Friday, January 15, 2010

beware of pollsters when results can make big news and big money

The internet is buzzing with the Suffolk poll that shows
Brown leading Coakley in the Massachusetts senate race 50%-46%.

I put this result in the Rasmussen category of "why waste the opportunity to make a good headline". However, unlike Rasmussen, sometimes Suffolk just gets things plain wrong. I remember back in the 2008 Democratic Primaries that their final results sometimes stood out from the others. This leads me to think that they are more honest than Rasmussen, who seems to get really close to the consensus by election day.

This pre-election polling is a tricky business because it is really two different models:

1. pre-election - there isn't any accountability, so the pollster can shade numbers, samples, etc. to fit the best narrative.
2. final survey- people judge their last poll with the results and determine their credibility

No comments: