reply @ Sykozis
That's exactly how the media and government operate. When they want to up play or down play a situation to suit their side of the arguement. They'll sample less or more people accordingly and word the questions in such a way to give more or less percentage to their cause.
For example, Obama recently heavily touted 40% of guns sold do not go through a background check. That was from a study done in 1993 just after background checks were implemented. The sample size was something like 250 people, yes only 250 people. And the question was "do you know if your firearms dealer had an FFL license?". That's like asking if i know if my mechanic is ASE certified, i would assume so, but i'm not really sure. Yet he wants to use this small misrepresentative study to make pivotal changes to the constitution.
The NSA is no different. They want to continue infringing on the 4th amendment, knowing full well they are dancing over the line and getting away with whatever they can. So yes, some groups will seek to discredit any study going against the NSA. But in the end, i believe that most given the option wouldn't want their "meta data" in a database to be searched and prodded.