[lucasjung]

I'm not sure what you're trying to say here. Are you questioning the veracity of the "likely voters" screen? If so, you should understand that this is not some arbitrary designation. Different pollsters have different methods for applying this screen, and none of them are perfect, but all of them are based on serious analysis by real experts. I'm not a polling expert, but I've read enough about it to have an appreciation for how complex the problem is, and how creative some of the pollsters can be in finding ways to make their polls more accurate. For anyone interested in building an "enlightened layman's" understanding of polling, a good place to start is by reading Sean Trende:

http://www.realclearpolitics.com/authors/sean_trende/

It's not necessarily that he's the best analyst out there (I happen to think that he's at least in contention for that title), but rather that he explains all of his reasoning and processes so clearly.

[powera]

Unless there is going to be a national plebiscite, whether or not a person plans to vote in some election (he doesn't say which) shouldn't really be relevant here. Congress and regulatory agencies are supposed to represent all the people, not just the ones that vote.

[lucasjung]

"Congress and regulatory agencies are supposed to represent all the people, not just the ones that vote."

That may be what government is supposed to do, but in reality congressional representatives and senators mostly care about the ones that vote, since that's how they keep (or lose) their jobs. Also, I don't think that's all bad: "How much have you voted in the past? And do you plan to vote again next time?" is a pretty good proxy for "How much do you care about current issues?" Not voting is a way for the less informed/concerned to defer to the more informed/concerned. In an ideal world everyone would be fully informed and concerned, and so everyone would vote, but humans aren't ideal and our system handles that fact pretty well.

Furthermore, while there generally is a diffrence of about 5 points between likely voters and registered voters, and another five points between registered voters and adults, there's no way that a 1:5 ratio is going to swing the opposite way due to the difference between likely voters and the general population. At best it would shift to 1:4 or 1:3, still an overwhelming majority.