There are a thousand topics..

[Tall Chai Latte]

Hello GradCafe,

'Tis the season to be preparing for prelims... Our prelim format is coming up with a research topic that interest you, write a proposal, and defend it in front of a committee.

Although, the rule is that the topic cannot be something you are or have worked on. Current projects, current projects your lab has, past rotation projects, even your past undergraduate research projects are not allowed. Looks like stepping out of the comfort zone is what's asked of us...

I'm totally new to defining a research project from a field that I have limited knowledge on. I do have one or two areas that interest me (that's outside of my past experience, expertise, and current interest), how does one even begin approaching defining a research topic after reading lots of background info? Any advice is great.

[fuzzylogician]

I usually stumble onto topics by coming across data I don't think is adequately explained in the papers I am reading or by coming up with new data of my own, usually following reaing a paper that I think makes some predictions that weren't tested. For more experimental work, I occasionally come across a theoretical paper that makes claims that I think it would be possible to test in the lab, and I go from there. I know some of my professors occasionally want to try out a new method and then try to come up with a problem that fits the method. Another way I used in undergrad for shorter papers was find papers on interesting topics that explicitly discussed "questions for future research" or however else they call those pesky problems that still linger in one's work these days. If it's up to you to define a topic from scratch, I can't think of a concrete way to get moving except use the readings and find corners where they seem unsatisfactory, but not hopelessly impossible. I assume you are allowed to consult with professors and peers once you've started zeroing in on a topic? working in a vacuum would be counter-productive and an unintuitive way to do research.

[Eigen]

I work best by looking for holes/unanswered questions or extensions of work that I'm reading. The more familiar you are with the field, the easier they are to spot.

I jot down anything that makes me think "I wonder if...". Sometimes it quickly becomes apparent why, or why it wouldn't work, but sometimes you find a cool idea. Those that fit in with my current work, I take farther- the others I keep written up for future work- post docs, job talks, etc.

Once you've got the idea it's about setting it up into phases- what do you want to show, and how do you go shout doing it? What preliminary work do you need to do? Is the problem a complex one that needs to be broken into several components before a final study?

Then it's about the how part- what studies are you going to perform, how do you want to set the up, what do you need? Controls?

Then it's about the what if stage- what can go wrong, how can you get around it, what do you do if you can't?

Thats how I approach it, at least. If you want to talk about more specifics, I'd be glad to in PMs.

[kotov]

The topic I'm planning to use for my dissertation was something that I read about in a study of one issue that admitted that it warranted a separate study.