Monday, October 6, 2008

Survey USA: Obama Wins Virginia

Just in. Survey/USA has released a new poll giving Barack Obama a 10 point lead over John McCain. Call it a rap folks, I mean 900 Virginians sampled can't be wrong?Come on, 900 out of a potential 5 million voters is simply an acceptable barometer of Virginia?

These 900 came predominately from the Washington DC suburbs and the urban areas of Virginia according the the poll breakdown. Of course, these areas are truly indicative of the overall Virginia sentiment in terms of voting.

Heck, based on these areas polled, John Kerry should be President right now. Out of potentially 5 million voters after todays registration ends, a poll of 900 is ridiculous regardless of whom it has leading.

7 comments:

Anonymous said...

LOL on this one. 900 people is some sample.

Is there a bias at all with this poller. I know that the PPP polls you have referenced in the past is a Democratic poller.

Anonymous said...

Considering you've misspelled your own user name, I'm not going to spend much energy on this, but, seriously, you should be aware that, first, this is a poll not of 900 adults, but of 666 likely voters, which is, in fact, a perfectly adequate sample for a state poll. Second, seeing as the full crosstabs of the poll are freely available, I don't understand why you're lying about the geographic breakdown of the survey. You want to make some points here, maybe question the party breakdown, which has been trending a little more Democratic with every release of this particular poll -- maybe there's something real there and maybe there's not -- but ignorantly complaining about sample size and geographic breakdowns isn't going to convince anyone of anything.

James said...

Welcome to Alter of Freedom, Anonymous.
First and formeost the terminology behind "alter" has been freely expressed and addressed many times here so I will not go into again except to say familiarize yourself with the original manuscripts inked by Lincoln himself where he uses the old english version of the modern "altar". Alter of Freedom frequently quotes Lincoln and other great Americans using the exact "historical" wording.

On your next point.

The facts appear clear.

Survey/USA interviewed 900 Virginians from 10/04 to 10/05 for this poll. Of those 832/900 were registered voters and 666 were determined to be likely voters.

Those are the facts and parameter metrics.

This poll was as stated concentrated in markets of Washington DC, Roanoke (City), Richmond (City) and the Tri-Cities as stated in the metrics.

So you have taken a poll in which the post stated that there were 900 respondents and reduced it to the 666 "likely" number though the start number was 900. Fair enough. But you just reduced the sample my friend even further.

Public Policy Polling, Rasmussen and the like are twice as many likely voters in their metrics and by far this poll is by far the least "adequate" sample either demographically or numbers.

If I recall Alter criticized the Christopher Newport University polling for the same reasons in terms of demographics.

The metrics have a 3.9% nargin or error as well and the makeup:

43% declared Republicans
53% declared Democrats
4% Other/Ind

was not even posted by Alter.

With all do respect, nothing in the post is "lying" in any stretch of the imagination. Everything is readily available at Survey/USA for review my friend.

When you consistantly have a ten point spread in your metrics should any other result realistically be possible?

Anonymous said...

Media sponsors doesn't equal concentration. It's a statewide poll.

Alter of Freedom said...

James:
Thanks for the support.
Semper Fi.

You touched on the merits of the poll as expressed by Survey/USA.

I have always stated that I prefer the Public Policy Polling on balance and that is a Democrat-leaning pollster as well so frankly I have demonstrated an open-mindedness regarding results.

I do not feel any of Survey/USA's polls are reflective of the State. Ten points is the largest margin of any recent poll, but we will see come November.

I have made no bones that the State has been leaning Democrat so there is no bias here on my part either in terms of the internals of the poll.

Those respondents certainly came from the markets that James expressed in his comment as pronounced by the pollster and given that the basis of declared affiliation is 10% points difference it is easily dicerned where the origins of the respondents where from.

Where do we find such wide margins in affiliation? Why in Arlington,Alexandria, Fairfax, Loudon, City of Richmond, Tri-Cities and to a lessor extent Roanoke. I would venture to say a large number of the Republicans came from the latter.

Of course this was a statewide poll. It just so happens it concentrated heavily in Democrat-leaning areas.

The pollster also made some pretty out on a limb statements like that Republicans have lost a 24 pt lead in the Shenandoah Valley region, but no other polling data supports that claim at all. Of the 666, how many came from the Valley in order for SUSA to make that claim I wonder? Roanoke does not include the Valley vote nor does it include the core Southwest Virginia localities either.

Anonymous said...

Rasmussen/Fox latest poll has Obama up by 2 pts. A far cry from 10 pts Anonymous. Obama has add at least a three point lead for several weeks by this pollster. It is also calling an Obama lead in OH so please refrain from screaming biased notions.

Anonymous said...

Public Policy polling released another North carolina poll yesterday giving Obama a six point lead.

Likely voters was over 1,200. How many people live in NC versus Va Anonymous?