SecretMethod70

Plurality voting is the technical term for what is commonly referred to as "first past the post" voting. Everything else you mentioned (proportional parliamentary representation, political parties, etc) has nothing to do with the math of the voting. All of that is what you do with the results of voting, but says nothing of how those votes are cast and tabulated.

Plurality voting gets its name because the winner is the person who gets a plurality of votes (more than anyone else) but does not have to necessarily get a majority of votes. It is incredibly simple, which is why it is so flawed. Voters vote for one person, and whoever gets the most votes wins. Second (or third, or fourth) opinions are entirely disregarded. In terms of determining the will of the voters, it's terrible. Case in point: the 2000 US presidential elections. The electoral college is a convenient scapegoat, since it had the most obvious and visible effect on the election, with the winner of the national popular vote not winning the election, but in reality that difference was only a symptom of the real problem. If Florida had used any other generally accepted system of voting, other than plurality voting, Al Gore would have won Florida and the electoral college would have elected him as president.

Let's take a look at a few, in order of increasing complexity (there are quite a few). First there's approval voting, in which voters do basically the same thing as plurality voting, only they can vote for as many candidates as they want. Vote for everyone you don't mind, don't vote for anyone else. If this had been used in the 2000 US presidential election, most of those Nader voters in Florida would have also voted for Al Gore, and so he would have gotten those votes as well, causing him to win Florida and be declared the winner. However, approval voting still has the very significant flaw that a candidate can be declared a winner even without majority support.

Then there's instant runoff voting. With IRV, voters rank candidates according to their preferences, and a candidate does not win until they have a majority of support, rather than just a plurality. Let's say there's a race with candidates A, B, and C, and there are 5 voters in this election. Two voters vote, in order, A, B, C. Another two vote C, B, A. And the fifth votes B, A, C. That gives A two first place votes, C two first place votes, and B one first place vote, but none of the three candidates have a majority. So, since B only got one first place vote, he's out of the race and then you look at the second choice of the B voters, which is candidate A. That gives A three votes to C's two votes, and A is the winner. If IRV were in place for the 2000 elections, Nader voters would have been able to declare their support for Al Gore, and he would have won Florida.

IRV gives particular weight to first choice preferences by only counting those first, and then only counting the second choice of ballots whose first choice candidate was eliminated. There is a somewhat similar voting method which works differently, called the Condorcet method. There are a lot of sub-methods available within the Condorcet method, but I'll stick with the basics. First, it's important to understand that the Condorcet method is so named because it is focused on satisfying what is called the Condorcet criterion, which is that the winner of an election should be the candidate that would defeat all the others in a head-to-head match. Voters in a Condorcet method election vote in much the same way as they do in an IRV election, by ranking the candidates. But then, the votes and rankings are looked at the find the candidate who would defeat all the other candidates in a one-on-one election. This method is less biased towards first choices and, instead, seeks to find the best overall choice. Using the previous example, with 2x(A,B,C), 2x(C,B,A) and 1x(B,A,C), the Condorcet method would see that A defeats C 3-2, but loses to B 3-2. It would also see that C loses to both A and B. Finally, it would see that B defeats both A and C in a head-to-head election. So, the winner of the election using the Condorcet method would be candidate B, which is a very different result from the IRV election!

No voting system is absolutely perfect. There are, for example, anomalies where there is no Condorcet winner, and there are a number of ways to deal with that. Also, the larger the electorate, the less likely an anomaly is to occur. Personally, I think the Condorcet method is the superior voting method, but IRV isn't terrible either. Between the two, it's a philosophical issue of whether or not you think first choice should trump overall preference or not.

Regardless of which voting method is best, there's one thing that is generally agreed upon: plurality voting is the worst. The really sad thing is that this isn't news. Marquis de Condorcet (who created the Condorcet criterion) lived in the 18th century. Personally, I've been reading about how bad plurality voting is since before the 2000 election even thrust the US election system into the spotlight, only to be disappointed to see that that spotlight is rarely shone on the actual problem of how we count votes (except in magazines like New Scientist or Discover, which is where I first read about the problems with plurality voting during the 2000 primary season).

Anyway, I hope that clears up the confusion. The parties and parliament and proportional representation, etc etc etc have nothing to do with the voting method. The voting method is just the math behind determining the winner of an election.

(sorry for the thread-jack)

__________________
Le temps détruit tout

"Musicians are the carriers and communicators of spirit in the most immediate sense." - Kurt Elling