What will be the result of Monday's Gallup Presidential Poll?

Would last week's question have settled differently if Tri's method had been used?

I thought all of them were based on rolling averages, but maybe that's not the "average" you are referring to.

If Sunday's poll had been the deciding poll, the winning prediction would have been different. I didn't "see" Sunday's result coming ;-)

Over 121 million voters participated in the US presidential elections in 2004. Gallup's methodology for this tracking poll is: "For the Gallup Poll Daily tracking survey, Gallup is interviewing no fewer than 1,000 U.S. adults nationwide each day during 2008.". Yet they claim a margin of >2% is statistically significant. Yeah, right!

They poll .00165% of voters in the most recent US presidential election, who are adults and may or may not be registered to vote, and claim a statistically significant margin is +/- 2%?

Reading further, of the, "at least 1,000 US adults", they poll, apparently all are not registered voters since, "The general-election results are based on combined data from June 14-16, 2008. For results based on this sample of 2,605 registered voters".

So after polling some small thousandths of a percent of possible voters, Gallup claims their degree of error falls within */-2%. Balderdash!

That follows standard rules for statistical significance. We went through it in detail during my graduate-level statistics course.

See here: http://en.wikipedia.org/wiki/Statistical_significance, "A statistically significant difference" simply means there is statistical evidence that there is a difference; it does not mean the difference is necessarily large, important, or significant in the common meaning of the word.

BTW, for those who don't wait obsessively until 1pm every day to see what the latest gallup poll is, its been Obama +2 for the last three days.

People can also sign up with gallup to receive email notification of the poll results.

Yes, but I've found the RSS feeds/email notifications take a little bit longer to come out. I'm too impatient. You'll typically find me maddly hitting refresh on gallup.com from 12:58 until its posted. :)

ahhh ... the F5 key. One of Al Gore's better inventions. ;)

Oh ... you guys don't have the "media outlet" early feed?

;-)

Just a thought for future polls, as I've seen them listed both ways, this question is actually about Sunday's poll, correct? (As in "What will be the result of Sunday's Gallup Presidential Poll?") ... it would be good to have them consistent to avoid any potential user confusion, notwithstanding the settlement details.

Its actually a rolling average from Friday, Saturday and Sunday surveys.

Yeah ... that's what I asked about in Comment #1 -- which was a response to what you asked in the market question background.

No doubt I'm not understanding something ...

Sorry, I should have clarified after your first comment. Each gallup poll release is actually an average of three days of surveys (we don't get access to the raw daily survey results). So anytime someone references a particular day's gallup poll (i.e. Todays) its actually 3 days worth.

Tri used to ask questions averaging the week's gallup poll surveys. For instance:

June 17 Obama +4
June 18 Obama +5
June 19 No Poll
June 20 Obama +2
June 21 Obama +2
June 22 Obama +2
June 23 Obama +3

Average for the week: Obama +3

<<(4+5+2+2+2+3)/6 days of polls>>

So if you suspend the question midweek, you know at least half the numbers are are predicting the others. Hopefully that's clearer now?

OK -- thanks! FWIW, I like your version more than the average of the week.