At long last- my voter file analysis of the 2018 House elections, which shows what mattered in 2018 was the demographic composition of the electorate became more friendly to Democrats (via turnout of Ds & Indies) not "kumbaya" voting w disaffected Republicans, is up!!! 
The article includes an interactive you can use to check out each district & how it changed between the 2014 & 2018 Midterms. Low turnout of Dem Coalition voters (women, younger voters, Latinos, Asians, African American voters) led to midterm shallackings for Ds in 2010 and 2014.
My expectation is (unless the Dem Party implodes on itself bc of a Sanders nomination, something I mentioned in my July forecast & will talk more about in the March update) similarly strong turnout in these districts put Dems in a dominant position for 2020, esp if campaigns
invest heavily in Dem Coalition turnout as part of their 2020 strategies (I am not saying do not invest in persuasion, but under-investment in mobilization has a $ and you can actually see it in the analysis). Along with demonstrating a turnout surge of Ds, I show a turnout surge
Indies plays a massive role in these D victories. Who are these Indies? They are lazy suburban millennials, likely left-leaning, who were sitting out midterms until the election of Trump. The activation of voters like this has had a massive impact on the fortunes of D candidates
in elections (and cycles) ever since Trump was elected, driving turnout rates up, on average, more than 10pts in elections held since 2016. This means we are talking about different voters, w different preferences, not Rs changing parties and in the forecast I will further look
for evidence of disaffected Rs in the electorate (there isn't much). Indeed, one of the most surprising findings, and the thing that took me by surprise in 2018 which I am accounting for in 2020, is that instead of declining like D turnout did when Obama was in power R turnout
increased in 2018, it did not decrease. There is very little evidence in the data that there are significant swaths of suburban voters leaving the R Party. Instead, the suburban swing is powered largely by activated turnout of latent Dem and left leaning Indies now motivated to
vote because of fear of Donald Trump. And if Ds make fear of a Trump 2nd term and what it means for American institutions and national security a main focal point of then election, in other words, if they do what they did not do in 2018 and lean into the referendum effect, they
are very well positioned to beat Trump, gain senate seats, and House seats. But if they make it about wonky policy debates or worse yet, turn it into a referendum on their own nominee, all bets are off. More of this in the forecast update.
