I could and will comment that a poll is not a sentient creature capable of lying, I know you know that and it isn't your point but the difference here isn't semantic in my eyes and I want to underscore that what stats tell people versus what they choose to hear are really very different things.

That's not really all that unlikely, these are usually 95% confidence, I play D&D way too much to consider a natural 20 unlikely Also I don't think it is really a semantic difference here, intentional falsehood vs misinterpretation vs method error in data acquisition vs intentional push polling etc are all very different things.

Well, I'd have to debate that, I've done a lot of local automated polling, it gave us a huge advantage being able to run 1000 person samples for under $100 a piece rather than around $10,000 but doing that required a lot of effort to determine what could skew the sample and what we could do to adjust the data rather than eliminate that skew. 'Executed correctly' is a very flexible term and not just in polling but in almost all science. As an example, if I give people:

Q1: Barrack Obama the democrat, press 1, Mitt Romney the Republican, press 2, Not sure, 3
Q2: Jane Doe D, 1, John Smith R, 2, not sure 3
Q3: Bob Jackson D, 1, Eric Carlson R, 2, not sure 3
Q4: Alex Park R, 1, Donald Hill D, 2, not sure 3

A lot of people who are hardcore Dem or GOP are going to seem to break party on Q4 because they though 1 meant Dem and 2 GOP rather than the order being alphabetical. Now I can toss that data entirely or I can salvage it and note the problem. People who pressed 1 across the board, or 2 across the board, can probably be assumed to have switched on Q4, and you can just reverse their data points. It adds an extra layer of iffy, but it isn't useless data. Same, if I get Obama over Romney by 20 points, the national poll has him by 5 points and historically we vote 3 points left of center in that area, I know I should be assuming the other R's got 12 points higher than they really got. It's touch and go and often more art then science but that it was still 'executed correctly' and truth be told that is perfectly common in a lot of science by necessity, ye olde fudge factor is normal enough.

True enough