There was a lot of garbage polls in South Carolina

[Lexblues]

NBC Marist and YouGov comes to mind. It makes me wonder what other polls are trash. It seems like nearly all the polls overstate Bernie's performance no matter the state. I'm really losing any faith in the polling these days. One pollster once said that Bernie is doing better in polls than in reality cause his supporters are fervent and they go out of their way to answer pollsters. I guess we'll see come Super Tuesday.

[still_one]

1. I really think that polls that have low ratings, and sponsered by those with a political

agenda need to be eliminated in the Calculus

[jmg257]

2. Could be - for all the "huge turnout" sanders infers,

It doesn’t seem to materialize.

[ZZenith]

3. But you have unshakeable faith in electronic voting machines?

[Bumperstickers]

4. Most of the polls were within moe for Sanders final numbers

But they did not account for the Clyburn endorsement which is huge and good for 10 pts at least and they also did not account for Styer deopping 10 pts that obviously went to Biden.

So South Carolina has quite a bit of intangibles to deal with unlike other states.

[squirecam]

5. All Steyer voters went to Biden?

You would expect then, as Bloomberg drops, for the same?

[Blue_true]

7. My guess is the Biden/Bloomberg dynamic has shifted.

I expect Biden to take a lot of Bloomberg votes in California. Voters not sending in their ballots yet implies that they were waiting for a clear signal on Biden, South Carolina gave that, IMO.

[Blue_true]

6. In 2016, polls overstated Bernie's appeal in California by around 14%.

It would not surprise me at all to see Biden win California and Texas, if Biden wins both or is close to Bernie in California, Bernie is done, the road forward gets brutal for Bernie in that he has only Wahington State, Oregon, maybe Michigan, maybe Indiana, maybe Rhode Island where he can gap Biden, all the other states except small population plains states, Biden most likely win by a good margin.

[Happy Hoosier]

8. So much of a poll is the population model.

And for a primary, the population model means the turn-out model, which can be notoriously hard to predict. And sample sizes are often small, so a bad sample can really fuck the results.

That's one reason why poll aggregation is better than individual polls.