their own capacities, believing themselves to be more able than they previously would have thought [50]. The struggles endured added to the sense of accomplishment. A trusted and trusting supervisor appears to aid in the PhD being a process of transcendence [62].

More broadly, the PhD also helped DRs to transcend personal tragedy, allowing immersion in a meaningful activity which begins as a means of coping and becomes something completely [39]. The PhD could also serve as a transformative selection process for DRs’ social relationships, with some relationships cast aside and yet others formed anew [39]. Overall, therefore, the very aspects of the PhD which were challenging, and distressing could allow DRs to transcend their former selves and, through the struggle, become something more.

Summation of results

The findings regarding the risk and protective factors associated with DR mental health have been summarised in Table 3 in relation to (1) the type of research evidencing the factor (i.e. whether the evidence is quantitative only, part of the meta-synthesis only, or evident in both results sections); and (2) the volume of evidence (i.e. whether the factor was found in a single study or across multiple studies). The factors in the far-right column (i.e. the factors found across multiple research studies utilising both qualitative and quantitative methods) are the ones with the strongest evidence at present.

Table 3 Risk and protective factors associated with DR mental health in terms of the type and volume of evidence

Discussion

This systematic review summarises a heterogeneous research area, with the aim of understanding the mental health of DRs, including possible risk and protective factors. The qualitative and quantitative findings presented here suggest that poor mental health is a pertinent problem facing DRs; stress appears to be a key issue and significantly in excess of that experienced in the general population. Several risk and protective factors at the individual, interpersonal and systemic levels emerged as being important in determining the mental health of DRs. The factors with the strongest evidence-base (i.e. those supported by multiple studies using qualitative and quantitative findings) denote that being female and isolated increases the risk of the mental health problems, whereas seeing the PhD as a process, feeling socially supported, having a positive supervisor relationship and engaging in self-care is protective.

Results in context

Stress

Stress can be defined as (1) the extent to which a stimulus exerts pressure on an individual, and their propensity to bear the load; (2) the duration of the response to an aversive stimuli, from initial alert to exhaustion; or (3) a dynamic (im)balance between the demands and personal resource to manage those demands [91]. The Perceived Stress Scale (PSS) [18, 19] used in our meta-analysis is aligned with the third of these definitions. As elaborated upon within the Transactional Model of Stress [92], stress is conceptualised as a persons’ appraisal of the internal and external demands put upon them, and whether these exceed their available resources. Thus, our results suggest that, when compared to the general population, PhD students experience a greater maladaptive imbalance between their available resources and the demands placed upon them. Stress in itself is not a diagnosable mental health problem, yet chronic stress is a common precipitant to mental health difficulties such as depression and posttraumatic stress disorder [93, 94]. Therefore, interventions should seek to bolster DRs’ resources and limit demands placed on them to minimise the risks associated with acute stress and limit its chronicity.

Individual factors

Female DRs were identified as being at particular risk of developing mental health difficulties. This may result from additional hurdles when studying in a male-dominated profession [95,96,97], and the expectation that in addition to their doctoral studies, females should retain sole or majority responsibility for the domestic and/or caring duties within their family [52, 76]. It may also be that females are more willing to disclose and seek help for mental health difficulties [98]. Nevertheless, the World Health Organisation (WHO) mental health survey results indicate that whilst anxiety and mood disorders are more prevalent amongst females, externalising disorders are more common in males [99]. As the vast majority of studies in this review focussed on internalising problems (e.g. stress, anxiety and depression) [37, 64, 79, 80, 83, 89], this may explain the gender differences found in this review. Further research is needed to explore which perspective, if any, may explain gender gap in our results.

Perhaps unsurprisingly, self-care was associated with reduced mental health problems. The quantitative findings suggest that all types of self-care are likely to be protective of mental health (i.e. physical, emotional, professional and spiritual self-care). Self-care affords DRs the opportunity to take time away from their studies and nurture their non-PhD identities. However, the results from our meta-synthesis suggest that DRs are not attending to their most basic needs much less engaging in self-care behaviours that correspond to psychological and/or self-fulfilment needs [100]. Consequently, an important area for future enquiry will be identifying the barriers preventing DRs from engaging in self-care.

Interpersonal factors

Across both quantitative and qualitative studies, interpersonal factors emerged as the most salient correlate of DR mental health. That is, isolation was a risk factor, whereas connectedness to others was a protective factor. There was some variability in how these constructs were conceptualised across studies, i.e. (1) isolation: a lack of social support, having fewer social connections, feeling isolated or being physically separate from others; and (2) social connectedness: multiple group membership, academic relationships or non-academic relationships; but there was no indication that effects varied between concepts. The relationship between isolation and negative health consequences is well-established, for example both physical and mental health problems [101], and even increased mortality [102]. Conversely, social support is associated with reduced stress in the workplace [103, 104]. Reducing isolation is therefore a promising interventional target for improving DRs’ mental health.

The findings regarding isolation are even more alarming when considered alongside the findings from several studies that PhD studies are consistently reported to dominate the lives of DRs, resulting in poor ‘work-life balance’ and losing non-PhD aspects of their identities. The negative impact of having fewer identities [105] can be mitigated by having a strong support network [106], and increasing multiple group memberships [107]. But for DRs, it is perhaps the absence of this social support, combined with identity impoverishment, which can explain the higher than average prevalence of stress found in our meta-analysis.

Systemic factors

DRs’ attitudes towards their studies may be a product of top-down systemic issues in academia more broadly. Experiencing mental health problems was reported as being the ‘norm’, but also appeared to be understood as a sign of weakness. The meta-synthesis results suggest that DRs believed their respective universities prioritise academic success over workplace wellbeing and encourage unhealthy working habits. Working in an unsupportive and pressured environment is strongly associated with negative psychological outcomes, including increased depression, anxiety and burnout [108]. The supervisory relationship appeared a particularly important aspect of the workplace environment. The quantitative analysis found a negative correlation between inspirational supervision and mental health problems. Meta-synthetic finding suggested toxic DR-supervisor relationships characterised by powerlessness and neglect, as well as relationships where DRs felt valued and respected—the former of these being associated with poor mental health, and the latter being protective. The association between DR-supervisor relationship characteristics needs to be verified using quantitative methods. Furthermore, DRs’ sense that they needed to exhibit ‘optimum suffering’, which appears to reflect a PhD-specific aspect of a broader academic performativity [109], is an important area for consideration. An accepted narrative around DRs needing to experience a certain level of dis/stress would likely contribute to poor mental health and as an impediment to the uptake and effectiveness of proffered interventions. Although further research is needed, it is apparent that individual interventions alone are not sufficient to improve DR mental health, and that a widespread culture shift is needed in order to prevent the transmission of unhealthy work attitudes and practices.

Limitations of the literature

Although we found a respectable number of articles in this area, the focus and measures used varied to the extent that typical review analysis procedures could not be used. That is, there was much heterogeneity in terms of how mental health was conceptualised and measured, as well as the range of risk and protective factors explored. Similarly, the quality of the studies was hugely variable. Common flaws amongst the literature include small sample sizes, the use of unvalidated tools and the incomplete reporting of results. Furthermore, for qualitative studies specifically, there appeared to be a focus on breadth instead of depth, particularly in relation to studies using mixed methods.

The generalisability of our findings is limited largely due to the lack of research conducted outside of the US, but also because we limited our review to papers written in English only. The nature of doctoral studies varies in important ways between studies. For example, in Europe, PhD studies usually apply for funding to complete their thesis within 3–4 years and must know their topic of interest at the application stage. Whereas in the US, PhD studies usually take between 5 and 6 years, involve taking classes and completing assignments, and the thesis topic evolves over the course of the PhD. These factors, as well as any differences in the academic culture, are likely to affect the prevalence of mental health problems amongst DRs and the associated risk and protective factors. More research is needed outside of the US.

‘Mental health’ in this review was largely conceptualised as a type of general wellbeing rather than a clinically meaningful construct. None of the studies were ostensibly focused on sampling DRs who were currently experiencing or had previously experienced mental health problems per se, meaning the relevance of the risk, vulnerability and protective factors identified in the meta-synthesis may be more limited in this group. Few studies used clinically meaningful measures. Where clinical measures were used, they captured data on common mental health problems only (i.e. anxiety and depression). Due to these limitations, we are unable to make any assertions about the prevalence of clinical-level mental health problems amongst DRs.

Limitations of this review

As a result of the heterogeneity in this research area, some of the results presented within this review are based on single studies (e.g. correlation data; see Fig. 5) rather than the amalgamation of several studies (e.g. meta-analysis/synthesis). To aid clarity when interpreting the results of this review, we have (Table 3) summarised the volume of evidence supporting risk and protective factors. Moreover, due to the small number of studies eligible for inclusion in this review, we were unable to test whether any of our findings are related to the study characteristics (e.g. year of publication, country of origin, methodology).

We were able to conduct three meta-analyses, one of which aimed to calculate the between-group effect size on the PSS [18, 19] between DRs and normative population data. Comparing these data allowed us to draw some initial conclusions about the prevalence of stress amongst DRs, yet we were unable to control for other group differences which might moderate stress levels. For example, the population data was from people in the United States (US) in 1 year, whereas the DR data was multi-national at a variety of time points; and self-reported stress levels may vary with nationality [110] or by generation [111, 112]. Moreover, two of the three meta-analyses showed significant heterogeneity. This heterogeneity could be explained by differences in the sample characteristics (e.g. demographics, country of origin), doctoral programme characteristics (e.g. area of study, funding status, duration of course) or research characteristics (e.g. study design, questionnaires used). However, due to the small number of studies included in these meta-analyses, we were unable to test any of these hypotheses and are therefore unable to determine the cause of this heterogeneity. As more research is conducted on the mental health of DRs, we will be able to conduct larger and more robust meta-analyses that have sufficient power and variance to statistically explore the causes of this heterogeneity. At present, our findings should be interpreted in light of this limitation.

Practice recommendations

Although further research is clearly needed, we assert that this review has identified sufficient evidence in support of several risk and protective factors to the extent that they could inform prevention and intervention strategies. Several studies have evidenced that isolation is toxic for DRs, and that social support can protect against poor mental health. Initiatives that provide DRs with the opportunity to network and socialise both in and outside of their studies are likely to be beneficial. Moreover, there is support for psychoeducation programmes that introduce DRs to a variety of self-care strategies, allow them to find the strategies that work for them and encourage DRs to make time to regularly enact their chosen strategies. Finally, the supervisory relationship was identified as an important correlate of DR mental health. Positive supervision was characterised as inspirational and inclusive, whereas negative supervision productised DRs or neglected them altogether. Supervisor training programmes should be reviewed in light of these findings to inform how institutions shape supervisory practices. Moreover, the initial findings reported here evidence a culture of normalising and even celebrating suffering in academia. It is imperative therefore that efforts to improve and protect the mental health of DRs are endorsed by the whole institution.

Research recommendations

First, we encourage further large-scale mental health prevalence studies that include a non-PhD comparison group and use validated clinical tools. None of the existing studies focused on the presence of serious mental health problems—this should be a priority for future studies in this area. Mixed-methods explorations of the experiences of DRs who have mental health problems, including serious problems, and in accessing mental health support services would be a welcome addition to the literature. More qualitative studies involving in-depth data collection, for example interview and focus group techniques, would be useful in further supplementing findings from qualitative surveys. Our review highlights a need for better communication and collaboration amongst researchers in this field with the goal of creating a level of consistency across studies to strengthen any future reviews on this subject.

Conclusions

The results from this systematic review, meta-analysis and meta-synthesis suggest that DRs reported greater levels of stress than the general population. Research regarding the risk and protective factors associated with the mental health of DRs is heterogenous and disparate. Based on available evidence, robust risk factors appear to include being isolated and being female, and robust protective factors include social support, viewing the PhD as a process, a positive DR-supervisor relationship and engaging in self-care. Further high-quality, controlled research is needed before any firm statements can be made regarding the prevalence of clinically relevant mental health problems in this population.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Abbreviations

CI:

Confidence intervals

DR:

Doctoral researchers

HESA:

Higher Education Statistics Agency

M:

Mean

PSS:

Perceived Stress Scale

SD:

Standard deviation

UK:

United Kingdom

US:

United States

References

  1. 1.

    Baron E. Eleven sketches inspired by the university mental health crisis—in pictures. The Guardian. 2017. Available from: https://www.theguardian.com/education/gallery/2017/jun/27/eleven-sketches-university-mental-health-crisis. Cited 2017 Oct 6.

    Google Scholar 

  2. 2.

    National Union of Students. 20 per cent of students consider themselves to have a mental health problem: National Union of Students; 2013. Available from: https://www.nus.org.uk/en/news/20-per-cent-of-students-consider-themselves-to-have-a-mental-health-problem/. Cited 2017 Oct 6.

  3. 3.

    YouGov. One in four students suffer from mental health problems. 2016. Available from: https://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/obtomdatp4/Survey_Results.pdf. Cited 2017 Oct 6.

    Google Scholar 

  4. 4.

    Universities UK. Minding Our Future: starting a conversation about the support of student mental health. London: Universities UK; 2017. Available from: https://www.universitiesuk.ac.uk/minding-our-future.

    Google Scholar 

  5. 5.

    The Guardian. Number of university dropouts due to mental health problems trebles. The Guardian. 2017. Available from: https://www.theguardian.com/society/2017/may/23/number-university-dropouts-due-to-mental-health-problems-trebles. Cited 2017 Oct 6.

    Google Scholar 

  6. 6.

    Thorley C. Not By Degrees: Improving student mental health in the UK’s universities. London; 2017. Available from: www.ippr.org. Cited 2017 Oct 6.

  7. 7.

    Oswalt SB, Lederer AM, Chestnut-Steich K, Day C, Halbritter A, Ortiz D. Trends in college students’ mental health diagnoses and utilization of services, 2009–2015. J Am Coll Health. 2018;68:41–51.

  8. 8.

    Auerbach RP, Mortier P, Bruffaerts R, Alonso J, Benjet C, Cuijpers P, et al. WHO world mental health surveys international college student project: prevalence and distribution of mental disorders. J Abnorm Psychol. 2018;127(7):623–38.

    PubMed  PubMed Central  Google Scholar 

  9. 9.

    OECD. Education at a Glance 2019. Education at a Glance: OECD Indicators: OECD; 2019. (Education at a Glance). Available from: http://gpseducation.oecd.org/Content/EAGCountryNotes/BRA.pdf. Cited 2020 Mar 26.

  10. 10.

    OECD. OECD Science, Technology and Innovation Outlook 2016. OECD Science, technology and innovation outlook 2016. 2016. Available from: http://www.oecd.org/sti/STIO 10 key technology trends for the future.pdf. Cited 2020 Mar 26.

    Google Scholar 

  11. 11.

    Higher Education Statistics Agency (HESA). Higher education student statistics: UK, 2016/17: HESA; 2018. Available from: https://www.hesa.ac.uk/news/11-01-2018/sfr247-higher-education-student-statistics/numbers.

  12. 12.

    NCES. Number of doctoral degrees earned in the United States from 1949/50 to 2028/29, by gender. 2019. Available from: https://www.statista.com/statistics/185167/number-of-doctoral-degrees-by-gender-since-1950/.

    Google Scholar 

  13. 13.

    Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37–46 Available from: http://psycnet.apa.org/index.cfm?fa=search.displayRecord&uid=1960-06759-001.

    Google Scholar 

  14. 14.

    Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159 Available from: http://www.jstor.org/stable/2529310?origin=crossref. Cited 2017 Mar 8.

    CAS  PubMed  Google Scholar 

  15. 15.

    Noblit GW, Hare RD. Meta-ethnography: Synthesizing qualitative studies: Sage Publications; 1988. p. 88.

  16. 16.

    Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews. BMC Med Res Methodol. 2008;8:45.

    PubMed  PubMed Central  Google Scholar 

  17. 17.

    Wilson DB. Meta-analysis stuff. 2011. Available from: http://mason.gmu.edu/~dwilsonb/ma.html. Cited 2017 Dec 3.

    Google Scholar 

  18. 18.

    Cohen S, Williamson GM. Perceived stress in a probability sample of the U.S. In: Spacapam S, Oskamp S, editors. The social psychology of health: Claremont Symposium on Applied Social Psychology. Newbury Park, CA: Sage; 1988.

    Google Scholar 

  19. 19.

    Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav. 1983;24(4):385–96.

    CAS  PubMed  Google Scholar 

  20. 20.

    Cohen S, Janicki-Deverts D. Who’s Stressed? Distributions of psychological stress in the United States in probability samples from 1983, 2006, and 2009. J Appl Soc Psychol. 2012;42(6):1320–34.

    Google Scholar 

  21. 21.

    MedCalc Software bvba. MedCalc Statistical Software. Belgium: Ostend; 2016.

    Google Scholar 

  22. 22.

    Viechtbauer W. Bias and efficiency of Meta-analytic variance estimators in the random-effects model. J Educ Behav Stat. 2005;30(3):261–93 Available from: http://jeb.sagepub.com/cgi/doi/10.3102/10769986030003261. Cited 2017 Mar 8.

    Google Scholar 

  23. 23.

    Hedges LV, Ollkin L. Statistical methods for meta-analysis. New York: Academic Press; 1985. Available from:. https://doi.org/10.1002/9780470743386.refs.

    Google Scholar 

  24. 24.

    Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to meta-analysis. Chichester: Wiley; 2009. Cited 2019 Jan 10. Available from. https://doi.org/10.1002/9780470743386.

    Google Scholar 

  25. 25.

    Fisher RA. Frequency distribution of the values of the correlation coefficient in samples from an indefinitely large population. Biometrika. 1915;10(4):507–21 Available from: https://www.statista.com/statistics/185167/number-of-doctoral-degrees-by-gender-since-1950/.

    Google Scholar 

  26. 26.

    Schünemann H, Brożek J, Guyatt G, Oxman A. Handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach: GRADE Working Group; 2013. Available from: gdt.guidelinedevelopment.org/app/handbook/handbook.html.

  27. 27.

    Lachal J, Revah-Levy A, Orri M, Moro MR. Metasynthesis: an original method to synthesize qualitative literature in psychiatry. Front Psychiatry. 2017;8:269.

    PubMed  PubMed Central  Google Scholar 

  28. 28.

    Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3(2):77–101.

    Google Scholar 

  29. 29.

    Braun V, Clarke V. Successful qualitative research: a practical guide for beginners. London: Sage Publications Ltd; 2013.

    Google Scholar 

  30. 30.

    Bhaskar R. The possibility of naturalism: a philosophical critique of the contemporary human sciences. London: Routledge; 2014.

    Google Scholar 

  31. 31.

    Fletcher AJ. Applying critical realism in qualitative research: methodology meets method. Int J Soc Res Methodol. 2017;20(2):181–94.

    Google Scholar 

  32. 32.

    Walsh D, Downe S. Meta-synthesis method for qualitative research: a literature review. J Adv Nurs. 2005;50(2):204–11.

    PubMed  Google Scholar 

  33. 33.

    Maton K. Reflexivity, relationism, &amp; research: Pierre Bourdieu and the epistemic conditions of social scientific knowledge. Sp Cult. 2003;6(1):52–65.

    Google Scholar 

  34. 34.

    Tong A, Flemming K, McInnes E, Oliver S, Craig J. Enhancing transparency in reporting the synthesis of qualitative research: ENTREQ. BMC Med Res Methodol. 2012;12(1):181.

    PubMed  PubMed Central  Google Scholar 

  35. 35.

    von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: Guidelines for Reporting Observational Studies. Ann Intern Med. 2007;147(8):573. Cited 2019 Sep 27. Available from. https://doi.org/10.7326/0003-4819-147-8-200710160-00010.

    Article  Google Scholar 

  36. 36.

    Acker S, Haque E. The struggle to make sense of doctoral study. High Educ Res Dev. 2015;34(2):229–41. Available from. https://doi.org/10.1080/07294360.2014.956699.

    Article  Google Scholar 

  37. 37.

    Appel ML, Dahlgren LG. Swedish doctoral students’ experiences on their journey towards a PhD: obstacles and opportunities inside and outside the academic building. Scand J Educ Res. 2003;47(1):89–110. Available from. https://doi.org/10.1080/00313830308608.

    Article  Google Scholar 

  38. 38.

    Devonport TJ, Lane AM. In it together: Dyadic coping among doctoral students and partners. J Hosp Leis Sport Tour Educ. 2014;15:124–34. Available from. https://doi.org/10.1016/j.jhlste.2014.08.002.

    Article  Google Scholar 

  39. 39.

    Enzor J. Friendship, mental health, and doctoral education: a generic qualitative thematic analysis: Capella University; 2017.

  40. 40.

    Kurtz-Costes B, Helmke LA, Ülkü-Steiner B. Gender and doctoral studies: the perceptions of Ph.D. students in an American university. Gend Educ. 2006;18(2):137–55.

    Google Scholar 

  41. 41.

    Bazrafkan L, Shokrpour N, Yousefi A, Yamani N. Management of stress and anxiety among PhD students during thesis writing: a qualitative study. The Health Care Manager. 2016;35:231–40.

  42. 42.

    Cotterall S. More than just a brain: emotions and the doctoral experience. High Educ Res Dev. 2013;32(2):174–87.

    Google Scholar 

  43. 43.

    Kaufman JA. Personal perceptions of stress and self-perceived need for social support among doctoral psychology students in a distance education university sample: Capella University; 2004.

  44. 44.

    Kenty JR. Stress management strategies for women doctoral students. Nurse Educ. 2000;25(5):251–4 Available from: http://pesquisa.bvsalud.org/portal/resource/pt/mdl-16646205.

    CAS  PubMed  Google Scholar 

  45. 45.

    Scrubb MM. An examination of the doctoral student stress survey (DSSS) as an instrument for measuring the effects of stress as perceived by doctoral students in a distance learning university: Walden University; 1997.

  46. 46.

    Usman Yousaf S, Akram M, Usman B. Exploring the causes of stress and coping with it amongst doctoral level students: highlighting the importance of information collection and management. Pakistan J Inf Manag Libr. 2016;18(2):12–25 Available from: https://search.proquest.com/openview/68246cbba8a9da0ec067d56631b479f1/1?pq-origsite=gscholar&cbl=54989.

    Google Scholar 

  47. 47.

    Sandelowski M, Voils CI, Barroso J. Defining and designing mixed research synthesis Studies. Res Sch. 2006;13(1):29 Available from: http://www.ncbi.nlm.nih.gov/pubmed/20098638%0Ahttp://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2809982.

    PubMed  PubMed Central  Google Scholar 

  48. 48.

    Bauer J. Personality factors, self-care, and perceived stress levels on counselor education and counseling psychology doctoral students. Dissertations: Western Michigan University; 2016.

  49. 49.

    Begun AL, Carter JR. Career Implications of Doctoral Social Work Student Debt Load. J Soc Work Educ. 2017;53(2):161–73. Available from. https://doi.org/10.1080/10437797.2016.1243500.

    Article  Google Scholar 

  50. 50.

    Benjamin S, Williams J, Maher MA. Focusing the lens to share the story: using photographs and interviews to explore doctoral students’ sense of well-being. Int J Dr Stud. 2017;12:197–217.

    Google Scholar 

  51. 51.

    Benesek JP. Stress and coping among psychology doctoral students: a comparison of self-reported stress levels and coping styles of PhD and PsyD students: University of Hartford; 1998.

  52. 52.

    Bireda AD. Challenges to the doctoral journey: a case of female doctoral students from Ethiopia. Open Prax. 2015;7(4):287–97.

    Google Scholar 

  53. 53.

    Bolliger DU, Halupa C. Student perceptions of satisfaction and anxiety in an online doctoral program. Distance Educ. 2012;33(1):81–98.

    Google Scholar 

  54. 54.

    Cole LJ. Academic worry and frequent mental distress among online doctoral students: Walden University; 2008.

  55. 55.

    Devine K, Hunter KH. PhD student emotional exhaustion: the role of supportive supervision and self-presentation behaviours. Innov Educ Teach Int. 2017;54(4):335–44.

    Google Scholar 

  56. 56.

    Drake KL. Psychology graduate student well-being: the relationship between stress, coping, and health outcomes: University of Cincinnati; 2010.

  57. 57.

    Dumitrescu GA. Self-efficacy, locus of control , perceived stress and student satisfaction as correlates of dissertation completion: Andrews University; 2016.

  58. 58.

    El-Ghoroury NH, Galper DI, Sawaqdeh A, Bufka LF. Stress, coping, and barriers to wellness among psychology graduate students. Train Educ Prof Psychol. 2012;6(2):122–34.

    Google Scholar 

  59. 59.

    Haynes C, Bulosan M, Citty J, Grant-Harris M, Hudson J, Koro-Ljungberg M. My world is not my doctoral program…Or is it?: Female students’ perceptions of well-being. Int J Dr Stud. 2012;7:001–17.

    Google Scholar 

  60. 60.

    Hill LM. Perceived stress, academic support, social support, and professional support factors as predictors of student success in distributed-learning doctoral education: Fielding Graduate University; 2010.

  61. 61.

    Holahan CK. Stress experienced by women doctoral students, need for support, and occupational sex typing: An interactional view. Sex Roles. 1979;5(4):425–36.

    Google Scholar 

  62. 62.

    Hunter KH, Devine K. Doctoral students’ emotional exhaustion and intentions to leave academia. Int J Doctoral Stud. 2016;11.

  63. 63.

    Kaufman JA. Stress and social support among online doctoral psychology students. J College Stud Psychother. 2006;20(3):79–88 Available from: http://search.proquest.com/docview/57185202?accountid=12253%5Cnhttp://man-fe.hosted.exlibrisgroup.com/openurl/44MAN/44MAN_services_page?url_ver=Z39.88-2004&rft_val_fmt=journal&genre=unknown&sid=ProQ:ProQ%3Aeducation&atitle=Stress+and+Social+Support+among+.

    Google Scholar 

  64. 64.

    Levecque K, Anseel F, De Beuckelaer A, Van Der Heyden J, Gisle L. Work organization and mental health problems in PhD students. Res Policy. 2017;46:868–79.

    Google Scholar 

  65. 65.

    Lonka K, Chow A, Keskinen J, Hakkarainen K, Sandström N, Pyhältö K. How to measure PhD students’ conceptions of academic writing - and are they related to wellbeing? J Writ Res. 2014;5(3):245–69.

    Google Scholar 

  66. 66.

    Lowe RL. The relationship between personality, self-care, stress, and perceived wellness in psychology doctoral students: Tennessee University; 2015.

  67. 67.

    Marais GAB, Shankland R, Haag P, Fiault R, Juniper B. A survey and a positive psychology intervention on French PhD student well-being. Int J Dr Stud. 2018;13:109–38.

    Google Scholar 

  68. 68.

    Martinez E, Ordu C, Sala MRD, McFarlane A. Striving to obtain a school-work-life balance: The full-time doctoral student. Int J Dr Stud. 2013;8:39–59.

    Google Scholar 

  69. 69.

    McGregor BA, Antoni MH, Ceballos R, communication BBBS. very low CD19+ B-lymphocyte percentage is associated with high levels of academic stress among healthy graduate students. Stress Heal. 2008;24(5):413–8.

    Google Scholar 

  70. 70.

    Nelson K. Academic progress in doctoral students: Levels of hope, subjective well-being, and stress: Walden University; 2014. Available from: http://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=psyc12&NEWS=N&AN=2015-99151-189.

  71. 71.

    Nottingham K. A predictive analysis of the psychosocial outcomes of doctoral students: Capella University; 2017.

  72. 72.

    Orozco AB. Self-care effects on anxiety in doctoral students: Fielding Graduate University; 2014. p. 35–61.

  73. 73.

    Peters BM. The relationships among physiological and perceived stress, quality of life, self-care and impairment in doctoral students: State University of New York at Buffalo; 2007.

  74. 74.

    Pifer MJ, Baker VL. “It could be just because I’m different”: Otherness and its outcomes in doctoral education. J Divers High Educ. 2014;7(1):14–30.

    Google Scholar 

  75. 75.

    Platt J, Schaefer C. Clinical psychological students’ subjective stress ratings during their doctoral training. Psychol Rep. 1995;76:994 Available from: http://www.ncbi.nlm.nih.gov/pubmed/7568619.

    CAS  PubMed  Google Scholar 

  76. 76.

    Pychyl TA. Personal projects, subjective well-being and the lives of doctoral students. Ottawa: Carleton University; 1995.

    Google Scholar 

  77. 77.

    Pychyl TA, Little BR. Dimensional specificity in the prediction of subjective well-being: Personal projects in pursuit of the PhD. Soc Indic Res. 1998;45(1–3):423–73 Available from: file://d/d/Texte/Guenter/Literatu/Original/Journals/SIR98-45-423-473.pdf.

    Google Scholar 

  78. 78.

    Rocha-Singh IA. Perceived stress among graduate students: development and validation of the Graduate Stress Inventory. Educ Psychol Meas. 1994;54(3):714–27.

    Google Scholar 

  79. 79.

    Scheidler JA. Effects of perceived stress and perceived social support on marital satisfaction in doctoral students: Walden University. 2008;20.

  80. 80.

    Sekas G, Wile MZ. Stress-related illnesses and sources of stress: comparing M.D., Ph.D., M.D. and Ph.D. students. J Med Educ. 1980;55:440–6 Available from: http://www.embase.com/search/results?subaction=viewrecord&from=export&id=L10000589%5Cnhttp://sfx.library.uu.nl/utrecht?sid=EMBASE&issn=00222577&id=doi:&atitle=Stress-related+illnesses+and+sources+of+stress%3A+comparing+M.D.-Ph.D.%2C+M.D.+and+Ph.D.+student.

    CAS  PubMed  Google Scholar 

  81. 81.

    Stubb J, Pyhältö K, Lonka K. Balancing between inspiration and exhaustion: PhD students’ experienced socio-psychological well-being. Stud Contin Educ. 2011;33(1):33–50.

    Google Scholar 

  82. 82.

    Stubb J, Pyhältö K, Lonka K. The experienced meaning of working with a PhD thesis. Scand J Educ Res. 2012;56(4):439–56.

    Google Scholar 

  83. 83.

    Ülkü-Steiner B, Kurtz-Costes B, Kinlaw CR. Doctoral student experiences in gender-balanced and male-dominated graduate programs. J Educ Psychol. 2000;92(2):296–307.

    Google Scholar 

  84. 84.

    Volkert D, Candela L, Bernacki M. Student motivation, stressors, and intent to leave nursing doctoral study: a national study using path analysis. Nurse Educ Today. 2018;61:210–5. Available from. https://doi.org/10.1016/j.nedt.2017.11.033.

    Article  PubMed  Google Scholar 

  85. 85.

    Waaijer CJF, Heyer A, Kuli S. Effects of appointment types on the availability of research infrastructure, work pressure, stress, and career attitudes of PhD candidates of a Dutch university. Res Eval. 2016;25(4):349–57.

    Google Scholar 

  86. 86.

    Wang C-H, Chen Y-W, Wu T-Y. Self-guided bibliotherapy: a case study of a Taiwanese doctoral student. Int J Humanties. 2010;8(1):413–22.

    Google Scholar 

  87. 87.

    Williams MD. HBCU vs. PWI: institutional integration at PWIs and Black doctoral student depression, anxiety, and stress: University of Minnesota; 2014.

  88. 88.

    Wright T. Issues in brief counselling with postgraduate research students. Couns Psychol Q. 2006;19(4):357–72.

    Google Scholar 

  89. 89.

    Scrubb MM. An examination of the Doctoral Student Stress Survey (DSSS) as an instrument for measuring the effects of stress as perceived by doctoral students in a distance learning university, vol. 58: Walden University; 1998. p. 3041. Dissertation Abstracts International Section A: Humanities and Social Sciences. Available from: http://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=psyc3&NEWS=N&AN=1998-95003-070.

  90. 90.

    Hadden BW, Smith CV. I gotta say, today was a good (and meaningful) day: daily meaning in life as a potential basic psychological need. J Happiness Stud. 2017:1–18.

  91. 91.

    Butler G. Definitions of stress. Occas Pap R Coll Gen Pract. 1993;(61):1–5 Available from: http://www.ncbi.nlm.nih.gov/pubmed/8199583%0Ahttp://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC2560943.

  92. 92.

    Lazarus RS, Folkman S. Stress, Appraisal, and Coping. New York: Springer Publishing Company; 1984.

    Google Scholar 

  93. 93.

    Siegrist J. Chronic psychosocial stress at work and risk of depression: evidence from prospective studies. Eur Arch Psychiatry Clin Neurosci. 2008;258(SUPPL. 5):115–9.

    PubMed  Google Scholar 

  94. 94.

    Marin MF, Lord C, Andrews J, Juster RP, Sindi S, Arsenault-Lapierre G, et al. Chronic stress, cognitive functioning and mental health. Neurobiol Learn Mem. 2011;96(4):583–95. Available from. https://doi.org/10.1016/j.nlm.2011.02.016.

    Article  PubMed  Google Scholar 

  95. 95.

    Ward M. The gender salary gap in British academia. Appl Econ. 2001;33(13):1669–81. Cited 2019 May 23. Available from. https://doi.org/10.1080/00036840010014445.

    Article  Google Scholar 

  96. 96.

    Howe-Walsh L, Turnbull S. Barriers to women leaders in academia: tales from science and technology. Stud High Educ. 2016;41(3):415–28. Cited 2019 May 23. Available from. https://doi.org/10.1080/03075079.2014.929102.

    Article  Google Scholar 

  97. 97.

    Gardiner M, Tiggemann M. Gender differences in leadership style, job stress and mental health in male- and female-dominated industries. J Occup Organ Psychol. 1999;72(3):301–15.

    Google Scholar 

  98. 98.

    Mackenzie CS, Gekoski WL, Knox VJ. Age, gender, and the underutilization of mental health services: the influence of help-seeking attitudes. Aging Ment Health. 2006;10(6):574–82 Cited 2017 May 12. Available from: http://www.tandfonline.com/action/journalInformation?journalCode=camh20.

    CAS  PubMed  PubMed Central  Google Scholar 

  99. 99.

    Seedat S, Scott KM, Sampson NA, Williams D, Kessler RC. Cross-national associations between gender and mental disorders in the World Health Organization World Mental Health Surveys. Arch Gen Pschiatry. 2013;66(7):785–95.

    Google Scholar 

  100. 100.

    Maslow AH. A theory of human motivation. Psychol Rev. 1943;50(4):370–96 Available from: http://content.apa.org/journals/rev/50/4/370. Cited 2019 Jun 10.

    Google Scholar 

  101. 101.

    Leigh-Hunt N, Bagguley D, Bash K, Turner V, Turnbull S, Valtorta N, et al. An overview of systematic reviews on the public health consequences of social isolation and loneliness. Public Health. 2017;152:157–71 Available from: https://doi.org/10.1016/j.puhe.2017.07.035.

    CAS  PubMed  Google Scholar 

  102. 102.

    Holt-Lunstad J, Smith TB, Baker M, Harris T, Stephenson D. Loneliness and social isolation as risk factors for mortality. Perspect Psychol Sci. 2015;10(2):227–37. Available from. https://doi.org/10.1177/1745691614568352.

    Article  PubMed  Google Scholar 

  103. 103.

    Viswesvaran C, Sanchez JI, Fisher J. The role of social support in the process of work stress: a meta-analysis. J Vocat Behav. 1999;54(2):314–34.

    Google Scholar 

  104. 104.

    Michie S, Williams S. Reducing work related psychological ill health and sickness absence: a systematic literature review. Occup Environ Med. 2003;60(1):3–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  105. 105.

    Brook AT, Garcia J, Fleming M. The effects of multiple identities on psychological well-being. Personal Soc Psychol Bull. 2008;34(12):1588–600.

    Google Scholar 

  106. 106.

    Thoits PA. Self, Identity, Stress, and Mental Health. In: Aneshensel CS, Phelan JC, Bierman A, editors. Handbook of the Sociology of Mental Health. Dordrecht: Springer; 2012. p. 357–77.

    Google Scholar 

  107. 107.

    Haslam C, Jetten J, Cruwys T, Dingle G, Haslam A. The new psychology of health: Unlocking the social cure. New York: Routledge; 2018.

    Google Scholar 

  108. 108.

    Faragher EB, Cass M, Cooper CL. The relationship between job satisfaction and health: a meta-analysis. Occup Environ Med. 2005;62(2):105–12.

    CAS  PubMed  PubMed Central  Google Scholar 

  109. 109.

    Macfarlane B. Student performativity in higher education: converting learning as a private space into a public performance. High Educ Res Dev. 2015;34(2):338–50.

    Google Scholar 

  110. 110.

    Daniels K. Perceived risk from occupational stress: a survey of 15 European countries. Occup Environ Med. 2004;61(5):467–70.

    CAS  PubMed  PubMed Central  Google Scholar 

  111. 111.

    American Psychiatric Association (APA). Stress in America: missing the health care connection. Washington: American Psychiatric Association (APA). 2013.

  112. 112.

    Twenge JM. Time period and birth cohort differences in depressive symptoms in the U.S., 1982–2013. Soc Indic Res. 2015;121(2):437–54.

    Google Scholar 

Download references

Acknowledgements

Thank you to the Office for Students for their funding to support this work; and thank you to the University of Sussex Doctoral school and our steering group for championing and guiding the ‘Understanding the mental health of Doctoral Researchers (U-DOC)’ project.

Funding

The present project was supported by the Office for Students Catalyst Award. The funder had no involvement in the design of the study, the collection, analysis or interpretation of the data, nor the writing of this manuscript.

Author information

Affiliations

Authors

Contributions

CH contributed to the conceptualisation, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, validation, visualisation, writing—original draft preparation and writing—review and editing of this paper. LC contributed to the data curation, investigation, project administration, validation and writing—review and editing of this paper. SV contributed to the data curation, formal analysis, investigation, project administration, validation and writing—review and editing of this paper. PR contributed to the funding acquisition, project administration, supervision and writing—review and editing of this paper. JN contributed to the conceptualisation, funding acquisition, methodology, project administration, supervision, validation, writing—original draft preparation and writing—review and editing of this paper. CB contributed to the conceptualisation, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, validation, visualisation, writing—original draft preparation and writing—review and editing of this paper. The author(s) read and approved the final manuscript.

Corresponding author

Correspondence to Clio Berry.

Ethics declarations

Ethics approval and consent to participate

Ethical approval was not needed for the present study.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hazell, C.M., Chapman, L., Valeix, S.F. et al. Understanding the mental health of doctoral researchers: a mixed methods systematic review with meta-analysis and meta-synthesis. Syst Rev 9, 197 (2020). https://doi.org/10.1186/s13643-020-01443-1

Download citation

Keywords

  • Mental health
  • PhD students
  • doctoral researchers
  • Postgraduate researchers
  • Systematic review