| There's been a whole lot of OMG! today online about a new polling memo from Republican pollster POS on behalf of American Crossroads, the new one-stop-shopping Karl Rove 527 emporium for attack ads, message testing, and the like. Taegan Goddard, in particular, seems to have fallen prey to the way the memo is deviously worded, titling a post "Senate Up For Grabs?" and saying the poll "shows Republican U.S. Senate candidates averaging a high single-digit lead over their Democratic opponents in 13 states with close races -- suggesting Democrats might lose control of the chamber in this fall's elections."
Um, yeah. Except that's not 13 Democratic-held seats. The poll is of five GOP-held seats and eight Dem-held seats, but without topline numbers for any seat. It's a poll of smaller-than-useful samples in a bunch of different competitive races, globbed together into bigger numbers. The numbers still seem outlandish at first, with a 45-37 advantage for the GOP in the GOP-held seats, and a 47-40 advantage for the GOP in Dem-held seats. But then, looking a little deeper, you notice that those Dem-held seats include Arkansas, Delaware, and Indiana, where the GOP currently has large leads, tipping the scales' balance against the Dems in those other Dem-held seats.
To unpack this, I looked at today's Pollster.com rolling averages of each of these races, and averaged those out within each category to see what it looks like. For instance, for their five GOP-held seats, you have FL-Sen (using Meek vs. Rubio) at 15-32, KY-Sen at 43-45, MO-Sen at 44-48, NH-Sen (using Hodes vs. Ayotte) at 39-48, and OH-Sen at 41-43. That averages out to a 36-43 deficit for the Dems in those races altogether, not much different from POS's 37-45 figure. And in the eight Dem-held seats, there's AR-Sen at 34-57, CO-Sen (using Bennet vs. Buck) at 43-45, DE-Sen at 36-50, IL-Sen at 42-41, IN-Sen at 31-50, NV-Sen at 45-43, PA-Sen at 41-45, and WA-Sen at 49-47. That averages out to a 40-47 GOP advantage... remarkably, the exact same number that POS comes up with.
So my response to all this is: so what? If you go by the Pollster.com averages (and I wouldn't necessarily, as these are heavily determined by Rasmussen numbers and rely heavily on the large disparity between Rasmussen and other pollsters in OH, PA, and CO), they show Dems poised to lose six seats (throwing in ND-Sen, so far gone that they didn't bother polling it)... and gain one, assuming Charlie Crist dances with them that brung him and caucuses with the Dems... which would be a Rasmussen-style-worst-case-scenario net loss of five. Which is simply no different from what any other prognosticator, including us, would expect. So Karl Rove's The Math (TM) looks like it may actually be correct for the moment... but is it telling us anything that we don't already know?