All polls that use a "likely voter" model calibrate their population sample based on their model of how they think the actual voting population will come out. Each polling outfit uses their own model and it can be just a top-line number (we predict that 53% of the voters will be Democrats and 47% will be Republicans) or it can be more nuanced, breaking down the population by age, gender and race (for example). A lot of these models will use sample data, where they call people and find out how confident they are that they'll be out voting on Nov. 6th and some will also use adjustments to populations from previous elections. Will the voting populace look like 2008 (higher than historic norms for turn outs of minority and young voters) or will it look more like 2004 or 2000? Are Republicans or Democrats more fired up to get out and vote? Will Independents say fuck it and stay home or do they feel motivated to cast a vote one way or another?
So, yes, it is true that the different polls have different samples for likely voter screens. A lot of polling outfits think that having Obama on the ticket and recent battles over student loans and such will keep minority and youth voting rates closer to 2008 levels than, say, 2004 levels. They adjust their likely voter sample to reflect those guesses. Republicans, of course, are betting on disillusionment because of the economy and the lack of movement in Congress to keep a lot of people home who came out in 2008 because they thought things could change. If they are right and those people stay home, it's true that the likely voter models would be over sampling for Obama. Undercutting the sampling argument, however, is the fact that national polls of registered voters (that don't resample for likeliness of voting) also show a modest cut consistent lead for Obama of around 3 points. Polls being done to assess voter enthusiasm also point to Democrats being as enthusiastic or more enthusiastic than Republican voters, so the argument that the LV samples are horribly wrong seem pretty weak to me. Possible, but unlikely. Cheers, Judah On Mon, Sep 17, 2012 at 12:17 PM, Jerry Barnes <[email protected]> wrote: > > "Most polls have Obama ahead, thought it's a slight margin. " > > Just came across this article about polling for this election: > http://www.theblaze.com/stories/this-graph-shows-why-obama-is-ahead-in-the-polls/ > > It has a nice graph that shows the lead along with the over sampling. Here > is the graph broken down: > > Poll,Lead,Over sampling > Rasmussen,-1,-1 > Tipp,2,5 > CBS/NYT,3,6 > Fox News,5,6 > Ipsos/Reuters,7,9 > ABC,1,10 > Dem Corp,5,11 > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Order the Adobe Coldfusion Anthology now! http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion Archive: http://www.houseoffusion.com/groups/cf-community/message.cfm/messageid:355190 Subscription: http://www.houseoffusion.com/groups/cf-community/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/groups/cf-community/unsubscribe.cfm
