Thursday, 21 April 2011

AV, status quo bias and definitions

One of the arguments given against the Alternative Vote system is (as laid out in this good but rather long post) that "Under AV the person who comes second can win."

Gowers points out in the linked article that this is not true - all it means is that the person who would have come second under FPTP can win. Of course, the whole point of the referendum is that a different person could win under AV than under FPTP. The reverse argument is equally true: the person who would have come second under AV might win under FPTP.

But why does this argument have such appeal? Even AV defenders are trying to make a rational case for why it may be more democratic for "the person who came second" to win. Instead, one might expect them to challenge the premise.

The reason seems to be that the status quo bias is very strong here. People who might think they don't suffer from status quo bias (in that they have no particular desire to keep the existing voting system) may fall into it unknowingly, through the definition of the terms in the debate. It is much more intuitive to define the word "winner" in the terms of the current system, rather than the potential alternative system.

This is a particularly insidious form of bias, because it's very hard to see - spotting it requires an understanding both of the bias itself and of how strongly language and framing affects our choices. And even once you know about it, it might take a heroic level of introspection to change your mind.

Definitional bias, as we might call it, is really important in trying to persuade people to make a choice. This insight is related to the idea of linguistic relativity - the words we use shape our worldview. And our worldview, of course, affects our behaviour. But by focusing in on a single phrase like "winner" it's easy to see exactly how the phenomenon works. We start with a definition of "winner" based on our current system; we agree that the winner should have power (it's the basis of representative democracy, after all); and so we fear that a new system, where the current "winner" will not win, will somehow be less democratic. One could instead argue that we're just changing the definition of "winner" to a better one. But that's a bit too abstract for political debate.

The bias will typically appear when something that's broadly good (a democratic electoral system) is being compared with something a bit better (a democratic AV electoral system). And because there are legitimately positive things to say about the existing system, it will become very easy to present the new alternative as a threat to something positive. It is easy to assume that it's good for the person with the most first-choice votes to be elected, because that's what we currently define as "the winner". And why shouldn't the winner, er, win?

A similar phenomenon includes the bias towards established brands (most people choose Coke, so Coke must be better than [insert newly launched brand] - which neatly bypasses the circular question of why most people currently choose Coke). No doubt there are many others.

The same bias could in theory be turned against the existing system by finding things that people agree are bad about it - for instance safe seats, corrupt MPs - but the problem is that AV won't eliminate those things, it just (might) reduce them a bit. A category distinction is a much more powerful influence on decision-making than a quantitative, incremental change. Is there a definitional term which could make the case for AV obvious? Maybe "fairer voting" is the best candidate, which is indeed the phrase that the Yes campaign has chosen to focus on. But it's still not viscerally obvious that the current system is intrinsically unfair, so this argument relies on intellectual reasoning rather than gut feel.

The gut is always powerful in politics; I think that most successful political arguments are either based on the gut or on a definitional change. Is there either a gut, or a definitional, argument for AV which might be persuasive?

1 comment:

Anonymous said...

We already have a term to describe this phenomenon: begging the question.

I'm not sure there's any benefit in recasting it in behavioural econ terms as status quo/ definitonal bias. It's not like we're coming up against the bounds of rationality; it's a simple logical fallacy that one would learn in any intro philosophy/logic course.

And I’m not sure it adds anything to the explanation of WHY people might happen to make this mistake. The answer is a failure to consider logical form; talking about how the words we use frame our thinking is just a fancy way of saying the same thing.

Steve