FL-St. Pete: Crist +1

[Donerail]

Hell yeah Charlie crew

[Donerail]

This same poll last year:

If this garbage poll is the best hope Democrats have in Florida I am not worried one bit.

Stupid gotcha — you cherry-picked their poll of Pinellas County instead of their final general election poll, which was 49-48 D (within the margin of error, and closer than most national pollsters). They're not perfect, especially when polling small geographies like a county or city, but they're a better outfit than most.

[Donerail]

Stupid gotcha — you cherry-picked their poll of Pinellas County instead of their final general election poll, which was 49-48 D (within the margin of error, and closer than most national pollsters). They're not perfect, especially when polling small geographies like a county or city, but they're a better outfit than most.

They predicted Biden winning Florida by 6.

My case is closed.

1. A poll is not a "prediction." It is a snapshot of the sentiments of the electorate, based on a representative subsample, at the time the poll is taken. If you do not understand that, I would recommend you get out of the polling subforums and stick to those more suited to your IQ, like US General Discussion.
2. Their poll in July was Biden 50-44. Their poll in November — the closest thing to a "prediction" — was 49-48.

[Donerail]

1. A poll is not a "prediction." It is a snapshot of the sentiments of the electorate, based on a representative subsample, at the time the poll is taken. If you do not understand that, I would recommend you get out of the polling subforums and stick to those more suited to your IQ, like US General Discussion.
2. Their poll in July was Biden 50-44. Their poll in November — the closest thing to a "prediction" — was 49-48.

So their predictions were

1) Wrong in November

2) Horribly wrong a few months out

So all this tells me is that this is a garbage poll that is especially inaccurate months before an election, let alone years before.

Polling months before an election is often inaccurate. That is because there are things that happen in the months between the election and the poll, known as "events," that can shift the sentiments of the electorate. That does not invalidate the poll's role as a measure of how the electorate is currently leaning.

[Donerail]

In 2018, St Pete overestimated Gillum by more than 5. In 2020, St Pete overestimated Biden by more than 4.

Fake news! They overestimated Biden by 1 and overestimated Gillum by 1. They certainly *under*estimated the GOP candidates in both races by a lot more, but if you're going to attempt to discredit the poll, at least be precise in your claims instead of arguing that 1>5

[Donerail]

In 2018, St Pete overestimated Gillum by more than 5. In 2020, St Pete overestimated Biden by more than 4.

Fake news! They overestimated Biden by 1 and overestimated Gillum by 1. They certainly *under*estimated the GOP candidates in both races by a lot more, but if you're going to attempt to discredit the poll, at least be precise in your claims instead of arguing that 1>5

You’re blatantly incorrect. Their final prediction had Gillum winning by 4-5 and Biden winning by 1-2.

So that means they consistently underestimate Republicans by 4-5 points, which would basically mean DeSantis is actually up by around 5 here.

You should google what the word "overestimated" means, as it applies to an individual number, and consider revising your post. Words have meanings. You could say "the poll was off by four," you could say "the poll overestimated the performance of this specific candidate by four" — those sentences do not have the same meaning.

Anyway, if you want to apply your novel method of "unskewing the polls," let's take a look at the only other independent polling we have of this race, Mason-Dixon back in February. Their last poll in 2020 was 48-45 Biden, actual result was 51-48, so using your crude method for calculating a house effect, theirs is D+6. Their last poll had DeSantis winning 52-41; using your methodology, that should actually be more like 55-38. So assuming arguendo that you're right, that the poll should really show DeSantis winning 47-43 — that's a massive collapse over the last few months!

[Donerail]

3) I don’t expect the margin of victory for DeSantis to be 15-20 points. Nor do I expect the race to be a tossup by any means. Is that reasonable to you?

No, it does not seem reasonable to think the gubernatorial election will be decided by more than, say, five points. That is our well-established normal result for elections in Florida, so you would have to produce substantial and reliable data to shift that prior. So far you've got, what, a couple internals and a Mason-Dixon from February?

[Donerail]

3) I don’t expect the margin of victory for DeSantis to be 15-20 points. Nor do I expect the race to be a tossup by any means. Is that reasonable to you?

No, it does not seem reasonable to think the gubernatorial election will be decided by more than, say, five points. That is our well-established normal result for elections in Florida, so you would have to produce substantial and reliable data to shift that prior. So far you've got, what, a couple internals and a Mason-Dixon from February?

Literally every poll except this one has DeSantis winning over 5- and even then, considering how St Pete’s consistently underestimates Republicans by an average of 5 every election, this means DeSantis is probably up by around 5 or so here.

So a couple R internals and a Mason-Dixon from February. Got it.

Anyhow you are rather dense, and clearly not worth talking to further, as my argumentative skills are more worthy to be used against more intelligent people. Please do not come up in my notifications again.

I'm in your notifs again. What are you gonna do about it? You gonna cry? Gonna piss your pants maybe?

[Donerail]

Bumping this thread to note that St. Pete Polls issued its final poll of the St. Petersburg mayoral election on Saturday.

Poll:
Welch 37%
Blackmon 27%
Rice 17%
Newton 6%
Boland 5%
Other 2%
(undecided 5%)

Final result, with ~90% of precincts reporting:
Welch 39% (+2)
Blackmon 28% (+1)
Rice 17% (±0)
Newton 7% (+1)
Boland 6% (+1)
Other/write-in 2% (±0)

For a poll of a low-turnout local election? That's pretty much dead-on.