Via Jim Geraghty I find that Michele Bachmann is playing fast and loose with the “statistical dead heat” terminology.
It’s bad enough when people pretend that the Margin of Error is a band of uniform probability, and not a bell curve where the extremes are unlikely. But what Bachmann’s email is saying is beyond belief.
Polling is not a difficult science to understand. It’s difficult to do well, but no in-depth knowledge of statistics is required to read and understand a poll. Just basic probability is all that’s needed. Once you can visualize what a probability density curve looks like, you know all you need to know in order to read a poll and get an idea for what it’s actually saying.
It pains me when a major Presidential campaign at best hopes the candidate’s supporters will buy this, or at worst the team doesn’t grasp it themselves.