That's interesting. I would expect you not to get many submissions from people who would expect themselves to do poorly on your exam. But, if it turns out that a sizable portion of people who expect to do well do not, that's indeed interesting.
Given the selection bias we would expect among respondents, have you thought about how you will interpret the responses?
which is exactly the point why this type of question is useless
I will always reference other sources (google,book,peer) if I am not sure and I am not sure for anything that I am not using frequently
Real test is how many people will get it wrong using all resources usually available in working environment
There's a lot of people who are very confidently giving wrong answers to questions, which weighs against that position.