No, dammit, that is NOT what you were told. Please. But at least you’re not misrepresenting quite as totally as RTF.
Yet the polls are the only data there is. As you’ve been told before, and not only by me, if you discard your only data, you can say whatever you like.
You do? I find it *mystifying *- that you could think anybody has said that.
The data has been there for you all along, and still is. If you don’t like E-V’s approach to compilation, and as you ought to know by know I don’t either, at least the *raw *data is posted there too. Would you like to discuss it, to repeat the only data there is, or stick with your previous nonsense?
Good to see you’re not terminally depressed. I’d say “welcome back to the land of the living,” except.
As Nate Silver has demonstrated for months over at Five Thirty Eight (which has been repeatedly pointed out to you), no they’re not. I hate to burst your little balloon here.
And even if we assume arguendo that they are, national as well as state polls are relevant data, and it goes without saying that polls besides the most recent poll in each state are also relevant data. Hell, the most recent one is far from guaranteed to even be the best single datum, but you seem perfectly content to use as your blunt instrument a tool that only takes into account that single datum from each state.
Claiming an evidence-based argument while deliberately ignoring piles of evidence is not my idea of a sound approach.
Projections are not data. Handwaving is not evidence. Denial is not rebuttal.
538, for instance, is a study in how much different reality would have for Obama to win. The fact that it uses numbers does not make that process objective. It is not “data”.
Such basic failure to understand the concept does help explain your basic misrepresentations of explanations to the contrary, though.
If you want to insist that they’re not, you’re on your own. (YOYO.)
You should abandon those debate techniques, then.
Electoral-vote.com’s site is vulnerable to a similar problem. It’s ‘objective,’ but that doesn’t make it meaningful.
I can sit here and flip a coin to decide which way each state is going to vote. It’s perfectly objective, but it would still be a crock of shit. Electoral-vote.com’s really not much better.
I’ve read the whole thread, and I don’t think anyone has pointed out that polling at this point is exactly as useless as “nothing” if we’re trying to use something as a data point to predict the election. Polling has not proven to be a more accurate way to predict elections than flipping a coin when the margins are this close.
I don’t think McCain is forked. I don’t know whether he is or not, and neither can anyone else. We have to wait a bit and see whether Obama gains momentum after the fallout from Hillary supporters settles down, and nobody can realistically predict what’s going to happen. McCain is not by any means a weak candidate, and would do well in an election against ANY democrat. I think to be convinced otherwise is foolish.
If Tuesday night’s speeches were any indication, McCain might just be forked.
It was just bad. Stilted, uninspired, and at times just really creepy. He used that weird green backdrop even though he was in New Orleans, a city with plenty of nice visuals despite it all. Of course, NO scenery would have brought up questions McCain doesn’t want to deal with right now, so he was pretty much screwed either way. Perhaps he should have considered another locale for such a high-profile speech.
Meanwhile, you had Obama: cool, confident, and inspiring, addressing a huge and enthusiastic throng, hitting all the right notes. This was the first night when the race was officially down to Obama and McCain, and the contrast couldn’t have been more stark.
Maybe it was an off night for McCain, but it’s not like it was an unusually “on” night for Obama. Maybe McCain’s “A” game compares more favorably to Obama, but can he bring it every time? The only smart thing for McCain to do would have been to stay home that night, and the fact that he didn’t doesn’t speak well for his campaign judgement.
It’s hard for the junkies among us to grasp that a lot of the American public just hasn’t seen much of McCain yet. If this is an indication of what they’re going to see, then yeah, he’s forked.
For the purposes of assessing the situation as it is, to find out what the candidates’ strengths and weaknesses are so that the strategies they can use can be defined, there’s nothing else available, though. The data do show that McCain has electoral support on the order of Obama’s. Up to Clinton’s withdrawal, they showed her in a much stronger position. From that snapshot, it is hardly unreasonable to “predict” that Obama has a tough job ahead of him, tougher by far than Clinton would have had.
Which is what I’ve been saying for some time, without getting through to a few folks.
Yep. That would be true even without data to show it. RTF, when you go find out what a Monte Carlo simulation is, let us know. The short version is that it is indeed a projection, NOT “data”. Data is what goes *into * it. The longer version is that it also depends on the inputted assumptions about how the data will behave. Those assumptions, also, are NOT “data”. They can, however, be adjusted to provide a desired answer, which is what the 538 guy is doing - he’s seeing what assumptions have to be made in order to “predict” an Obama win.
I realize how important it is to you to hold onto a bit of numerology that purports to give you the result you desire, but come on, enough of this shit is enough.
The data didn’t necessarily show that. That’s one interpretation of the data. Another interpretation is that Clinton had a harder time holding on to traditional Democrat states, and that Obama was more likely to challenge some traditionally Republican states. The data was never clear as to who was in a stronger position vs. McCain, because different people define strength differently. Clinton was extremely vulnerable in states that Obama won solidly, and more Red states became “purple” when comparing McCain to Obama than when comparing McCain to Hillary.
Your response to the previously polled close race with McCain leading by slim margins in a few states was:
…and…
What I just read in this thread in regards to Obama holding a few slim leads was:
Now, granted, you’re two different people and so I did not directly reference you in my original comment. I did notice that the same people saying “He’d have to depend on McCain self-destructing” aren’t responding to the latest polls with “McCain would have to depend on Obama self-destructing” to win.
RCP also shows Obama leading based on polling data without additional manipulation. they’re much more tie-happy than other sites I’ve seen but whether you eliminate toss-ups or not, Obama is in the lead.
Moi?
If you’re that resistant to the notion that what goes into a mathematical procedure is not the same thing as the procedure itself, then yes, there’s no point trying to explain it to you.
Measure, same thing I told him - a projection based on raw data is not the data itself. The assumptions that underlie the projection method have to be considered and understood, and even then the results are projections, not data.
Sure it was. Add the electoral votes, the only thing that ultimately matters, in the states each was leading and that’s what you had - Clinton ahead nearly 2/1 and Obama neck and neck. Speculate on what might happen to change the situation in the futre and you’re no longer “clear”, sure - but, again, speculation is not data.
A clearer summary of Poblano’s statistical projection methods, for those readers who do know the basics. Just look at all the “ifs” that go into his Monte Carlos. IF the black vote goes up 10 percent more, IF the youth vote goes up 20 percent, IF the Latino vote goes up 40 percent, IF those groups go with Obama by some other IF percentage …
Just a minor chime in on the difference between “data” and “projections”
The “data” such as it is are polls collected at various past dates and various demographic statistics. Data includes behavior in past elections and what the candidates state are their plans for competing.
Electoral-vote.com, fivethityeight.com, and various others, each make projections based on some of that data. The model could be a simple one such as just using the latest poll as your sole predictor, or one more complex based on past identified behaviors of various demographics, but both are projections.
Either both of those are “data” or neither are.
Both are however likely to be as reliable as me predicting the weather on November 4th based on today’s temperature. Or who will win the World Series this year based on today’s records, maybe. (Go Cubs! Subway series maybe?)
I was talking about use of demographic data like this.
The idea that, in the aggregate, people vote like other people who are like them, is hardly an earthshaking idea. But if that’s too much for you to handle, that’s not my problem.
