Even though there are scientists who have the kind of expertise that might help us to better understand this pandemic, there’s a tendency to suggest that it would probably be best if they stayed in their own lane. Although I do have some sympathy for this, I think it’s too simplistic; researchers should be free to study what they wish, and those within a discipline should be willing to listen to people with relevant expertise from outside their discipline. However, there are certainly examples when researchers have tackled problems outside their core area and not made particularly constructive contributions. The problem, though, isn’t that people don’t stay in their lane, it’s that they don’t do their homework properly when they move outside their lane.
One of the most difficult things about doing research isn’t the technical aspects, it’s being very familiar with the details of a topic, and knowing what questions are worth asking. Just taking some data and throwing some analysis method at it isn’t very useful if you don’t understand how the data was collected, it’s limitations, or the significance of the analysis in this particular context. For example, the impact of this virus is almost certainly going to depend on the strategy that is employed. Hence, you can’t really infer anything about an alaysis if you don’t take into account what strategy has already been employed and how this strategy might evolve. There are, of course, many other factors that should also be considered; it’s clearly not simple.
What motivated this post was a recent post by Andrew Gelman that highlights how to be curious instead of contrarian about COVID-19. It was itself motivated by an article by Rex Douglass that provided Eight Data Science Lessons, using an article written by a rather contrarian, and not very curious, lawyer to illustrate what not to do.
In my view, the key thing is that even though these are unprecedented times, it doesn’t mean that we should take research short-cuts. As the articles above highlight, we should be familiar with the topic, care about the research questions, be careful about the design of the research project, be willing to revise our understanding if the model doesn’t match the data, or if new data becomes available, and be very clear about assumptions, uncertainties, and the overall context. I would add that we should also be willing to “trust” other experts. If we want to live in a world where people listen to us when our expertise is relevant, we should be willing to listen to other experts when their expertise is relevant.