As (social) scientists, sometimes we all experience a project that just never works. Sometimes I underestimate the initial investment and it never really gets off the ground (e.g., I thought I could get these resources but I can’t). Sometimes you do the whole thing and it looks great, but when you try to replicate it you can’t. Sometimes the data just don’t tell a coherent story and you need to go back and think about your methodology, predictions, theory, etc. I have had each one of these happen to me, some more than once.
How do you recognize when a project is failing, and what do you do with it afterwards?
Possible things to do with it: set the data aside, maybe they will be useful for something later on. Revise your procedure and start over. Let a student have it for a small project that you know won’t get published anyway. Shrug and chalk it up to a learning experience, taking the long view that my career is decades long and not every study has to pan out. These are all painful, but necessary in the case where you don’t want to lose more time throwing good effort after bad data.
One more choice: write an article about the methodology instead of the results, giving advice to other researchers who may want to do something similar. This approach has worked for me at least once.
Each time something fails after an initial investment of time and energy, it hurts. It sucks. I could always just choose to do projects with lower initial startup costs and lower risk of total failure (p.s. design your study so you will still get some usable data even if your main hypothesis is uninterpretable). However, I don’t want to stick with the safe and easy all the time. I did a study my first year on the tenure track that was easy, fast, data collection went great, it looked simple — but then the data analysis turned into a bear and we’ve just now gotten the manuscript out. I want to keep trying new things.
I’m trying to think that if you don’t fail at some projects, you aren’t trying enough things. If you don’t get rejections on your paper, you aren’t aiming high enough. (Thanks, CPP, for reinforcing this idea!)
I tell myself, you are allowed to suck. Indeed, you must suck. Get all your sucking out of the way so you can move on. Fail better.
Any advice from the Grumpeteers on when to cut your losses?