Poll biases: It’s not just the sampling weights
As some national polls show President Obama widening his lead in his race for another term, much has been made about the sampling weights that pollsters use. Analysts on the left insists the polls are accurate. nalysts on the right say the polls are inaccurately favoring Obama by assuming his supporters will come out on election day in the same numbers as they did in 2008.
But its not just weighting that reveals a pollster’s bias. The way the question is asked also makes a difference.
In a Monmouth University/Asbury Park Press poll about the 2013 New Jersey gubernatorial race released this morning, pollster Patrick Murray asked if voters were “bothered” with how Governor Chris Christie interacts with his critics and detractors.
Thinking about Chris Christie’s style and not his policies, does the way he speaks to or about people who disagree with him bother you personally or not bother you? [If BOTHER: Is that a lot or just a little?]
63% of respondents said they weren’t bothered by Christie’s style. 23% said they were bothered a LOT and 11% said they were bothered a LITTLE. Given the way Murray asked the question, one could conclude that 74% of New Jersey voters are indifferent about Christie’s style.
In his narrative of the poll, which sets the tone for how much of the lazy lefty media covers it, Murray highlights his spin on Christie’s style.
“NEW JERSEY ON CHRISTIE’S STYLE: ‘MEH!’ ” is Murray’s headline. His opening sentence:
Governor Chris Christie’s job approval rating has ticked up a few points in the latest Monmouth University/Asbury Park Press Poll and few New Jerseyans are particularly bothered by the way he deals with people who disagree with him.
Notice the use of the word few.
Christie’s numbers are the highest they ever been in a Monmouth poll. 55% of registered voters approve of the governor’s performance. 36% do not approve.
Yet Murray spins the results to read that a few people like him better and a few people are bothered about how he talks to people who don’t agree with him. The few who are bothered take top billing over the fact disclosed but not reported that Christie’s numbers are better than ever in Murray’s poll.
What does that tell you?
The Asbury Park Press’s coverage of the poll leads with the “bothered” question.
The headline at NJ.com for an Associated Press story is Christie’s approval rating up slightly, poll says.
To their credit, PolitickerNJ cut through Murray’s spin and covers the poll results very well. They reported the real news of the poll results; New Jersey’s sagging economy is not hurting Christie’s popularity with voters and that of potential Democratic challengers in 2013, only Newark Mayor Cory Booker and former acting Governor Richard Codey have sufficient name recognition to be considered credible candidates for governor next year.
What if instead of asking if voters were bothered by Christie’s style, Murray asked if they liked his style? If Murray had done that, the headline would be:
CHRISTIE’S APPROVAL AT ALL TIME HIGH
New Jersey voters like his style
Distortions are common in all publications that report statistics to consumers. If the writer does not understand statistics, he is not able to write an accurate interpretation. If he does understand statistics, he might (knowingly or unknowingly) skew the article with his bias.
You would have to study statistics to fully understand the game of manipulation. But there is a well written, easy to understand book on the subject for us little people:
http://www.amazon.com/How-Lie-Statistics-Darrell-Huff/dp/0393310728/ref=sr_1_1?s=books&ie=UTF8&qid=1348756808&sr=1-1&keywords=how+to+lie+with+statistics
That Sits On Their Butts, On Welfare & Want Freebies. ABW (Angry Black Woman) DOES Indeed Apply Here
http://www.youtube.com/watch?feature=player_embedded&v=tpAOwJvTOio
Good numbers for the big guy especially since the one question had a bit of a negative connotation.
I thought that “MEH” was a bit weird to put in the headline…truth be told I had to look up what it meant. The stories themselves seemed to put a very positive light on the info (as they should) for the Gov.