Gaff or insidious plan to lull democratic voters into a false sense of security? Probably neither. But you decide!
1 Like
I’ve thought for a while now that if Trump got rid of the fake tan, let himself be honestly bald, and wore a well-fitted suit, he’d look much better, more dignified and more presidential.
8 Likes
It wouldn’t help. As soon as he starts to talk, you would know.
I’ve heard lots of leaders talk, all of them have this “thing” about them. Where you feel confident in what they’re saying and they project a level of confidence.
Trump doesn’t have that, not even close. His style feels like a locker room looking for acceptance of the joke just told, not like a leader inspiring people to move in the direction he’s pointing.
12 Likes
But that 80% assignment was arbitrary. If you look just at what the polls were saying the odds should have been a 99.9% chance for Clinton to win. The only way you downgrade that is by adding other factors, like how reliable the polls are, and other factors that have nothing to do with the polls. So even to get to 80% you have to say that “maybe the polls are wrong” and “let’s add factors that dilute the poll results because maybe they’re wrong”. But the problem was that the polls were wrong.
Sam Wang and Nate Silver had a long back and forth about how much the polls should be taken at face value and what the probablility was that the polls were wrong. Once you establish which candidate the polls are saying should win, then the rest is just the level of certainty. And you can peg that just about anywhere you want, based on the probability you assign to there being polling error.
So yes, in the end, the polls were wrong.
1 Like
No they weren’t.
Many pundits were treating Clinton’s projected victory as a sure thing, but the pollsters and statisticians were just treating it as the most likely statistical outcome.
15 Likes
Right, but 538’s aggregate model assumed a higher chance of the polls being wrong. If you assumed that the polls were right, there’s no way the odds would be that low.
Polls are strictly of entertainment value at this point, so we should be looking for ways to enhance that.
What would be FUNNY is if all the other pollsters now came out with Trump-positive polls, leaving Fox hanging in the wind as the only “Trump-haters”.
This could be strung out for weeks, with MSNBC doing favourable stories on the “Trump Magic” and wondering aloud why “Fox is trying to sabotage the Trump Avalanche”, begging him for interviews to discuss the “Fox Betrayal”.
Then flip it around a few weeks later and accuse Trump of manipulating all the other polls, and humbly congratulating Fox on getting it right in the first place. Accuse him of stuff he HASN’T done! That’ll be a novel experience for him, though his denials not-being-believed will be familiar.
It’s time we all got together to gaslight Trump, not for political gain, just to enjoy the reality show that results. It’s fair play in his case.
13 Likes
Every time I see poll results now, I imagine a giant salt boulder rolling down a hill made of pundits. Polls, like Dump himself, are stupid and useless, except to drive traffic to your bullshit of choice.
1 Like
The chart you are displaying is not what the polls were saying. It is 538’s algorithms based on polls, PLUS chance of polling error and other non-poll factors.
So don’t say it reflects what the polls were saying, it does not.
After the debates, it kinda looked like a sure thing up until Comey’s “October Surprise”.
1 Like
IMHO, polling “error” is exactly what resulted in this mess. Or more accurately, targeted voter roll manipulation (which has never been denied by the administration, they only claim that no cast ballots were modified).
1 Like
538’s aggregate model correctly noted that many results in important states were within the polls’ margin of error given the polling sample size and the quality of the polling, which is not the same thing as the polls being wrong.
7 Likes
I have previously expressed my frustration with statistics being “maniplulable” (not a word but should be!) but the truth is the polls were not really wrong, HRC won the popular vote by about as much as was projected. The distribution of a small number of votes in a few states which (IIRC) were not all that well polled as they were thought to be solid blue was what decided the election, combined with the nuttiness that is the electoral college. I don’t feel we can blame the Nate Silvers of the world for the shitshow we are currently enjoying.
7 Likes
And often far outside the polls margin of error. That’s where they differed from Sam Wang. When you give weight to the chance of polls being outside their margin of error, you are saying the polls could be wrong.
For the nation as a whole the polls were pretty accurate, maybe a few points away but a 1-2 point difference is normal. And I don’t blame Nate Silver, he tried to push back against what the polls were saying, really, by softening the impact of strictly going by the state polls and their margin of error.
But I think the polls were flawed, for many reasons. I think cell phones have really screwed up the way polls have worked for decades. Also polling firms use different values for “likely voters” which may no longer be accurate in the internet age. And they of course don’t account for a late shift in sentiment, if the polling period is over 7 days and something happens in the last few days.
I think all of these things, plus the fact that Trump was a bizarre candidate that the polls never had to deal with, meant that polls in key states were wrong, and had they been correct, it would have been easy to see that the election on an Electoral College level was essentially up for grabs, even if Trump’s chance to win the popular vote was close to zero.
3 Likes
Tell you what, why don’t you show us these polls that supposedly showed Clinton having a 99.9% chance of victory. Because I maintain that no serious polling agency ever made such a claim.
7 Likes
It is important that we all get out and vote. You can rest assured that the nutcase Trump fans will all be out in support of their orange messiah. Trump worshipers may be a minority but they are a dedicated minority.
2 Likes
This is SOP. They report low poll numbers to fire up the base
2 Likes
Not only “not the same thing”, but if that’s the case, it’s the opposite. (results being within a margin of error would be ‘right’ not ‘wrong’).
However, here’s my tedious quibble: the margin of error is calculated for a truly random sample with this simple relation: 1/sqrt(N) for a sample size of N; so watch the news and there will be endless examples (in tiny writing at the bottom): 3%MOE. This means they’re declaring a (truly random) sample of 1000 respondents. But you may be nearly certain of one damning fact: the sample isn’t remotely random.
A truly random sample would be one where every voter had an equal chance of being part of that sample; and that’s utterly not the case. They probably have to phone up ~10,000 folks before they get ~1000 who will answer an un-identified caller(ID) and then bother to take the poll, (also note that the size of refuse to answer pile are never factored-in to the resulting MOE). And (i assert) that skewed sample skews significantly to an older less technically inclined sub population. So… that whole 1/sqrt(N) calculation is bogus. That is, next time you see a news poll with a 7% lead between politician A and B, mumble to yourself: “yeeeah plus or minus 10% on each percentage”, or no significant difference. …oh, and just to be age-ist: older folks would tend to prefer Biden.
2 Likes
7 Likes
