So now that we know what happened last Tuesday, the work begins to understand... what happened last Tuesday. Or, more accurately, how it happened. We often rush to draw conclusions from limited and flawed datasets, and these early false narratives can be hard to shake.
The good thing about this election is, with so many votes cast early and by mail, we will have the individual-level vote history in many states far earlier than we normally would. I've been digging into the turnout data in CO, OR, and NV, where the data is largely complete.
First, age. In all 3 states, voters <30 saw their vote share increase over '16 and '18 levels, as did voters over the >65. This is the "boomer-zoomer" coalition the Biden campaign talked about. Meanwhile, voters age 50-64 plummeted in vote share, Trump's strongest group. 
Race/ethnicity: similar patterns emerge here. Surges in turnout among white, college educated voters, Latinos, and Asian voters. African-American turnout (not a large population in these 3 states) held steady. White non-college turnout couldn't keep pace.
Party registration: In all 3 states both Democrats and Republicans saw their electorate shares diminish, relative to '16. Meanwhile, unaffiliated voters saw a huge surge in turnout in all 3 states, accounting for a much larger share of the electorate.
All 3 states had huge surges of voters who didn't vote in '16. These voters were overwhelmingly unaffiliated voters (almost half), and tended to be younger.
Keep in mind, these 3 states were functionally all mail which could skew the turnout results relative to other states. But the initial evidence is that Dems outperformed GOPs in turnout, driven by younger voters and voters of color, but Trump benefitted from persuasion efforts.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
