This fact perfectly illustrates my point. The Republican pipe-dream on PA is that it is so close that it will flip if the election is close, or if they win the popular vote. It's simply untrue, and has been proven so. Even when they win the popular vote they still lose PA.
And your (and the Republicans') continued faith in the "rust-belt", (which basically means Appalachian whites, right?) is misplaced. PA has about 12 million people. The Philly-metro area accounts for around 6.5 million of that, and the Pittsburgh-metro area contributes another 2.5 million. The T only has 3. So even with 50% turnout, the urban 9 million outvotes 100% turnout in the rural 3 million by a mile. That is why PA will not flip, the "rust-belt" just doesn't have the warm bodies.
The other reason(s) I am so confident is anecdotal. I have a ton of family in OH and western PA (Pittsburgh) and I've spent my entire life driving across that state, visiting small towns, crossing the Alleghenies, driving through the tunnels, staying in hotels, eating at small town restaurants, sleeping at rest stops, listening to the truckers' CB chatter, etc. I have probably been to as many or more PA towns than the towns of any other state, including states I've lived in. I have also as I've said, been to Pittsburgh countless times, and I lived in Philly for many years.
Based on that experience, I'm telling you... there just aren't enough people in "upstate" PA ("the T") to overcome the MASSIVE populations of Pittsburgh and Philly... PA is not like an east coast/Atlantic state where the cities are surrounded by smaller cities and endless moderately populated suburb after suburb. The state is Philly, Pittsburgh and a lot of endless uninhabited mountains, farmland and forest, and not much else. The Republicans will win the T as they always do, but there's hardly anybody out there, so it won't flip the state under any circumstances. Yes to the second thing, but that's normal for Republicans. To the second thing... No, for the reasons I've already given. In short, there just aren't enough people in the T to overcome the metropolis.
I looked at the data from the last four elections, and it doesn't really support this. Defining Philadelphia's metro area (within PA) as Philadelphia, Bucks, Montgomery, Chester, and Delaware Counties, I found that only 34.1% of the state's total votes in 2012 lie within that metro area. The Philadelphia metro area is slowly becoming more Democrat leaning and slowly growing relative to the rest of the state, but the rest of the state (including the Pittsburgh area) is on average becoming more Republican even faster. Over the past four elections, the net Republican shift of the rest of the state has outweighed the relative growth and slight Democratic shift of the Philly area, so that the state has become slightly more Republican compared to the national average over the past four presidential elections.
The relative margins statewide were 3.66%, 4.97%, 3.05%, and 1.52% more Democratic than the nation as a whole in 2000, 2004, 2008, and 2012 respectively. A crude extrapolation of the vote trends in the Philly area compared to the rest of the state along with its relative growth would predict it to go 1.33% more Democratic than the nation in 2016, and if anything this may be an overestimate, given that it seems Trump is unusually popular among Rust Belt voters.
"Rust Belt" as a term is usually used to refer to the western part of the Northeast and the eastern part of the Midwest. Roughly, it's everything west of the BosWash megalopolis, east of the Mississippi, and north of the Ohio and Potomac except that West Virginia is often included. It's the old manufacturing heartland of the US, most of which has suffered a slow and grinding decline with deindustrialization over the last few decades. All of PA except the Philly metro area and maybe some immediately adjacent areas would be included in most definitions, I believe. I'm talking about all of these voters including urban ones, not just the rural areas of Appalachia. Of course the people Trump has been winning over have been nearly all white, so when I talk about voters swinging I'm implicitly talking about the white ones, but this includes white voters in Pittsburgh and its suburbs as well as in smaller cities.
Here is the graph. It's four crude Excel plots of only four points each, so I wouldn't pay very much attention to the equations of the trendlines or their correlations. The top line is Philly metro vote share as a percent of the state total, the upper middle one is the average margin in the Philly area compared to the nation as a whole (the (D-R) percent for the region, minus the (D-R) percent for the nation), the bottom one is the same margin for the rest of the state, and the lower middle one is that margin for the whole state. You can see how the non-Philly line dips rapidly more negative (R-leaning) over the past two cycles, as the other two D-leaning trends pull the average up more slowly, so the average trend is slightly Republican.
As you can tell, I'm thinking in an almost purely data-driven way about the election, partly because I just like stats and partly because this approach has been pretty successful so far with US presidential elections. I don't doubt your anecdotal experiences, but I don't put a whole lot of weight on anecdotes (including my own) when there are stats to play with.
