Thread by @tbonier

This is incorrect, Republicans are not outregistering Dems in any of the states listed here. The analysis is flawed. I'll explain, and share the actual data.

axios.com/trump-is-regis…

Start with PA, Dems have a 15 point advantage among those who have registered since Nov of 2016.

Also, note the large share of unaffiliated voters. Who are they? Younger voters are much more likely to register unaffiliated, and are reliably Democratic when it comes to candidate choice. Here's the modeled party of those newly registered unaffiliateds in PA.

I'll share the same info for FL and NC shortly, and explain how these false conclusions are being drawn, but need to do some work! More later this afternoon.

Here's NC. Dems have a 5 point advantage among new registrants since the '16 election, but a plurality are unaffiliated voters.

And here's the modeled partisanship for unaffiliated new registrants, since '16, in NC - just as was the case in PA, they skew heavily Democratic.

And finally, Florida, where Dems have a 2 point registration advantage among those who have registered to vote since the '16 election, and again, unaffiliated voters have a plurality.

So how does the Axios analysis reach a different conclusion? They compare registration now to Eday '16, which lags, thanks to purges, party switches, deaths, etc.

The best indicator of recent trends in the state is to look at who has registered during that time.

I should add, because I realize my hastily assembled Excel charts may not be the most clear data visualizations out there - the bar graphs for unaffiliated voters show the distribution by ranges of modeled Dem ID likelihood - the higher the number, the more likely Dem.

Anyhow, the shorter version:

Don't panic, the GOP has not been out-registering Dems, but you should work as hard as you would if that was happening.

One more note - credit to the reporter and Axios for listening and updating, not just the headline, but the story as well.