16 January 2010

Safe Seats Disenfranchise Voters

One of the outcomes of a first-past-the-post electoral system is that it creates safe seats. These are seats where one party has such a large lead over their opponents in their constituency that the chances of being overturned is slim.

It seems obvious that people in these seats will be less minded to go out and use their vote. The statistics back this up.

This graph compares the size of the majority (the scale of the win) in the 2001 general election to the size of the turnout in 2005 in each constituency.
(Excludes Northern Ireland)

There is a significant trend (PMCC -0.71). As the majority at the last election gets bigger, the turnout tends to goes down.

If we are to get people voting again, there needs to be an end to the first-past-the-post electoral system that gives MPs unassailable majorities.

UPDATE 26/01/2010: Timothy asked if the scatter plot could be colour-coded by party. Happy to oblige. Each dot is colour-coded by the winning party in 2001.
So it does seem that the seats with the big Labour majorities 2001 subsequently had low turnouts in 2005. I'm not sure this shows much, since the Tories had no heavily safe seat in 2001. I'll definitely remake this plot after the next general election to see if the same pattern occurs.

6 comments:

Timothy (likes zebras) said...

I'd be *really* interested in seeing that plot remade with blue dots for Tory-held seats and red dots for Labour-held seats, and separate stats calculated for both populations.

My impression had been that turnout was generally pretty good in Tory seats, so I would be interested to see if the relationship still held.

Duncan Stott said...

Nice idea Timothy; I'll look into it this evening.

Statto said...

Er, correlation ≠ causation, dude. :)

Duncan Stott said...

I have shown how an occurrence in 2001 is correlated with an occurrence in 2005.

One thing occurs after another.

That's causation innit?

Statto said...

I don't think you've demonstrated it conclusively—there are still parts of the causal chain unproven.

Are voters aware of the overwhelming majority and, if they are, does that influence their decision about whether to vote or not? Proving those two links would be the minimum required to demonstrate that there was something causal going on.

Most likely it's a bit more complicated than that.

Malcolm Todd said...

There's a likely indirect causation effect: parties are less likely to campaign actively in seats where they have little or no chance of winning, and to some extent even in seats that they are almost bound to win. In other words, it could be strategists'/activists' awareness of where seats are safe that depresses turnout, rather than voters' perception of the relevance of their vote.