I realize it's not the same thing, but BITD half the 8 year olds on my block in DC owned "Daisy / Red Ryder" BB guns that were advertised on the back cover of comic books. They're not deadly or anything, but they can sure put some kid's eye out. They still make them, and I have no idea whether or not gun laws apply to them, but they sure didn't back then.

Exactly. Kings and tyrants have been known to be deposed.

BITD half the 8 year olds on my block in DC owned "Daisy / Red Ryder" BB guns ...they can sure put some kid's eye out

Of course, the other point is that it's still an Obama lead, and in line with his lead in most other polls of LV. Including the ones that you say are passing too many of the respondents.

Anyhow, I've heard the arguments for the free market plenty of times, I guess. I'm more curious out of morbid fascination than anything else to see if a $30 per gallon price helps anyone or anything in NJ. Not as convinced, obviously.

My heart's with Christie, but my head's with Stossel. It's hard to imagine how anyone would want to transport gas from faraway places to New Jersey if they thought they couldn't make a profit on it, especially considering that under normal circumstances New Jersey's gas prices are usually about the lowest anywhere in the East.

What could happen is a low turnout. Obama still carries both states easily, but his national popular vote total takes a hit, leading to Repub cries of illegitimate.

I wasn't suggesting what we know now about Benghazi should lead to impeachment. But it has been put forward that there was malfeasance and cover-up.


IF - IF - there was, then that is the sort of topic and behavior that I think should be the basis for impeachment. Malfeasance in the duties of president. Not ordinary citizen-like crimes.

I haven't followed the Nate Silver saga, but what added value does he claim to bring over just looking at the polls themselves?

I think it's more likely than not. The polls are likely missing the enthusiasm/turnout gap, many by a substantial margin. The fundamentals are just too unfavorable for Obama.

He should avoid precise percentages when talking about events that are too infrequent to estimates percentage odds reliably. It's like quantifying a teams chance of winning the World Series in percentages in April, it's stupid. Stick to toss-up, leans-O/R, likely O/R, very-likely O/R.

Correct. We can compare them, and assess a correlation. My only point is that with a working sample of 4 or 5 elections, maybe 20 if you bring in Gubernatorial and Senatorial races, the correlation just can't be that strong.

Have you no soul, that you would so flippantly derail a Monty Python hijack?

Spoken like someone on the vege of losing an election. The polls themselves are extremely specific, and routinely result in several decimal places.

But they're rarely reported that way. One of this week's polls that was reported as 49-48 Obama was actually 48.56 Obama to 48.49 Romney — a difference of one or two respondents.

Are you advocating for more detailed results (48.56-48.49) or less detailed results (Nate rounds up)?

No, my proposal is don't tell me the guy is 4 for his last 9, and base a prediction for today's game on it.

I don't think the detail adds any value beyond an illusion of precision.

Nate got burned three times on close Senate elections in 2010. It's interesting the amount of faith people have in Nate's ability to divine the outcome of the presidential election based on a similar batch of close polling numbers in Ohio and 1-2 other states.

But given the data we have, it seems that betting on polling error of 2-3 points or more in Romney's direction - which is effectively what calling the election "a toss up" is - that is not a bet I would make.

You seem to put a lot of faith in those three mistakes, without including the fact that he was correct on many other results.

Nate's track record in the closest races is what's truly important here, and that track record isn't as good as some people seem to believe.

What was his track record in states with close margins in 2008 and 2010?

In the 2010 Senate races, Nate was 2-for-5 in the five closest races. He missed on Buck, Angle, and Miller.

That's not what I asked, and it makes me suspicious that you might be cherry-picking the data points to make Nate look bad.

2010 wasn't a presidential election year, so there were no "states" for Nate to call.

If you can't acknowledge that after a day when Mr. Obama leads 19 out of 20 swing-state polls, then you should abandon the pretense that your goal is to inform rather than entertain the public.

I know. That's why I said "states with close margins in 2008 and 2010." 2008: Senate, House, Presidential. 2010: Senate and House. And not "the five closest" but "all close margins."

Nate's increasing defensiveness seems to betray some uncertainty.

Hasn't it been well established that no pollster/aggregator really projected the degree of GOP's 2010 blowout? And that Silver was by no means an outlier?

I took a look at the election results and it appears there were 8 close races and not 5.
Colorado
Alaska
Nevada
Illinois
Illinois
Pennsylvania
Washington
Wisconsin

Nate got 5 of those 8 right.

I have a hard time believing that you're actually so dense and so hyper-partisan that you can't/won't understand why 96% of poll respondents a few days before an election are labeled as likely voters. If it's not party ID it's LV screens or biased polling or whatever, anything to argue that Romney is actually leading the race. Never mind that the reason that Obama might be outperforming fundamentals is that Romney is a terrible candidate running on a terrible platform of more money for the rich, less help for the poor, more wars in the Middle East, interference in women's healthcare, and regression in social progress. If the fundamentals were neutral for Obama he would be crushing Romney because the right wouldn't be able to use scary economic doom-saying to hide how their platform is crap for the majority of Americans.

You know this though, you have to know. You're just ignoring it because you can use horribly faulty logic to twist it into another claim of poll bias and ignore the fact that your candidate is losing.

Not to worry - Michael Barone calls it 315- 223 for Romney!

Hasn't it been well established that no pollster/aggregator really projected the degree of GOP's 2010 House blowout? And that Silver was by no means an outlier?

To this excellent post I'll add only that I'm not quite as surprised that density or hyperpartisanship is blinding someone to plain math. After all, the inability to obtain and marshal facts in an objective manner is part of what leads someone, these days, to vote Republican.

You counted Illinois twice, while Washington and Wisconsin were 5-point margins.

Both of those were closer than Harry Reid's 5.7 margin of victory over Sharron Angle.

On a related note, what would have been the reaction if, say, Elizabeth Warren threatened to start killing politicians if she wasn't elected the way Angle did?

I took a look at the election results and it appears there were 8 close races and not 5.
Colorado
Alaska
Nevada
Illinois
Illinois
Pennsylvania
Washington
Wisconsin

Nate got 5 of those 8 right.

Was Illinois so close it counts twice?

I would think the MoE of for midterm senatorial elections is greater than for Presidential elections.

Nate's model for the 2010 senate probably had a 95% chance of happening as predicted. In the end he got 92% of his picks right. If Nevada had been polled better he would have gotten 95% of his picks right but that is simply a what if not a justification that he was still right. 92% is pretty darn good.

I know little about statistics or polling, but I'm going to say because if actual turnout is around 72%, the "likely voter" figure is 96% of that amount. I'll assume I've screwed it up, so someone just let me know how.

Giving Nate credit for being right on 30 Senate races that any political neophyte could have predicted is little more than stats-padding. All that really matters is Nate's track record in the closest elections.

Did Dick Morris get them all right?

If that's the case, shouldn't Nate's model account for that? He shouldn't have the same probability of a candidate winning with a +2 in a thinly polled Senate race that he would a presidential candidate with a persistant +2 across a large number of polls the week before election.

The tracking polls seem to show the presidential race tightening. Rasmussen numbers released this morning show Barack Obama ahead of John McCain by only 50 percent to 47 percent—the narrowest margin in Rasmussen polls for more than a month and the first time McCain has been over 46 percent since September 24 (nine days after the collapse of Lehman Brothers and two days before the first debate). Gallup's tracking released yesterday showed McCain behind by only 49 percent to 47 percent on its traditional-turnout model but behind by a much larger 51 percent to 44 percent on its expanded model.

John Zogby's numbers yesterday showed a 49 percent-to-44 percent margin for Obama, considerably smaller than the one he had a few days earlier.

Against these numbers one must weigh polls in battleground states showing Obama doing just about as well as he has been since the financial crisis hit in mid-September. And the respected Pew Research poll shows Obama out in front by a whopping 53 percent to 39 percent. I'm reluctant to dismiss Pew as an outlier, but if you do, the realclearpolitics.com average of recent polls looks like Obama 49.6 percent, McCain 44.8 percent, a margin of 4.8 percentage points—the smallest Obama margin in RCP averages since September 30.

This sounds plausible to me. Hypothesis: The McCain "Joe the Plumber" high-taxes issue is having some impact. So are doubts, kindled by a typically verbose Joe Biden comment, about how Obama would handle an international crisis. Is this the beginning of a McCain surge? Possibly. It's beginning to look like something more than statistical noise. But it's hard to be sure at this point.

Is this the beginning of a McCain surge? Possibly. It's beginning to look like something more than statistical noise.

Michael Barone, October 29, 2008, six days before the election...

This whole thing is nothing but troll stupidity. Joe is either pretending he doen't understand how polls work or is genuinely stupid when it comes to polls. Joe thinks that every single person that gets contacted is part of the poll. He doesn't understand that the only people that get counted in the poll are people who actually take the poll and that at this point in the game the only people taking part in polls are people very much likely to vote.

If a polling agency calls me and I don't pick up the phone my not picking up the phone isn't counted as part of the polling group. If I pick up the phone and they ask me to take part in a poll and I say no, again I am not counted as part of the poll. Joe does not understand that or is pretending to not understand that. You go back and look at all the polls that are done by phone and you'll see that almost all if not all of the polls have a higher LV% than actually what happens and that as we get closer to the election the LV% increases. Not taking the poll is part of the screening process. Is it perfect? Of course not but not for the reasons that Joe is trying to use.

McCoy, I know Biden could, but was curious as to whether the Senate had only the choice between Biden and Ryan. I can't find specific info on that.

Why aren't pundits proven consistently wrong shamed out of the business? Hell, in Italy, they'd be imprisoned.

The only issue I have with this is that the criteria for selection should be a race that was predicted to be within 4 points and not the other way around.

(smile)

I added it later but if the House deadlocks Biden would become the President.

The sitting VP does not get to break a tie in the case of picking the VP. You need a majority of Senators to vote to pick a VP.

Dick Morris is a pundit/storyteller. He's not giving odds to four decimal places.

If a faithless elector casts a ballot for someone other than Obama or Romney, the House can choose that person

It's funny how things work around here. If a poll shows a D+8 advantage in 2012, an even higher Dem advantage than 2008, we're told that the poll result is representative of the electorate, and anyone who questions it is a "poll truther." But now, we're told that a poll that shows 96 percent of RV to be LV is not representative but it's somehow still accurate. Or something.

I think I sort of half agree with this criticism. There's not really all that much to the model beyond the basic concept of aggregating all the available polling data. If the data is crap, the model can't fix it. So if Nate is right, it mostly means that the polls were right, and not all that much more than that.

but it's hard to name many people who might know more about U.S. politics — or have been right more often — than Michael Barone.

I know little about statistics or polling, but I'm going to say because if actual turnout is around 72%, the "likely voter" figure is 96% of that amount. I'll assume I've screwed it up, so someone just let me know how.

Either the polls are systemically and widely incorrect, or Romney is losing. Snapper is showing a decent way to argue as if you just believe the polls are all wrong.

A close senate race is by definition a 50-50 proposition. So Nates expectation should be to miss 2-3/5 or 3-4/7.
If he got 7/7 he would have just have been lucky.

Of course the OPPOSITE is true for the Ec. It's not a Coin flip because Romney needs several states, not just Virginia

Oh and McCoy explains the "lv" gap quite well.

I explained this to Joe (and McCoy) several pages ago, but he's still repeating this nonsense. A polling firm found (in one poll) that 96% of people who say they are registered to vote are likely to vote. While it's true that only ~72% of people who are actually registered to vote actually vote, the correct comparison is to the percent of people who say they are registered to vote who actually vote.

Since ~90% of people who say they are registered to vote actually voted in 2008, the pollster's likely voter screen appears to be reasonable. The likely explanation is that some people don't realize they're registered to vote, and the people who do realize they're registered are much more likely to actually vote than those who are registered but don't realize it.

A note on RCP. They have a mostly arbitrary, possibly somewhat partisan poll screening method that gives them a rightward lean in their averages compared to Pollster, TPM, and 538. They don't include about 25% of publicly released polls, and their selection methods appear entirely subjective. I think that by making use of every existing poll not released by a campaign, the other poll averages are better options. If there were an RCP that handled all polls equally, I'd be interested to see how it compared to the more complex aggregation models at 538 and Pollster.

Either the polls are systemically and widely incorrect, or Romney is losing. Snapper is showing a decent way to argue as if you just believe the polls are all wrong.

Yes, only likely voters answer the phone for 90 to 180 days before elections. We can call this the Watson Rule.

I did know RCP was right-leaning. Which to me is a feature, not a bug, since Nate is obviously left-leaning.

In the Gallup poll on which you're relying, over 40,000,000 registered voters — 40,000,000 voters — went missing.

Beyond that, explaining away these 40,000,000 registered voters by surmising they're "registered but don't realize it" is a huge leap that's unsupported by any evidence. Are you seriously claiming that almost a quarter of the American electorate is unaware they're registered to vote?

He thinks the following NEVER EVER happens:

**ring**ring**

Citizen: Hello?
Pollster: Hi, we're conducting a poll for....
Citizen: *click*

Personally, I think that probably happens quite a bit, and furthermore, the person more likely actually take the poll is the person who is probably actually interested in voting, what someone may call a "likely voter".

Wait, this is his reason for thinking LV screens are wrong? Holy christ, that is either sheer stupidity or just big time trolling.

Personally, I think that probably happens quite a bit, and furthermore, the person more likely actually take the poll is the person who is probably actually interested in voting, what someone may call a "likely voter".

By the way, I just noticed that one of the tags for this thread is "mr president".