It turns out this is one of the strongest results - if we whack all of our variables into a regression, Brexit vote, along with age, is one of the most robust predictors. ¯\_(ツ)_/¯. We'll have a hint why in a second. 17/n

Ben Ansell Intriguing! Could be a trust effect? Do you control for any trust variables?

Jane Green We have a bunch of trust in government questions (trust in scientists too). I wanted to avoid these for the first analyses since they are also 'values' expressed in the survey as opposed to previous info on panelists that YouGov have. But we'll be on it!

Ben Ansell Cool! I can't think of a political reason Remainers are more pro-vaccine that isn't driven by other demographics (you control for already), other attitudes/values-differences, or by geography (where they happen to live).

Jane Green Must be trust I think. We also find that Remainers have substantially higher trust in science and in governments' chief scientists. More to come...

Ben Ansell That makes sense, and is very interesting if that effect holds controlling also for different education levels.

Jane Green It does in this baseline analysis, probably because education itself is surprisingly not correlated with vaccine attitudes. Who knew?

Ben Ansell Jane Green Aye but I have a quibble/Q on that: you say educ. doesn’t correlate but educ. is age dependent is it not? (ie younger cohorts educ. >> older.)
And if you have an effect here with age (self interest of older gen?!) then is there possibly age cancelling educ?
1/

Iain MacGilleBhràth Jane Green Good point. I have been wondering if the 'finding' that people whose education stopped at 18 (A-Levels or equiv) had increased most is actually proxying for age. But there doesn't seem to be an interaction between the 2. Here are 2 simple models

Ben Ansell Iain MacGilleBhràth Be good to know this all holds with a categorical education level var.

Jane Green Iain MacGilleBhràth Ask and ye shall...