Polite Dissent

As I was grabbing a quick breakfast of Rice Krispies in the hospital this morning, another physician came into the doctors’ lounge and turned the TV to CNBC. One of their reporters was doing a piece on whether Bush’s speech last night swayed any undecided voters. To determine this, he gathered a sizable group of undecided voters and sent them to a “speech party” that a local Republican was holding. Afterward, he interviewed four of the attendees. The first one said she was leaning towards Kerry. The second one said she had firmly decided on Kerry. The third interviewee said that he was leaning towards Bush. The final person interviewed, who was the Republican who hosted the party, said that he was firmly in the Bush camp. The reporter then summed up by saying that his “informal poll” showed that undecided voters were “split evenly” between Bush and Kerry.

• Informal poll or not, interviewing 4 people is not enough draw any conclusions.
• The reporter did not interview 4 undecided voters, he interviewed three undecided voters and a member of the Republican Party, invalidating his entire premise.
• You could also argue the point that having the undecided voters attend a Republican speech party was not a neutral environment and could skew the results as well.

Now I’m not picking on Republicans here. I’ve seen both major parties frequently skew statistics. It’s not just the big two either; the Reform and Green Parties handle statistics just as poorly.

CNBC should know better though.