I did an extended investigation last year, though I didn't do too bad I don't think! My research question was on comparative literature studies on the nuclear art and Japanese literature post-war era. My written + oral piece got an A+ but my sacs got scaled down a lot so hence why I didn't get a great score! My advice for you would be to develop a question you're very interested in -- just start reading now and then it should fall into place. I'm pretty sure I didn't even have a question until the last (two weeks)

. I can't remember what my research question was on but it was basically 'How did Japanese literature evolve and influence art, post-second world war?' maybe a little more compact than that, I can't really remember for the life of me! Unfortunately my computer with the essay broke down on me, so I can't get it up.
If you're doing anything on the vicinity of literature/classics I might be able to proof-read for you if you like! To answer your question, your research question and report doesn't have to be whacky to be great! Just make sure you're succinct and have someone to read over your report! Technical language is desirable if you're doing science for the added sophistication. But I'd strongly advise against writing it for the sake of it, when familiar phrases could lend for better understanding. The reports really just come down to how well you can clearly present an argument/research (call it whichever you like). This means proof-reading and as well as the format in which the paper's laid out. There are marks allocated just for that!
As for science-maths kind of questions, I'm not too familiar with it, but I'd hazard a guess that it'd be advisable if you avoid complicated questions that may be too hard to understand (ie. some physics questions out there haha). Stick to questions that are interesting and easily translatable on paper.