StatCounter

Tuesday 6 November 2012

Will the polls be right?

THE US Presidential election polls have been fascinating me, especially the way the headline figures are so often quoted without any analysis of the figures or how the poll was taken. regardless of the result tonight take a look at these links for some interesting polling background.

Here's Breitbart
'The last three polls Marquette has conducted of the presidential race have had a sample of D+2-4. Their polls throughout the year have ranged D+2-6. Their poll right before the June recall election of Scott Walker used a sample of D+5 and accurately predicted the election results. The sample for this latest poll? D+11. '
Here's another Breitbart
'One of the chief reasons the media is so mistaken in their zeal to proclaim Obama the election winner is that most of the polls they're basing their predictions on are wildly inflating Democrat turnout. The overwhelming majority of media polls are an assuming an electorate that is at least as Democrat as the one in 2008, with many polls assuming a higher Democrat turnout. This is just daft. The latest CBS poll of 3 swing states is the most recent to show that GOP has a huge turnout advantage over Democrats. '

Of course the question as to why the polls and the media seem so keen to push Obama's prospects is another matter.

No comments: