Supervision award
Last night I received the Vice-Chancellor’s postgraduate supervision award at a function at Government House. I am deeply honoured that my students thought to nominate me for the award. I think I was as surprised as anyone to win, and some people have asked me what I did to deserve it. Actually, I’m not sure that I did deserve it, but I can tell you what I told the award committee who chose me.
I was asked to write a document explaining my approach to supervision. My students and colleagues also had to contribute comments, which are not reproduced here. This is what I wrote . . .
The ultimate supervisor is one who is involved with all aspects of their students’ research including settling in a new country, framing the research problem, searching the literature, writing papers, giving seminars, meeting other researchers, computer programming, mathematical typesetting, attending conferences, etc. I see postgraduate research students as “apprentice researchers” with the supervisor aiming to help them become fully independent researchers over the course of the degree.
Of course, I am not yet the ultimate supervisor, but I am always learning how to do the job better. As Michelangelo said, Ancora imparo (“I am still learning”). I believe that I do help my students develop into complete researchers. It is my goal to be a guiding hand in all aspects of a student’s research development. Many of my supervision ideas were developed as I perceived a need. The following sections outline various problems that I or my students have encountered, and how I have tried to address the problem.
1 The Forecasting Research Team
Once the number of research students I supervise grew large enough, it became apparent that I was spending a large amount of time saying similar things to each student. Furthermore, they were tending to work in isolation rather than as a cohesive group. As a result, I thought it would be useful for them and for me to have some group meetings rather than only the one-on-one meetings that I had been having. Hence, the “Forecasting Research Team” was born. This has been enormously beneficial in sharing ideas, developing a spirit of camaraderie, and providing a forum for research planning. The students feel much less isolated, and are more likely to help each other as a result of being part of a identifiable research team. My research team consists of all PhD students (currently numbering six) plus my two postdoctoral research fellows. I have tried to develop the group as a strong and cooperative team rather than have them each working individually. Once each fortnight, the whole team meets to discuss issues of mutual interest. Several other HDR students also join us as they have seen the benefits of a research team and want to be part of it. At these meetings, the discussion can be wide-ranging. Some of the things we have talked about in the past year are how to give a good research seminar, explaining your research to people in other fields, new software tools, debugging computer programs, using LATEX for writing papers and a thesis, how to subscribe to RSS feeds for journals, writing a referee report, using some of the new search tools in research, etc.
2 Web page
Students often spend a lot of time on the web trying to track down information about software or research papers, and do not always know where are the best places to look. To help with this process, I have developed a very extensive website robjhyndman.com with a large range of resources for my students. It provides advice on the best search engines to use for different types of searches, useful freeware, how to install some research software at Monash to avoid proxy problems, and so on. On most days, the page receives between 2000 and 3000 hits and so it obviously meets a need for other researchers as well as my own research team. Largely as a result of this website, I receive many requests from potential overseas students wanting to come to Monash.
3 Getting to Monash
My involvement with students often starts long before they arrive at Monash. When a student approaches me as a potential supervisor, if they appear to be suitable for our program, I begin working with them on a research proposal. I also discuss various practical issues with them such as sitting the IELTS test, satisfying the university regulations, and so on. Occasionally, this process has taken more three years before the student finally arrives in Australia.
4 Settling in
Almost all of my students are from other countries, and so the first few weeks often involves finding accommodation, learning where the nearest shops and other facilities are, meeting people of the same ethnic background, sorting out computer accounts and library access, and so on. When a student arrives at Monash, I assign one of my other students or research fellows as their helper for the first few weeks. Where possible, I ask someone from the same or similar background to act in this helper role. Often they will invite the new student to stay in their own house. As well as providing immediate practical assistance, having one of the research team act as a helper means that the new student gets to know at least one other member of the research team very quickly.
5 R and LATEX issues
Two key pieces of software in my area of research are R (for computational statistics) and LATEX (for mathematical typesetting). Every new student needs to spend some time learning these software environments. To make the process easier, I have constructed a webpage for each package providing a resource for my research students with instructions on installing and maintaining the software as well as links to helpful online resources. Both of these pages are now very widely used by people outside Monash as well.
6 Research blog
After the research team had been in existence for a year, we found that we needed to address topics again as new people had joined. To avoid a lot of replication of work, I started a “Research tips” blog containing a summary of some of the key team discussions. This has become part of my web page at robjhyndman.com and provides a permanent record of some of our discussions for future reference. It also makes them available to researchers elsewhere. This has become quite widely used, with several entries having received over 1000 hits each.
7 Seminar preparation
Before a student presents a seminar or a conference paper, I will get them to give their talk to the group and encourage feedback from other students. Sometimes this leads to lively debate about what is appropriate, and how best to express the key ideas. By the time a student gives the conference paper or seminar, the talk has been well-practised and the flaws addressed.
8 Team teaching
Postgraduate research can be a very lonely and isolating existence, with little opportunity for working in a team environment. Consequently, I sometimes ask a student who has developed some specific skills or has an interest in a particular area to teach the group about what he/she has learned in that area. This also gives them confidence and experience in a team environment, which is often lacking in postgraduate research in my discipline. In the past year, members of the research team have taught the group about diverse topics such as debugging tools in the R computer language, using animated graphics in seminar presentations, and statistical tests for comparing forecast models.
9 Forecasting wiki
At some of our research team meetings, students commented that the time series and forecasting material on Wikipedia was particularly poor, often being inaccurate or non-existent. We discussed what could be done and decided to set up our own forecasting wiki site (www.monashforecasting.com) which anyone on the research team can contribute to as they choose. We are currently in the process of setting this up, and we hope it will become a permanent and useful on-line resource for other researchers. One of the benefits of setting up our own wiki is that students will receive some experience at writing about their research for a general audience. It will also not be subject to editing by any readers (unlike Wikipedia) and so we will be able to control the quality of its content.
10 Individual meetings
I allocate 45 minutes to each student for fortnightly individual meetings. In this time, the student updates me with what they have been working on and what problems they have encountered. Sometimes I will suggest alternative approaches or relevant papers that they should read. I try to lead the student to solve the problems themselves rather than do it for them. This means the research takes longer than it would otherwise, but the student becomes a much stronger and more independent researcher as a result.
11 Ad hoc interactions
Some things can’t wait until a scheduled meeting, so I encourage my students to drop into my office with ideas they want to share, or to email me with problems they have encountered. As a result, I have regular interactions with students via ad hoc conversations. Some students prefer to interact via email between meetings and these often develop into long and intense conversations. I generally aim to respond to student questions within a day (often in less than an hour) if they are relatively simple. Questions requiring more detailed responses are usually answered within a few days at most.
12 Research library
The Monash library has quite a good collection of statistical and forecasting books, but I found that students often wanted long-term loans which wasn’t possible via the Monash library. Other students would just need to quickly check a result in a book, but may not bother as it meant walking to the library. So over the last few years I’ve developed an extensive library of books of use to students. These are available for any of my research team to borrow, and most students will have several of my books on their shelves at any one time. When it became difficult to keep track of where the books were, I introduced a borrowing system via my website so that any student could see which other student had a book on loan.
13 Conferences
I consider attendance at international conferences a vital component of research life, and encourage all of my students to attend the best international conferences to present their work and to meet other researchers working in similar areas. Every research student in our department is funded for one overseas conference during their candidature. For my students, the best conference to attend is the International Symposium on Forecasting (ISF) which is held annually. Every year for the last five years, the Monash contingent has been the largest from any university at the ISF. This has been an enormously valuable experience and has allowed the students to meet other researchers working on similar problems. As Editor-in-Chief of the International Journal of Forecasting (which is associated with the ISF), I know most of the key researchers in the field, and I try to introduce my students to the people they will find most helpful in their own research. This often leads to ongoing correspondence and an interchange of information for several years after the ISF. Many of my students have attended two and even three ISFs. After the first one (funded by the department), the students have been very resourceful in obtaining funding from other agencies so they can attend additional ISFs.
14 Writing issues
One of the issues that all of my students seem to face is a difficulty with writing in English. Consequently, we cover English grammar and other writing issues in my research team meetings. As the research team grew, I was spending too much of my time correcting basic English errors. So I now employ a research assistant with expertise in copyediting to help the students with their English. When a student has completed a draft chapter or a draft paper, my copyeditor then goes through it carefully and provides suggestions and feedback to the student. I only look at the writing after the student has revised the piece of writing in the light of the comments from my copyeditor. Not only does this free up a substantial amount of time I used to spend on correcting English, it has also given my students more confidence when showing me their work because they know that it has already been checked by an expert in English. An additional advantage is that it has allowed the feedback process to be sped up. The copyeditor provides a rapid turnaround-less than a week for any chapter or paper, and less than three weeks for a whole thesis. When I see the piece of writing, I can provide faster feedback because I don’t have to spend time on the English issues. I fund the copyeditor out of money earned through consulting and contract research projects, so it doesn’t cost my students anything.
16 Financial difficulties
Often a student will find that their scholarship is insufficient for their basic living, especially if they are also supporting a spouse and children. Where possible, I try to find some kind of employment for my students (within the scholarship restrictions) to help them financially. Sometimes this will be as a research assistant for me; other students have been employed by colleagues in other universities for research assistance and data analysis; others will work as sessional tutors for our undergraduate courses.
17 Conclusions
As the above summary shows, my supervision is often about discovering the problems that students face and trying to generating helpful solutions. As a result, I feel like I am a much better supervisor now than I was a few years ago. And I hope I can continue to improve in this area.
15 Social interaction
There is always an inevitable distance between a professor and a new research student. So I try to engage socially with the students from time to time. For example, I have an annual trip with those students who are interested to see an international cricket match at the MCG. Some of my students have also enjoyed dinner at my house. On occasions, students have also invited me and my family for dinner at their homes. We have enjoyed many Turkish, Bangladeshi, Pakistani and Greek meals.