Yesterday, I examined a Ph.D. It’s not an unusual experience – and huge congratulations to the candidate who had a well-deserved pass! But every time I go through this process I spot a number of weaknesses in the UK examination system that really should be put right. These reflections are not specific to the thesis I just examined. They are, rather, a broader policy reflection on the process.
Examining Ph.D.s is time consuming and badly paid. There is no requirement to examine Ph.D.s in my academic contract. They do serve as “esteem markers” in various promotion documents and you can use them as part of a narrative. Ph.D. examining is a paid activity, on top of an academic salary. However, it is badly paid. I calculated this out for a previous examination and found that, if I worked at minimum wage, that is, if I worked at the absolute lower limit of what I should legally be paid (which I do not), then I would have just 4 minutes to spend on each page of a Ph.D. thesis. That doesn’t include the time spent at the viva, filling out forms for associated admin, travelling etc. If this is supposed to be a serious part of the highest examination system in the country, then should remuneration not be equal to the task?
How to fix it: Ph.D. examination should be paid at a far higher rate than it is, commensurate to the work involved. This payment does not even necessarily have to be directly to the academics involved. It could, for instance, actually be part of a formal workloading system and go to the employing institution (although, to be honest, I am very sceptical of that approach as I believe that institutions would not properly account for such activities and we’d just end up doing it on top of everything else, with no allowance.)
There are perverse incentives to pass candidates. There are several perverse incentives to be lenient. I would say that most academics are not unduly swayed by these and act with integrity in the examination system. However, it is also well known that junior academics are thought to be harsher examiners. Presumably this is because they have not yet learned about the problems of failing or referring a Ph.D. examination.
These problems are many. For one, referring a Ph.D. thesis means that you will end up re-reading the whole thing, re-examining it, with all the perils in the above section. So, referring a thesis means that you end up doing the poorly paid, time consuming work, again.
For another, referring or failing a Ph.D. means that you are likely to fall out with the supervisor. Often, in small sub-fields, this can cause serious inter-personal friction. We are supposed to avoid conflicts of interest, but this can be very difficult when you know everyone who could examine work.
How to fix it: sort the remuneration issue above. Change the pressures on supervisors (see below).
Supervisors feel/are held overly responsible for the performance of their candidates. Obviously, we want to identify poor supervision practices and ensure that candidates are treated well. I have seen examples of supervisors who do not treat their students well – not reading work etc. This is clearly unacceptable. However, there is, actually, only so far that a supervisor can go in advising a candidate whether to submit or not. By the time of submission, the candidate should be the master of their field and know more about it than the supervisors. It is also always possible that the viva will take an unexpected turn; at the highest level of research there are often conceptual differences held by different schools. The outcome, therefore, is not always predictable, even to an advanced and successful supervisor.
However, Ph.D. supervisors are judged on the number of successful completions that they obtain. Just as with education elsewhere, teachers are made to feel responsible for the success of their pupils. As I said, this is a delicate balance. There must, surely, be some link between bad teaching/supervision and outcome – at least in some cases. But even if you are supervising responsibly, it is possible to get an outcome that is beyond your control. It is, also, ultimately the candidate’s responsibility to submit and to be accountable for their result.
If supervisors were not appraised in this way, the above problem of “falling out with the supervisors” would be lessened. At the moment, failing someone else’s Ph.D. candidate has negative repercussions for the supervisor, as well as the candidate.
How to fix it: reappraise the relationships inherent in supervisor accountability. Remove the “number of successful supervisions” criterion from promotion and appointment criteria documents.
Time pressures to submit create perverse expectations and incentives. Increasingly, in recent work I have examined, I have seen candidates write of “running out of time” to, say, obtain full knowledge of the secondary literature. I understand that the Ph.D. should be time limited. There should be a reasonable expectation of what is achievable within the period of study. However, this is not how this is being used. Some candidates are able comprehensively to cover the secondary literature in this time period while others are using the time limit as an excuse for not conducting the basic work.
By standardising the time expectation – rather than the standard of the work – we create a problem. It is entirely possible that two different candidates will reach two different standards if they work over the same time period. This should not be the standard for the highest-level degree, which must have some absolute criteria against which the work is judged.
How to fix it: the “what is reasonable in the time period” criterion needs to go and should be replaced with measurable standards that are discipline-specific but also not bound to particular candidates.