I had an email exchange w @Nate_Cohn that *may* have partly inspired him to look at this. Really interesting stuff. would love to see which surveys these are to see if the more urban/suburban districts are the 1's most under-estimated.
I think every cycle is a learning opportunity for analysts. I've learned so much in the 3 cycles I've been preemptively analyzing voter behavior (though again, I did not put out anything formal other than my polling & my verbal "hot takes" about my belief about massive turnout
surge & its impacts on the demographic composition of the electorate, and what that would mean for the level of competition between Rs & Ds during Trump's tenure). But from that 1st cycle I learned the surge's capabilities had much greater potential to overcome poor campaign
strategy (so natural surge potential) than I originally expected. From 2018 I learned that my "surge & decline theory" was like many other aspects of our political system: asymmetrical. Surge & decline is really a problem Democrats face bc even when controlling the entire
government, not only did R turnout not decline in the 2018 midterm, it actually improved over its 2014 numbers & as I'll show in this forthcoming article, in most places, Rs STILL outvoted Dems in 2018. And in the VA 2019 election I learned that although PVI is the more powerful
variable in my model, that also significant college education variable really does matter bc it predicts exactly where those D surges will be big and where they will be more modest and also suggests that strategy for coalition turnout should probably account for this variation.
I've also learned from both '18 & '19 that threat surges are not uniform among racial groups- whites & Latinos are more sensitive to threat than black voters who, thanks to the research of @Davin_Phoenix we now know are less responsive to threats like Trump bc of
institutionalized racism & what Phoenix calls "resignation." Thus, my theory is so much richer than when I first laid it out in my 2018 post revealing my 2018 House predictions.
