The May 3 Scotland Parliamentary Election used a mixture of single-member constituencies and proportional representation. 129 seats were up. 73 members were elected from districts, the old-fashioned way that has always been used in the U.S. (in other words, whichever candidate gets the most votes, is elected). The other 56 seats were filled proportionately.
Voters voted for a district member. Then, in a different part of the ballot, they voted for their favorite party, for the non-district seats. The results: Labor received 37 district seats but only 9 more proportionate seats, for a total of 46. The Scottish National Party won 21 district seats but 26 more seats from the proportionate part of the ballot, for a total of 47. The Conservative Party won only 4 district seats, but 13 more proportionate seats. The Liberal Democrats won 11 district seats and 5 more proportionate seats. The Green Party won 2 seats, and one independent was elected.
Even though the Scotland Parliamentary elections had used this same system in 2003, this time there was confusion. Whereas in 2003 there had been one ballot for the district seats and one district for the at-large seats, this time there was only one ballot for both types of seats. Unfortunately, the directions said “vote twice”. The people who wrote the directions meant to explain to voters to vote both for a district seat, and an at-large seat. But some voters tried to vote for 2 candidates for a district seat, or tried to make 2 choices for the at-large seats. Approximately 100,000 ballots were spoiled.
Completely separate from the Parliamentary election, were elections for local council, held all over Great Britain except for London. In Scotland, these elections used Single Transferrable Vote, which had not been used in Scotland before. It worked well.
Great post, Richard.
But I thought the error rate in the choice voting (STV) side of things was only about 3%.
It was closer to 10% with the MMP, no?
The first press report I found says that a major cause of invalid MMP ballots was people ranking the candidates, as if the vote for the regional parliament were also an STV election.
More generally, the press coverage seems vague on which problems affected which election(s). As might be predicted, the press is more concerned with making events sound out of control than with getting any of the detail right themselves.
It cannot have helped that this was (I think) the maiden voyage of an optical character recognition system that is supposed to read handwritten numeric rankings as well as X’s in a box.
I think that both the Electoral Reform Society (in the U.K.) and FairVote (in the U.S.) need to mount major damage control operations.
Matt Shugart has an intelligent discussion of the counting problems here: http://fruitsandvotes.com/?p=1230.
As an observer of these elections, I disagree with much of the analysis of the problems with the voting. Without any doubt, by far the biggest factor was that the government moved from holding the mixed member PR elections on two ballots to one, and had a large-type instruction on the top saying “you have two votes.” If they had used the old ballot design on two ballots, adjusting it slightly to be able to be counted on the optical scan machines, the error rate would have plunged. You can see my fuller comments in a few articles, including one in the Scotsman at:
http://news.scotsman.com/politics.cfm?id=701122007
The STV elections indeed did go smoothly, and produced remarkable shifts in the direction of fair representation, by the way.