Running a lab and doing science are hard. I owe thanks to the Twitterverse for directing me to a thoughtful piece written by Uri Alon in 2009 which is still relevant today. The title is “How to choose a good scientific problem” and it’s a very quick read, but it articulates very clearly the challenge in selecting scientific problems for yourself and your trainees. He relates that any scientific problem can be mapped out on two axes which are feasibility and interest. He argues that you want to avoid spending time in the quadrant containing problems that are hard and yield a small gain in knowledge. The efforts that you’d have to rustle up in order to solve a problem in that quadrant won’t pay off much. In contrast, putting a new student with little experience on a problem in the quadrant where the project is easy and produces a small gain in knowledge is a smart choice. As trainees gain experience and confidence (e.g. senior graduate students and post-docs), you can move them into solving problems that are a bit more challenging and lead to larger gains in knowledge. I really liked this approach for selecting good problems to work on and how to assign them to particular trainees. I also like his idea of making his trainees “take time”. He makes his trainees wait for 3 months or more before they commit to a particular problem. During this time his trainees read, plan, and question and come up with a solid problem to solve before they dive into research. I also take this approach with my trainees when they first enter the lab and there are certainly times when we both feel that we are wasting time by not producing results immediately. I will make the argument that this initial investment in time pays off in terms of my trainees better understanding their research and being more motivated and engaged in solving their defined problem. The other powerful observation that he makes is contrasting two different schema when it comes to visualizing what the research process looks like. The Scientific Method is taught in classes as a series of linear steps, which I think is wrong. Perhaps because of this false structure, many scientists view research as a series of sequential steps (e.g. that you must go directly from A to B). This leads to a lot of frustration because in my experience research never directly goes from A to B, but meanders all over the place. Alon suggests that it is better to start with a nurturing schema for research that expects that meandering will occur and takes steps to nurture students while they are stuck in “the cloud” (i.e. when everything goes wrong and your assumptions prove to be false). This schema accepts and embraces the possibilities for new research directions and personal and professional growth.
I’m often guilty of biting off more than I can chew with my own research problems, but I try to protect my students as best I can from this tendency. Alon’s short essay has given me some new things to think about and has confirmed some of my conclusions about choosing a good scientific problem that I have made during my first 5 years as the head of a research lab. I recommend reading his piece and seeing if it influences how you choose your future research problems.