On many occassions I’ve heard the phrase: Epidemiology has too many methodologists. The idea being that we have too many people developing methods and not enough people applying them. There is a lot of methods development and not all of it is used. Many methodological developments sit in a journal collecting proverbial dust. But this isn’t necessarily a bad thing. Science requires pushing multiple boudaries without knowing ahead of time where the real advances will come from. We can’t move things ahead without hitting some dead ends.
I think if we really want to improve the quality of epidemiologic research, we need, to some extent a redeployment of methodologists.
I have this thought experiment where I compare what epidemiology looks like today to what it would have looked like today if, five years ago, we banned the development of new methods and instead, took all existing methods and tailored them to be applied to specific questions that applied researchers face. A tweet from Professor Sherri Rose brings up the same point:
I had some things to say in this interview with @laura_tastic.
— Sherri Rose (@sherrirose) August 5, 2020
For example, my thoughts on applied statistics and data science in academia…https://t.co/3wkkfvzlBu đŸ”“ #JSM2020 pic.twitter.com/TQVavahtRj
Our departments tend to prioritize hiring faculty who specialize in theory or methodology that could be applied to real problems but isn’t.
There are a wealth of methodological research topics out there that could use existing methods and tailor them to work to answer common but more specific research questions.
One clear case that comes to mind is genetic epidemiology which often asks questions about the mechanisms through which genes operate to affect a phenotype. Epidemiology has done a lot to develop the theory around causal mediation but this theory remains unread/underappreciated/misunderstood by many genetic epidemiologists who have questions that related directly to mediation analyses. And the general theory around mediation does not always directly answer the specific questions researchers can have.
I don’t want anyone to come away with the idea that we should stop focussing on methods development which is crucially important. But we should also be aware that there is an important gap between theory and application that can be filled with methodologists that can speak both the language of theory and of applied research. I’d like to see redeployment of methodologists, to some extent, to fill in those gaps.
