Anyone else hate the opinion poles? I can't stand them and with the political season cranking up since the Republicans have now virtually settled we are going to start seeing them dominate news all over.
So tonight ABC started talking about a new pole that surprised them. They did a pole to find out which candidate the public believed in more, the result was that Romney had a slight edge (something like 2%). The thing that surprised them is that just last week they did a similar pole with the result giving Obama a 12% edge.
What surprises, and annoys, me is that they were surprised by it! I thought that I had written about it before but I can't seem to find the link. The problem is with the poling process itself. They pole some small sample in one city and expect that to represent the entire population. How is poling 10k people in New York City going to represent opinion of the entire population of the United States? Come on, you have offices in every major market in the country! Have each one of those offices pole 10k people and maybe you will have a better representation. The other issue with poling just New York City is the people that you are poling, think about it. Even if you stand in the same place at the same time a week later you will probably talk to 9500 different people than you talked to the previous week. So how accurate are your results?
And that question leads to the remainder of the problem I have with poles. They always state somewhere on the screen that the results could be plus or minus 10%. What that does is give them a 20% variance in their results. I didn't take statistics in college but that is a pretty big variance! That is almost as large a variation as the weather predictions, but that's a different rant...
No comments:
Post a Comment