re: #45 The Mongoose

Hi Charles,

Pew’s previous accuracy is part of what has me so flabbergasted at the horrific skew in this poll. I’m combing through the entire poll (your link appears to be the HTML version of the PDF I’m reading) and I see nothing to indicate that they’ve normalized for the D/R/I being so far off. Correct me if I’m wrong but they’re showing D+9 among registered voters and D+7 among likely.

Gallup, Rasmussen, and AP/GfK all have this within one point. SurveyUSA has Obama up 3. Pew is wrong.

I’m not sure you really understand this polling thing.

FIrst of all, the numbers you put up for the D/R/I divide don’t add up. They consistently have more I than D. Can you please cite where you’re getting your number from?

Second of all, if this is the result that Pew got, than this is the result they got. It would be ‘wrong’ for them to adjust the polls because it’s an outlier. If you can point to a problem in their methodology, that’s fine— and there is definitely a big discrepancy between R and D affiliation shown in this poll— but to say the poll is wrong because of its results is to not understand what a poll is. A poll isn’t right or wrong, a poll is a collection of data from questions people have answered.