Heretic;1308995 wrote:My insider poll is working on the most important questions pertaining to the election. Namely that depending on who wins, will either Gibby or Quaker off themselves in disappointment/disgust? And if so, how much will that improve the average IQ of American citizens?

If only a third party could win and they both killed themselves. Pretty sure our average IQ would at least triple.

Ok, so we're clear...most of the polls are garbage either because they're so hopelessly biased and/or corrupt working for their client the democratic party or hopeless incompetent and therefore makes the aggregation of those polls a worthless pursuit.

In the apparently unlikely scenario where Obama wins the states he's currently projected to win in say the RealClearPolitics No-Toss Up...do we still adhere to the belief that the aggregation of these garbage polls is worthless?

BoatShoes;1309172 wrote: In the apparently unlikely scenario where Obama wins the states he's currently projected to win in say the RealClearPolitics No-Toss Up...do we still adhere to the belief that the aggregation of these garbage polls is worthless?

Aggregation does nothing to remove uniform bias. If most or all oversample, an aggregation doesn't change that. That's the essence of garbage-in, garbage-out.

I think most of the polls are terrible because they're based on landline calls. You're not going to get a good representation of the country with a landline based demographic. There's other problems sure but that's one of them.

I Wear Pants;1309184 wrote:I think most of the polls are terrible because they're based on landline calls. You're not going to get a good representation of the country with a landline based demographic. There's other problems sure but that's one of them.

Haha, good point. Didn't even think of that one.

But I think most of the polls (even Rasmussen and Gallup, until very recently) have calibrated to the 2008 turnout because it's defensible. They'll say "look, our job is to poll people in an unbiased manner...we are not in the business of projecting turnout". From that perspective, their data is good. It's not particularly hard to take likely voter results and overlay that on a turnout model, run different scenarios. And that's plenty useful for campaign strategy because you can see where you need to move the needle and where you just have to rally turnout/base.

So give me a poll that lets me waive to the media "look! my candidate is winning!". Internally I can dissect it to see what the real implications are.

I Wear Pants;1309136 wrote:If only a third party could win and they both killed themselves. Pretty sure our average IQ would at least triple.

Screw my prior post; I hereby will my full share, $34,000, of new obama debt to this hapless little progressive shill, IWP. No one deserves more obama debt than him, or her, or it.

gut;1309177 wrote:Aggregation does nothing to remove uniform bias. If most or all oversample, an aggregation doesn't change that. That's the essence of garbage-in, garbage-out.

I understand what you're saying...you're confident in a Romney victory because most polls are hopelessly biased/garbage and that polling aggregation is a worthless pursuit and therefore the projection from that data of an Obama win can't possibly be right.

BoatShoes;1309194 wrote:I understand what you're saying...you're confident in a Romney victory because most polls are hopelessly biased/garbage and that polling aggregation is a worthless pursuit and therefore the projection from that data of an Obama win can't possibly be right.

I'm confident in a Romney victory because I don't believe the turnout is anywhere near the 2008 numbers, and many polls are actually trending that way. Early voting results are also confirming that. 2010 also gives indication of that. I think the polls are garbage because when you look at breakdowns and cross-sections it really doesn't jive - internally inconsistent = garbage.

You obviously don't understand statistics. Polls are easily rigged to produce a result. And they are either being rigged to suit an agenda, or they are being unintentionally using a poor baseline because it's defensible, as opposed to putting a reputation on the line pegging a number with a huge margin of error.

QuakerOats;1309193 wrote:Screw my prior post; I hereby will my full share, $34,000, of new obama debt to this hapless little progressive shill, IWP. No one deserves more obama debt than him, or her, or it.

QQ some more why don't you.

gut;1309199 wrote: You obviously don't understand statistics. Polls are easily rigged to produce a result. And they are either being rigged to suit an agenda, or they are being unintentionally using a poor baseline because it's defensible, as opposed to putting a reputation on the line pegging a number with a huge margin of error.

:rolleyes: Ok so you're betting on a Romney victory because of a widespread agenda in the polling industry in favor of Obama. Fair enough. What do you think it would mean in the apparently unlikely event that Obama wins in line with the projections of other people who do, by all accounts, understand statistics?

BoatShoes;1309240 wrote::rolleyes: Ok so you're betting on a Romney victory because of a widespread agenda in the polling industry in favor of Obama. Fair enough. What do you think it would mean in the apparently unlikely event that Obama wins in line with the projections of other people who do, by all accounts, understand statistics?

You are reaching gibby and isadore levels.

BoatShoes;1309240 wrote::rolleyes: Ok so you're betting on a Romney victory because of a widespread agenda in the polling industry in favor of Obama. Fair enough. What do you think it would mean in the apparently unlikely event that Obama wins in line with the projections of other people who do, by all accounts, understand statistics?

The turnout is already dropping by the day, which is why polls are all narrowing.

But like I said, the majority of these polls produce a result for their client.

It's really quite simple. When you see Obama's favorability and % declining virtually across the board, across almost all demographics and issues....The only way your inputs go one direction and the results go the other is by oversampling, and in some cases it was even the other direction (BETTER turnout for Obama in 2012 vs. 2008). The polls are getting better, but still plenty hokey about many of them. I don't think there's a reasonable or rational explanation for any poll showing Obama doing as well as he did in 2008, much less better - we can comfortably call such a poll garbage, especially when we see oversampling.

I don't know what the agenda or purpose is, but normally when you have inconsistent results, and inputs going one direction that usually dictates adjusting your baseline. Like I said, it's easier to defend a bad forecast because you used 2008 then it is to defend a bad baseline. They are hedging their bets - if for some miraculous reason we see 2008 turnout, then they will be right. If not, they'll say "well no one predicted those turnout numbers, and with the correct actual baseline we were spot on".

gut;1309252 wrote:The turnout is already dropping by the day, which is why polls are all narrowing.

But like I said, the majority of these polls produce a result for their client.

It's really quite simple. When you see Obama's favorability and % declining virtually across the board, across almost all demographics and issues....The only way your inputs go one direction and the results go the other is by oversampling, and in some cases it was even the other direction (BETTER turnout for Obama in 2012 vs. 2008). The polls are getting better, but still plenty hokey about many of them. I don't think there's a reasonable or rational explanation for any poll showing Obama doing as well as he did in 2008, much less better - we can comfortably call such a poll garbage, especially when we see oversampling.

I don't know what the agenda or purpose is, but normally when you have inconsistent results, and inputs going one direction that usually dictates adjusting your baseline. Like I said, it's easier to defend a bad forecast because you used 2008 then it is to defend a bad baseline. They are hedging their bets - if for some miraculous reason we see 2008 turnout, then they will be right. If not, they'll say "well no one predicted those turnout numbers, and with the correct actual baseline we were spot on".

Which polls are showing this?

I Wear Pants;1309184 wrote:I think most of the polls are terrible because they're based on landline calls. You're not going to get a good representation of the country with a landline based demographic. There's other problems sure but that's one of them.

And then you have people like me voting for Obama on those phone calls because I'm voting for Johnson; and a vote for Johnson is a vote for Obama.

I Wear Pants;1309271 wrote:Which polls are showing this?

Ty has posted several. Go peruse and then go look at the 2008 results. It's been pointed out a few times. Just one example - Obama won OH by 4.5pts in 2008, so go find some polls showing Obama +5, +6 and more. Shouldn't be hard, at least not a few weeks ago.

Or put on your thinking cap and see Obama won MI by 17(?) in 2008. Some are saying it's now a toss-up, but even the most generous polls have it +8 or +10 Obama (again, a few weeks ago). Now does it make the least bit of sense to you that MI could move that much and OH doesn't budge, or actually goes up?

Like I said, you look at subs of the polls and Obama has declined across the board, and then you see a result that's better than 2008. There's only one way to get such a result (well 2, I suppose, if you count incompetence or fraud).

Relevant to turnout:

sleeper;1309278 wrote:And then you have people like me voting for Obama on those phone calls because I'm voting for Johnson; and a vote for Johnson is a vote for Obama.

I actually saw a Gary Johnson sign today which made me happy.

I Wear Pants;1309296 wrote:I actually saw a Gary Johnson sign today which made me happy.

You drive by my house spying on me again?

justincredible;1309299 wrote:You drive by my house spying on me again?

Do you live on the on-ramp of 534 and 76?

I Wear Pants;1309305 wrote:Do you live on the on-ramp of 534 and 76?

Maybe. Maybe not.

justincredible;1309348 wrote:Maybe. Maybe not.

Well I've been watching whoever lives there sleep for the past three months. I hope I haven't wasted my time.

se-alum;1309246 wrote:You are reaching gibby and isadore levels.

We can't all offer the grounded, reasoned insight of se-alum I suppose

gut;1309252 wrote:The turnout is already dropping by the day, which is why polls are all narrowing.

But like I said, the majority of these polls produce a result for their client.

It's really quite simple. When you see Obama's favorability and % declining virtually across the board, across almost all demographics and issues....The only way your inputs go one direction and the results go the other is by oversampling, and in some cases it was even the other direction (BETTER turnout for Obama in 2012 vs. 2008). The polls are getting better, but still plenty hokey about many of them. I don't think there's a reasonable or rational explanation for any poll showing Obama doing as well as he did in 2008, much less better - we can comfortably call such a poll garbage, especially when we see oversampling.

I don't know what the agenda or purpose is, but normally when you have inconsistent results, and inputs going one direction that usually dictates adjusting your baseline. Like I said, it's easier to defend a bad forecast because you used 2008 then it is to defend a bad baseline. They are hedging their bets - if for some miraculous reason we see 2008 turnout, then they will be right. If not, they'll say "well no one predicted those turnout numbers, and with the correct actual baseline we were spot on".

So this is what this Nate Silver guy posts on Twitter: @fivethirtyeight: 7 polls released in Ohio in past 48 hours: Obama +2, Obama +3, Obama +3, Obama +3, Obama +5, Obama +5, Obama +5. #notthatcomplicated

Rasmussen, on Nov. 2 of 2008, had Ohio a tie at 49-49 between McCain and Romney. the RCP average was Obama +2.5 averaging similar "garbage polls." The final result was 51.5 for Obama and 46.9 for McCain. Obama is +2.3 in the RCP average this year with "garbage polls". I could be wrong but I think it exhibits a pretty transparent bias to just discount polls you think are no good rather than just consider the average. You weren't on the site that shall not be named from what I remember but I feel like there were people parsing the polls in the same way you are here....suggesting that there's widespread polling failure overestimating democrat turnout.

I think you're making assumptions that aren't substantially supported but what do I know?

I Wear Pants;1309350 wrote:Well I've been watching whoever lives there sleep for the past three months. I hope I haven't wasted my time.

BoatShoes;1309369 wrote:So this is what this Nate Silver guy posts on Twitter: @fivethirtyeight: 7 polls released in Ohio in past 48 hours: Obama +2, Obama +3, Obama +3, Obama +3, Obama +5, Obama +5, Obama +5. #notthatcomplicated

Rasmussen, on Nov. 2 of 2008, had Ohio a tie at 49-49 between McCain and Romney. the RCP average was Obama +2.5 averaging similar "garbage polls." The final result was 51.5 for Obama and 46.9 for McCain. Obama is +2.3 in the RCP average this year with "garbage polls". I could be wrong but I think it exhibits a pretty transparent bias to just discount polls you think are no good rather than just consider the average. You weren't on the site that shall not be named from what I remember but I feel like there were people parsing the polls in the same way you are here....suggesting that there's widespread polling failure overestimating democrat turnout.

I think you're making assumptions that aren't substantially supported but what do I know?

Let me ask you a simple question. Obama has declined 8-10pts or more in other swing states from 2008. Why would OH be flat, or even UP?!? Why is OH different? Past results don't guarantee future success. Just because averaging polls gave good results once (and Nate has done this for a grand total of one presidential election, so hardly the track record of a Rasmussen or even Gallup) doesn't mean a whole lot. You can't average garbage into a good number, you cannot.

Of course OH is not different or special. Michigan and Indiana - right next door, they're affected by the UAW payoff, too. Very similar, yet the OH numbers don't move from 2008? Do you really need Nate Silver to spell that out for you, or can you think for yourself?

I've been saying all along these polls don't make sense. There's actually a lot of support for it - much of it I've detailed, and which you can readily confirm by thinking and researching for yourself (but that requires you to seek information from places other than the left).