Currently the response rate is 9%. That's such a small number a dedicated group can inform its members/adherents/coreligionists/whatever to keep on the look out for calls from pollsters ~ AND ALWAYS ANSWER THEM ~ or, alternative use their CALL BACK option.
Done in earnest a group can have its opinion AMPLIFIED 11 times.
We know National Education Association and other public employee unions, as well as the abortion industry and homosexuals (of all kinds) can be depended on to respond the same way to certain questions. If their response rate is above 9% their opinion ~ which is a minority opinion ~ will seem to the pollsters to be more than half the public response.
They are doing it already ~ then there's the problem of wealthier folks having more than one personal phone number.
And yet the 2012 election polling, particularly the state polls, were dead on accurate. So accurate in fact that poll aggregators, such as the NY Slimes 538 blog, got all 50 states correct in 2012. The people that were wrong were the folks that refused to believe the polling that showed Hussein being re-elected and Democrats winning most of the big competitive Senate races.
Polls are getting more accurate, not less so. The science and methodology is getting better, not worse. Obviously you need to see the poll questions on specific issues to make sure they aren't loaded to achieve a desired result, but most of these polls with neutral questions are very accurate - especially when you take an aggregate of polling for things like generic ballot, presidential approval, etc.
Even that stupid unskewed polls guy admitted the pollsters were right and he was completely wrong. It was sad watching so many conservatives buy into his BS, but at least he had the stones to admit his assumptions about polling were wrong. Dick Morris is another one that refused to believe the polling and has since admitted the turnout model the pollsters predicted was absolutely accurate and he was completely wrong.