On the day of the Scottish independence referendum, it is interesting to think about how large collections of people should make decent decisions on big issues. Voting isn’t a bad way forward, but when issues are big and likely to be irreversible (at least for a while), there is a fear that a bad decision might be made. In particular, there is always a fear that some minor slip-up, or some temporary surge of feeling, might distort the result.
One approach to this is to require a “supermajority”. That is, the change needs the approval of more than 50%, for example needing 66% support or 80% support. Surely, the argument goes, if a decision is that important, it oughtn’t to depend on the whims of a few people around the borderline. This approach brings a bias towards the status quo—it sees the change as the problem, whereas we might want to say that the decision not to change might be just as momentous a decision. Put another way, once something has been fixed one way, it means that a small minority can keep it that way.
Instead, I propose multiple votes over a reasonable time scale. One of the problems with the single vote, even with a supermajority, is the “morning after” effect; a rush of enthusiasm for one side or the other, or a single screwup by one side, can mean that people might make a capricious decision on the day. By repeating the vote a number of times and averaging in some way, these effects could be smoothed out.