Sep 18th, 2008 at 7:33 am
Register
RSSLogin
Advertisement
Visit Our Affiliated Sites
Books By Matthew Yglesias
Topic CloudView topics sorted by popularity
Afghanistan
Basketball
Budget
China
climate
Congress
DC
Denmark
Economics
Economy
education
Energy
Environment
Finance
Gender
Germany
Health Care
History
International Law
International Relations
Iran
Israel
Media
Monetary Policy
National Security
NBA
Pakistan
Palestine
Philosophy
planning
Political Reform
Proliferation
Public Health
Public Opinion
Race
Regulation
Russia
Science
Sports
Stimulus
Sweden
taxes
Technology
Trade
transportation
Featured
Subscribe to the Progress Report
Contact Matthew Yglesias
Use this form to contact blog author Matthew Yglesias.
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
September 18th, 2008 at 7:39 am
Congrats. You’re only 48 hours behind in figuring this out, Matthew.
September 18th, 2008 at 7:46 am
Petey, man, what’s with the attitude?
September 18th, 2008 at 7:51 am
Petey, why don’t you get together with Al, O’Neal, and Mixner and have your own blog? Seriously.
September 18th, 2008 at 8:02 am
“Petey, man, what’s with the attitude?”
I’ve got no truck with the groupthink.
Obama as President is going to be an unprecedented clusterfuck for the left. I’ll vote for him, but I’ve got no desire to be “helpful”.
The truth seems a refuge in the storm.
September 18th, 2008 at 8:09 am
I heart graphs with no titles. WTF?
September 18th, 2008 at 8:11 am
This is misleading junk. From the same NY Times article Matt quotes:
” The contest appeared to be roughly where it was before the two conventions and before the vice-presidential selections: Mr. Obama had the support of 48 percent of registered voters, compared with 43 percent for Mr. McCain, a difference within the poll’s margin of sampling error, and statistically unchanged from the tally in the last New York Times/CBS News poll, in mid-August.”
In short, Obama is not up five, but statistically tied with McCain. I remind you: hope is not a plan and optimism is just as delusional as pessimism. The reality is that Obama/Biden are tied in a race to the White House with the most ideologically extreme and most experience-challenged opponents ever.
That should be cause for serious alarm.
September 18th, 2008 at 8:22 am
The truth seems a refuge in the storm.
yeah yeah. and only you know it. and everybody else is stupid and deluded. blah blah blah.
really, you need your own blog.
September 18th, 2008 at 8:26 am
1) The American People are going apeshit over whether their life savings are going down the tubes — and Obama can’t pull ahead of a 26 year Republican Senator. Pathetic.
2) That’s very bad news. Even if he wins, he will be a disaster as a President, giving the Jimmy Carter bullshit we’re seeing from him now. 3 Days after he enters office, Fox News and the Republican propaganda machine will start screaming about the Obama Depression and they won’t let up.
All signs now are that Obama will humpered up and whimper.
3) There is one thing WORSE than a Republican. It’s a Democratic leader who cracks under pressure, flees the battlefield, and leaves his grassroots to be fucked. I KNEW the Clintons were like that — and several Democratic leaders of Congress as well. I had hoped Obama would be different.
4) He’s starting to look like fatally flawed Condi Rice. Both weakened by the experience of being Black.
Always hoping that if they act White enough –and behave properly –, the elites will give them an A on their report cards and accept them. Too timid to ever express resentment or leadership.
5) I guess all the Black men of courage are dead or in prison. Starting to look like only the bitches are left.
September 18th, 2008 at 8:26 am
It would be cause for serious alarm – but Democrats are seriously cognitive dissonance challenged.
Meanwhile, the “October Surprise” is coming – whether it be bin Laden on a stick, an Iran attack, or a Pakistan attack, or some other crap is not certain yet (I’m still voting for Iran, but anything is possible). McCain will get his five points and Obama will go home loser in November. And the Democrats will wander around for the next four years saying things like “Obama should have gone more negative” or – worse – “We should have picked Clinton” (and then in 2012 they’ll pick Clinton and lose again.)
And Matt and the rest here continue to pick on that stupid bitch from Alaska as the easiest target while we’re heading for a disaster at Warp Nine.
September 18th, 2008 at 8:34 am
tristero,
You and many others don’t understand statistics. If a poll shows that candidate X is 5% ahead, the BEST ESTIMATE of the difference in the race, according to that poll, is a 5% difference. This may not be statistically significantly different from a population difference of 0% (i.e., a population difference of 0% or less). But it remains that a 5% difference in the population is the best guess, far better than the 0% (or less) population difference guess.
The NYT’s quote is also partially misleading. Had the new poll been outside of the margin of error from the last poll, there would be 95% confidence that the poll numbers were changing in Obama’s direction. The change doesn’t support this. But it certainly doesn’t support the apparent remaining hypothesis that there has been no change. The likelihood of “no change” is close to zero. There must have been some change. But, according to the poll, you can’t say, with greater than 95% confidence, that there has been a change in Obama’s direction. Statistical assertions of “no change” are virtually impossible.
September 18th, 2008 at 8:35 am
Quinnipiac shows Obama +4. It does seem as though both convention bounces (or Palin bounce in lieu of RNC bounce, supposedly) might completely recede and we may be right back to where we were before late August.
September 18th, 2008 at 8:36 am
Quinnipiac has a +/- of 3.1%, btw.
September 18th, 2008 at 8:41 am
Petey’s problem with Obama is the same as Lynn Forester de Rothschild’s problem and it starts with the letter n.
September 18th, 2008 at 8:43 am
Re cube’s comment “tristero,
You and many others don’t understand statistics”
—————
And you don’t seem to understand that white noise is not heavy metal rock — although they sound the same.
Different polls give different results because the pollsters are GUESSING are who are the LIKELY VOTERS — because those are the only ones that count.
Lots of people thought Bush was a flaming shithead in 2000 — the number was actually rising — but those fuckers were too lazy to turn off the TV, lay down the hash pipe, and go vote.
The Republicans have their flaws but they have always been able to turn out their goosestepping morons for the Nuremberg Rallies.
September 18th, 2008 at 8:45 am
If McCain gets elected this will prove to the rest of the world that America is full of hypocritical, deeply racist retards.
It will be fun to watch how sensationally fucked up things can get under more Republican rule (this crisis is just an apetizer, believe me).
September 18th, 2008 at 8:53 am
Utterly goddamned meaningless information without labels. Impossible to interpret.
You know better than this, man.
September 18th, 2008 at 8:54 am
Re the “Likely Voter” problem , see
http://www.pollster.com/blogs/Erikson,%20Panagopoulos%20and%20Wlezien.pdf
September 18th, 2008 at 9:07 am
“Lots of people thought Bush was a flaming shithead in 2000 — the number was actually rising — but those fuckers were too lazy to turn off the TV, lay down the hash pipe, and go vote.”
Uh, Don. Bush *lost* the popular vote in 2000. So enough of “those fuckers” did go vote. They just weren’t living in Florida.
September 18th, 2008 at 9:16 am
too lazy to turn off the TV, lay down the hash pipe, and go vote.
And what is the one thing Obama is doing different this year?
September 18th, 2008 at 9:18 am
Re “They just weren’t living in Florida”
————
You don’t think there’s lots of bong-sucking TV watchers in Florida? Place is a monument to the breed.
Excuses are easy –losing debates to a fucking bonehead newcomer like George W Bush (who thinks math is “fuzzy”) takes real genius.
Although its possible there was a last minute tsunami of support flowing from bong-sucking TV watchers to George W — due to a belated recognition of a fellow slacker and cocaine afficianado.
September 18th, 2008 at 9:48 am
The poll confirms what has been obvious since yesterday. McCain got a convention bounce but has not retained the momentum.
Bush got enough momentum from his last convention to gain the election by a narrow margin. Kerry actually beat him in all the debates and got a considerable bounce-back but it was not enough.
Seems out GOP friends are depending on an “Act of God” in October to win the election. But maybe God wants to vote Democrat, too. We’ll see.
September 18th, 2008 at 10:24 am
In addition to people misunderstanding the implications of the margin of error (I personally think we should outlaw the phrase “statistical tie” and its variants), it should also be noted that with lots of polls available, the effective sample size for the aggregation of those polls is dramatically increased. So taking all the recent polling together, it is extremely well-confirmed at this point that Obama has retaken a small lead in the polling.
Of course it remains true that polls are merely a snapshot in time, and things could still change by the election. And it is also true that due to likely voter modeling issues, the aggregate of the polls could be systematically off. On the other hand, both of those factors count both ways.
September 18th, 2008 at 10:45 am
cube,
Of course, there’s been a change, but it’s not significant. “Obama up five” is thoroughly misleading. The candidates are, in any meaningful sense, tied. Of course you or I, if we buy a lottery ticket, have a very real chance to win the Big Prize. But it’s not significant.
I fully agree that I’m not an expert on statistics. My point is: neither is Matt and his chart and headline is grossly misleading.
September 18th, 2008 at 11:15 am
Petey’s problem with Obama is the same as Lynn Forester de Rothschild’s problem and it starts with the letter n.
I think Petey is as an pretentious ass as much as the next guy, but I think this is pretty clearly not true. I think he’s just obsessed with Obama’s perceived policy failings on healthcare.
September 18th, 2008 at 11:25 am
Obama’s starting to pull away. It looks like everyone caught up in convention bounces will have to rethink their outlook. Looking at the polling today, Obama could be up 3-5 points nationally by the weekend. In virtually every poll that he was down 5 a week ago, he’s up 3-4 points now, and that’s with cyclying in some of last weeks data.
September 18th, 2008 at 11:28 am
CNN’s “poll of polls” today confirms that “Obama again holds the slim but clear advantage he held in August before both parties’ conventions — one month ago, Obama held the same 2 point margin over McCain”: http://politicalticker.blogs.cnn.com/ . (10 days ago, MCain had the 2-point edge.) The six polls in the Poll of Polls all have a pretty good reputation, except the eccentric American Research Group. (The GOP is damn lucky to have nominated McCain: Obama would be pureeing anyone else at this point. And I still think that McCain’s gain was due to his acceptance speech — which I recognized at the time as pretty good — rather than to Palin.)
September 18th, 2008 at 12:00 pm
tristero,
Even ignoring all the other polls, it is simply incorrect to say a poll being barely within the margin of error means the people are tied. I know people say this all the time in the media, but they are wrong about what a margin of error implies.
The basic situation is this. In a properly designed poll, if one person comes out ahead at all, that means there is more than a 50% chance they are in fact ahead in public sentiment. The greater the margin, the greater the percentage chance the person leading in the poll is actually leading in public sentiment. When the percentage chance based on the margin crosses 95%, the poll is said to be outside the margin of error.
But what that means is that a person up in a poll but by less than the margin of error could be a favorite by anything from about 51% to 94%. And it doesn’t make much sense in a case where, say, the person up in the poll has an 85% chance of actually leading and the person down in the poll only has a 15% chance of leading to say the people are “tied”.
So people really should just eliminate the notion of a “statistical tie” from their vocabulary.
September 18th, 2008 at 2:22 pm
Wow, lots of hostility and fear in these comments.
We’ve got a lot of time before the election. The Obama campaign has built an incredibly strong ground operation. The Palin magic bullet is starting to turn in mid-air, back towards McCain. We’re a few days before the total public perception of this crisis gets fully reflected in the polls.
September 18th, 2008 at 4:28 pm
DTM,
Correct if I’m wrong here, but we don’t know the margin of error. It could be 5 percent. It could be 10%. So it “could be” that up five means close to being significant. it “could be” that up five means very insignificant.
Also, if I understand this properly, what falling within the margin of error means is that the difference is not considered statistically significant. Agreed that “statistically tied” is a crummy term. But pointing to “up 5″ as something of significance when we know that that is within the margin of error AND we don’t know the actual margin of error is utterly misleading.
September 18th, 2008 at 4:35 pm
Question; Are cell phones called in a national poll?
September 18th, 2008 at 5:01 pm
the trend for this particular poll may be more enlightening:
If by trend you mean Obama’s upswing between August and September, I think you should realize that you’re just looking at 2 points on a plot (one poll in August, one poll in September). That’s not much to draw a trend line on, especially since we all know how variable the numbers have been over the last few weeks due to the conventions.
September 18th, 2008 at 5:33 pm
Correct if I’m wrong here, but we don’t know the margin of error.
The thickness of the line on the NYT chart is the margin of error (+- 4% for the latest poll). But you have to realize that the margin of error is a statistical probability term. McCain 47/Obama 44 is within the 95% confidence interval of the poll, but it’s not equally likely to the reported value of Obama 48/McCain 43. And when you consistently see one candidate polling 2-3 points ahead of the other you can conclude that that candidate is significantly more likely to win the race even if all the polls have MOE larger than that.
That said the more pertinent section of the paragraph you cited is that this poll is “statistically unchanged” from the poll taken in August.
March 13th, 2009 at 4:52 am
Very interesting site. Hope it will always be alive!
March 14th, 2009 at 9:20 am
I want to say – thank you for this!
xanax
March 22nd, 2009 at 7:22 am
tramadol
Incredible site!
April 2nd, 2009 at 9:38 am
I bookmarked this site. Thank you for good job!
buy cheap viagra
April 8th, 2009 at 3:21 am
Great site. Good info
viagra
April 16th, 2009 at 3:43 pm
I want to say – thank you for this!
viagra