The first thing is to address your frustration with the interviewer. They are not allowed to deviate from their script. The script is written to control for any sort of design bias. They are instructed to read it in a flat voice so as not to give any cues to the respondant. If you think it's frustrating to be on the receiving end of a poll (and I agree that it is) you should be on the other end sometime.

A second thing to state is that all professional pollsters use mathematical models to correct for the difference between their sample and the actual known demographics of the voting population. This is because it is impossible to randomly draw a representative sample.

A third thing to state is that I typically dismiss out of hand any poll that has less than a thousand respondants. The reason is that any error – intentional or not – is magnified by orders of magnitude as the sample size shrinks. After a thousand people, the numbers don't change the percentage of error (significantly) until you get to the millions. For Kos' poll, having 400 respondants in each of the primary will yield a mathematical Margin of Error of 4.9%. That is simply unacceptable.

A fourth thing is that we aren't told the response rate. That is, we don't know how many people refused to answer various questions. People like me – who know the ins and outs of polling and despise it – often refuse to answer polls. But I vote quite often. It skews the results.

Finally, no single poll is noteworthy. They are only worthwhile in showing trends over time – which means the exact same poll has to be conducted over and over again. Comparing one poll result to another, while it makes for good column writing and punditry, is bad science (unless one is an example of how not to poll). Even within the run of a single poll, one poll that falls badly outside of the others is not any reason for alarm – or even much interest. If Kos could produce the exact same poll four or five times with similar results, it might be worth examining. As it is, it is an outlier or a statistical deviation. In other words, there is probably something wrong with the poll – and that would be true no matter who runs it or what their sample size was.

It's not only the original bill that matters here, obviously, but also what's in the amendments to it.

It's a handy way of hiding things, in fact, which wouldn't be surprising from these Democrats.

“It's not in the [original] bill” is about as respectable as “If you [will still have, and] like your doctor…”

http://energycommerce.house.gov/Press_111/20090…

The lunatics are running the asylum in Sacramento again, as they currently are in Washington, DC.

But she wasn't asking about your KNOWLEDGE of any abortion laws, she was asking your OPINION on abortion. Furthermore she did exactly what she was supposed to do when you challenged the question – just restate it. Robotically. If she had gone into explaining any nuances of the question she would be corrupting the study.

It's true that polls can be used to manipulate and even spread bad (false) data. That's why, as a journalist, you need to take note of the polling firm's reputation, how they did their sampling, and exactly how each question was asked.