Thread by @chrishanretty

Now that the #c4brexit show's over, time for some background charts focusing on trends in different areas (1/n)

As you might expect, there's a strong association between 2016 vote and how LAs would vote now (2/n)

Switches away from Leave are:
- more pronounced in areas that started out Leavier;
- more pronounced in areas that voted UKIP in the 2014 EP elections;
Switches away from Leave are:
- less pronounced in areas which voted Cons. in the 2014 EP elections

There's quite a lot of dispersion within regions -- but it's predominantly Labour areas in the North that have seen greatest change.

These are all aggregate changes, and we should be cautious about telling individual stories on this basis. There's a lot of Leave->Undecided and Undecided->Remain here.

There's also something of a turnout story. It's low turnout areas which have switched away from Leave. High turnout, high-leave areas have stuck.

Where do we find low turnout areas? Areas with younger electors....

You can say, "these areas would never turn out in a real referendum", but then everyone's turnout predictions in #euref were borked

.@keiranpedley and @chris__curtis are already arguing about magnitudes

https://twitter.com/chris__curtis/status/1059551121797111809

The main focus of my work was on producing local authority estimates (you wouldn't do a 20k sample otherwise), and those were mostly for the Leave/Remain fresh referendum question. But...

When you try and produce estimates for some of the other acceptable/not acceptable stuff, you see how some of this stuff does or doesn't map on to the 2016 vote

(This is mostly a consequence of an individual level fact: no one has a scooby about trade or goods checks)

More along in the next couple of days, including Westminster estimates. There will be confidence intervals... <fin>

p.s. Credit to the peeps at Survation for giving me nice clean data and time to do some of the backend stuff over the summer.