In his case, the shiny tool is the "original position." In a nutshell, it means a fantasy-place where we make desisions on our life paths while being denied any data upon which to make our decision. He then uses a snazzy game-theory method, the "minimax" criterion, without explaining why minima and maxima would make sense in the absense of any data. His underlying metaphysical premise is the old fave of the liberal academician: randomness. How we can decide that randomness is correct without any data upon which to base this assumption? No answer.
Methodologically, he got away with it because mainstream statistics assumes that an event for which we have no data has a 50/50 chance of occurring. That's what makes the math work. But, it's only a methodological assumption. It's not an axiom, and has never been justified as such; it's only a convenient assumption. Great for a professional statistician, but not quite so for a philosopher.
Because he has to smuggle in assumptions that implicitly contradict the very conditions of his "original position" to make his schema work, Rawls' "original position" is just another burst of sophistry. Like other sophistries, it's used to justify positions arrived at earlier - pre-judgements, if you will. One of the findings of classic symbolic logic is that a contradition can be used to "prove" anything.