Why do certain news/polling organizations oversample Democrats?
When, writes Michael Barone, “you get the Gallup and Rasmussen tracking polls and the Politico/George Washington University Battleground poll all showing Mitt Romney leading Obama by 1 point, an Obama victory seems far from inevitable.”
Some polls, however, don’t show a contest nearly that close. Yesterday, Allahpundit wasn’t all that worried about a Reuters survey that put
Obama up seven on Romney nationally — thanks chiefly to a sample with a 12-point Democratic tilt (excluding leaners). Elections guru Jay Cost of the Weekly Standard had a laugh over it on Twitter earlier and noted that even in a poll this blue, O’s job approval stands at just 50 percent.
Not a good sign for the president if his approval stands at just 50 percent in a poll with a greater sample than the electorate that turned out in November 2008.
Seems that our friends in the media may well want to great the impression that Obama’s reelection is, well, inevitable.
2 Comments
RSS feed for comments on this post.
Sorry, the comment form is closed at this time.
Hi Dan,
You raise a good point about over-sampling. However, you don’t raise (or deal with) similar issues of statistical significance, on polls that favour Repub talking points. So, good for you for pointing out the problems in one set of polls. As the IPSOS data suggests, “With a sample of this size, the
results are considered accurate within 2.9 percentage points, 19 times out of 20, of what they would have been had the entire adult population in the U.S. been polled.” Well, a 95% confidence level is not alwyas that great…
Comment by Cas — May 9, 2012 @ 2:10 pm - May 9, 2012
Because it’s not about reporting the news. It’s about SHAPING it.
Comment by Bastiat Fan — May 9, 2012 @ 6:56 pm - May 9, 2012