Bad Poll Questions

Club for Growth highlights a CBS News/New York Times poll question.  The results are interesting, but the bigger issue is how poorly worded the question is.

Here's the question:

How much of the time do you think you can trust the government in Washington to do what is right: just about always, most of the time, or only some of the time?

The problem is that this question is absolutely horrible.  Everyone has a different image of "what is right".  Are they asking "morally right"?  Politically right?  Diplomatically right?  Economically right?  What if what is right in terms of national interest is morally wrong?   What if what is morally right is offensive to the diplomatic norms?  Liberals and conservatives could have the exact same answer, but come at it from totally opposing positions.  Heck, even two different conservatives could have the same answer, but for different reasons.

This is the problem with poll results.  They mean absolutely nothing, because they do not -- can not -- capture the nuance of the answer.

Example:  ask me if I approve of the job President Bush is doing.  My answer:  absolutely not.  Now, ask Ted Kennedy.  His answer:  absolutely not.

But my answer is based on very different reasons.  I abhor the amount of spending that Bush has launched.  No Child Left Behind, Medicare Part D, earmarks, pork, etc.  I abhor his conduct of the war-- not the war, but how he has pussyfooted around with an incompetent Iraqi government, let the Iranians bitch slap us around, and generally tried to use a Marshall Plan approach to Iraqi reconstruction. I abhor his handling of the Valerie Plame manner, his appointment of Judge Chertoff of DHS, his trust in FEMA director Brown, his policies on Immigration.

Ted Kennedy's reasons are probably much different.

But the poll wouldn't show that.

Polls aren't scientific exercises in uncovering truth.  They are the spin-doctors' magic brew.  Keep that in mind.