Here is the Fall 2011 poll report:
http://wpr.org/announce/survey1111/2011f-survey-1.pdf
Here is the Spring 2011 poll report:
http://www.wpr.org/announce/survey1104/survey1104recalls.pdf
Here are my criticisms:
1. The sample size is rather small (400-500) and is from a “random” telephone sample. Unlike the larger national polls, the St. Norbert college poll doesn’t appear to document how “random” the poll actually was. There is no mention of what percentage of the respondants were republican, democrat or independent.
2. As there was no measure of the randomness in the poll, the pollster was free to take its “random” samples in any way that they wanted to (i.e. they were free to skew the results). For example, they could have focused their “random” sample on the predominantly democrat Madison area. The report doesn’t indicate otherwise.
3. I’m not sure what this artifact means, if any: The spring poll was made of landline phone owners. The Fall poll was made of landline+cell phone owners.
Didn’t they just tried to recall a whole bunch of state senators and lost?