What I found interesting about this election was not issues with the electoral college system, but why the polls were as wrong about the outcome as the Brexit polls. That is something that needs to be addressed. I suspect it has to do with the hostility of these two campaigns. People were afraid to admit to someone that they were voting for Trump or voting for Brexit. That assumes people are being directly asked who they are voting for either on the phone or in person. If they could have sensed they were anonymous I don't see how being afraid to admit who they were going to vote for would have mattered. However, I don't know how the polling was done.
Regarding the rural/urban issue it seems that it must have been changes in the
urban population that allowed the upset to occur since there are more people there. I assume that counties with large populations that went for the Democrats still went for the Democrats but with smaller percentages.
I was looking at the two exit polls from 2012 and 2016 wondering if I could see in which demographics the shift occurred: 2012:
http://www.cnn.com/election/2012/res...ace/president/ and 2016:
http://www.cnn.com/election/results A few areas stood out from a quick comparison.
In 2012 people aged 18-29 voted 60% for Democrats. This dropped to 55% in 2016.
In 2012, black and latino votes went 93% and 71% for Democrats. This dropped to 88% and 65% in 2016.
In 2012, people making less than $50,000 went 60% for Democrats. This dropped to 52% in 2016.
Although the Democrats won all these groups, they did so with a smaller base.