I spent 10 years in Marketing Research and know about Gallup and their methods. They tend to use criteria that bends to the right, and their polls are usually "outliers" in the curve when compared to other polls.
A lot depends on the question and how it's worded. I can write two surveys that "generally" ask the same thing, but by wording the questions slightly I can get two competely different results.
Been there, done that. Research is like sausages...you really don't want to see it made.