Polling for substance

So, if someone asks you if you support elected representative-X, how much merit does your polled data have?

Often times the things we support are because we believe we ought to support them, or because an authority outside of us has instructed us that we ought to support said thing. People often mirror their support for things based on people they respect or admire, their parents, their hero's, their pastor/priest, and even in opposite; They support X because their noisy, good-for-nothing neighbor supports Y.

If taken into the sample poll, the social connections would really reveal the underlying truth. Assuming the six degrees is accurate, then in any poll taken, the representations of a sample would really just boil down the two or three highly connected and influential people in the social network. They could be the saint or the devil, depending how their opinion was spun, and then everyone else would simply reflect that.

If, of 1000 people polled at random, 95% of them support elected representative-X, but 95% of them also watch the 700 club, how valid is that 95% based on the singular social connection.

Also, in any poll, there's going to be the random noise. People who really don't have a clue and just go with the first answer that pops into their heads.

So, I guess, that if there was to be an accurate poll constructed reflecting my opinions and beliefs, it would have weighted against my social network, with varying degrees of weight placed on the proximity to other polled in that network.

Maybe in the future, electronic voting for elections will just be on Myspace or Friendster?