Just a brief note. There was a column in today's Washington Post looking at a recent GAO report that examined various studies comparing public to private sector pay. What I find notable is that the column briefly mentions that the GAO report doesn't tell us which methodology is correct.
I disagree with the claim that this leaves the debate at the same place it was before the report. A null result is a significant finding. If the perception of difference relies on the method used to perceive that difference, and all methods are methodologically sound for the question under examination, than the difference is too small to be worthy of basing public policy on (or the methodologies being used are either flawed or improper for the question at hand, but this is not the finding of this report).