You know, if it weren't for this race I'd have no excuse to write diaries. Thanks, Rick.
Allentown Morning Call/Muhlenberg College. Sept. 18-25. MoE 4.5%.
Head-to-head
Bob Casey 37%
Rick Santorum 29%
Undecided 31%
If those numbers seem strange to you, you're not alone. More on the flip side.
The last two polls recent enough to be comparable were
this from Keystone and
another from Strategic Vision (R). They were conducted from Sept. 8-15 and Sept. 10-12, respectively, and showed
Casey 50% - Rick Santorum 37% - Undecided 13%
and
Casey 52% - Santorum 38% - Undecided 7%.
Notice the differences in those numbers. Both Keystone and Strategic Vision show Casey and Santorum to be at roughly 10 pts higher than Muhlenberg, and the number of undecideds are both far, far lower in these polls: 7 and 13 as compared to Muhlenberg's 31.
With the exception of a Keystone outlier from March, most of the polls on this race have been fairly consistent - they've generally shown Casey to have the same lead (all three July polls showed an 11-pt gap between he and Santorum), and they've all shown a steady progression from a low-single to low-double digit lead for Casey. It's alo worth noting that none of the outfits watching this race have reported a number of undecided voters as high as Muhlenberg's.
Obviously, Muhlenberg has very different methodology, especially when assigning undecideds. It's probably this difference that explains Casey's lower lead in the this survey. Muhlenberg's internals aren't available yet, but when they are I'll update with further analysis.
And finally, Tim Tagaris of SwingStateProject has taken a look at these numbers, and came to a very different conclusion: namely, that every other firm is wrong and Casey is slipping. He specifically says that "Bob Casey's lead over Rick Santorum is down to eight points."
I have one question for Tim: Down from what? This is the first time that Muhlenberg has surveyed this race, so how can Casey's lead be down from anything?
Unless of course, he's trying to mislead readers into thinking that every polling firm is identical, and that their numbers can be mix-and-matched at will. But he would never do something so dishonest.
Update 1: In my haste to write up the h2h numbers, I forgot to mention that the Morning Call comes to the same conclusion that I diaried a few days ago: tying Santorum to George W. Bush is a winning strategy. My favorite tactic is to keep Rick talking about Bush's Social Security scam, but suggest your own in the comments.