I don't think there are really good pollsters and bad pollsters*. They're all just guessing

[Renew Deal]

I don't think any of them are particularly good or bad. They're all a different shade of the same color. They're all playing the Price is Right where they try to get the closest without going over. The organization that does that is the best for four years. Just think of the "good" pollsters over the last bunch of years. It was Zogby for a while. Then Rasmussen in '08, then IBD in 2012. This year it will be someone else. It's not because they are that much better than anyone else. They just get lucky.



*Some of the academic polls that don't do a lot of polling are clearly and obviously bad.

[brooklynite]

1. I think you're wrong

Polling and market research are mathematically sound. Some people do a good job, some people don't.

[Foggyhill]

7. Man, the sampling part has nothing to with math

You need random representative sample of the population your polling, say likely voters, for the math part to even enter in consideration

If it was all math, well all polls would give you the same results wouldn't they

[uponit7771]

15. the "LV" part is the uber variant in their analysis sometimes based on what the pollsters "think" an

... and that's been my issue with a lot of these polls that have ignored 2008 and 2012 as if 2006 demo change among PoC voters didn't happen

[zenabby]

5. Polls are sound

But as a person who did GOTV calls in NC, what I heard was very very annoyed voters. Many didn't want to talk, said "oh, you again" and most were not at home or were busy. Many told me that they have already voted and someone had already called so why was I calling again? Also a ton of wrong numbers - and they said that they had informed about wrong number so why was it not fixed in the data base?

So, I'm wondering (a) are they getting a good sample set by phone
(b) Is GOTV really useful? It seems to me we don't care about the annoyance factor of swing voters.

[uponit7771]

13. nooooo, some of them take some really good data and make them part of the conclusion while others

... jus make shit up and pull stupid non fact based numbers out of their asses or let numbers that make no sense that they've gathered be part of their analysis

[mahina]

18. Really, no. Data science is a thing now. Just because I might not understand (don't understand)

All the statistics and equations and models and all that, doesn't mean it's a guess.