Yes, in fact, a presidential sample more Republican than the recall electorate is an odd result.
It wasn't that odd in 2004 when it was R+3. That would mean the presidential sample was more Republican that year than the 2012 recall electorate.
Dane County put up extremely high turnout in the 2012 recall (Madison was even higher), and Milwaukee County was right at the state average. To pretend that there is some hidden reservoir of votes is to ignore facts. With the exception of Menominee the lowest turnout counties were Walker Counties.
Morning in America has to go back 8 years to find a comparable electorate. That says all we need to know. There are tons of voters who only vote in presidential elections.
I'm not in a place to run the number, so help me out. How much higher was this "extremely high" turnout in Dane County in the recall vs. the presidential election in 2008? How many tens of thousands of votes over the presidential election in 2004?
I use the voter registration statistics posted here.
http://gab.wi.gov/publications/statisticsDane County, 2012: 253k votes cast
Dane County, 2012: 345k registered voters
Dane County, 2012: 73% turnout
Within Madison City, jsonline indicates that turnout was 5-6% higher than Dane County at large. I don't know the exact figures.
Milwaukee County, 2012: 394k votes cast
Milwaukee County, 2012: 549k registered voters
Milwaukee County, 2012: 72% turnout
Wisconsin, 2012: 2511k votes cast
Wisconsin, 2012: 3453k registered voters
Wisconsin, 2012: 73% turnout
Waukesha put up comparable figures to Madison city at 79% turnout.
I would love to find registered voter counties from 2008 to see comparable figures. The problem was that the GOP base was not energized in 2008 and McCain performed extremely poorly in Waukesha and neighboring counties. How do you think they will perform with a local hero on the ticket?