Researchers have found that merely asking people to assume their initial judgment is wrong, to seriously consider why that might be, and then make another judgment, produces a second estimate which, when combined with the first, improves accuracy almost as much as getting a second estimate from another person.
