Monday, November 3, 2008

Do Polls Make the News?

A properly done poll reflects the attitudes of the population by drawing inferences from a sample of that population.

An improperly done poll reflects the biases of the people who operated the poll.

When we have polls with such a wide disparity in results, you can ask whether this is simply differences in method, but when you look at some of these polls closely, they seem to be intentionally skewed.

Polls that oversample one party, polls that ask leading questions, and polls that ignore known polling issues (e.g., Republicans are always under-represented in weekend polls), are published as if they were perfectly valid polls. Why?

Are the polsters -- or the people reporting these polls -- trying to shift opinion? In the current presidental election, are they hoping to depress turnout on one side by making the other's win seem assured? Or are they trying to excite them to action by reporting that they're behind -- or slipping? Perhaps they're trying to create a bandwagon -- get the undecideds to join the winning team?

I think most people these days know they need to take polls with a grain of salt. We also need to steel ourselves to not be swayed by the polls. Let's do our best to keep the polls from making the news.

No comments:

/* -------------- -----analytics code */