Mitt Romney 10 percent lead in unskewed data from ARG poll

[Ballygrl]

http://www.examiner.com/article/mitt-romney-10-percent-lead-unskewed-data-from-arg-poll


Mitt Romney 10 percent lead in unskewed data from ARG poll

September 7, 2012
By: Dean Chambers

The American Research Group (ARG) poll of the race between Mitt Romney and President Obama released today shows a 49 percent to 46 percent lead for the former governor of Massachusetts. The poll of 1200 likely voters, surveyed between September 4-6, has a margin of error of 3 percent. Over-sampling of Democrats by eight percent, the survey sample included 38 percent Democrats, 34 percent Republicans and 28 percent independent voters.

The sampling of the ARG poll differs with the partisan data measured from hundreds of thousands of voters by Rasmussen Reports, which measures the partisan percentages at 37.6 percent Republicans, 33.3 percent Democrats and 29.2 percent independents. This indicates a degree of over-sampling of Democrats by eight percent, a plus four margin for Democrats as opposed to the plus four margin of Republicans among the likely voting electorate.

If this data is weighted for the appropriate percentage of independents as shown by the Rasmussen data, the survey indicates a far larger and growing lead for Mitt Romney. Analysis of the data by those criteria would lead to a result showing Romney leading with a 53 percent to 43 percent margin over President Obama. That would be a lead of 10 percent, larger than any for Romney reported by any recent national poll.

This survey is not the only such poll recently to be skewed by over-sampling Democrats to skew the results in favor of Barack Obama. Earlier this week, the latest CNN/ORC poll was similarly skewed. Last month on the Fox News segment “Campaign Insiders” today, Democratic pollsters Pat Caddell and Doug Schoen both confirmed their belief that major polls are skewed in favor of the Democrats by over-sampling of Democratic voters when the surveys are conducted.

[Bad Dog]

As I've said before, they have to keep the race close to generate the need for polls & news coverage.  Who is going to pay for polling if it is an obvious blowout.  Oh, and then there is the partisan hack thing too.

[NHSparky]

Couple this with the internals of the IPSOS poll posted on Drudge last night that has Obama up 4, and it's becoming more and more apparent the polls aren't a reflection of what the American public thinks but rather a desperate attempt by media organizations to try to keep up interest in the election for another eight weeks.

[Ballygrl]

I just sent this poll to Drudge, let's see if he posts it.

[BigTex]

If Rasmussen over samples republicans by 4% doesnt that mean their results will be a bit off too?

[Zeus]

If Rasmussen over samples republicans by 4% doesnt that mean their results will be a bit off too?

I think the real over sampling is the so called Independent pool.

[Kyle Ricky]

Romney has been around 10 points ahead among independents for a while now..

[Ballygrl]

This is the latest breakdown of party affiliation that Rasmussen uses, this is from wikipedia:

June 2012, 35.4% identified as Republicans, 34.0% as Democrats and 30.5% were unaffiliated

[BigTex]

This is the latest breakdown of party affiliation that Rasmussen uses, this is from wikipedia:

this is from Rasmussen's website

During August, 37.6% of Americans considered themselves Republicans. That’s up from 34.9% in July  and 35.4% in June. It’s also the largest number of Republicans ever recorded by Rasmussen Report since monthly tracking began in November 2002.

http://www.rasmussenreports.com/public_content/politics/mood_of_america/partisan_trends

 

[DumbAss Tanker]

Saw Schoen on TV today; in the discussion, it was subtly admitted that the 'Likely voters' numbers were not reliable because it appeared the pro-Romney "Likely voters" were a whole lot more likely to vote than the pro-Obama "Likely voters," any over- or under-sampling aside.