On health and Brexit

A while ago, I saw this graph of health and % swing in vote to Trump in the 2016 US presidential election.

I wondered if the same graph could be plotted for the results of the referendum on EU membership by English local authority.

Here is a graph:

Healthy life expectancy data is for 2012–14 and comes from the Public Health Outcomes Framework. Voting data is available from the Electoral Commission website.

For the nerds

A simple linear regression suggests that healthy life expectancy is a statistically significant predictor of the % voting to leave the EU. The gradient is -0.99 (p < 0.01), suggesting that for every year more of healthy life expectancy in a local authority, you’d expect about a 1% decrease in the % voting to leave the EU. Less healthy places voted more strongly to leave.

The relationship is pretty weak though, and explains less than 10% of the variance. The linked article in The Economist used a mixed basket of health indicators (which may raise questions about overfitting). Regardless, there’s a lot of variation left to explain. Other factors, like education and age are likely to have an effect. And health may just a proxy for deprivation.