Tuesday, June 17, 2014

Scrambling yourself together and making big changes in your research

I recently wrote a post on How to know when to change course in your research, and I got the following question from a reader:

interesting (and helpful) comments. however, once you recognize that there is a 'problem' how do you dig yourself out a hole?

Indeed, it is not enough to know that you drove your wheels into a muddy patch - the real deal is then figuring out what to do about the situation.

Here's a number of strategies that you can use to get back on track:

1. Trace back your trail

Follow the trail that you developed back, until you find the point where you went off on tangents. For problems in STEM fields, you might like to go back to your basic assumptions, and test if these are valid. You might find that your research led to the conclusion that one (or more) of your basic assumptions were not fulfilled. And that's not a catastrophe - it's an important aspect to recognize. But if the basic assumptions aren't fulfilled, you'll have to sit quietly for a moment and wonder why they were not fulfilled, and how your approach changes because of these assumptions not being fulfilled. And once you start pondering on these questions, I reckon you will start to see new possible roads to follow, and before you realize it, you'll be up and out of the hole again.

2. Gather your troops

You lost a battle, but not the war! So take a deep breath, and take stock of your "army". What work that you did can you still use, in one way or another? Which literature sources are key to your resource? Which experiments did you do, what did you learn, and what needs to be figured out still? Try to come up with an honest assessment of your current situation. Start by thinking about your research question (that's the key to everything!) and then write down / sketch / draw / mindmap / whatever for yourself where you are. Which subquestions have you answered? Did you move away from your original research question, and did that maybe cause you to find yourself at a dead end? Do you need different methods to come to a better answer of some subquestions of your research question? Consider yourself a general who is counting all the remaining soldiers in his army after a battle, so that he can come up with the best strategy for the next round.

3. Take a break

If you feel overwhelmed, escape. Really, you just can't think clearly if your brain is in panic- and stress-mode. So take a day off, take a weekend off, spend a week long watching movies from your bed every night, get out and hike a long trail - whatever works for you to get the worries off your mind and prepare yourself to come back after a few days with fresh eyes to reassess your situation. You'll most likely see that it is not that bad, after all.

4. Assess what was useful

If you spent a couple of months following a lead that didn't work out - not all is lost. Most times, you will have produced something. And research showing what does not work, in my opinion, should be shared and published too - if only to warn other researchers not to go down the rabbit hole. It's a fact that negative results barely ever get published - a worrying statistic that leads to the so-called "publication bias". You might not make it into a big journal by writing about your failures, but somewhere you owe it to your research community to communicate what does not work: present it at a conference, blog about it, put out a report into the public domain - but do something with it.

5. Ask for help

If your thoughts keep running in circles, don't feel ashamed. There are always senior faculty members who are willing to sit with you for a moment and follow your logic - and then you where it sounds flawed. Academia is for the independent, but often we could do things faster and more profound by asking for someone else's input. If you're a PhD student, you are an apprentice, and it's perfectly fine to tell your supervisor that you hit a roadblock and are wondering how to escape. But even if you are an early career researcher, or maybe a senior faculty member, it's still OK to ask someone to sit with you and discuss the topic you got stuck on. And often, you'll get a new perspective even from talking about the problem to someone else. Fleshing out a knot out loud brings you new insights.

6. Ask yourself critical questions

Let's go back to the red flags for being stuck in the wrong direction in your research. Now we can turn these red flags into critical questions to help you find your way out of the black hole you got into:
- What are the basic assumptions? Which ones are violated? Why are these violated? Which changes do you need to make? (something we discussed at length in #1).
- How can you return to basic principles from your approach? Can you study a benchmark problem and see what is missing? Which steps do you need to take to make simple cases suitable for your work?
- What are the boundaries of the theories that you are applying? Are you using any (semi)-empirical formulas that are tied to the range over which the parameters were tested? Can you revisit the theories you applied and check the boundaries?
- Can you simplify the methods you are using? I once heard a professor say that, if you simplify your approach down to an easy hand calculation method, and you can have results within 20 to 30% of the "true" (or experimental) values, that you are on the right track. The beauty of science often lies in simplicity. Try to find where the simplicity in your solution might be.

By all means, don't beat yourself up over getting down on the wrong trail. You might have lost a few months' of time, but you have come out the other end with a few more skills, and lessons learned!

What do you do when you need to get yourself out of a black hole? Do you agree with these techniques? Please let me know in the comments below!


  1. I enjoyed this article. I start my PhD program in September and it's always good to find great blogs about research and dissertations. I know the journey will be a tough one, but I will make it. With great tips on this blog, I know I will be just fine,

    1. Thanks again! Feel free to ask me questions along the way :)