Rassmussen translated "Don't panic when Bush is up by 4-5 again"

[Frodo]

http://www.rasmussenreports.com/Poll%20Differences%20Sept%206.htm

Implying that (all else being equal) Bush will be up by 4-5% tomorrow.

They make it clear this is NOT a new "bounce", but rather the result of a single day of bad data over the weekend that threw off the numbers.

Nothing to worry about. But it might be fun to watch how the freepers overreact.

[a new day]

1. And don't forget

The problem with cell phones and <35 voters.

It's dead even now.

[Fionn]

4. No it isn't. That's rubbish.

Pollsters (good ones at least) weight their samples appropriately with age. And if you've seen a poll which gives Kerry a huge lead with cellphone users, then you are unique in that respect.

Call a spade a spade. Kerry is 5-6 points down. But this ain't over.

[a new day]

9. Not according to polling companies

http://www.zogby.com/soundbites/ReadClips.dbm?ID=4864

"Increasingly, people are relying on cell phones and don't even have a home phone" <said Tom W. Smith of the National Opinion Research Center at the University of Chicago.>

"The current approach to cell phones (in polling) is to exclude them," he said. "You can do that with the current numbers. But it becomes less and less acceptable as more households have only cell phones."

"Up the road, we definitely need to address these issues," said Schulman <Mark A. Schulman, president of the American Association of Public Opinion Research>. "But I wouldn't pin this (election's polling results) on cell phones or caller ID."

(11/7/2002)
- By Bob Dart , The Daily Sentinel Grand Junction, Co.

[Frodo]

10. What in your post refutes his point?

The article you linked was trying to explain why pollsters undermeasured republican turnout in the last election. But you're using the point to assume they are now undermeasuring democrats?

Did the cell phone using population change parties in the last two years?

[a new day]

13. Not a matter of D vs R, a matter of Bush

[Frodo]

16. ???

Cell phone users don't like Bush?

I mean, my "sample size" of TWO agrees with that point, but I can't see how it can be statistical. Has Bush done something specifically to cell users?

[MoonRiver]

14. Lol!

[Frodo]

7. I'll need more than assumptions before I buy that.

The "cell phones" issue is just the newest wrinkle in why polls are difficult to conduct.

An unspoken secret in the polling industry for YEARS has been the incredibly LOW resonse rate to polling calls. But I haven't seen a compelling reason why one party or the other is benefiting from it. Especially for that majority of decent polling firms who rebalance by party affiliation.

[a new day]

11. If cell phone users are excluded, how can it not skew results?

that's what the quote from my previous post says they do with cell phone users.

http://www.mlive.com/news/aanews/index.ssf?/base/news-9/1086360055141730.xml

The study showed 25 percent of respondents under the age of 25 only use a cell phone, while 21 percent of 25- to 34-year-olds rely solely on their cells. As a group, African Americans were three times as likely as whites to report their cell phone as their primary number, the study stated.

[Frodo]

12. Easily.

They rebalance figures by age and party affiliation.

So if it's harder to reach 18-25 year olds they overweight the ones they DO reach to ensure a correct sampling. Unless you have a way of saying 22 year olds WITH only cell phones vote differently from those who do not.



I added the following to that last post:

"The article you linked was trying to explain why pollsters undermeasured republican turnout in the last election. But you're using the point to assume they are now undermeasuring democrats?

Did the cell phone using population change parties in the last two years?"

[a new day]

15. I stand by may statement

"As a group, African Americans were three times as likely as whites to report their cell phone as their primary number, the study stated."

That should give you some Democrats.

And,if the polling companies say the exclude these voters, how do they know how to "balance" for them?

[Frodo]

17. Again,

You're point relies on the assumption that African Americans WITH cell phones vote differently than African Americans WITHOUT cell phones.

On what basis do you make that assumption?

[a new day]

18. -

So, you exclude the analysis of 25% of the population and are not making assumptions?

[Frodo]

19. No.

Have you ever done any statistics?

Have you ever studied how polls are conducted?

They rebalance for missing populations ALL THE TIME. And the African American population is OFTEN the one they need to account for. Higher proportion of poverty means a lower population of phones (of any kind). So you overweight the members of that group that you DO reach.


As for this most recent point? You don't make any sense. That isn't "25% of the population", it's 25% of the 18-25 year old population which is itself a small portion of the voting population because 18-25 year olds rarely vote. And their partisan balance is not that great.

So even assuming your point, it could only throw the results off by a small fraction of one percent. NOT enough to put a five point race into the "dead even" category.

[papau]

2. I choose to panic! -- but I expect the bounce to be gone in a week!

"If we drop the Saturday sample from our data, Bush is currently ahead by about 4 percentage points in the Rasmussen Reports Tracking Poll.

That's still a smaller lead than shown by Time and Newsweek. Those polls appear to have the mirror image problem of a Los Angeles Times poll in June reportedly showing Kerry with a huge lead. That LA Times survey included too many Democrats in their sample. Today, it seems likely that Time and Newsweek included too many Republicans. <snip>

Four years ago, 35% of voters were Republicans, 39% were Democrats, and the rest were unaffiliated. If you apply those percentages to the Time internals, you find Bush up by about 3 percentage points. If you do the same with the Newsweek internal numbers, you find Bush with a six point lead. Those results are very close to the Rasmussen Reports data (excluding the Saturday sample).

All of this leads me to conclude that the President is currently ahead by 4 or 5 percentage points.<snip>

[Frodo]

3. You don't strike me as the "panicky" type.

It can't be gone in a week because Kerry doesn't have a news cycle opportunity until after 9/11. The next few days will be all about Bush going down to "personally save people trapped in buildings" in the wealthier parts of Florida.

If Ivan doesn't change course, Bush gets a trifecta of news coverage being presidential and Kerry doesn't get a cycle until the debates.

[papau]

8. I agree the 911 cycle will be media forgetting the intel disaster and just

praise for Bush - a RNC version of the world.

But the RNC convention with the phallic symbol podium with the cross in wood in the podium was seen by 3 million more than saw the DNC convention (27 to 24 m). I have to believe that all that button pushing will have a short term only effect.

But I agree the debates will be the win or lose moment - indeed it will be the first debate, since I do not see folks cutting Bush slack like they cut Reagan in 84, so he could joke his way back to a "why not Reagan" position after the disaster of the first 84 debate.

[featherman]

5. That may reflect the "bounce" but is iffy.

Rasmussen has been trending up for Bush in the past week or so. I'm still of the opinion that a plus 3 for Bush by 9/15 is gonna be ok for Kerry. Determined, perhaps, by an average of 4-5 of the better horserace polls. After that Kerry needs to start reeling him in to get it back to about even. I am guessing "even" polling translate into a Kerry lead in the final (horserace) election results and in the Electoral College. Anything better than even for Kerry heading into the final week or so will be a very good sign indeed.

[tedoll78]

6. I'll add my own "don't panic" advice..

First, Gore was down in most polls right before Election Day in 2000. If we're within 5% of Bush before November 2nd this year, the same story may develop.. a real chance at winning.

Second, in 2000, a poll asking Dems if they were "excited" about voting for Gore got the "Yes" answer from about 35% of those polled. If you'll recall, our turnout that year was pretty nice. This year, that same question has been asked, and the "Yes" rate was near 70%! Extrapolate the turnout trend from there.. :)

Furthermore, Bush got a chunk of the undecided vote in 2000. We can pretty much expect that swing to go to us this year.

I'd guess that, for being right after the GOP convention, we're in pretty damn good shape.

By the way.. did I mention turnout? :D