- http://greatresearch.org In particular, I recommend:
- Eloquent Science: A Practical Guide to Becoming a Better Writer, Speaker, and Atmospheric Scientist; Should be available as an e-book from your UVic account.
A PhD indicates that a scientist is able to carry out independent research, is an expert in their immediate research topic, and has enough knowledge and background to teach undergraduate students in their field. In the case of my students this would be in physics/oceanography. I’ll elaborate on each of these in inverse order.
Background knowledge and ability
A PhD indicates that a scientist understands their field at an expert level such that they would be able to (with some preparation) teach an undergraduate course in their field. This implies that they have some general knowledge and are familiar with the methods for approaching and solving problems.
This is ability is generally absorbed through course work, and general exposure to the student’s field. It is difficult to test, since of course we all forget explicit details of various problems. If necessary this can be tested by asking the student to prepare a lecture on a short time frame.
Being an “expert” in your immediate field
This usually means reading, and more reading of the literature. In my field, if a PhD student has not read a couple hundred papers, and understood most of them, it is unlikely that they will have acquired the expertise to earn a PhD. A good PhD thesis will have over 100 references, properly referred to.
At the end of the PhD the student should know more than the advisor about their particular subject area. The advisor does not have time to understand all the details of every student’s thesis, and by the end of the thesis the student should be teaching the advisor.
Being an independent researcher
This is the hardest for students, and differentiates technicians from researchers. Students must be able to identify and advocate for important problems, devise “experiments” to test the problems, execute the “experiments”, and then communicate the results. This process happens on different levels in a good thesis. The overall problem should be convincingly advocated for and progress made on understanding it. The overall problem is often broken up into smaller sub problems (call them the chapters of the thesis if you like) and within those subproblems, there are a number of smaller steps that all require this process.
I put “experiments” in quotes because I mean the process of testing something. The whole thesis can be thought of as an experiment, but each individual analysis can as well. i.e. I hypothesize that the variability in the internal wave field has a seasonality due to the winds, so I devise an “experiment” to test that by calculating the variance of u’ in two-day windows. If there is a signal there, I refine the “experiment” by seeing if there is a correlation to the wind strength, or wind variance, etc etc. So, even if I didn’t collect this data, each foray into the data is an “experiment” that has a hypothesis that is related to my overall problem.
In order to earn a PhD, the student needs to be generating a substantial portion of these “problems” and “experiments” themselves. The overall thrust of the thesis should be largely guided by the student, and they should garner the supporting material to argue that solving the problem will be a significant scientific advance. Again, this goes all the way down to individual steps within each chapter. If an analysis is performed on the data, the merit of the analysis should be justified in the context of the larger problem, and its role in advancing the argument made clear. Again, a PhD student should be generating the bulk of these analyses and justifications.
Now, of course, it would be a very poor piece of science if this all happened in isolation, even for expert scientists, let alone students learning the process for the first time. The role of the advisory committee, and other scientists the student shows the material to, should be to give feedback on the merits of the effort, and to point out where the effort has flaws. This allows the student to go back and refine their arguments and improve their analyses. This feedback can take many iterations; a very respected colleague counted over 80 drafts of his thesis before his advisor accepted it. This process should be respected, and engaged in eagerly.
During this process the advisors should be providing the student with ideas, but it should not be the advisor supplying the student with “next steps”. Students have to be independent.
The process of getting this feedback will be greatly helped by good communication skills. Students need to learn how to write and give good presentations. A common pitfall here is that students provide a literature review with little context, and then explain their plots with little context, and then wrap up with a summary of their plots. This is poor communication. Introductions should be essays explaining the point of the paper, and why it is an advance. Presenting the result of “experiments” should be in the context of answering the questions posed. The conclusions and discussion should indicate how they question has been moved forward by the work, major caveats of the work, and what remains unanswered. This very brief outline applies to chapters of the thesis, quite often to subsections within the chapters, and it definitely applies to oral communications.
Being able to articulate the point of your research and why it is important is a strong indicator of being able to carry out independent scientific research. Going through the process of arguing the merits of your research always helps improve the story, and more often than not, leads to ideas for new analyses that again improve the final product. If the effort to add the context is not engaged in, the result will be amorphous and of questionable value.
Advisor Student Relationship
As I indicated above, the student’s advisors should be sounding boards for the student’s ideas, and helpful critics of their efforts. Criticism is necessary to improve the student’s work, and should not be viewed as a personal affront or a final judgement by the advisors on the student’s ability.
That said, final judgements do have to be made. The General Exam and the Final Exam are two points where these judgements are made, though in my opinion students who are not going to pass these should not be encouraged to take them.
PhD Thesis Proposal
Departmental suggestion about the content of the PhD thesis proposal is suprisingly lacking:
- SEOS suggests:
This should: (i) provide the scientific background and rationale for the proposed thesis research; and (ii) propose the methodology for undertaking the research that will constitute the thesis. This may include brief discussion of preliminary results if appropriate.
- Physics suggests… 20-40 pages, and nothing else.
This lack of specificity leaves it up to graduate committees to decide what should be in the thesis proposal. Unfortunately, this does not lead often lead to a specific set of guidelines either. Here is what I suggest and why.
The proposal should be like most proposals and convince the reader that the student has an interesting research topic that will lead to significant progress in their field. Making this case is more involved than most students expect.
- What: You must clearly state the research problem. This might take a couple of paragraphs.
- Why: You discuss, pretty extensively, why the problem is important. This should be both to your field, science in general, and ideally to society as a whole. This section should be 2-3 pages.
- You discuss whats already been done on this problem. This should be written not as a straight literature review, but as an argumentative essay that will point out what approaches have been tried and their pluses and minuses. After all if this subject was fully figured out, you shouldn’t have it as a thesis topic. This sections is another 3-4 pages.
- You argue that you are going to add something to this problem and explain what you are going to do. This section is 4-5 pages.
- At the very end, you have a page on work completed (your comittee has already see this, right?), and second for your timeline.
With figures and references you now have something that is about 15-25 pages.
All these sections should be heavily referenced and convince the reader that you are an expert in the field. But more importantly, each section should have arguments to make.
The most important thing to note is that the proposal is not an information dump. You will convince us of your knowledge in the context of the above arguments, not by just by writing an extedend wikipedia article.
Why should it be this format?
The problem I’ve seen with a number of PhD proposals is that they try to impress me with a level of knowledge rather than a level of thinking. Writing the proposal as above will ensure you have thought about the point of your research. If you know the point, then you will be a much more efficient researcher. You will also find it a lot easier to write your papers and thesis.