« Another Truly Insane Ad | Main | Robo-Polled »
October 30, 2008
And One More Faked "Tightening" Poll
It's as though the different rep-leaning pollsters are trading off on who will take the bullet for using their poll to lie about the race each day. It really is becoming fascinating.
Andrew Sullivan kicks around today's Fox News poll with Barack Obama now leading John McCain by 47 percent to 44 percent among likely voters, down from 49 to 40 last week.
Wow.
Until you look under the hood...
Last week's poll was of 401 Democrats, 345 Republicans, 148 Independents
This week's poll was of 379 Democrats, 364 Republicans, 146 Independents
894 – 44.9% Dem, 38.6% Republican, 16.6%
889 - 43% Dem, 40.9% Republican, 16.4%
So, if you gave Obama back his 1.9% of the vote, he’d be back up to 48.9% and if you took away McCain’s increase of 2.3% of the vote, he’d be back down to 41.7%... a 7.2% spread… or pretty much what most polls are showing.
Doug Henning would be proud... except that he dies of AIDS... which the Republicans wouldn't fund research for... oh well.
Posted by dpoland at October 30, 2008 02:55 PM
Comments
How is that faked? Random samples vary randomly. That's how they work. You can argue the sample skews the result (which can then be controlled for), but arguing that the survey is therefore faked isn't a very compelling argument. It is, in fact, the same kind of shoddy thinking you are usually so critical of.
Posted by: Blackcloud ![[TypeKey Profile Page]](http://www.mcnblogs.com/thehotblog/nav-commenters.gif)
at October 30, 2008 03:31 PM
It interesting how you'd make a case against a poll showing a 3% margin, and never mention CBS/NYT poll, which has been a consistent outlier.
The RCP poll average is about 6%. The Fox News poll is 3% off that average. The CBS/NYT poll has Obama by 11%, which is 5% off the average. Which poll is more egregious?
Posted by: mysteryperfecta at October 30, 2008 04:22 PM
I am happy to concede that any poll with Obama more than 9% ahead is an outlier on that side. The difference, of course, is that the Obama campaign is not desperately trying to make the case that they are a mortal lock to landslide. And you haven’t seen me tout a single wide margin poll on this site.
Unlike Blackcloud's false assumption, no major polls are completely random. Professional pollsters do weight their polls. "Random" is not accurate. This is why the internet polling is so worthless... it's opt-in, which, even though hacks keep attaching traditional polling standards to them (x people = x variation inaccuracy possibility), they are not close to accurate.
And when a pollster allows or willfully changes their metrics in favor of one side, it is a skewed poll. A few days ago, it was the 51% evangelical poll… today, it is a significant swing towards Republican voters. These details are very, very significant, but the headlines they bring are not interested in the detail.
FiveThirtyEight says that NYT and Zogby are the two worst national polls, in terms of “pollster induced error.” So, point to you, mystery.
And of Fox News polling, PoliPundit notes: “Opinion Dynamics Corporation has a constantly updated database of demographic information about the national samples and, if one should deviate significantly from past averages, weighting is used to bring the sample into conformity with other samples”. In English, that suggests that Fox will weight some polls, but not others, which is a strike against consistency.”
This analysis was done in 2004, but it seems to be just as true tonight.
Posted by: David Poland at October 30, 2008 05:15 PM
dp -- just curious...where did you get the notion that henning died of aids?
Posted by: scooterzz at November 1, 2008 10:34 AM
Post a comment
Thanks for signing in, . Now you can comment. (sign out)
(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)