Several press account yesterday touted a new Washington Post-ABC News poll showing President Obama leading Mitt Romney 51-43 percent. The media used this poll to make cases for everything from Mr. Obama being more likable than Romney to being more in-tune to the needs of the middle class.
Conversely, the poll showed that Romney’s message resonated more so than the president’s and that people believed the former is more able to handle the economy and energy prices.
While I know the media love polls, and they are of great value to campaign strategies, and indeed polling keeps folks like Scott Rasmussen and my good friend John Couvillion employed, I have to question the legitimacy of this particular poll. I found something interestingly skewed in its methodology.
The poll, conducted April 5-8, had a margin of error of +/- 3.5% because of its sample size of 1,103 adults. This is a very common national sample size and resulting margin of error. However, this particular polled, in my opinion, did not gain an adequate sample of respondents from the two major political parties. The Post & ABC polled a sample of 34% Democrats to only 23% Republicans, when the actual voter registration number are roughly 36-32%, Democratic. This severe under-sampling of GOP voters drastically skews the numbers in favor of Mr. Obama.
The value of a skewed poll is nil and shouldn’t be given any credence; however, the media will not look at that if it doesn’t suit their story line. But now you, the well-informed, know the truth.