Introduction

University students globally have been adversely affected by the COVID-19 pandemic [1]. Mental health indicators, in particular, have demonstrated the emotional toll on students. For example, Ma and colleagues reported a prevalence rate for acute stress, depression, and anxiety symptoms of 34.9%, 21.1%, and 11.0%, respectively among a sample of 746,217 college students in China during the COVID-19 outbreak [2]. Similarly, a survey of students at a large American public university [3] showed that 71% reported increased stress and anxiety due to the COVID-19 outbreak. Specific stressors included fear and worry about their own health and of their loved ones (91%), difficulty in concentrating (89%), disruptions to sleeping patterns (86%), decreased social interactions due to physical distancing (86%), and increased concerns about academic performance (82%) [3].

For higher degree by research (HDR) students, the demands of undertaking a research degree can be stressful and challenging even in the best of times [4–8]. Reported stressors include financial pressures, a lack of institutional support, inadequate training initiatives, administrative processes [4], difficulties establishing a researcher identity [5, 6], the pressures of meeting milestones, such as confirmation of candidature [7], as well as struggling with feelings of social isolation, lack of motivation, and challenges with supervisors and the academic environment [8]. The challenges of research study often affect the mental health of HDR students [4], with more than 40% reporting “depression … or high levels of stress” [9] during their studies.

A comprehensive review by Sverdlik et al. [10] of 163 empirical studies that had investigated factors influencing the mental health of doctoral students highlighted the complexity of the doctoral experience. The authors concluded that four factors external to the student (i.e., supervision, personal and social lives, departmental support and socialization, financial opportunities) and three internal factors (i.e., motivation, writing competencies, academic identity) interact to shape the overall experience of doctoral study and the attendant risk to mental health.

The COVID-19 pandemic has resulted in additional impacts for HDR student wellbeing. Although some research has recorded positive effects for doctoral researchers, including increased work-life balance and better productivity due to work from home arrangements [11], the majority of findings have shown negative impacts for research students as a result of COVID-19. This has included increased anxiety and stress [12, 13], poorer general health and academic functioning [14], and disruptions to overall wellness [15], mental health [16], and mood [14].

Collectively, these impacts represent a significant deleterious effect on the mental health and associated wellbeing of HDR students. For example, Anwer et al. [12], in a study of 108 full-time postgraduate students in Hong Kong, found most students reported high levels of poor sleep (83%), mild to severe levels of anxiety (76%), and moderate levels of stress (89%) during the pandemic. Similarly, Paucsik et al. [13] reported significant increases in stress, anxiety, and depression, and decreases in engagement and wellbeing of doctoral students in France, during a one-year period of the pandemic. Pyhalto et al. [16], in a mixed methods study of 768 Finnish PhD students, found that COVID19 had negatively impacted student progress and mental health, with the highest risk for full-time students, those studying in the natural sciences, and those in the middle stages of their degree.

Assessment of mood is commonly used to screen individuals for risk of mental health issues. The mood measure used in the present study, the Brunel Mood Scale [17, 18] includes subscales to assess Tension, Depression, Anger Vigor, Fatigue, and Confusion. According to the Mental Health Model [19] a mood profile characterized by a high score for Vigor and low scores for Tension, Depression, Anger, Fatigue, and Confusion is indicative of positive mental health, whereas a profile characterized by a low score for Vigor and high scores for Tension, Depression, Anger, Fatigue, and Confusion indicates risk of mental ill-health. The former profile has been termed the iceberg due to its shape when plotted graphically [19] and the latter profile is known as the inverse iceberg [20].

More recently, four other mood profiles have been identified in the literature [21]. The most negative of these, referred to as the inverse Everest profile is characterized by a low score for Vigor, high scores for Tension and Fatigue, and very high scores for Depression, Anger, and Confusion. The high to very high scores on negative mood dimensions indicate elevated risk of mental ill-health and are associated with a range of psychological disorders [22–24]. The shark fin profile, which is characterized by below average scores for Tension, Depression, Anger, Vigor, and Confusion, combined with very high Fatigue scores, has been linked to athletic injury [25] and poor adherence to safety procedures in high-risk vocations [26]. The surface profile, which is characterized by average scores on all mood dimensions, can be considered to represent a typical mood, and the submerged profile, which is characterized by below average scores on all mood subscales, may be beneficial in activities that place a premium on remaining calm and unemotional [27].

Mood profiling is used globally to assess for risk of mental health issues among a wide variety of populations. Examples include screening for risk of post-traumatic stress disorder among military personnel in South Africa [24]; evaluating population-level mental health and monitoring the psychological wellbeing of cardiac rehabilitation patients in Brazil [28, 29]; managing performance anxiety and preventing injuries among ballet dancers in Japan [30] and assessing youth for elevated suicide risk in the USA [31].

The aims of the present study were (1) using quantitative methods, to compare the moods of HDR students with pre-COVID population norms [n] and the moods of a population sample assessed during the pandemic [n]; (2) to establish whether six distinct mood profile clusters identified in the general population [n] were also evident among HDR students, their relative prevalence, and the associated risk of mental ill-health; and (3) using qualitative methods, to explore the experiences of HDR students during the period of the pandemic.

Methods

The present study used a sequential mixed-methods approach, with an initial online survey followed by a series of focus groups among HDR students to better understand their experiences. The survey therefore informed the development of the discussion points and participant composition for the focus groups. The reporting structure for this mixed-methods study is informed by the recommendations of Tariq and Woodman [32], whereby the two parts of our study are reported separately in our methods and results sections and then integrated in our discussion section.

Results

Part 1: Online survey

Mood scores are shown in Table 1. Standardized mean scores for BRUMS subscales for the whole sample varied from population norms (M = 50, SD = 10), ranging from 48.16 for Vigor to 53.12 for Tension. This pattern of mood scores is referred to as an inverse iceberg profile and is seen as indicating elevated risk of mental health issues [33, 40]. Fig 1 shows the mean scores for HDR students plotted against BRUMS norms. To provide an additional point of comparison, mean BRUMS scores gathered during COVID lockdown [41] are also plotted.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Mood scores of HDR students compared to BRUMS norms and COVID mood.

https://doi.org/10.1371/journal.pone.0279698.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. BRUMS subscale scores for participants (N = 231) grouped by demographic variables.

https://doi.org/10.1371/journal.pone.0279698.t001

Comparison of mood scores grouped by demographic characteristics showed significant differences between groups (Table 1). Females reported higher Fatigue scores than males. Younger students reported higher Fatigue and Tension scores than older students. International students reported higher Tension, Depression, Anger, and Confusion scores than domestic students. Full-time students reported higher Tension and Confusion scores than part-time students. Between-group differences explained between 6.4% (older vs. younger) and 14.5% (domestic vs. international) of the variance in mood scores (Table 1).

Cluster analysis

Results of a seeded k-means cluster analysis showed that all six mood profiles described in the literature (i.e., iceberg, inverse Everest, inverse iceberg, shark fin, submerged, and surface profiles) were clearly distinguishable among our sample of HDR students. Fig 2 shows a graphical representation of the six profile clusters and their prevalence rates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Graphical representation of the six-cluster solution.

https://doi.org/10.1371/journal.pone.0279698.g002

Cluster prevalence.

Significant variations in cluster prevalence were found for several demographic variables (Table 2). For gender, females were overrepresented, and males underrepresented for the inverse iceberg and shark fin profiles, whereas males were overrepresented and females underrepresented for the submerged profile. Older participants were overrepresented and younger participants underrepresented for the iceberg profile. For study mode, full-time students were overrepresented and part-time students underrepresented for the inverse Everest profile and the surface profile, whereas part-time students were overrepresented and full-time students underrepresented for the shark fin profile.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of clusters by demographic variables (N = 231).

https://doi.org/10.1371/journal.pone.0279698.t002

For learning mode, on-campus students were overrepresented and online students underrepresented for the inverse Everest profile, whereas online students were overrepresented and on-campus students underrepresented for the shark fin profile. For citizenship, international students were overrepresented and domestic students underrepresented for the inverse Everest profile and surface profiles, whereas domestic students were overrepresented and international students underrepresented for the shark fin and submerged profiles.

Discussion

As a group, our sample of HDR students reported moods that, in terms of mental health, indicated they were more at risk than the general population under normal circumstances but less at risk than during COVID-19 induced lockdown [41, 42]. Scores for negative aspects of mood tend to be positively skewed, with a large proportion of the population scoring at or close to the minimum with a much smaller number scoring towards the upper end of the scale; and this was indeed the case with our sample. As a result, most participants reported mood scores below the population mean for Tension, Depression, Anger, Fatigue, and Confusion, suggesting a reduced risk of mental health issues, with only a minority of participants in the elevated risk category.

However, it is evident that 52 participants (22.5%) reported an inverse Everest or inverse iceberg profile, meaning that between 1 in 4 and 1 in 5 of the HDR students in our sample were at risk of experiencing mental health issues. This prevalence of the most negative mood profiles is higher than for the general population, where typically 15.7% of people (about 1 in 6) report either an inverse Everest or inverse iceberg profile [33]. On a more positive note, our sample compared favourably with mood profiles gathered during pandemic-induced lockdowns, which indicated that about 1 in 3 people reported increased risk of mental ill-health [41].

Between-group comparisons showed that the characteristics of HDR students most associated with risk of mental ill-health and reduced psychological wellbeing are female, younger, international, full-time, and on-campus. Aristovnik and Keržič [1], in a global survey of 30,383 undergraduate students from 62 countries, similarly found that females studying full-time were generally affected more by the pandemic than other groups in terms of their emotional life and personal circumstances.

Moods have been shown to vary according to gender, with men tending to report higher Vigor and lower Tension, Depression, Anger, Fatigue, and Confusion scores than women [40]. Moreover, the prevalence of mood disorders in Australia is substantially higher in women than in men [43]. Gender comparisons in our sample showed that female HDR students reported significantly higher Fatigue scores (Table 1) and were more than twice as likely to report an inverse iceberg or a shark fin profile than male HDR students (Table 2). Biological explanations for mood differences between men and women typically emphasize hormonal and reproductive functions [44, 45], although there are many socio-psychological reasons why women might report more negative moods than men, especially related to the disadvantage women face in several aspects of life, such as education, family responsibilities, and careers [46, 47].

Age-related differences in mood profiles were also evident, with younger HDR students reporting significantly higher Tension and Fatigue scores than older HDR students (Table 1). Moreover, they were twice as likely to report the most negative profiles (i.e., inverse Everest and inverse iceberg) and half as likely to report the most positive iceberg profile, compared to older HDR students (Table 2). Similar age-related differences in mood have been found among the general population, both prior to [33] and during the pandemic [41], which have been explained in terms of adaptive coping strategies being more common and better developed among older adults compared to their younger counterparts [48, 49].

In terms of citizenship, international HDR students were about four times more likely to report an inverse Everest profile compared to domestic HDR students. Although international students face similar challenges to domestic students in terms of completing an HDR degree, they experience additional stressors as temporary migrants and are at increased risk of experiencing poor mental health, with isolation from families and culture, language barriers, financial stress, and academic pressures among the key drivers [50]. Further, international students have been found to be less likely to seek help for mental ill-health than domestic students [51].

Study mode was also shown to be related to mood profiles, with full-time HDR students reporting significantly higher Tension and Confusion scores (Table 1) and being five times more likely to report an inverse Everest profile compared to part-time HDR students (Table 2). Qualitative data illustrated that both full-time and part-time students feel the stress of balancing the demands of life and study, although logically the stress may be more acute for full-time students who have less non-study time available to devote to other responsibilities. Part-time students may find it easier to maintain a sense of normality outside of HDR study and have access to a more extensive social support network than full-time students operating in an intense study environment, which may provide a buffer against the stressors involved [52].

Similarly, study location was linked to mood responses, with on-campus HDR students reporting significantly higher scores for Depression, Anger, and Confusion (Table 1), and being three times more likely to report an inverse Everest profile, compared to HDR students studying online (Table 2). Although some evidence implicates online study as contributing to increased social isolation and threats to mental health [1, 2, 53–55] in situations where HDR students out of necessity must balance the demands of family responsibilities with the opportunity for higher degree study, the greater flexibility afforded by online study may alleviate rather than exacerbate the stressors involved, as suggested by the present findings. It should be noted that the university in which the present study was conducted has a long history of providing successful online education, and hence the findings related to study location may not generalize more broadly.

The qualitative data helped to unpack the survey results by providing insights into some of the specific stressors experienced by HDR students that might be responsible for the increased risk of mental ill-health indicated by mood profiles. Previous evidence has shown that doctoral students struggle with a wide range of stressors that include social isolation, time pressure, lack of motivation, uncertainty of doctoral processes, financial concerns, doubts about academic competence, work-life balance, and challenges with supervisors [8]. The results of our focus groups endorsed all of these stressors, whilst highlighting the three themes of feeling competent, balancing life demands, and isolation and peer support. It was clear from all three focus groups that individual stressors associated with a questioning of own competence was challenging for students. The typical support mechanisms were frequently not in place to mediate this, particularly for international students. Furthermore, the challenges of balancing other life tensions were clearly articulated by the participants, indicating that the reflective experiences for the sub-group of HDR students in this sample supported the scores and profiles measured by the BRUMS.

The challenge of feeling competent in an academic environment may be explained by the big-fish-little-pond effect (BFLPE) [56, 57]. Emphasising social comparison theory [58], and accentuating the multidimensional nature of self-concept [59], the BFLPE proposes that academic self-concept is shaped by evaluative comparisons with peers who provide a favourable or unfavourable frame of reference by which to gauge level of academic competence. While commonly investigated in school-aged children and adolescents, the BFLPE has been found to extend to higher education settings cross-culturally [60]. Social isolation may hinder the development of a positive academic self-concept due to segregation from relative target models. HDR students already boast high-level academic skills and abilities although may inadvertently rank themselves against practised supervisors to gain verification. Such misdirected efforts would only serve to reinforce an unfavourable evaluation of academic competency, and further strengthen the perception of being a little fish in a big pond. This notion is supported by a meta-analysis by Gerber et al., which found that individuals were 76% more likely to make upward comparisons in lieu of access to equivalent peers [61]. Of note, academic self-confidence is reciprocally related to academic accomplishment [62], and has implications for student well-being, making salient the importance of providing adequate supports [63].

Supporting information