Moving the Electoral Needle

Labour took one hell of a beating, and they’re electing a new leader. Those in favour of continuity sometimes produce figures aiming to show that continuity is fine, and can bring victory.

So here are four graphs of things that might matter, versus the number of seats won. One of these is not like the other three.

When looking at these graphs, notice the R2 value. R-squared measures the amount of change in one variable predicted by change in the other variable. A value of zero means that one variable is unrelated to the other, and a value of one means that all of the change in one variable can be explained by change in the other variable.

England + Wales Results in General Election, 1950-2019

Firstly, what proportion of the total Labour and Liberal votes did Labour get, versus how many seats did Labour win. There is a positive relationship, but only a very slight one. Taking votes off Liberals might feel good, but it doesn’t correlate with Labour forming a government. That shouldn’t be surprising, since taking votes off Liberals would normally only win seats where Liberals were previously in first place and Labour in second.



Next, the relationship between total votes cast for Labour, and total seats won by Labour. There’s a relationship here, too, and it’s a little stronger than in the previous graph. Votes are good, after all. But population and turnout change this number without changing election results, so total votes by itself doesn’t mean a lot for electoral success.

Votes cast for Labour versus all other parties matters still more. Obviously if you can get towards 50% of all votes, you will win a lot of seats. But you can still win a lot of seats with 36% of the vote, if other factors go your way.

Finally, here’s the magic: the proportion of Labour votes versus Tory votes has a very strong correlation with the total proportion of seats won by Labour. Most seats in England and Wales have Labour and the Conservatives in first and second place, in either order, so this is expected.






I took general election data from the site, and filtered the data for general election results from 1950 onward in England and Wales. The excel file I used is here; note that I took data from a separate file for the 2019 results, hence the change in populated columns.



There’s nothing new here, I think, just a demonstration of the obvious: many factors that are cited as evidence of future electoral success for Labour are very weak. Pretty much the only thing that counts is the proportion of votes they get versus the Tories, or almost equivalently, the percentage gap between Labour and the Tories.

I’m not claiming any causation here, I’d hardly know how to start backing up that claim. But if someone is trying to tell you that everything will be great because of some fact x, consider whether having x is even relevant to whether things are great.

Leave a Reply

Your email address will not be published. Required fields are marked *

I accept that my given data and my IP address is sent to a server in the USA only for the purpose of spam prevention through the Akismet program.More information on Akismet and GDPR.

This site uses Akismet to reduce spam. Learn how your comment data is processed.