In a generic election I would agree with you, but the Obama camp did such a good job framing the auto bailout in places like Ohio, Wisconsin, and Michigan that I feel like he had more of an inherent advantage there than in Pennsylvania. And the kind of affinity the bailout created - the kind where you are sort of indebted to the president for your continued livelihood - makes the voters' persuadability extremely low, meaning Romney would be less of a threat there unless he really pumped the money in. Pennsylvania was a bit more open to persuasion in 2012.
I guess 2012 could be the exception but I just think that Pennsylvania is fools gold for Republicans for the time being. They always float it as a place where they could "expand the map" but always lose by big margins.
Not to nitpick, but even with the factors you mentioned Ohio was still much closer then Pennsylvania was it not? I just don't see how even with said factors Pennsylvania is a "better bet" for Romney than Ohio/Wisconsin. Don't get get me wrong there was a bit of a panic on the Democratic side about Pennsylvania in October because Romney hadn't been defined there yet, but it turned out to be hot air. IIRC Ohio was the second closest state in the Union next to Florida on election night, so isn't it still more plausible that Romney wins Ohio (still losing the election, mind you) then Pennsylvania in some huge upset?
I definitely see why Ohio seemed like more of a swing state due to the closer margin on election day, but that number doesn't always signify how safe or unsafe a given state is. There are three factors that determine the election - voter persuadability, affinity, and likelihood to vote. Most of what I or others do involves quantifying these qualities and devising strategies to move them in a favorable direction.
Persuadability - how easy/hard is it to convince a voter to change their vote
Affinity/Support - which way the voter is leaning already. Usually measured on a 1 - 5 scale, with 1 being Strong Support and 5 being Strong Opposition. Some campaigns use 10 point scales or more complicated algorithms for this. Local campaigns can get away with a 5 point scale.
Likelihood - how often does this person turn out to vote? This is determined by looking at their voting record. Voting in 3 out of 3 past elections would make you highly likely, 2 out of 3 likely, and 1 out of 3 sporadic. It gets more complicated when factoring in mid-terms and the like, but that's irrelevant for now.
Anyway, I'm saying all of that to make this point: a state is more favorable with a) 52-47 support split and b) extremely low persuadability than it is with a) 60-40 support split and b) extremely high persuadability.
After those two are pushed as far as possible, your GOTV operation ensures that any issues with your supporters' likelihood to vote are irrelevant. Door knocks and phone calls aid with pushing their likelihood way up.
In 2012, Ohio had a) pretty narrow supporter split in favor of Obama, b) low persuadability, and c) a way better field/GOTV operation that Pennsylvania since it is "the" swing state and Obama had invested heavily there years in advance.
By that analysis, PA was more up for grabs than OH by quite a bit.
Put more simply: it would cost far less money to persuade 10% of Pennsylvanians than it would 2% of Ohioans.
Edit follow-up: it was about a 150,000 vote margin, I think, and those 150,000 votes were extremely secure; moreso than the margin of voters in PA, if that makes sense. Looking just at how small the margin is can be deceptive; think Karl Rove insisting that the Romney campaign had plenty of room to win Ohio when all the other analysts were just like "...yeah no." The likelihood of the votes and the partisanship of the voters was much higher than the folks Romney was counting on to cancel them out at the district/precinct level.
