There is a lot of guidance out there about how to provide academic supervision to students writing dissertations (see References). Much of the advice relates to studying for, or supervising, the PhD degree. Clearly a dissertation being prepared as part of a third-year undergraduate course is a more modest affair, but even here, a "contract" exists between the student and the supervisor which needs to be understood. Little that follows is original; it is gleaned from the various sources listed, with a small amount of added personal experience.

Reasonable expectations

A supervisor can expect a student preparing a third-year undergraduate dissertation to:
1. Turn up to appointments, prepared for them.
2. Write regularly, and share the draft material - but not too often!
3. Tell the truth about work done and not done.
4. Keep in touch - practically (holidays, sickness, change of address etc.) and academically.
5. Most importantly, do the [research] tasks that have been mutually agreed and scheduled.

In return, students can expect from their supervisors:
1. Regular supervision: a student can reasonably expect to see his or her dissertation supervisor about four times, for a private one-to-one discussion of the dissertation.
2. Written feedback: a student can expect to have draft material read, and returned with written comments within two weeks.

Recommendations from one source, with annotations

Students expect to be supervised
A truism, but it important that you are clear and explicit about stating that it is you, and nobody else, who is responsible for providing the student with supervision for his/her dissertation. Clearly, if there are colleagues to who the student can be referred for further specialist advice - well and good - but ultimately, if you are the named supervisor, you (and the student, of course) are where the buck stops.

Students expect supervisors to read their work well in advance

You must establish an agreed programme - making clear what you will read (and reading that promptly), but not raising expectations that you will promptly read innumerable different drafts.

Students expect their supervisors to be available when needed
These days, e-mail is the preferred mode of contact for most students and most supervisors. Respond reasonably promptly to messages from students, and make yourself available reasonably promptly if a student wants to see you. If you plan an absence from Cambridge of any length, tell the student whose dissertation you are supervising about your plans.

Students expect their supervisors to be friendly, open and supportive.
Students may feel quite uncertain and vulnerable when presenting their work; beware of confidence-sapping measures.

Students expect their supervisors to be constructively critical.
Presented in the right spirit, criticism is much more welcome than bland uninformed approval. Discussions about dissertations can be an opportunity for real dialogue.

Students expect their supervisors to have a good knowledge of the research area.

This is true of a PhD supervisor, but need not be so much the case for the supervisor of a third-year dissertation. If a student is really keen to follow up some particular enthusiasm - on which there may be no great expert locally available - that's OK - just as long as the student is left in no doubt about the kind of support that the supervisor can provide. Writing a dissertation is as much about presentation and organisation as it is about getting all the facts and their interpretation "right"; if you don't know a lot about the topic, say so, and focus on helping the student to find out how to find out.

Students expect their supervisors to structure the tutorial so that it is relatively easy to exchange ideas.

Again a truism - the student should reach a stage quite early on the process when he/she knows more about the topic than the supervisor. So let the interaction be two-way.

What the student needs to write a good dissertation

• The most authoritative sources. Read around the field, and don't rely on the first source you find. Find out which authors' work is referred to most commonly by other leading authors in the field. Remember not to regard photocopying as a substitute for reading.
• The most up to date references. Searches using computers can help you to find if the prominent people in the field have written things which may not yet have appeared in libraries - or are in fairly new copies of journals which are "at the binders" (a regular bugbear!).
• Accurate references. It is amazing how often references are incorrectly cited by other authors. Write out the reference in full when you find them. Refer to references accurately, and in the correct way.
• Confidence that you have covered all the necessary work in your chosen field, and have not failed to refer to some important contribution that the person assessing your work may know about. If you have an external examiner, it's worth discovering what their area of expertise is, and knowing what they have done in it.

Stages of the process

The initial stage

My starting point for students is to say "What is it in four (or however many) months' time you'd like to give to me, because it's something that interests you?" There's a question there that you don't know the answer to and you want to spend time answering, so that you can give me something written which, if I asked you that question, you could confidently say "Read my dissertation".

The first stage involves the student preparing a rough draft that sets out "This is what I think, and what I want to explore". There can be no reason, these days, for most of the work on a dissertation not to be performed on a word-processor. Reinforce the importance of keeping enough up-to-date backup copies of computer files. A dissertation may be chosen from a menu of titles offered by the course organiser, or the title/topic may be one of the student's own choosing. This preliminary planning is a very important stage. The scope of the dissertation, given the word limit, will need to be narrow in focus. If it is broad, it could turn out simply as a re-hash of an article in a Trends journal, New Scientist or Scientific American. It should normally draw mainly on original scientific papers, and should not be synthesis of secondary sources (books, review articles and the like).

The structure

There needs to be a clear "setting out of the stall" in the dissertation, making clear what is to be discussed, why it is important, and broadly how the dissertation is to be structured. In a research-based scientific subject there needs to be space left for a reasonably full section in which "future research" should be discussed. Students can be encouraged to think themselves into the position of someone preparing a Wellcome Project Grant Application, setting out a feasible programme of work over the coming five years - setting out what needs to be done, how and why, and the insights that might be expected to emerge.

Locating and reading the literature

One can search BIDS, Medline or whatever using keywords, progressively refined - clearly a necessary step. Adding the word "REVIEW" to the search terms can be helpful. It may well be possible to identify a paper written within the last 10 to 15 years which has been absolutely critical to the development of the chosen field - so much so that anyone writing in the area since then is virtually bound to have cited that earlier paper in their reference list. This is where citation searching is particularly useful. Identify that early key paper, and see who has cited it subsequently; this feature is available using BIDS.

The rough draft

After the student has become acquainted with the literature (and should by now be more expert on it than the supervisor), the student should construct rough draft with topic heading, and some summary of what each section will consist of. A meeting should be arranged to discuss the rough draft. This will be an important session. The plan of the dissertation will need to be discussed, and joint decisions taken about what should be attempted, and how long it should take.

The major effort of assimilating and writing

It is between the rough draft and the penultimate draft that the student can be expected to make the greatest strides in terms of becoming an authority on the subject matter, and developing independent ideas about how to interpret what has been achieved in the area, and what further work is needed. The penultimate draft should be a polished piece of work - with a proper summary, bibliography (set out in accordance with proper publishing rules). Such disciplines should not be left to the final draft. Decide on a "house style" - from an appropriate established Journal, and encourage the student to stick to it. Make sure that cited works are given full references. Students may need to be reminded of the adage: "The excellent is the enemy of the good." There will always be improvements that can be made to a dissertation, but remember the Law of Diminishing Returns.

Reading the penultimate draft

Supervisors are not expected to read and comment on more than one draft of any student's dissertation. Students should be expected to produce a penultimate draft of a reasonably high standard. This draft should:

• be free of spelling mistakes
• include appropriate diagrams (diagrams can be reproduced from published work with appropriate acknowledgement of their source)
• be within the number limit, explaining the chosen topic in a way that does not require a very highly specialised reader
• include, if need be, a glossary of terms when specialist language is being used

At this stage, the supervisor should point out examples of sloppy work, but should not feel obliged to correct all errors of spelling, grammar etc. After all, it is the student whose work is being assessed, not the supervisor.

In most subjects, the dissertation must be handed in to the Examiners by the first week of the Easter Term - supervisors should confirm with the student the deadline for handing in their dissertation. The supervisor should, therefore, aim to have read and provided written and verbal feedback on the penultimate draft, by the time the student leaves Cambridge for the Easter vacation - or by one week after the end of Full Lent term, whichever is the earlier. For subjects with a different deadline, feedback on the penultimate draft should be provided early enough to allow the student to produce a final draft.

The final draft

It is helpful to students, if possible, to offer them Departmental facilities for binding their work; expensive "posh" methods of binding are not necessary and should not be encouraged.

Administrative matters

The notification form

Each student in NST Part II Biological and Biomedical Sciences (BBS) is individually responsible for presenting a form to the Course Organiser of the subject in which they wish to do a dissertation. This needs to be signed by the supervisor and their Director of Studies and submitted by the Division of the Michaelmas term.

Payment for supervision

The Senior Tutors' Committee have indicated the importance of Colleges continuing to receive reports on the progress of students in their third and/or final years of study. Such reports are essential for monitoring the progress and wellbeing of those students and also for use in any appeals that may be made to the Applications Committee. There are already arrangements for project supervision in some subjects, such as Engineering and certain of the Arts and Humanities. It is not intended to change these existing arrangements. There are certain science subjects (including NST Part II BBS) in which dissertation and project work is supervised within the department and for which, currently, no report is issued to the Colleges. The Committee have agreed that, in these subjects, Colleges will pay for a nominal one hour of supervision, in a group of one, for a report on the dissertation or project work of students. Such reports should be submitted through CamCORS to the Colleges and payment will be approved in the normal way.

References

1. Supervising the PhD: a guide to success by Sara Delamont, Paul Atkinson and Odette Parry (1997) (The Society for Research into Higher Education and Open University Press, Buckingham).

2. Chapter 11 (How to supervise) from How to get a PhD: a handbook for students and their supervisors by Estelle M. Phillips and D.S. Pugh (2nd ed., 1994), Open University Press, Buckingham.

3. Chapter 3 (Going for gold in assessed coursework) from How to get a good degree: making the most of your time at university by Phil Race (1999), Open University Press.

4. Chapters 8 (Dissertations (I): starting) and Chapter 9 (Dissertations (II): analysing and writing) from Successful study for degrees by Rob Barnes (2nd ed., 1995), Routledge.