With the Australian federal election a few days away, all eyes seem to be on (among other things) daily opinion polls.
To generate data for those polls, people are asked "Who would you be most likely to vote for if an election were held this Saturday?". Chris Lloyd, one of our professors at MBS, argues that there is a lost opportunity in asking this question because "binary data gives much less information than continuous data – roughly 5-10 times less information." As a result, he argues, these opinion polls don't capture the swing vote, even though that's where all the action is. I agree with him completely and wish they would get all that extra data for us. However, I have three theories as to why they don't do a more comprehensive survey.
The first is the obvious one (though a guess on my part): the economics aren't worth it. That is, the comprehensive poll will cost too much because it'll both take longer and will be more complicated to execute.
The second is the less obvious one: it'll add too much uncertainty into our lives (or, put another way, it'll give us too much information). Because these votes are swing votes, they'd change much too often for the general public's liking. If the numbers jumped up or down by, say, 5% every other day, the polls would be perceived as being overly sensitive and voters as being overly fickle. Eventually, because the polls would seem almost erratic, the public would lose interest in them (they would seem less relevant as accurate gauges of solid public opinion), the media would stop buying them from polling agencies, and everyone would lose money. Yes, this reason comes down to money as well.
The third is the least obvious one: people would prefer not to know (or, put another way, too little uncertainty isn't fun, either). Because, if they did know (with a reasonable amount of accuracy) who was going to win beforehand, it would make the actual voting on the election day much less exciting. And, in the same way that no one wants a boring, predictable sports tournament final, no one seems to want a boring election day either.
At the end of the day, though, people have to vote with their feet and not with their survey responses. And that little bit of uncertainty -- that "anything can still happen" feeling -- keeps people interested and much, much more motivated. Both of which, ultimately, lead to a better election process and a better execution of a democracy because everyone will have been interested in it and everyone will have participated in it.