follow us in feedly

Friday, February 12, 2010

Wording and Survey Design

It doesn't take much to make a poll on the same topic give very different results. Case in point:

I don't have the "n's" for the 2 polls, but it doesn't take a rigorous test to see that these differences are very unlikely to be random sampling error.

Monday, February 8, 2010

Economist Humor

Blog Archive