I was looking at today's polls and was intrigued by the Public Policy Polling results from Wisconsin that showed Obama ahead of Romney by 52% to 45%. This is quite a big swing from the last poll in that state by the same organization that showed Romney ahead by 48% to 47% one month ago. Of course, since Public Policy Polling always seems to skew its results towards the Democrats, I decided to look a bit further. It did not take long to find that in August, PPP polled 3% more Republicans and 3% fewer Democrats than it did in September. Since both Republicans and Democrats are supporting their nominees by better than 95%, that sample change means that there would be a 3% swing in the August results for that reason alone. In other words, by changing their sample, PPP took results that showed a Romney 48 to 47 lead and made them an Obama 50 to 45 lead. In that context, the results today of 52 to 45 are alomst the same as a month ago.
There is no easy way to know the proper breakdown between Republicans and Democrats in Wisconsin. Party registrations are at best approximations. Probably the closest figure to use is the party breakdown in the exit polls from the recall election last June. Under that test, the latest PPP poll has 4% too few Republicans. In other words, the figures in August were close to the recall numbers, while the latest figures are substantially skewed towards the Democrats. Were this another polling organization, I would be prepared to believe that the skew is inadvertant. With Public Policy Polling, however, I have to suspect that this is all intentional.
How close is Wisconsin? Obama is probably ahead by 2%. The PPP poll is just garbage.
No comments:
Post a Comment