How about the "statistic" that 90% of Americans want the bill to pass? I don't know anyone who was asked. I did read yesterday how the poll was taken and I'm sure it was a very select group who were asked to speak for all Americans.
To be fair, a poll with sample size of about 1,100 or so *can* accurately reflect the opinions of a much larger population to a high degree of accuracy and precision, IF the sample is selected randomly, AND those polled provide truthful answers.
As an example, if you have a population of 300 million, and you want to be 95% certain that your poll results reflect the population as a whole, plus or minus three percentage points, you only need a sample size of 1067 people.
It's the same reason why even though TV shows live and die by their Nielsen ratings, and there's huge money involved, there are only about 25,000 "Nielsen families" used to represent over 100 million TV-viewing households in the U.S. And they could easily get by with even fewer if they didn't need to further break up the sampled households into demographic groups like "male/female", "18-34", "18-49" and so on, while still maintaining an adequate sample size for each sub-group.