I love getting interviewed. Mostly I get interviewed about bicycling. Sometimes I’m interviewed about science culture. Less often I’m interviewed about my research. Interviews have been for tv, newspaper, articles, and student projects. Today I was interviewed by a student in a Master’s of Public Health program for her research class. The assignment was “Interview a researcher”. I had a grand time talking for an hour and a half. She sent me the questions the day before so I had time to think about it.
Q: How do you start a research project?
A: Ideally it starts with a question. Then you spend a long time whittling down the question into one you can answer. For example, “How do I cure cancer?” first becomes “Can I cure cancer with crystals?” Since cancer is many diseases, you further refine that to “Can I cure estrogen responsive ductal carcinoma in situ with crystals?” Of course you can’t just go around giving crystals to women with breast cancer– suppose the question you started with had been “Can I cure cancer with toxic sludge?” You can’t give crystals to women to cure their breast cancer until you’ve shown that it works in models (animal models, or cell culture) AND you’ve shown HOW IT WORKS. So those crystals are never going to be in a clinical trial until you’ve gotten past “Magic” and moved on to “crystals selectively induce apoptosis in tumor cells by upregulating caspase 3”.
However, all of that is merely the ideal start. The latest project I’ve started began with 4 tools we had available, then we crafted a question that could be answered with those 4 tools. I have expertise in reproductive physiology and rat models of disease. My collaborator has expertise in cartilage and post-injury osteoarthritis. Our institution has a special interest in and promotes research in manual therapy and mind-body which includes exercise. Put these together and you are testing the efficacy of manual therapy and exercise to prevent post-joint injury osteoarthritis in a female rat model.
Sometimes it starts with the question. Sometimes it starts with the tools. One of my thesis advisers liked to tell the joke about the drunk looking for his keys under the streetlight. Another fellow starts helping him look. They look and look, and don’t find the keys. Finally the other fellow says, “Are you sure you dropped them here?” The drunk answers, “No, I dropped them over there, but this is where the light is.” That describes research very well. This technique won’t answer the question we have, but this is what we have. This is where the light is.
In grant-ese terms, it’s a balance between the significance of the research question, and the feasibility of the proposed research.
It’s great to collaborate and get ideas from other people. Sadly, some people start their research project by stealing someone else’s idea.
Q: What specific tools do you use?
A: I prefer Medline for my literature searches. It’s really important to find out what has been done in this area and how it was done, before embarking on any project. That will keep you from falling into the same traps as before, or re-inventing the wheel. Or re-inventing the wheel and then rolling it headlong into the pitfall, to completely mix the metaphors.
More importantly, talk to the people who have published these papers. I have made numerous contacts, usually by phone, sometimes I start with an email, to people who have done something similar or worked with similar approaches to what I want to do. I can’t stress this enough, how important it is to talk to other people.
The statistician is one person you should consult…BEFORE the project. Our statistician has in her email sig this quote:
To call in the statistician after the experiment is done may be no more than asking him to perform a postmortem examination: he may be able to say what the experiment died of. –Sir Ronald Fisher
This came up yesterday. She very tactfully and gently explained that we have to throw out an experiment because of a design flaw. I said, “This is where we should have talked to you before doing the experiment, like in your sig!”
Q: How do you gain your expertise with the various tools you use?
A: It’s important to keep up on new technology. I find the older I get, the harder this is to do. It’s so easy to get set in your ways. You find something that works and why take the time to learn something new? But this is a trap I see a lot of older researchers fall into. Nowadays it’s all microarrays and high throughput, and they’re still doing old school research one gene at a time. You have to learn new things, and you have to keep up your skill at learning new things. Your expertise isn’t enough to carry you through more than a few months.
Q: What are some important experiences you suggest for a novice researcher?
A: Fail, and laugh at yourself. Fail again and again and again. Probably the most important thing my PhD has given me is the ability to make dumb mistakes with confidence and without losing any self worth over them. A year ago I presented data at a meeting and someone said, “Did you pH the saline? Because anything acidic will cause a pain response.” Oh. I don’t know if I can explain to a non-scientist how that is such a basic, rookie, stupid mistake. It would be easy to feel ashamed of having missed such an obvious thing, to try to hide that I had done that. I didn’t do that. I freely owned up to it and repeated the experiment with properly pH’d saline. I called it what it was–a dumb mistake.
Aside from dumb mistakes, most of research is failure. What’s the legend about Thomas Edison and the lightbulb, that he tried hundreds of filaments before finding the one that worked? (A quick google search reveals that he tried 100, 1000, or 10000 materials. Or 6000. Whatever, the point is he had a LOT of experience at failing.)
Again putting aside the experiments that “failed” because the data was negative, there are the failures in techniques. These are a little different than the dumb mistakes. These are when you try a technique for the first time. I have a theory that if an assay works the first time I do it, it will never work again. So far I have not had an opportunity to test my theory. So, plan to do a dry run a few times (depending on the expense and time commitment) before testing something.
In research, most of success is failure. Only we prefer to call it “negative data”.
Laugh at yourself. Don’t take your mistakes and failures personally. Everyone makes mistakes. That was a mantra for me, for several years as a grad student, because I would beat myself up over every mistake and it felt horrible. Everyone makes mistakes, everyone does dumb things. (That used to be a GEICO commercial. “That’s why there’s GEICO.”) Eventually I internalized the message because now I wear my mistakes with pride, as an example to everyone around me that they don’t need to be ashamed of their mistakes either.
Q: If I wanted to be learn how to become a competent researcher, what specific tools would you suggest I work with?
A: Other people, a statistician, failure, and humor.