Landslide Lyndon
px75
Atlas Star
Posts: 27,119
|
|
« Reply #1 on: April 08, 2011, 10:57:26 AM » |
|
From Political Wire's comments section:
The poll is extremely dubious on countless levels.
Start off with the sample size of 360 - pathetic for a statewide poll, producing a huge MOE.
Then consider that the poll asked an overwhelming 32 separate issue questions before they asked the Senate preference question - a sure-fire way not simply to throw the results off by influencing responses, but also to weed out lots of potential participants who simply don't have the patience to sit through a survey that long. As an issue poll design, this sort of thing is frowned upon if you want to get accurate results on the big candidate choice question that is being asked at the end of such a long string of potentially leading ideological questions.
But more importantly, the racial/ethnic demographics are completely off - 85% white, 10% African-American, 5% "other". In 2008, the turnout was 70% white, 20% African-American, 5% Latino, 3% Asian, 2% other.
Same on the age - 27% are over 65, compared to 11% of the electorate in 2008, and 62% are over 50, compared to 37% in 2008.
The also appears to have under-sampled Northern Virginia (21% in this poll, 26% in 2008), and over-sampled SW VA/Shenandoah Valley (23% in this poll, 17% in 2008).
Given all these issues, anyone who looks at it as a realistic snapshot of that state of the 2012 Virginia Senate race is relying on a very flawed piece of research.
|