Obama Gets ‘Big Bounce’ in Gallup Poll
In the first poll since the start of the Democratic convention, Gallup gives Barack Obama what RealClearPolitics calls “a big bounce,” taking a six point lead over his rival, John McCain.
Democratic candidate Barack Obama has gained ground in the latest Gallup Poll Daily tracking average from Monday, Tuesday, and Wednesday, and now leads Republican John McCain among registered voters by a 48% to 42% margin.
The latest three-day Gallup Poll Daily tracking average (Aug. 25-27) is directly coincident with the first three days of the Democratic National Convention in Denver, and is no doubt beginning to reflect the typical convention “bounce” that Gallup has observed in most party conventions in recent decades.
Another way of looking at the same data is that the race continues to be a tie fluctuating within the range of normal sampling error:
Indeed, the only other national poll conducted during this period, Rasmussen‘s, shows Obama and McCain tied 47-47. And, unlike Gallup, Rasmussen samples likely voters rather than mere registered voters.
Now, I fully expect Obama to get a bounce here, probably in the range of 4-5 percent. Thus far, though, there’s no evidence that it’s happened.
You say this is insignificant fluctuation, but if one scrolls down three or four posts on your own blog, one sees another article predicated on the significance of a similar-sized drop in Obama’s lead. Either they are both insignificant, or they are both significant. It can’t be that the drops are significant and the bounces are not…
Which post are you speaking of? Every post on polling I’ve had in the past two weeks has said that the race is essentially tied with random fluctuaton.
He’s probably referring to this:
Though in that post you weren’t clear if you were merely talking about the perspective of the pundits, or even how far back you were going in your reference to Obama’s impressive lead.
Sorry, should have linked to the article in the above post: https://www.outsidethebeltway.com/archives/what_obama_needs_to_do/
Michael is correct. In your own words you said that Obama’s “lead in the polls has all but vanished.” That’s not some pundit’s quote. A few hours later, when the same lead reappears, it’s sampling error.
Don’t misunerstand me. I’m firmly in the random fluctuation camp. I’m just bemused by the fact that the same evidence leads to opposite conclusions when it suits the author.
Analyzing poll results seems about as meaningful as trying to interpret tea leaves.
Just because Rasmussen has them tied does not mean they actually are. Rasmussen indicates that the bounce is may be coming based on their day-by-day polling data. Also, just because Rasmussen uses a likely voter model does not make their polls more reliable. Mark Blumenthal‘s answer to which polling method is better:
Ah: Yes, Obama had a meaningful lead months ago. It’s been statistically tied, though, for quite some time. A 3 point fluctuation in a single poll is just random error.
I was reading an interview with David Pouffle (Obama’s compaign manager) who said they don’t care about the national polls. He said they care about 18 states and that Obama is where they want him to be in those states. The race Obama is running seems to be on a different track than the national media thinks it is.