The problem with using bad samples in polling

The Conventional Wisdom, according to the WaPo:

The dominant narrative since the beginning of 2012 has been that President Obama has regained his footing after a rocky 2011 and is trending upward. Ask 10 political types who will win in November, and eight of them (or so) will say Obama.

The piece then goes on to discuss conflicting polls, including the flagrantly manipulated CNN poll the other day that had Romney losing by nine points. Meanwhile, a “new Rasmussen Reports national telephone survey finds that 46% of Likely U.S. Voters would vote for the Republican in their district’s congressional race if the election were held today, while 36% would choose the Democrat.” That is surely a significant data point.

It only serves Democrats’ interests to use phony samples and adults, not likely voters, if legacy media outlets control the narrative. Otherwise, it’s an unhelpful exercise in self-delusion.

Perhaps the problem is even worse than that. The NYT’s Tom Friedman: “Obama, who has a plan to cut, tax and invest — albeit insufficiently — could lead, but, for now, he seems preoccupied with some rather uninspiring small ball, preferring proposals like ‘the Buffett tax’ over comprehensive tax reform that would lower all rates, eliminate deductions and raise more revenue.” If even Mr. Friedman can see through the malarky that comes from this crew in Washington, things must be bad indeed.

Leave a Reply