I wonder if DU will call this an outlier, too.


Bush Leads Kerry by 11 Points in Newsweek Poll After Convention

Sept. 4 (Bloomberg) -- President George W. Bush leads his Democrat opponent John Kerry by 11 percentage points according to a poll immediately after the Republican National Convention in New York, Newsweek magazine reported.

Bush is supported by 54 percent of the 1,008 registered voters surveyed Thursday and Friday, compared with 43 percent support for Kerry, a four-term Massachusetts senator. Independent candidate Ralph Nader polled 3 percent. The poll has a margin of error of 4 percentage points, Newsweek said.

Bush, 58, accepted his party's nomination Thursday after convention speakers said he's strong on national security and in fighting terrorists. Kerry, 60, said in a speech after the convention ended that ``it's time to move America in a new direction'' and that Bush's policies have failed.

The president's job-approval rating rose to 52 percent, the first time it's been above 50 percent since January, Newsweek said. A 53 percent majority wants to see him re-elected, the highest since May of last year, the magazine reported.

quote.bloomberg.com/apps/news?pid=10000087&sid=aWMeOSOk0kic&refer=top_world_news

All I hope is that it's accurate and holds (if not grows!) between now and November. While I'll take any win I can get, I hope Dubya can win by as larger margin as possible, just to show the left how full of crap we really think they are.

Still, there's lots of time left for bad stuff to happen, the debates, attack ads, etc. I doubt we'll see this lead last, at least not as large as currently being seen. But I sure hope I'm wrong. Still, I'll take being ahead by 11 to being behind by 11 any day of the week!

I give it a 5 banana rating!

Goody, I don't have to bother to vote now!

"NOOOOO!  LIES!!!! THIS CANNOT BE TRUE!!!  DAMN YOU, KARL ROVE! YEEEEEEARRRGHHH!"

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=744894&mesg_id=744894&page=

How accurate can 1000 participants be in a poll?

Quoted: How accurate can 1000 participants be in a poll?

This is the average number of people that are used in most polls

Quoted: How accurate can 1000 participants be in a poll?

The weird thing about statistics is that they don't become more accurate with larger sample populations.  Or at least, that's what my professor told us in econometrics.  Of course, a sample of 1000 is better than 10, but the standard error of say, 100,000 versus 1000 might not be large enough to justify the costs of polling such a large population.

My thought is, if you aren't careful, you may get 600 from large cities, and 400 from rural, and skew your results.

Quoted: How accurate can 1000 participants be in a poll?

It's the standard sample for these types of polls, generally accurate within 3-4 %. 1000 participants is about as accurate as you can get economically.

Rassmussen and some others use tracking formulas, they do a poll every day, and use a 3 day average, so its less likely to be effected by variances in the sample. Seems to be pretty accurate, and less prone to large swings in the results.

But remember, the only poll that matters is in november, can't slack up yet, even if keery looks like he is imploding. A 10% lead can be eroded in weeks easily.

Stay vigilant.