A question about a weakness in PR-STV

I think I’m turning into a political wonk. (Yes it’s a word! Google it or check an online Dictionary.) Another comment that wouldn’t fit in the 1,000 character limit – the election has given me a dose of verbosity!  This was writing a comment on RTE’s news report about the election count, first recount and second recount (re-recount?) here in Galway West (not very interesting in itself and I won’t bother linking to it, because there are better reports elsewhere).

I need a clarification if anyone out there knows. I had always understood that when a surplus is distributed, the whole of the candidate’s vote is recounted and then the surplus is distributed according to the proportions of the total vote. So for instance if I was on 11,000 votes and the quota was 10,000, the next prefs on all of my 11,000 votes would be counted. If candidate A was next on 50% (5,500) of them, he would get 50% (500) of my surplus.  However, I  read yesterday (OK, on Twitter, but from a source who has seemed to know his PR-STV stuff over the weekend) that they take a random 1,000 votes from my 11,000, and just give them to whoever is next on the ballot paper.

The problem with this is that a random selection cannot be a true and accurate sample of a group, and therefore there is a random sampling error introduced every time a surplus is introduced. Over a 5-seater count, when multiple surpluses are distributed, the random sample error is compounded.

And the upshot is that a small difference in the final vote (e.g. 17 votes, as with Galway West or #gyw as it is now known) is not statistically signficant and does not reflect the will of the people.  It’s a thought, isn’t it?

In case you are wondering why the random sample doesn’t give a different result in a recount (or a re-recount): the random sample selected at the first count is kept separate, and is the only part of the elected candidate’s vote to be examined when their surplus is distributed.  This seems particularly stupid to me, because recounts are only called when the margin of victory is tight. Surely when every bloody ballot paper is being looked at by solicitors, they could look at all the next preferences and get a true picture of the voters’ intentions.

I’m quite disappointed at this discovery as you can see.  Does anyone know anything further on the subject?


~ by Lynn Duffy on March 2, 2011.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: