Best polling practices?

I've started a poll to see if we can characterize the 1.8L cold-start engine problem. What I try to do is have 3-4 independent questions with 3-5 answer ranges:

• 3-4 hypothesis - trying to find out what elements are key to a problem
• 3-5 range per hypothesis - this allows:
  • low, mean, high - trying to confirm a range
  • low, mean-low, mean, mean-high, high - a finer range

In a perfect world, a poll would have 4x5 or 20 options. However, the existing 10 poll responses pretty well narrows it down to 2x3 and 1x4 which limits how much information the poll can return.

Granted, I'm into diagnosing problems so my polling request may be beyond the 99% but I wanted to share why more polling questions could help in the future.

Yes, I could post multiple, polls, but the thread overhead gets complex quickly. Also, I don't have a good answer for how to share a poll across multiple models. I'm open to suggestions but wanted to share what it looks like to me.

Thanks,
Bob Wilson