“If These Numbers Are Right...”
That’s the problem. What trust has been earned by any poll? And it may go beyond pathetic statistical analysis (extrapolating 1000 answers to 200 million people, as you pointed out). What conclusions were made after the 2004 exit polling fiasco? Is there intentional skewing of data? If so, I want very public light shed on that...
I think in 1997, after a HORRIBLE round of polling in which every single major poll had Dole losing by much more than he actually lost, the Republicans in Congress tried to investigate this and were basically stone-walled with “we used established polling methods,” blah, blah. HOWEVER, for the next four years (three election) they straightened up, and were actually pretty accurate in 1998, 2000, and 2002.