Swings to Trump among white Obama voters

[jman123]

Which area of the country swung towards trump due in part to white Obama voters who voted for Trump in 2016? Where were these white voters, if any, concentrated in?

[White Trash]

Eastern Iowa and northwestern Illinois had huge swings to Trump from Obama.

[mieastwick]

A lot of White Obama voters must have voted for republicans (and certainly some for Trump) in the primary. Others must have swung to Trump between the primaries and the general. Those in Eastern Illinois mostly flipped during the primary, it appears.

[Matty]

Southeastern minnesota had out-of-this-world swings to trump.

Mower county minnesota...holy sh**t go look at that swing.

[Lourdes]

Parts of white-majority PA swung heavily for Trump. Probably the biggest I've seen was in Luzerne County, which was 52-47 for Obama in 2012 and 58-39 for Trump in 2016.

[Steam Boat Willie]

There's a reason Obama won Iowa twice and it's not because he had a lot in common with the voters there, but because he knew how to connect with them.  Hillary Clinton had no capability to do this.  Consider the Midwestern trend that took place in Missouri as well.  It's not moving throughout the upper Midwest.  Democrats are being seen as more and more anti-white which tells the story nationwide. 

[dirks]

There's a reason Obama won Iowa twice and it's not because he had a lot in common with the voters there, but because he knew how to connect with them.  Hillary Clinton had no capability to do this.  Consider the Midwestern trend that took place in Missouri as well.  It's not moving throughout the upper Midwest.  Democrats are being seen as more and more anti-white which tells the story nationwide. 

Obama alienated those people though after his 2012 campaign. His inappropriate response to Ferguson and attacking the police. The whole Clockboy crap and his ridiculous tweet when the clock actually did look like a damn bomb. The terrorist attacks in Orlando, San Bernardino, and Boston and his continuous reverence for Islam with absolutely no effort at all to curtail it as it festered in America. A totally open southern border. Invoking the Crusades at a prayer breakfast and completely neglecting the fact that they were in response to Islamic aggression. Allowing BLM to grow unchecked...v2 of the Black Panthers movement.

This is what drove them to Trump. They saw worsening national security and increased social tensions at home. If anything Trump should've done better among the midwestern white Obama voters. The only reason he didn't is because Trump was his own worst enemy about 1000x during his campaign

[Beet]

I understand the backlash to radical Islam and BLM, but what I don't get is the hatred of otherwise law-abiding undocumented immigrants. There was no surge of illegal immigration under Obama, it actually went down. And there was no surge of crime by Mexican immigrants or any high profile incidents. Yet it seems they're the ones being punished the most under Trump.

[NOVA Green]

Which area of the country swung towards trump due in part to white Obama voters who voted for Trump in 2016? Where were these white voters, if any, concentrated in?

So your question is both extremely specific "what area of the country swung towards Trump among white Obama voters", while also somewhat vague with your secondary question "where were these white voters if anywhere concentrated in".

Additionally both questions  infer that white voters who supported Obama did so in both '08 and '12, which is obviously not the case, when looking at detailed county and precinct level results nationally.

1.) white Obama voters (Either/And/Or) Obama '08/'12 voters in every region in the country, and State in the Country. Now if we are talking about Obama '08/ Romney '12/ Clinton '16 voters, sure there are a ton of them out there, and heavily concentrated in large Metro areas, Tech/University cities/towns and precincts.

If we look at where regionally the Obama/Trump voters were most concentrated in, as others have noted the Upper Midwest and the "Rust Belt" in general are certainly the correct answers.

2.) "Where were these White voters concentrated in?"

So here's the reality, is that outside of the Deep South, and larger cities of the Northeast and West Coast the swings among WWC voters between '08 and '16 wasn't that dramatic among regions or even states, once one adjusts for the large cities....

The swing against Obama among WWC voters was actually a significant item in '12, but because Obama won as result of a solid level of support in heavily white suburbs, this didn't receive that much air play.

Between '12 and '16 the bleeding that was already evident in '12 achieved critical mass, and even in many ancestrally white Democratic union and mill towns the floor completely dropped out under candidate Clinton.

Rural areas outside of these decaying mill towns had even worse news for the Democratic Party at a Presidential level.

For example, even in heavily Democratic Oregon, I can pull up so many examples of swings towards Trump of +10-15% (Or even more in some cases) between '12 and '16.

I could rewind the clock and look at the numbers between '08 and '12, and in many cases see an even deeper drop in Dem numbers between the two Obama elections...

Now, here's the key factor--- how much of the various Obama voters swinging to Trump was really just George W. Bush voters from '04 consolidating back, after giving Obama a try in '08, and then some of them again in '12?

[Mr. Morden]

I mean, Trump got a smaller %age of the total vote than Romney did (and if you believe the exit polls, not just among all voters broadly, but among white voters as well).  He got more total votes, but that's mostly just population growth, as it's not like there was a massive surge in overall turnout compared to 2012.

So I haven't really seen a comprehensive study of how much of the 2012->2016 swings were due to either D->R or R->D, how much were due to swings to 3rd parties, and how much were due to different groups of people turning out.  Even in counties where there was a significant swing and overall turnout didn't change much, how do we know how much of it was people flipping parties, and how much was different groups of people turning out to vote in that county?

I summarized these questions in this other thread:

BUMP for a re-evaluation with more accurate information/exit polling studies

Do any of the studies attempt to give some kind of solid #s to the fundamental questions?:

-What % of Clinton voters voted for Obama?  What % voted for Romney?  What % voted 3rd party in 2012?  What % didn't vote at all in 2012?

-What % of Trump voters voted for Obama?  What % voted for Romney?  What % voted 3rd party in 2012?  What % didn't vote at all in 2012?

-What % of Johnson/Stein/McMullin/other voters voted for Obama?  What % voted for Romney?  What % voted 3rd party in 2012?  What % didn't vote at all in 2012?

I guess that would answer it, if we had that information.

I also pose those questions in a different way here:

https://uselectionatlas.org/FORUM/index.php?topic=261463.msg5588503#msg5588503

[Arbitrage1980]

There's a reason Obama won Iowa twice and it's not because he had a lot in common with the voters there, but because he knew how to connect with them.  Hillary Clinton had no capability to do this.  Consider the Midwestern trend that took place in Missouri as well.  It's not moving throughout the upper Midwest.  Democrats are being seen as more and more anti-white which tells the story nationwide. 

Obama alienated those people though after his 2012 campaign. His inappropriate response to Ferguson and attacking the police. The whole Clockboy crap and his ridiculous tweet when the clock actually did look like a damn bomb. The terrorist attacks in Orlando, San Bernardino, and Boston and his continuous reverence for Islam with absolutely no effort at all to curtail it as it festered in America. A totally open southern border. Invoking the Crusades at a prayer breakfast and completely neglecting the fact that they were in response to Islamic aggression. Allowing BLM to grow unchecked...v2 of the Black Panthers movement.

This is what drove them to Trump. They saw worsening national security and increased social tensions at home. If anything Trump should've done better among the midwestern white Obama voters. The only reason he didn't is because Trump was his own worst enemy about 1000x during his campaign

I agree completely with this analysis. If Trump had a modicum of discipline, he would have won Minnesota, Maine outright, New Hampshire, Colorado, Nevada.

[NOVA Green]

So... it appears that nationally there was not a swing towards Trump among wealthier White voters residing in large Metro Areas, Medium size Metro Areas, and even middle-size cities...

Here's a thread where myself and others contributed pulling individual city and precinct level results that includes many major Metro areas from uber Republican precincts in Texas and Tennessee, to heavily Liberal parts of the West Coast, as well as some data from the Midwest and Northeast...

https://uselectionatlas.org/FORUM/index.php?topic=259050.msg5580099#msg5580099

Now, it does appear looking at the national county swing maps, and certainly precinct level data, that there was a major swing towards Trump in many heavily White Rural and Small Town communities across the country that had voted for Obama in either/or '08/'12.

Even if we look at the level of 3rd Party voting, it's pretty clear that in these heavily White "rural" counties there were a significant number of Obama '08/'12 voters that shifted over and voted for Trump.... Seriously, we're not talking about places here that have a ton of demographic change, other than old people passing on and a small spurt of younger voters jumping into the electoral ring for the first time. I haven't yet seen any serious evidence or argument indicating a massive "first time voter rural White voting surge for Trump argument", but would certainly would be willing to consider, based upon whatever evidence is presented.

As I have noted elsewhere, in many parts of rural and small town America, Obama's numbers dropped by a larger percentage between '08 and '12, than the collapse of Clinton's numbers vs Obama '12 numbers.

In many of these communities, that was likely Obama '08 voters who were a bit disenchanted that they had not yet seen recovery from the Great Recession, who were not into voting for Romney as "The Solution".

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Most of these individuals were already voters, who simply skipped the 2012 election....

That being said, it's pretty clear the Trump's message resonated very well in many traditional mill towns across America, especially with smaller population centers.

Even in many traditionally Democratic or leaning Democratic mill town in Oregon (A state I know best politically), there were +12-20% Republican swings between '12 and '16. Even if we adjust for the 3rd Party voting scene, there were definitely a significant number of Obama '12/ Trump '16 voters, with the situation even more evident in terms of Obama '08/ Trump '16 voters....

Now, among the "huge" discussion" regarding WWC voters, is the reality that many live in Metro and Large Metro areas... These White voters are harder to capture than in rural and smaller town areas, let alone upper-income suburbs. This creates particular complexity when looking at the "White Vote" in larger Metro areas.... We can't simply dissect voting precincts by race/ethnicity, Occupation, Age, Education, and income as easily.

[Adam Griffin]

There's a reason Obama won Iowa twice and it's not because he had a lot in common with the voters there, but because he knew how to connect with them.  Hillary Clinton had no capability to do this.  Consider the Midwestern trend that took place in Missouri as well.  It's not moving throughout the upper Midwest.  Democrats are being seen as more and more anti-white which tells the story nationwide.  

Obama alienated those people though after his 2012 campaign. His inappropriate response to Ferguson and attacking the police [blacks!]. The whole Clockboy crap and his ridiculous tweet when the clock actually did look like a damn bomb [muslims!]. The terrorist attacks in Orlando, San Bernardino, and Boston and his continuous reverence for Islam with absolutely no effort at all to curtail it as it festered in America [muslims!]. A totally open southern border [mexicans!]. Invoking the Crusades at a prayer breakfast and completely neglecting the fact that they were in response to Islamic aggression [muslims!]. Allowing BLM to grow unchecked...v2 of the Black Panthers movement [blacks!].

Word of advice: never let your own neurosis be conflated with the mindset of the majority. No meaningful number of people whatsoever went into the ballot box fuming about all of those niche, meaningless events who wasn't already going to vote for Trump and who didn't also vote for McCain/Romney. All of the sh**t you just listed is pure racist/bigoted/xenophobic blabber that gets sh**t out daily by the right-wing outrage machine to keep the base angry, agitated and mobilized. Normal people don't care about it and certainly don't decide who's going to be Commander-in-Chief as a result of it.

[mencken]

There's a reason Obama won Iowa twice and it's not because he had a lot in common with the voters there, but because he knew how to connect with them.  Hillary Clinton had no capability to do this.  Consider the Midwestern trend that took place in Missouri as well.  It's not moving throughout the upper Midwest.  Democrats are being seen as more and more anti-white which tells the story nationwide.  

Obama alienated those people though after his 2012 campaign. His inappropriate response to Ferguson and attacking the police [blacks!]. The whole Clockboy crap and his ridiculous tweet when the clock actually did look like a damn bomb [muslims!]. The terrorist attacks in Orlando, San Bernardino, and Boston and his continuous reverence for Islam with absolutely no effort at all to curtail it as it festered in America [muslims!]. A totally open southern border [mexicans!]. Invoking the Crusades at a prayer breakfast and completely neglecting the fact that they were in response to Islamic aggression [muslims!]. Allowing BLM to grow unchecked...v2 of the Black Panthers movement [blacks!].

Word of advice: never let your own neurosis be conflated with the mindset of the majority. No meaningful number of people whatsoever went into the ballot box fuming about all of those niche, meaningless events who wasn't already going to vote for Trump and who didn't also vote for McCain/Romney. All of the sh**t you just listed is pure racist/bigoted/xenophobic blabber that gets sh**t out daily by the right-wing outrage machine to keep the base angry, agitated and mobilized. Normal people don't care about it and certainly don't decide who's going to be Commander-in-Chief as a result of it.

"Those CNN exit polls show implausible levels of support for Trump with Hispanic voters"

"Hurr durr this CNN exit poll shows Obama is popular with 2016 voters"

Obama was unpopular for most of his second term until his two most likely replacements emerged, who most Americans found even more detestable (thus making the incumbent look good by comparison).

[Adam Griffin]

"Those CNN exit polls show implausible levels of support for Trump with Hispanic voters"

"Hurr durr this CNN exit poll shows Obama is popular with 2016 voters"

Obama was unpopular for most of his second term until his two most likely replacements emerged, who most Americans found even more detestable (thus making the incumbent look good by comparison).

You can scrutinize sub-samples in exit polling without calling into question the broader accuracy of the overall results. Discrepancies in 2016 Latino voting prefs and exit polls are likely (mostly) explainable by MoEs in the end; even if not, sampling Latino opinions has always been a hell of a lot more difficult than sampling the country at-large. We've seen it in polling for several cycles, in ways that have never existed in polling for the country as a whole. I don't think it's a stretch to say this could also affect exit polling. However, he's implying that there was some sort of huge backlash specifically because of things Obama did that wouldn't at all be explainable with a comparable discrepancy in exit polling.

As far as the second point: so? From what he's implying, Trump's approvals should have went up, because he was spewing a contrary view to Obama with all of the noxious talking points on issues like that: the ones they supposedly wanted to hear en masse, according to him. There's a pretty good overlap between what made Trump unlikable and his opinions/statements on those issues. Yet they not only failed to improve his standing, but they made Obama's approvals increase as a result?

I totally get what you're saying here and don't necessarily disagree with your statement alone, but I just don't buy it in conjunction with his argument ("these Obama policies made Obama unpopular, which then made a bunch of people vote for Trump while simultaneously developing a better opinion of Obama").

[Mr. Morden]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

[uti2]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

It's simple. The Romney -> Clinton voters were geographically skewed to the heavily Mexican-American SW, and weren't geographically distributed well enough to matter electorally.

[Mr. Morden]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

It's simple. The Romney -> Clinton voters were geographically skewed to the heavily Mexican-American SW, and weren't geographically distributed well enough to matter electorally.

Again, that map ignores both 3rd parties and any #s on voters who turned out in only one of the two elections.  Fact remains that the Republican vote shares in 2012 and 2016 were almost the same.  Biggest shift was in Democratic vote share dropping and Indie vote share increasing.

[uti2]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

It's simple. The Romney -> Clinton voters were geographically skewed to the heavily Mexican-American SW, and weren't geographically distributed well enough to matter electorally.

Again, that map ignores both 3rd parties and any #s on voters who turned out in only one of the two elections.  Fact remains that the Republican vote shares in 2012 and 2016 were almost the same.  Biggest shift was in Democratic vote share dropping and Indie vote share increasing.

Trump was able to turnout additional votes from non-regular voters along the Appalachia/Florida corridor, if that's the subject of your inquiry. The dynamics of why Trump won MI/WI are different from PA.

[mencken]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

It's simple. The Romney -> Clinton voters were geographically skewed to the heavily Mexican-American SW, and weren't geographically distributed well enough to matter electorally.

Again, that map ignores both 3rd parties and any #s on voters who turned out in only one of the two elections.  Fact remains that the Republican vote shares in 2012 and 2016 were almost the same.  Biggest shift was in Democratic vote share dropping and Indie vote share increasing.

My county analysis of Wisconsin suggests that, at least in that Midwestern state, the largest switches were Obama->home, Obama->Trump, Romney->Johnson/other, and home->Trump, roughly in that order. Based on a state analysis, it seems that there is a category of home->Clinton voters on the national level as well (a disproportionately Latino category)

[Mr. Morden]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

To follow up on this...

Here's a story on the Dems taking an internal look at the data, and concluding that "2/3rds" of the swing can be explained by Obama voters going for Trump:

http://www.mcclatchydc.com/news/politics-government/article147475484.html

But I'm not sure what that means exactly.  I guess it means that Obama->Trump was larger than Obama->home or Obama->third party or home->Trump.  But other than that, I don't know.  It also doesn't talk about how big the offsetting Romney->Clinton category was.

It's also tricky to define these things because a single Obama->Trump voter has double the impact on the 2-party swing that a single Obama->home voter does.

[Fuzzy Bear]

I understand the backlash to radical Islam and BLM, but what I don't get is the hatred of otherwise law-abiding undocumented immigrants. There was no surge of illegal immigration under Obama, it actually went down. And there was no surge of crime by Mexican immigrants or any high profile incidents. Yet it seems they're the ones being punished the most under Trump.

What happened during Obama's second term, and during Hillary's campaign, is that the Democrats came out in the open with their real immigration policy, which was to simply not enforce immigration laws.  The clear message Democrats sent on immigration was something on the order of "If you can swim the Rio Grande, Welcome Home!".  That tack is unsettling to folks in the middle who are not overly partisan, but who do believe in the rule of law as the thing that differentiates us from the unstable democracies of the World.

The Ferguson thing is another example of this.  Folks were demanding a police officer be indicted without regard to facts.  The man who was shot and killed was shown on video committing a strong-arm robbery on a frail female storekeeper minutes before he was shot resisting arrest.  The robber was eulogized as a victim, and the police officer was hounded out of his career, despite being found justified in his actions, and while SJWs see this as fine and dandy, law-abiding folks in the middle see this as the world turned upside down.

[NOVA Green]

So, again I'm not totally sold on the concept that Mr Morden alluded to as a variable, in terms of the "non-voting Trump surge" or people so disengaged from the political process that they crawled out of the woodworks just to cast a vote for their hero for the first time....

Oh, I'm not saying that there weren't plenty of Obama->Trump voters.  There were surely quite a few of them.  But there were also Romney->Clinton voters, Obama->3rd party voters, Romney->3rd party voters, didn't vote in 2012->Trump voters, didn't vote in 2012->Clinton voters, didn't vote in 2012->3rd party voters, etc., etc.

I just don't have a clear picture in my head of how big each group was nationally.  I've seen people try to address this at the individual precinct or county level, but I'd love to see someone build a model that attempts to quantify those numbers nationally.  Fact remains that Trump's share of the vote was about the same as Romney's, so whatever gains he made (either Obama voters backing him or getting people who didn't vote in 2012) have to be offset by corresponding losses (Romney voters going with Clinton, 3rd party, or not voting).  Without knowing how big each of those groups was, I don't feel like I have a good sense of what happened.

To follow up on this...

Here's a story on the Dems taking an internal look at the data, and concluding that "2/3rds" of the swing can be explained by Obama voters going for Trump:

http://www.mcclatchydc.com/news/politics-government/article147475484.html

But I'm not sure what that means exactly.  I guess it means that Obama->Trump was larger than Obama->home or Obama->third party or home->Trump.  But other than that, I don't know.  It also doesn't talk about how big the offsetting Romney->Clinton category was.

It's also tricky to define these things because a single Obama->Trump voter has double the impact on the 2-party swing that a single Obama->home voter does.

Thanks for the link!!!

So this specifically identifies Obama '12/ Trump '16 White voters....

I would posit, that potentially there was a greater swing between Obama '08 White voters to Romney in '12, than there was between Obama/ Trump White voters in '12 and '16.

Now, I suspect that there were not only US regional, but also other demographic variables, including age, Size of Community (Urban/Suburban/ City / Town / Rural), Education and Occupation/Industry that helps explain the differences between areas where there was a much greater Obama '08/Romney '12 swing vs Obama '12/Trump '16 swing.

My still current operating theory is that what Trump was able to do in '16 was to bring back many of the 1x/2x George W. White voters, Obama '08 was an exception to the norm, many of the upper income Obama '08 voters swung towards Romney in '12 and Clinton in '16. Meanwhile in many WWC areas among these swing voters it was more like an Obama '08/ Obama '12 (but less of them), and heavy swings towards Trump in '16 among older voters and younger voters <35 moving more towards the 3rd Party Protest option.

[Coolface Sock #42069]

Upper Mississippi Valley, Pennsyltucky, and a few other rural areas.