Sabato: Comparing the exit poll and the Cooperative Congressional Election Study

[Virginiá]

http://www.centerforpolitics.org/crystalball/articles/another-look-back-at-2016/

You must be logged in to read this quote.

Table 1: 2016 vote by gender, exit poll vs. CCES



Tables 2 and 3: 2016 vote by race, exit poll vs. CCES





Table 4: 2016 vote by race and gender, exit poll vs. CCES



Tables 5 and 6: 2016 vote by age, exit poll vs. CCES





Tables 9 and 10: 2016 vote by education, race, and gender, exit poll vs. CCES





It does seem like there were some noteworthy differences. Clinton may have won white college educated voters. Trump may have won slightly more non-white voters, and Clinton may have won more young voters.

[RaphaelDLG]

Very interesting! The gender gap is way less pronounced than initially thought

[Beet]

Nothing for income?

[Skill and Chance]

Very interesting! The gender gap is way less pronounced than initially thought

My understanding is that exit polls almost invariably oversample 18-29 year old women and undersample >65 year old men, so this isn't entirely surprising to me.  I wonder why the exit polls were apparently missing minority Trump supporters and oversampling white Trump supporters, though? 

 

[100% pro-life no matter what]

Very interesting! The gender gap is way less pronounced than initially thought

My understanding is that exit polls almost invariably oversample 18-29 year old women and undersample >65 year old men, so this isn't entirely surprising to me.  I wonder why the exit polls were apparently missing minority Trump supporters and oversampling white Trump supporters, though? 

 

If that's true, it should mean the reverse in the CCES (Trump doing better with youngs, better with whites, and worse with minorities).  I don't think it's possible to know which is best, just like individual polls where the crosstabs can sometimes get a little weird even if the topline is OK.

[DC Al Fine]

Very interesting! The gender gap is way less pronounced than initially thought

My understanding is that exit polls almost invariably oversample 18-29 year old women and undersample >65 year old men, so this isn't entirely surprising to me.  I wonder why the exit polls were apparently missing minority Trump supporters and oversampling white Trump supporters, though? 

Why is that?

[Skill and Chance]

Very interesting! The gender gap is way less pronounced than initially thought

My understanding is that exit polls almost invariably oversample 18-29 year old women and undersample >65 year old men, so this isn't entirely surprising to me.  I wonder why the exit polls were apparently missing minority Trump supporters and oversampling white Trump supporters, though? 

Why is that?

Think about it.  Exit poll volunteers are generally dominated by political science undergrads and grad students.

[jaichind]

The exit polls rings more true to me than CCES.  First, if you do the top-line calculations the exit poll seems to yield a Clinton lead of of around 2% while the CCES seems to yield a Clinton lead of around 2.9%.  Since Clinton won in reality by around 2.1% in the PV the exit poll numbers fit better with the real result.  Also the CCES result of the Black share of the electorate staying the same as 2012 does not seem to jive with a lot of local results where turnout dropped in heavy Black districts while Trump actually registered a net swing in the same districts which should imply Black turnout decreased.

[Person Man]

They both, but particularly the CCES, that Hillary did as well as Obama amongst those likely to turn out but the election was decided by "missing voters".

[The_Doctor]

This isn't good news for the GOP.

The minority bloc is on its way to making up 33% of the election by 2020-2024 (closer to 2020). They vote roughly 75-80% for the Democrats, presenting the Democratic Party with nearly half of their votes at the starting gate. The Democrats, by winning 35% of the white vote, have a minimum base of 48% of the vote.

I'm specifically lowballing the demographic numbers for the Democrats. 35% would be the lowest number among whites and 75% is considered on the low side for minority support for the Democratic Party. If you do 40% and 80%, that's roughly 53% of the vote which is equivalent to Obama's victory margins.

Republicans are never going to probably get 70% of the white vote. For a whole myriad of reasons (women favoring abortions, gay men favoring the Democratic agenda, upscale coastal whites preferring their liberal enclaves to the socially conservative GOP agenda, etc etc) the GOP is probably going to see the Democrats have a base of 35% whites and 75% minorities. Trump is probably going to, in the long run, Pete Wilson the Latino voting bloc and ensure that by being culturally anathema among this crucial bloc, that the Democrats can count on the African American, Latino, and Asian voting blocs. 

I just can't see the Republican Party abandoning its socially conservative agenda because the grip of the southern rural and Midwestern rural areas is extremely white and strong; although I could see a more populist Republican Party, but that runs against the grain of the Reagan Revolution and the core of the neoliberal socially evangelical Republican Party.