Owen Yang
I did not invent them but I could not remember the exact quote, but this was almost like a catch phrase that I really liked when I studied in Education: relevance, validity, and novelty.
I quite agree that these are the three elements with which a research ‘deserves’ to be a doctoral theses.
Relevance is why the topic matters
In my opinion, a researcher should be allowed to answer this question from different perspectives, but the answer does reflect us as a person. In medicine if you do not feel particularly creative, I do see these answers following a common pattern (genuine or not). This is to start with how important the disease is, such as how many people have it and what the economic or moral burden it may cause, and then narrow down gradually to say why this topic is important in the disease.
Dementia is important because it causes massive burden and costs money, and in an ageing society it is more so. Delaying the progression of dementia is important. Molecule X1 is important because it is known to be associated with dementia. Molecule X2 is important because it is known to affect molecular X1.
I usually laugh at this point because by the time you get to molecular X2, it might only affect 0.01% of the original massive ‘burden’ and cost more money than the original cost. But I understand the spirit of it and so I would let it pass.
But the point here is as a responsible researcher we should understand the relevance. What the examiners tend to confuse themselves is that we should oversell the relevance. There is really no need in an ideal world. It is just we are not living in an ideal world.
Validity is whether the question is answered
This is something that really gets to know the examiners and examinees as a researcher. What both of them tend to jump in is what the types of validity there are, or what the types of biases there are. I was a little disappointed when I was asked this question, and my examiners clearly did not know what the question was about.
I have full confidence that most people have some capability of going into the technical details of the types of validity and biases, but the broad question about validity is really about a summarising assessment about the extent to which your research aim is achieved after going through all the experiments, spending all the money, and endless bowing of personal dignity to beg for research grant.
The list of technical validities and biases can be very long, and we really need to know what matters. Nowadays, out of no where, it becomes common to be asked whether mediators have been mis-treated. This can be a good question, but in some circumstances it was asked out of context because the person who asks clearly does not share the same background that can understand the nature of the medical question. For example, when you investigate whether BMI can be associated with the risk of heart attack, they might ask you whether waist circumference can be the mediator. I do not know whether you could see the problem in this question because it is hard to explain in words, but there is a huge misunderstanding between the two people about the assumption behind this question. Instead of going through these details, one needs to focus back to the basics and think what the research purpose or the research question really is.
But again I would like to emphasise that in an ideal world, what we need to know is the honest extent to which the research question was answered. There seems to be a culture, or an interview technique, that we need to over-sell out validity. This is very sad and I would try to avoid subscribing to this culture.
Novelty would best be based on methods, not results
This is my personal opinion, but it does not mean the results are not important. I am just saying novelty as a criterion of the quality of a doctoral thesis, in an ideal world, needs to be from a new approach that has been developed. Of course it is great if you just discovered semiconductor or invented machine learning, and congratulations for making the history. But you kind of have covered that in your relevance section, and hopefully validity.
The novel method does not have to be ground breaking, but has to go through a process where a researcher takes responsibility and develops the details of it. It can be a study protocol, a questionnaire, or a framework with which we assess a product.
Nowadays many researchers are hired to follow a protocol because there is just so many data mining and exploratory experiments to populate. It is obviously something that needs someone to attract grants to hire someone to do, (i.e. it has to be done). But it would be a shame to pretend this is some sort of in depth education to become a doctor of philosophy. It would be perfectly fine as a Master level project.
Nothing needs to be ground breaking
My belief is ground breaking findings should be incidental (accidental) and should be built on small progresses.
