This study was funded under the Department of Education, Science and Training’s Research Evaluations Programme.
The Deans and Directors of Graduate Studies Committee enabled access to PhD supervisors.
The electronic survey was conducted via Central Queensland University’s Population Research Laboratory, with the assistance of Associate Professor Kerry Mummery, Dr Grant Schofield and Mr Lindsay Greer.
Dr Janet Grice, Postdoctoral Research Fellow, The University of Queensland, provided statistical analysis and advice.
Background and limitations of the study
Two claims prompted this research. They are:
-
a research training environment associated with poor supervision, inadequate levels of departmental support and limited access to quality infrastructure
-
high attrition rates and slow rates of completion for research students (bullets in original White Paper, DETYA, 1999, p.2).
These claims mirror growing global investigation into a range of influences on rates and times for research higher degree (RHD) completions in, for example:
-
the United Kingdom (Rudd & Hatch, 1968; Phillips 1980; Elton & Pope, 1987; Phillips & Pugh, 1987; Wilkinson, 1989; Wright and Lodwick, 1989; Rudd, 1987, 1990; Burgess 1994; Dunkley & Weeks, 1994; Hockey, 1994, 1995, 1996; Delamont, Parry & Atkinson, 1997, 1998; Pole & Sprokkereef, 1997; Pole, 1998; Wright & Cochrane, 2000; Tinkler & Jackson, 2000; Deem & Brehony, 2000; Haksever & Manisali, 2000);
-
Nordic countries (Kyvik & Tvede, 1999; Linden, 2000)
-
Australia (Connell, 1985; Moses, 1994; Green & Lee, 1995; Taylor, 1995; Aspland, Edwards & O’Leary, 1999; Dinham & Scott, 1999; Knowles, 1999; Grant & Graham, 1999; Kiley & Liljegren, 1999; Spear, 1999; Bartlett & Merger, 2000; Johnson, Lee & Green, 2000; Latona & Brown, 2001, DETYA, 2001).
Much of the cited literature identifies a relationship between RHD supervision and RHD completions. However, research that specifically investigates PhD completions, times to submission and the influences that give rise to them is limited. The research underpinning this report was therefore designed to add some empirical evidence to the national higher education knowledge base. The research focuses on both individual and institutional influences on PhD completions derived from two sets of data:
-
a two-phase national survey of 5450 and 1032 supervisors who supervised PhD candidates over the period 1990–97 in 26 State and private universities across all Australian states and territories
-
in-depth face-to-face interviews with 83 PhD supervisors and 26 present or former PhD candidates across 17 universities.
The data provide a limited picture of PhD supervision across Australia. The total number of current PhD supervisors actively engaged in PhD supervision across Australia at the time the research was conducted was unavailable from any source. This probably still is the case and is likely to be a function of inconsistencies between the Department of Education, Science and Education’s (DEST) databases, universities’ record keeping systems and the statutory reporting obligations of both. There is no requirement that universities report to DEST about some matters noted in this report. The apparent difficulties experienced by some universities in providing whole-of-university contact details for supervisors additionally suggest that some institutions’ data collection systems may not readily lend themselves to aggregation.
Further, it is an essential ethical condition of all research involving human beings that participants must participate in the research on a voluntary basis and must be free to discontinue their participation in the research whenever they wish. The initial intention of the research was to conduct a census of current PhD supervisors across Australia but this ethical condition made it impossible to do so.
Nonetheless, 28 universities participated in the study. Of these, 16 furnished lists of contacts for PhD supervisors that were believed to be complete at the time. The remaining 12 enabled contact with some PhD supervisors in their university and it is likely that the number of supervisors contacted was less than the total number of supervisors currently engaged in PhD supervision within those universities.
The survey included four Go81 and 24 non-Go8 universities and two and 14 of these types of universities respectively furnished full lists of contacts. It covered state and private universities in all states and territories. Numerically and in the time-period of the study, this investigation is larger than any research previously undertaken in the area of PhD supervision in Australia.
In addition, the surveys were conducted with the intention of identifying and interviewing supervisors with apparently strong and weak records of PhD completions. Eighty-three PhD supervisors as well as 26 present and former PhD candidates were interviewed across 17 universities, including four Go8 and 13 non-Go8 universities. Supervisors who had fewer than seven completions were excluded from the potential interview sample.
The survey findings about PhD completions and times to submission are similar to the only other Australian findings from research of a comparable national scale (Martin et al, 2001). That study’s findings were derived from a different data set, suggesting a degree of diachronic reliability between the figures reported in it and the present study. The present study’s findings are also consistent in many respects with domestic and international research literature.
These factors combined represent sufficient grounds for treating the overall data sets as adequate for providing a contribution to understanding supervisory influences that contribute to the timely completion of PhD candidatures.
The report is structured for readability. The summary of findings is followed by the main findings of the research and a concluding section that considers matters arising from the study. Research methods, statistical and interview data overviews are included as separate appendices, with detailed discussion of interview data included in the interview data appendix. A summary of findings is now presented.
Footnote 1. Go8 refers to Australia’s eight ‘research intensive’ universities.
Research cultures
University type and research discipline influence the timely completion of PhD candidatures. However, research discipline has more influence than university type. The PhD candidature appears to be a rite of passage into distinct research cultures that manifests in discipline-specific completions and times to submission.
PhD completions
Sixty-four per cent2 of PhD candidates supervised over the 1990–97 period were conferred with the award of Doctor of Philosophy.
-
more Go8 candidates (69%) received the award than non-Go8 candidates (61%).
-
comparatively more candidates in the Natural Sciences (75%) received the award than in the Social Sciences (52%), the Humanities & Arts (54%) and Other3 disciplines (61%).
Times to submission
Sixty-eight per cent of candidates submitted their dissertations for examination:
- 40 per cent submitted in four4 years or less.
- 57 per cent submitted in five years or less.
A greater percentage of Go8 candidates (73%) submitted than non-Go8 candidates (64%).
- 45 per cent of Go8 candidates submitted in four years or less in comparison with 36 per cent of non-Go8 candidates.
- 64 per cent of Go8 candidates submitted in five years or less in comparison with 54 per cent of non-Go8 candidates.
Comparatively more Natural Science (79%) candidates submitted than in the Social Sciences (55%), the Humanities & Arts (59%) and Other disciplines (64%):
-
48 per cent of candidates in the Natural Sciences submitted in four years or less compared with 30 per cent in the Social Sciences, 28 per cent in the Humanities & Arts and 41 per cent in Other disciplines.
-
69 per cent of candidates in the Natural Sciences submitted in five years or less compared with 44 per cent in the Social Sciences, 46 per cent in the Humanities & Arts and 52 per cent in Other Disciplines.
Footnote 2. All reported percentages are approximations, because it is possible that individual participants who work in the same organisational element reported supervising the same candidate. In this sense the figures may be over-estimates. However, completion data do not include candidates who were still enrolled at the time of the study and who have since completed or will complete at some time in the future. Similarly, the same caveat applies to a lesser extent for submissions. While full-time candidate submissions are accurate to four years, some part-time candidatures in progress will fall within the four or five full-time equivalent years timeframe if they are completed by 2004 or 2006. In this sense the figures for completions and submissions may be an under-estimates.
Footnote 3. Survey respondents self-identified disciplines. Survey and interview data indicate that those who identified as ‘Other’ disciplines tend to conduct trans-disciplinary research.
Footnote 4. All reported submission times are full-time equivalent years.
Completions and times to submission reflect disciplinary research and publications customs, orientations toward and success in earning external research income, and associations between these factors and supervisors’ mean success rates.
-
Natural Scientific research culture is collaborative in its orientation to the publication of research and is highly oriented and successful in the pursuit of external research income for further research or to fund candidatures.
-
Humanities & Arts research culture is individualistic in its orientation to the publication of research and somewhat indifferent and unsuccessful in the pursuit of external research income for further research or to fund candidatures.
-
Social Scientific research culture exhibits a blend of Natural Sciences and Humanities & Arts characteristics and orientations, with mixed results.
These factors are associated with supervisors’ success rates. In mean terms, supervisors who have been supervising for longer times, have candidates who submit within five years, publish and present papers with present or former PhD candidates, win larger numbers of Australian Research Council (ARC) Large and Small grants, have full-time candidates who do not change supervisors or topics or take leave of absence and examine more PhD theses, have better success rates in terms of timely completions. Supervisors with this profile are found predominantly in the Natural Sciences.
At least five relative advantages that Natural Sciences research culture affords PhD candidates in comparison with the Social Sciences and the Humanities & Arts additionally explain aggregate disciplinary differences in PhD completions and submission times. Natural Sciences research culture offers:
- a more attainable credential
- more collaborative research support
- more effective levels of stakeholder investment in candidates’ success
- safer candidate selection criteria
- a more established supervisory pool.
While individual supervisors’ practices tend to reflect their respective research cultures, there are practices that individual supervisors engage in irrespective of university type and across disciplines that can be called ‘good’ because they contribute to the timely completion of candidatures.
Some supervisors take a ‘hands off’ approach to supervision that leaves candidates largely to their own devices. Except in a minority of cases where beginning candidates are already self-confident, independent, knowledgeable, skilled, organised and socially adroit, ‘hands off’ approaches tend to be associated with slow and non-completion.
In contrast, supervisors who are more ‘hands on’ in their approach to supervision tend to be associated with faster and more completions. The main reason for this is that most commencing PhD candidates do not possess all of the ideal qualities that are often expected as pre-requisites to successfully undertaking a PhD. ‘Hands-on’ supervisors accept this situation and their relatively interventionist approach to supervision is more effective than ‘hands off’ approaches.
The association of ‘hands on’ supervisory practice with more and minimum time completions is primarily attributable to an interventionist pedagogic approach to supervision. ‘Hands on’ supervisors actively assist commencing candidates to structure their candidatures. This involves explicitly negotiating with candidates a firm timetable for completing the candidature, especially in relation to:
- available support and project logistics
- institutional quality checks
- project specific milestones such as the production of thesis text
- the presentation and publication of conference and journal papers.
By assisting candidates to structure their candidature, ‘hands on’ supervisors demystify the PhD exercise. In addition, in the process of structuring the candidature ‘hands on’ supervisors establish consistent and viable relationships with candidates. An important basis of these relationships is the achievement of early and lasting agreement between supervisors’ and candidates’ expectations of each other, coupled to action consistent with agreements. Agreement is reached and the relationship maintained by an ‘open door’ consultation policy combined with supervisors regularly initiating contact with candidates.
In particular, ‘hands on’ supervisors get to know their candidates well enough for a personal dimension of trust to exist within an otherwise professional relationship. Trust enables supervisors to detect whether and why candidates are experiencing difficulties and thus to make timely and appropriate interventions themselves, or to refer candidates to more appropriate sources of advice and assistance. Trust also enables candidates to approach their supervisors with confidence. ‘Hands on’ supervisors acknowledge that the supervisory relationship is one of unequal power between supervisor and candidates and use their superior position to mentor candidates’ professional development with a view to the candidate establishing him or herself as a peer.
The first year of candidature
The first year of candidature is crucial. During this period ‘hands on’ supervisors negotiate a mix of formal and informal interactions between themselves, their candidates, other candidates and relevant sources of expert advice. This encourages self-confidence in the candidate and simultaneously monitors progress. Text production is imperative from the outset and is vital throughout the candidature, because it is the basis on which supervisors give advice. In the first year of candidature interactions are of high frequency and entail rapid turnaround of text because candidates require regular and timely feedback to help them to decide whether or not they are making progress and what to do next. ‘Hands on’ supervisors also encourage candidates to work on more than one task at a time, because this prevents candidates from becoming bogged down by an apparent lack of progress in one area of the research.
Teamwork
‘Hands on’ supervisors use variations on a generic modus operandi for supervision, namely, teamwork. Teamwork approaches to supervision:
- foster collaborations between candidates via things such as informal coursework and the organisation of candidates into face-to-face and electronic cohorts
- involve academics and other experts additional to the supervisor in candidates’ research
- integrate candidates into supervisors’ broader associations with research groups and teams as well as industry networks
- enhance the candidate’s professional development via activities such as joint preparation of conference presentations and journal papers.
Duration of the candidature
The frequency of interaction between supervisor and candidate then fluctuates during the candidature, tending to decrease at the candidate’s discretion as the relationship becomes more like a peer relationship. However, ‘hands on’ supervisors attach great importance to times of peak candidate activity, especially writing. ‘Hands on’ supervisors consistently encourage and assist candidates to draft thesis text, and to publish and present their research in journals and at conferences, sometimes by the end of the first year and usually in its second, third and fourth years. ‘Hands on’ supervisors:
- vary the amount and level of input they provide into theses and publications, providing more input earlier in the candidature and less input later
- go through a number of iterations of thesis and publication drafts with candidates
- negotiate authorship protocols with candidates that reflect the respective contributions made by supervisor, candidate and any additional authors.
A number of inter-related considerations about improving PhD supervision and the future of the PhD exercise arise from these findings. At the level of federal policy the main issue is the broad mix of state and private funding and incentives attached to research, the PhD and research training. For universities, the mix of research and non-research degrees on offer, the conventional academic career structure and academic workload are pivotal. As far as research disciplines are concerned, cultural traditions that tacitly govern the conduct of research and PhD supervision within disciplines matter. The capacity of the PhD exercise to meet the demand of the global knowledge economy for rapid knowledge production is an increasingly important concern. These considerations are dealt with in detail in the concluding section.
In order to elaborate these findings, the influence of research cultures on the timely completion of candidatures is now discussed.