Don't Try This At Home

It is often dangerous when an idiot tries to interpret a poll. Sometimes that idiot happens to be the writer for the newspaper which commissions the poll.

"Poll results show the backlash from the [Wellstone memorial] service, which was broadcast live on radio and TV, may make its mark on the election's outcome. [Para.] Nearly a quarter of the 929 likely voters said the service made them more likely to vote for Coleman, while 16 percent said it made them more likely to vote for Mondale. An additional 53 percent said the service will make no difference in how they vote."

Does this mean the service changed one vote? Of course not. The "nearly a quarter" of voters who said the service made them more likely to vote for Coleman could all be Coleman supporters already, since Coleman already has the support of 41 percent of voters. Of course, Coleman supporters would say that the service made them more supportive of Coleman. How many of the 929 likely voters switched their vote from the DFL candidate to Coleman? Apparently the question was never asked.

(The article does quote one man who's going to switch from Mondale to the Indpendent candidate, Moore, which benefits Coleman, but doesn't make the man one who's "more likely to vote for Coleman.")

And Mondale wasn't in the race before the service, so how could 16 percent say the service made them "more likely to vote for Mondale" than before the service? No one was going to vote for Mondale before Wellstone died. Many of the respondents probably understood the (apparent) intended meaning of a poorly-written question, but those who answered the question as asked could not have said they were "more likely" to vote for Mondale.