Fun Fact: PA is Just as Much of a Swing State as OH, IA, or CO

[Torie]

[Torie]

We've seen Pennsylvania's PVI trend rightward before, it really doesn't mean as much as people here are making it out to be.



In fact PA's PVI has shifted rightward three times over the last 50 years and it hasn't crossed the threshold once.

It seems like the main reason it goes right is that the NATION's vote shifts to the left faster than PA does.    It would explain why the nominal vote totals in PA have favored Democrats in the last few elections (2004 vs 2012) despite the PVI moving away from the Dems.

What exactly does the bolded language mean please?  It makes no sense to me. In 2000 and 2004 the nation shifted rightward, and PA less so. In 2008 the nation sifted leftward and PA less so. In 2012, the nation shifted rightward, and PA more so. I fail to see any pattern of elasticity or whatever your point is. More to the point, is whether or not the trend to the Pubs in the last two elections is a blip or representative of a longer term trend. Charts won't help you with that.

[Torie]

The problem with that notion is that Pennsylvania has areas that trended Republican, but also areas that trended Democratic or otherwise have not been elastic or are just static. Southwest Pennsylvania's trends are about maxed out at this point and it's not a large enough region to swing the state Republican enough to flip. The areas that have trended Republican are also declining in population overall.

The Philly burbs trended a couple of points to the Pubs in 2012 and that area is growing disproportionately in PA. Ditto for Lancaster and York Counties. On what basis are you asserting that the Pub trend in Southwest PA has maxed out? Is Cambria County in SW PA by the way?

So many crystal balls here, none really data based, so little time. Is this a syndrome of youthful enthusiasm or what?

[Torie]

The problem with that notion is that Pennsylvania has areas that trended Republican, but also areas that trended Democratic or otherwise have not been elastic or are just static. Southwest Pennsylvania's trends are about maxed out at this point and it's not a large enough region to swing the state Republican enough to flip. The areas that have trended Republican are also declining in population overall.

The Philly burbs trended a couple of points to the Pubs in 2012 and that area is growing disproportionately in PA. Ditto for Lancaster and York Counties. On what basis are you asserting that the Pub trend in Southwest PA has maxed out? Is Cambria County in SW PA by the way?

So many chrystal balls here, none really data based, so little time. Is this a syndrome of youthful enthusiasm or what?

The word is spelled crystal. With that said, with Southwest Pennsylvania losing population, that doesn't leave much room for significant growth to swing the state. Lancaster and York have trended Democratic from where they were in 2004, and the trends in 2012 in the suburbs towards Republicans were not enough to flip the state. Considering that there are lot of moderates in Philadelphia's suburbs and Republicans have moved further right, those areas are going to remain at best static for Republicans against a conservative candidate for President.

That's nice. Still no actual data however to back up your asserted theories that the recent Pub trend is PA is coming to a halt.  And I don't blame you for not adducing any data, because frankly dear, there is none. It's all speculation. Thanks for catching the typo by the way.

[Torie]

OK. So based on four data points, and getting rid of one of the four that proves inconvenient, so we label that a noisome "outlier," and toss it in the dust bin, we find that PA does not swing overall at all much, be damned what the nation is doing, because it is a law of Newtonian physics, that in PA, while internally, things are swinging all over the place, the net effect must be close to a wash, sort of like the speed a light is constant, or where there is action in one place in PA, there will be nearly an equal opposite reaction elsewhere. Some might call that "data mining."

Again, I don't mean to single you out as a particularly egregious perp in this regard. Date mining is the secular religion of loads and loads of posters around here, and I am using this topic as an excuse to explore and expose this little phenomenon that so many of us have made into an almost obsessive hobby. Awareness of ignorance is the first step to knowledge. It's tough, and hard to do, but the first step is to say, hey, I have no idea really what will happen next. There just isn't any model of any statistical significance, with a reasonable explanation to back it up, that allows for any such extrapolations, that is other than GIGO.

So let's both just say it together, both of us, in unison - we just don't know what will happen next in PA. It's all speculation.  There, we did it. Don't you feel better now?

[Torie]

Seems to me, and I've said it before, PA tends to vote reliably a little bit Democratic, and then on top of that, a little bit against the incumbent President's party, if there's an incumbent running for re-election. Look at 1992 through 2012 and tell me this isn't consistent with the data.

That was true for 1996 (the Perot factor kind of messes up the data for 1992 and 1996), 2004 and 2008. It was not true for 1984, and 1972 looks flat. I guess it was also true for 1976 if one considers Ford an incumbent. Again, however, there are far too few data points, and each election has its own dynamics, with a ton of other factors in play other than incumbency, to draw any statistical significance from all of this, to wit, the PA voters are a bit grumpy, and unhappy, and thus tend to consider incumbency at the margins a negative, or less of a positive,  relative to the nation as a whole.

I think we need Muon2 to give us all a seminar on the topic of statistical significance, with Jimtex as his assistant. Humility is a virtue, hubris a sin. And the more you know, the more you know how much you really don't know. I'm old enough to finally know, that I'm just butt ignorant in general. I only predict with much confidence at all when drunk or stoned, or both,  typically at this point.

[Torie]

OK. So based on four data points, and getting rid of one of the four that proves inconvenient, so we label that a noisome "outlier," and toss it in the dust pin, we find that PA does not swing overall at all much, be damned what the nation is doing, because it is a law of Newtonian physics, that in PA, while internally, things are swinging all over the place, the net effect must be close to a wash, sort of like the speed a light is constant, or where there is action in one place in PA, there will be nearly an equal opposite reaction elsewhere. Some might call that "data mining."

Again, I don't mean to single you out as a particularly egregious perp in this regard. Date mining is the secular religion of loads and loads of posters around here, and I am using this topic as an excuse to explore and expose this little phenomenon that so many of us have made into an almost obsessive hobby. Awareness of ignorance is the first step to knowledge. It's tough, and hard to do, but the first step is to say, hey, I have no idea really what will happen next. There just isn't any model of any statistical significance, with a reasonable explanation to back it up, that allows for any such extrapolations, that is other than GIGO.

So let's both just say it together, both of us, in unison - we just don't know what will happen next in PA. It's all speculation.  There, we did it. Don't you feel better now?

Torie, the real question here is: Are there enough Republican votes in Pennsylvania to beat the Democratic vote there? Is 2.7 million votes max for pubs, in which case the only way pubs can win there is if dems fail to turn out.

Based on a few data points, as to which there are a plethora of different variables in play with respect to each of the data points, there are 2.7 million Pub votes in PA in the same way that the speed of light is 186,000 miles per second? The law of physics here is that no matter what the total vote total in PA, the Pubs get about 2.7 million votes, albeit they often are quite different voters, with their subtotals bouncing around in various parts of the state, but it all nets about to about 2.7 million, and so the only real variable is how many who don't vote Pub bother to go the the polls at all. Makes sense to me!

[Torie]

OK. So based on four data points, and getting rid of one of the four that proves inconvenient, so we label that a noisome "outlier," and toss it in the dust pin, we find that PA does not swing overall at all much, be damned what the nation is doing, because it is a law of Newtonian physics, that in PA, while internally, things are swinging all over the place, the net effect must be close to a wash, sort of like the speed a light is constant, or where there is action in one place in PA, there will be nearly an equal opposite reaction elsewhere. Some might call that "data mining."

Again, I don't mean to single you out as a particularly egregious perp in this regard. Date mining is the secular religion of loads and loads of posters around here, and I am using this topic as an excuse to explore and expose this little phenomenon that so many of us have made into an almost obsessive hobby. Awareness of ignorance is the first step to knowledge. It's tough, and hard to do, but the first step is to say, hey, I have no idea really what will happen next. There just isn't any model of any statistical significance, with a reasonable explanation to back it up, that allows for any such extrapolations, that is other than GIGO.

So let's both just say it together, both of us, in unison - we just don't know what will happen next in PA. It's all speculation.  There, we did it. Don't you feel better now?

Selectively using PVI and ignoring all other data points to show a Republican Trend is MUCH MUCH more of an example of data mining than anything I posted.

I just drew a frigging PVI chart, and then extrapolated the line based on its recent slope, and the did the same for the trend line, and then said I don't have a clue what will happen next, and it's all speculation, and you call that data mining? OK, whatever. But yes, I do consider PVI to be more salient a number as to any given election, that that the Pubs get about 2.7 million votes no matter what, or that PA really doesn't swing much, because it doesn't care what the rest of the nation thinks, or that it hates incumbents, yes, I do admit that. But just because changes in PVI is more salient, does not mean that it has much value about predicting the future, because it really doesn't.

[Torie]

In the long run, if Pubs don't start doing better with persons of color, ultimately PA will have a nice big Pub PVI, and Pubs will lose every election. In the long run, we will all be dead. Meanwhile, in the short run, there is an election in 2016. In recent times, I suspect much of the fall off in population in Western PA, is due to dying or departing Democratic voters. So if that continues, population can continue to fall (at least as a percentage of the population of the state), and the Pubs get bigger absolute margins in that area. White voters in the Philly burbs in most areas there had one of the biggest trends to the Pubs in the nation for whatever reason in 2012.

Speculate away. I have totally failed in deflecting you from that penchant, admit defeat, and now wave the white flag in surrender. We shall see if the Pubs get about 2.7 million votes in 2016, or lose by 200,000 votes, or the PVI trend of late stalls, or reverses, due to a lack of a Dem incumbent to lash out at, or whatever your metric of choice might be, leading to inevitable Pub defeat in PA, and/or a Dem PVI not smaller than what came before, or at least some Dem PVI (the latter one of course having the highest probability of all of the choices listed)- yet again. Yes,  there is a good chance one of the above will be true! Just as there is a good chance it won't be true. Be well.

[Torie]

And you can go back to reading the National Enquirer, hoping that Hillary dies so Republicans have a better chance at winning the election. I'm not going to getting a flame war about Pennsylvania, I'll just bookmark this thread for next year and settle the beef then.

I posted that 1) I hope Hillary does not die, and 2) I hope the article is false. Did you read that? But as to your second sentence, great idea!

[Torie]

This thread is great. Top thread of the week.

That says enough about the lowly quality, overall, of threads nowadays on Atlas Forum.

The Forum Democratic Hive Mind is to blame for this.

The main cause is all the objectively awful blue avatars. The centre-left members are much less active than we used to be.

That is where the "quality posters" are drawn from?  I feel just so left out.   And then there are all those other non center left posters who often post good stuff, that are apparently just chopped liver too.

Anyway, I think this board has improved myself, but that's just me.

[Torie]

Yeah - but given what you've been posting recently, and its occasionally shrill tone, you would think things have improved.

Ouch!

[Torie]

That is perhaps why PA chat gets so contentious.

[Torie]

Following up on what you said TJ, I assume that you have seen this quite interesting article, which asserts that there are a lot of unregistered voters out there in the Pub friendly/trending hinterlands of PA, as compared to say Philly (where the Dems do an outstanding job getting most breathing Dem friendly voters to the polls), and that rather than just dump advertising dollars on the state, what the Pubs really need to do, is what Obama did in 2008 so well - actually change the composition of the electorate so that it moves a bit in the Pub direction. So armies of folks need to wander around Johnstown, or Somerset County, or wherever, and register, register, and register some more - and then get them to the polls.

I noticed the same thing, when messing around with maps in VA over the redistricting lawsuit there (gosh I just love redistricting law - it's so complex, and f'ed up, that it is just gets me really hot and bothered trying to decipher it, and predict it). The turnout in western VA is about 5 points below the turnout in black precincts. Those working class white voters there in the hills and valleys near the TN and KY and WV border, just don't vote in the kind of numbers that suggest any concerted effort to register them and get them to the polls. If the Pubs want to get back in the game more in VA, they need to get those folks registered, and voting! The Dems have done it with their base. The Dems know how to play the game, and work hard on it, while the Pubs sit on their butts, and bore folks to death chatting about their interpretation of the Constitution that SCOTUS has rejected or something.  That is not the way to win elections. Who knew?

[Torie]

Look up the definition of "fact." Pennsylvania could be competitive, but it's absurd to assume that it will be as competitive as OH, IA, and CO. Evidence that PA is trending Republican is mixed at best, and I see no way that it will be critical to the electoral math. If Republicans win PA, they've already won easily. It's certainly not more competitive than FL.

PA has a D+1 PVI. It's called a safe blue state.
NH has a D+1 PVI. Rated Toss-up.
CO has a D+1 PVI. Rated Toss-up.
IA has a D+1 PVI. Rated Toss-up.
OH has a R+1 PVI. Rated Toss-up.

I know PVI isn't everything, but do you notice something?

I think the perception is largely based around Pennsylvania bordering Democratic strongholds like NY, DE, MD, and NJ.

Then there's no reason to assume that NH (which borders the most Democratic states in the country) is a Toss-up either (which I don't think it is, btw).

I do find it rather odd however, that you have such a different attitude towards NH vis a vis PA. I mean NH trended to the Pubs too in 2012, albeit by not as much as PA did. I tend to agree that perhaps PA has more potential for the Pubs than NH, but that is based on untapped potential as discussed above, that needs to be tapped to make a difference, and given the chaos in the Pub party at the moment, they might not be focused enough to do what needs to be done (maybe team Toomey will help do some of it for them himself).

The other big unknown, as TJ suggests, is turnout, and particularly black turnout, in Philly. Another aspect, is that it is conceivable, that blacks vis a vis the nation will trend a bit Pub in 2016. If so, that pushes PA a bit towards the Pubs relatively speaking, in a way that does not obtain in NH, which of course has few blacks. Mittens himself I read someone got 12% of the young black male vote. If that continues, and the young male blacks who jumped the Dem ship in 2012 keep voting Pub, that cohort as a percentage of the black population, will expand as time goes on. So many speculations, so little time.

[Torie]

Well, the first step is for the Pub to get close to 50% of the national vote. Unless PA trends Dem, that closes most of the gap as PA swings with the nation. The remaining 1% gap or so, then gets down to the host of factors you mentioned, as well as closing the turnout gap in places where the Pubs have upside potential arguably. A little bit of this, a little bit of that, or a lot of one thing. If Pubs suck wind even more than usual with the Hispanic vote in 2016 (think Trump as the nominee heaven forbid), that would tend to push the PVI to the Pubs in PA all by itself, I might add, which is light on Hispanics.

[Torie]

I do find it rather odd however, that you have such a different attitude towards NH vis a vis PA. I mean NH trended to the Pubs too in 2012, albeit by not as much as PA did. I tend to agree that perhaps PA has more potential for the Pubs than NH, but that is based on untapped potential as discussed above, that needs to be tapped to make a difference, and given the chaos in the Pub party at the moment, they might not be focused enough to do what needs to be done (maybe team Toomey will help do some of it for them himself).

What matters is the long-term trend. And even you have admitted that NH is gone.

You must be logged in to read this quote.

https://uselectionatlas.org/FORUM/index.php?topic=164324.0

NH is a cruel state, isn't it? Always threatening to go to the GOP, but it never materializes. Perhaps its residents are lying to the pollsters to trick the GOP into wasting its money there?

Gosh, I quite like my prose there.   Anyway, since I wrote it, NH trended Pub in 2012 by about 40 basis points, so the progress of the green revolution in NH was held at bay some for that election. But yes, I agree, that given the current brand images of the two parties, the long term prospects for the Pubs in NH are not too good. They need to become a bit more, well "Torie-like," to reverse that, but then of course, places like Georgia might go down the drain. It's never easy to find the perfect objective function to electoral success, given how disputatious and divided into little factions, the Fruited Plain is these days.

[Torie]

Look up the definition of "fact." Pennsylvania could be competitive, but it's absurd to assume that it will be as competitive as OH, IA, and CO. Evidence that PA is trending Republican is mixed at best, and I see no way that it will be critical to the electoral math. If Republicans win PA, they've already won easily. It's certainly not more competitive than FL.

PA has a D+1 PVI. It's called a safe blue state.
NH has a D+1 PVI. Rated Toss-up.
CO has a D+1 PVI. Rated Toss-up.
IA has a D+1 PVI. Rated Toss-up.
OH has a R+1 PVI. Rated Toss-up.

I know PVI isn't everything, but do you notice something?

If Republicans win PA, they've already won easily.

How? The next time a Republican wins the presidency, PA will flip. PA and VA (and perhaps CO as well) are the ultimate bellwether states in 2016.

I don't think anyone except the biggest Democratic hacks consider PA a "safe blue state." I would call it a Lean D state, as it's less elastic than states like CO, IA, and NH, and I would expect at least the first two to go Republican before PA. The PVI of one election is not the be-all, end-all. Consider this:

CO: Obama +5.4, rated toss-up
VA: Obama +3.9, rated toss-up
OH: Obama +3, rated toss-up
NC: Romney +2, rated Lean/Likely R

I'm aware that Obama won the popular vote fairly easily, but why should we assume the popular vote will be a tie?

Just because PA almost ended up being the tipping point state in 2012 doesn't mean that it will be in 2016. As I said, it's not safe for the Democrats, but in an extremely close election (think 2000), I'm 99% confident that it will go Democratic.

You wrote a lot of words that ended up stating that you are 99% certain that PA will have a Dem PVI, without really saying why. What I have been saying, is that it is foolish really to have much certainty one way or the other as to whether PA will have a Dem PVI. Sure, odds are that it will, but the odds are hardly anywhere near 99%. Maybe 60% or 65% seems more like it to me.

[Torie]

It is kind of interesting, that so many maps have VA as going Dem, in a very tight election. Why? Because despite it having a dead even PVI, it's trending Dem. And there is not that much controversy about it. But when it comes to Pub trending PA, and in the last two elections, trending at a faster rate than VA in the Dem direction, with about a .75% Dem PVI, it causes fire fight as to whether PA is in real play in a very tight election. Sure 75 basis points, which is the difference, is not chopped liver, but all the sound and fury about it all, just doesn't make that much sense to me.

I think some of it is due to this perception, that the Pubs will have trouble closing the 4% margin by which Romney lost. True, if that is the margin, the Dems are going to win. They are very, very probably going to win with a 2% margin (even a 1% margin is very likely to be enough). So if the Pubs cannot close the popular vote gap nationwide, it's curtains for them. But if they do, and given that swings vary, if it is not PA where they close the gap more than the nation, where is it going to be? Some states need to trend Pub. All states don't swing equally. Which states might that be, if not PA? And I say this in the context, that absent the Dem candidate being a liberal black (a secular one to boot), one might posit that the Dem candidate might do better in some places than whatever the national swing in many places in the South. So if one posits a Dem trend in much of the South, that means that there needs to be even more of a Pub trend elsewhere, in order to get to the overall national swing.

In summary, much of the heat about PA to me seems to be conflating how much the swing will be, as opposed to the trend. It's two different issues.

[Torie]

Franklin & Marshall had Toomey up by 12 which was proven false. But Clinton will win by 2.5 and Sestak by .5 or McGinty.

I just love your precise percentage predictions. Are you sure the margin by which Toomey will lose is 0.5%?  Why wouldn't it be 0.2675%?  What's Clinton's national margin going to be while we are at it? I want to know PA's PVI, and I want to know it now!

[Torie]

Well this topic has been beaten to death, and again, nothing is 99% certain. Sure, it is more likely that in 2016 Ohio will have a more Pub PVI than PA based on the past (even though recently Ohio has trended Dem, and PA Pub). In the present however, Hillary is doing better in the polls in Ohio than PA. But polls can change, and will, and it's early. My main beef is that there is so much certainty exuded in so many posts, when the main theme of this election cycle, is a lot of uncertainty.