How can we can encourage questions to show proof of initial research?

Question

I think that better questions show some at least rudimentary level of initial research.

This might just be:

• "I've read the first hit of a Google search, and I received conflicting answers."
• "I did a search on Google Scholar, and read journal article X, but something still wasn't clear."

At present, I guess there is some question seeding occurring. I can even think of some of my own questions which I've posted partially to seed the site as being a bit lacking on this requirement. I'm also not saying that we should close questions that lack initial research, and sometimes it is implicit that the person has searched for an answer already.

Rather, I'm wondering

• How can we encourage users asking questions to show the initial research they have done?
• What guidance can we give, perhaps in the FAQ, about the kinds of initial research that is encouraged?

Update

I've made a separate question for What level of initial research should we expect?

Answer 1

Reward them.

This is what the voting system is for. Up vote good questions which don't raise any additional questions and show sufficient effort has been done to formulate it.

Down vote questions which are overly broad and raise plenty of other questions. Leave those questions as a comment. In my opinion it is very important - especially during the beta - to leave comments for users on how to improve their questions.

Ideally, track those questions and remove the down vote, or even up vote once the question has been improved.

Redirect them to meta.

For now we can't redirect them to a FAQ, since no consistent idea of what this site should be like is formulated yet. That's the purpose of the private beta.

Instead, by linking them to relevant meta discussions we encourage them to participate in defining the site.