Yeah, that was definitely interesting.
I just read through most of this Brookings paper about the political geography of Pennsylvania:
The Political Geography of Pennsylvania: Not Another Rust Belt State
Fascinating stuff. 28 pages of info about the regions of the state, what characterises them in terms of labour market, educational profile, racial make-up and of course voting trends, and what the trends look like. Co-authored by Ruy Teixeira, of "Emerging Democratic Majority" fame. Most striking finding: the Democrats have been doing well in the parts of the state that are still, or again, dynamic and relatively upwardly mobile, and badly in the most stagnant/declining parts.
What surprised me is that, on the one hand, the general patterns in voting trends can be easily correlated to the educational/professional and age/race make-up of the regions, right - that's basically at the root of the whole paper. But the paper also reveals some striking differences in voting patterns or trends between regions that are otherwise quite or wholly similar. That's the only thing I was a disappointed about; that there was little analysis about how come.
For example, the Allentown/Scranton/Reading region in the northeast and the Harrisburg/York/Lancaster region in the south/southeast are by any standards of the given tables similar. Same racial make-up, same educational make-up, roughly the same labour market. Both, overall (Scranton excepted) fast-growing regions. Yet the Harrisburg/York/Lancaster region is solidly Republican, while the Allentown/Scranton/Reading region is fairly to strongly Democratic, depending on whether you look at presidential or congressional elections. Both are distinctly trending Democratic, but the Allentown/etc region more rapidly, so the difference will only grow. Why? No idea.
By ways of another example, both areas are still largely (44%) blue collar, with a quarter actually working in manufacturing and production; low-income, with 60% earning less than $35,000; and have a solid majority without college. Both are overwhelmingly (90%) white, and have a similar share (20%) of 65+ voters. All of this also goes for the West of the state (minus Pittsburgh itself). Pittsburgh's suburbs and exurbs and neighbouring towns, the county that's part of the Youngstown/Warren agglomeration across the border in Ohio, and Erie, together make up a region with the same demographic profile, just a little bit more so still (95% white, 60-70% under $35,000). The only difference is that this area is losing residents while the ones in the east are gaining them.
But the political difference is stark. The Western region used to be a Democratic bulwark, voting for even Dukakis by >10%, but is strongly trending Republican, and voted for Bush in '04. The Allentown/Scranton/Reading and Harrisburg/York/Lancaster regions in the East, meanwhile, are trending Democratic, in part strongly so. The Allentown/etc area voted for Bush Sr. over Dukakis by >10% and still preferred Dole over Clinton, but broke even between Bush and Kerry. Why the opposite trend? I dont know.
Anyway, after reading about all of that I thought I'd gotten something of a feel for the state's regions, Cyclo, and then I clicked on your link. Whoa! I'd still never expected that the regional differences within the state would be SO pronounced. No surprise that Pittsburgh's suburbs go for Clinton, but by 68% to 26% in at least one of those two polls? And the northeast, that Allentown/Scranton region, which you'd expect to be a little less Obama-hostile (what, with the growing economy and population there), 65-69% for Hillary? Dayumn!
He'd better do
real good in Philly and its suburbs, score wins in the city of Pittsburgh and Centre county (where State College is), and keep things close in Erie and some of that Harrisburg/York/Lancaster region...