If you want to do good research, it's important not to know too much. This almost sounds contradictory but really if you know too much and you get an idea, you will sort of talk yourself out of trying it because you figure it won't work. But if you know just the right amount and you get enthusiastic about your project, you go ahead, you do it and if you're lucky things'll work out.
To me the greatest moment in an experiment is always just before I learn whether the particular idea is a good or a bad one. Thus even a failure is exciting, and most of my ideas have of course been wrong. But this time it worked!
If I have learned anything as a scientist it is that one should not make things complicated when a simple explanation will do.
My own beliefs are that the road to a scientific discovery is seldom direct, and that it does not necessarily require great expertise. In fact, I am convinced that often a newcomer to a field has a great advantage because he is ignorant and does not know all the complicated reasons why a particular experiment should not be attempted.