The pollsters have surely been scrambling since the MI earthquake to make up for the identifiable deficits in the MA and MI pre-election polling in the upcoming states. Those deficits have varied from one state to the next, but fall within a two identifiable categories:

  1. Underestimate of the 18-29 percentage of voters and overestimate of the >64 portion.  
  2. Slight undersampling of male voters and oversampling of female voters.

In MI, there was an oversampling of AA voters.  The drop-off of AA primary voters from ’08 has been consistent across most states that have already voted.  So far, hasn’t mattered in  states with  an AA population much higher than the national average and VA and TX.

Those factors can be adjusted for, at least to some extent since MI.  But would that correct for the wide variances in polls to actual seen in other states?  The only state so far that behaved normally is Texas.  The “undecideds” or other were 5.7% and they split proportionally to HRC’s and Sanders numbers (2.9% and 1.2%).  The observable DEM primary variables (gender, race, and class) were consistent with the polls.  And they split in accordance with three variables: race, gender, and class.  Not only are racism and sexism are not in operation, there is a strong preference among AAs for Obama’s designated successor and woman (who significantly outnumber men in the DEM electorate) prefer the female.  The contaminating variable is class.  But other than among the highest income earners, it’s relative to where one was (or perceived one and or one’s family to be) yesterday compared to today.

An example may help clarify that last point because it’s not straightforward.  If one is either a woman or an AA and one’s class self-idenfication hasn’t changed in the past decade or two,  HRC gets that vote six or more times out of ten.  If class identification has dropped from say middle class to lower middle class, Bernie wins by a similar margin.  So, why is that not appearing earlier in the polls?  It’s not easy for some people to acknowledge publicly that they’re feeling like a loser in this economy.  Especially difficult when the media, and very specifically by DEM party and large numbers of ordinary DEM partisans, trumpets how well the economy is doing.

Over the decades after WWII, more people began to self-identify as middle class and fewer as working class.  Part of it was an authentic assessment of their personal financial well-being and part of it was  delusion.   Millenials seem not to be buying into the delusion.  And reality is breaking through to those living and working in regions that are slipping economically.  Boston metro is doing okay but outside that region life is tougher.  And that’s how they voted.

Consider the following:

MA – age 18-29: HRC 35% Bernie 65% and age 30-44: HRC 55% and Sanders 44%

MI –  age 18-29: HRC 19% Bernie 81% and age 30-44: HRC 42% and Sanders 53%

Older people in both MA and MI preferred HRC, but by much stronger margins in MI than MA (that’s not a typo).

So, what can this tell us about today/tonight?  Here’s what the pollsters have given us:

FL: 61.1% HRC and 31.5% Bernie
IL: 48.3% HRC and 46% Bernie
MO: 46.5% HRC and 43.5 Bernie
NC: 57% HRC and 33% Bernie  
OH: 50.8% HRC and 42.5% Bernie

FL was hit harder than many of states in the Great Recession, but suspect the self-identified class status hasn’t significantly changed.  It could  break as TX did (undecideds 2:1 for HRC), a regression to HRC as in TN (59% polls to 66% actual) and VA, or something like MA.  The first two scenarios seem more likely to me than the last, but would very much like to be wrong.  HRC wins.  (HRC with 61% to 65%)

IL  has that unique Rahm factor going for it.  The “undecideds” will break more like MA than MI and HRC more likely to hold onto her current polling number.  Sanders wins.  (Sanders with 50.1% to 55%).

MO will follow IL’s lead.  Sanders wins.  (Sanders with 50.1% to 52%.)

NC is looking too much like VA for it to deviate from that pattern.  HRC wins. (HRC with 65% to 69%)

OH is the toughest one of all to project.  The economy never seems to play as much of role in OH (and MO for that matter) election outcomes as would be expected.  The MI primary results appear to have impacted the latest OH polling numbers with downticks for HRC and upticks for Sanders.  HRC’s seven to eight point winning margin could now be baked in.  My head says, HRC with 54% and my gut says Sanders with 50.1% (hate it when my head and gut disagree and in such situations and neither has a superior track record to the other).

What I’ll be looking for tonight is if it is happening across the board.  By that I mean that Sanders beats my lowest estimates of his share of the vote in four out of the five states.  

Another great ad from the Bernie Sanders campaign.

UPDATE – Well, that was fun. Not. Still a few more votes to count but the percentages won’t change by much. Life must be good for all those middle class Americans that earn less than $50,000 a year in FL, IL, NC, and OH (and possibly MO). Perhaps one day it will dawn on them that $50,000/year doesn’t make them middle class.

IL — projected 50.1/49.9 Sanders; actual 50.4/48.8 HRC

FL – projected 65/35 HRC; actual 64.5/33.3 HRC

MO – projected 50.1/49.9 Sanders; actual 49.6/49.4 HRC

NC – projected 65/35 HRC; actual 54.6/40.8 HRC

OH – projected 54/46 HRC; actual 56.5/42.8 HRC

The kids (version 2.0 — “never trust anyone over the age of 44”) get it as far as the right thing to do, but continue to miss by letting the Boomers and Gen Q punch above their weight with a higher voter participation rate. Yet, so close.