Skip to comments.Very good review (and encouraging) review of the statistical analysis for 2008 POTUS Polls
Posted on 10/24/2008 2:41:08 PM PDT by NYC_BULLMOOSE
First, we need a short primer on basic statistics. Real Clear Politics offers an unweighted average, or mean, of the polls. As long as there is more than one poll in the average, we can also calculate the standard deviation, which is one of the most important concepts in inferential statistics. The standard deviation simply tells us how much the polls are disagreeing with one another.
(Excerpt) Read more at realclearpolitics.com ...
<< So, we have made three observations: (a) relative to 2004, the standard deviation for Obama and McCain’s polls are high, indicating more disagreement among pollsters at a similar point in this cycle; (b) the shape of the distribution of each candidate’s poll position is not what we might expect; (c) multiple polls are separated from the RCP average by statistically significant differences.
Combined, these considerations suggest that this variation cannot be chalked up to typical statistical “noise.” Instead, it is more likely that pollsters are disagreeing with each other in their sampling methodologies. In other words, different pollsters have different “visions” of what the electorate will look like on November 4th, and these visions are affecting their results.
It will not be the first time someone got burned by the “tails” on a standard deviation. Ask hedge fund LTCM about a decade ago. They ignored the tails that the Russian market could possibly meltdown and their hedge fund almost blew up.
less accurate...but ALL...even after this analysis...show the Big O ahead.
This election does have A LOT on intangibles, and it’s throwing the polls off.
I think it will be close.
TURNOUT IS THE KEY!!!!!!!!!!!!!!!!!!!!!
Why didn’t the polls have the same degree of variation from each other 4 years ago, assuming they haven’t changed their methods? Perhaps the big difference in newly registered Democrats over Republicans accounts for it. The polls may make different assumptions about their turnout. Just one possibility.
I was just talking about that last night...Greenspan only looked at 3 standard diviations on derivitaves and ignored the rest as impossible....rut-ro
He makes one very bad assumption in his article: he assumes that at least one of the polls reflects the actual electorate. This leads him to the conclusion that Obama is ahead. Since the polls seem to vary from 6% extra Democrats through 13%, and the normal electorate historically favors the Republicans, I submit that the entire process is flawed this time.
mm-m-m...yep...that’s a good thought.
I think Zogby predicting McCain’s victory a month ago wasn’t just his subversive way of discouraging the Republican vote (although I’m sure he did have the motive)
I think him saying “comfortable, old shoe” was a good way to hedge his bets so he doesn’t lose as much credidibilty as the rest will come Nov. 4th
What is the biggest turnout disparity ever shown in modern day presidential elections? I cannot believe 6%, 2-3% maybe at most. Are any polls using closer turnouts like that?
Garbage in, garbage out.
No more cmplicated then that.
This year there is simply way more garbage then normal.
Not a sane person would argue against that.
Thanks for the post!!
The biggest dummy advantage in the past several elections was 4%. It was 3% in 2006 and 0% in 2004.
I think the polls are flawed this time because the pollsters have built in the Bradley and ACORN effect in their polls. I think their models are a mess. They have added exceptiosn to the historical models they have used.
I think the pollsters are suffering from the Bradley effect. They are afraid to say the black guy is tied or losing. The adjust their sampling with bigger Dem turnout theories to not appear to be racist.
I have another theory which almost always answers every question. Money.
Polling firms with the exception of IBD/TIPP and Battleground usually have TV network or newspaper/mag patrons. Many of these mags and newspapers are owned by TV networks and the networks and both networks and newspapers own local TV affiliates.
Obama has spent $650 million and probably $380 million in advertising? He has made their year. Many of them are struggling especially newspapers. The pollsters make their money during elections. Batlleground is funded by a university and IBD is a business newspaper.
The point is the others are whores. Zogby, ABC, CBS, CNN or any sponsored polls. They want that Obama ad money. They are adjusting their sampling to keep the networks, newspapers and magazines they do the polls with happy.
McCain has far less money but he is spending it now. He may get better polls now too.
The pollsters and MSM will be pushing Obama til the last week. I think the MSM pollsters in the last week will have to stop being whores and do an honest poll or their reputations will be ruined.
Sure hope they're oversampling dims because of folks like me and folks that Rush encouraged to cross over for Operation Chaos.
I still think that the party Lbama's® planning in Grant Park in Chicago is going to be one of the biggest stinkers of all time.
“McCain has far less money but he is spending it now. He may get better polls now too.”
And that will be the key to victory. Obama’s lead has been fueled partially by a huge advertising advantage that is closing in the final days. McCain may yet pull neck and neck by the time election day arrives.
Turnout is not as big of a key as defection is. Defection is the number key to winning elections. Democrats are much more likely to defect than Republicans—they are a much more fragmented party. The last two elections have not had a lot of defection—they have been pretty partisan, and pretty close elections. I see no reason why this election won’t be close either. This 400+ electoral vote stuff is just absurd—unless EVERYTHING aligns for Obama, which I don’t suspect it will. Obama is up in the polls, that can’t be disputed, but like I always say a poll isn’t an election. I know I’m going to vote on the 4th (and I’m in Missouri, so I NEED to get out there and vote—my parents already voted for McCain absentee, so I know he has at least two votes here). If anyone else is curious about Missouri, the McCain radio ads just hit the airwaves today. It has been 24/7 Obama for months, but the McCain ads have in fact started—they are targeting the “liberal Congress” as well.
But ..... there is no discussion about checking for normality of the distribution. If the distribution isn’t normal then standard deviations mean nothing and the tests used to compare the distributions are worthless. Sigh - lies, damn lies, and statistics.