Pollster.com

Articles and Analysis

 

Getting Sleepy 'Outliers'

Topics: Outliers Feature

Gallup conducts a split-sample survey on regulation of "Wall Street" vs "financial institutions.

Chris Bowers updates his Senate forecast with and without Rasmussen and Research 2000.

Reid Wilson finds many Democratic members of Congress are holding off polling their districts.

Jeffrey Jones charts Obama's quarterly approval.

Renard Sexton continues his discussion on the Liberal Democrats' surge in the UK.

Pew Research Center finds coverage of the Brown-Coakley race focused heavily on polling.

OKTrends breaks new frontiers in data-driven blogging (via Nyhan).

FlowingData charts times of day spent asleep from infancy to retirement (via Sullivan).

 

Comments
Westwoodnc Westwoodnc:

Reid Wilson finds many Democratic members of Congress are holding off polling their districts.

ITA, the Dems who have done polling are not releasing their numbers. Cripes! Those must be some awful numbers! The DKostards went after Firedoglake for polling the truth about the awful chances of Dems in competitive districts. Hamsher4layf!

____________________

GARY WAGNER:

In typical lefty manner, Chris Bowers starts his article off with a completely false claim about Nate Silver's article and then continues on with faulty numbers to prove how Rasmussen is so biased.

In order to reach the numbers he shows, he would have had to have taken all of the rasmussen daily polls out compared to taking only the occasional Research 2000 poll out. That is removing far too much weight that a daily poll would have. There is no other way that Rasmussen by itself could cause as much as a 10.7 point difference in all polls.

Yes - Rasmussen right now is showing a right lean compared to other pollsters. They also have a track record of coming extremely close with polls near the elections.

But, you have at least a half dozen other polls (CNN, Daily Kos, R2000, YouGov, Washington Post, CBS/Times) showing a consistent left bias. Even Gallup with their daily poll usually leans left of average about as far as Rasmussen leans right of average. It has only been in recent weeks that Rasmussen and Gallup have been running almost exact with each other.

If you want to see how just two polls have a disproportional effect, go to the national congressional ballot poll and remove the internet polls (Harris and YouGov). When you include them, it is 45.4%/43.2% - a 2.2 point advantage for republicans . If you remove them, it changes to 46.1%/39.7% republicans over democrats - a 6.4 point advantage for republicans.

Then, with those two removed if you remove Rasmussen, the numbers change to 45.8%/42.4% - a 3.4 point spread - within the margin of error.

But yet it all comes down to a liberal grasping at straws and playing with number differences that are much smaller than the margin of errors to put things in the best light possible.

The conclusion of all of his number manipulation? The senate might only lose 7 seats instead of 8 seats if you don't include the rasmussen polls. OK. If it makes him happy if he can't see november as well with one eye closed as he can with them both open, then he should dismiss the polls he doesn't like and only include the ones that show the results he wants.

To be fair, he said he is going to keep including the rasmussen polls at full wait but only until he can prove more conclusively that the rasmussen exert too much conservative influence on his poll collection.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR