What can explain such divergent results which are nearly statistically impossible given the margins of error (+/-3% in each poll)?
"As you may know, the tax cuts passed during George W. Bush's administration lowered taxes by reducing maximum income tax rate for all Americans. These tax cuts are set to expire at the end of 2010, meaning tax rates would go back to what they were before the Bush tax laws. Congress is currently considering whether to let these tax cuts expire or extend them. Which of the following comes closest to your own view on what action Congress should take?:
-Let the tax cuts expire for all Americans
-Let tax cuts expire only for people who earn more than $200,000 a year
-Extend the tax cuts for all Americans"
CNN found opposition asking: "As you may know, the tax cuts passed into law when George W. Bush was president are set to expire this year. Unless a new bill is passed, federal income tax rates will rise to the level they were at when those cuts were enacted. Which of the following statements comes closest to your view:
-Those tax cuts should continue for all Americans regardless of how much money they make
-Those tax cuts should continue for families that make less than 250 thousand dollars a year, but taxes should rise to previous levels for families who make more than that amount
-Taxes should rise to the previous level for all Americans regardless of how much money they make"
Perhaps the respondents didn't understand that Reuters was referring to individuals and CNN to families thus respondents to the Reuters poll felt the tax cuts for high earners benefited wealthier taxpayers than did the respondents to the CNN poll
Perhaps respondents have a bias toward possible answers that they hear later in the poll
Perhaps the survey methodology was in some way flawed.
It's hard to know, but this ought to remind us not to believe any single poll we read.