Statistical Democracy
Tuesday, January 03, 2006
Over the holidays, I had a chance to discuss the incumbent Canadian election for Prime Minister with a few friends.
These conversations led me to start thinking about statistics and survey theory and concluded in my theory on Statistical Democracy.
First of all, Statistical Democracy doesn't exist. It is something I invented over the last couple of days--perhaps invented is too strong of a word. It is more of an after thought after a good dinner.
Statistical Democracy would work like this:
- Get a sample population, for every election, that is representative of all Canadians.
- For each election, school the sample population on the various platforms that are available.
- Have the "schooled" sample population elect a governmental head of state.
Before anyone objects, we should try to understand why we even have democratic elections--leave aside the two thousand years worth of history, and the struggles our brothers in arms in the greater part of Europe suffered to overthrow feudalism, oligarchies, and kingships out of the window. Essentially, we have groups making decisions because research has shown that groups of people "think" better than individuals do, and, in general, come up with better solutions to difficult problems.
Choosing our leaders in a free democracy does not require creativity, where individuals come up with better solutions, hence we elect our leaders with the idea that a "whole" population will do a better job at electing a representative that speaks for the majority of the people--sometimes, this is even the case.
However, there is a fallacy in our democratic process. The problem is, what I call, "uninformed democracy."
Speaking from personal experience, I don't consider myself well equipped to choose the next leader of our country, as I am not well versed with all the intricacies of each Candidate's political promises. I, then, wondered if the average Canadian is in the same boat I am (for all intent and purposes, I consider myself an average Canadian). Thus, I conducted a little survey: I asked unknowing participants (my friends) to explain to me why one candidate's stance on a particular issue was different to the other candidate's stance (the details are not important), and the response was the same: "I'm not very familiar with the issues."
It appears to me that we can't all be experts in the details of every platform; and we have our own bias when it comes to political thought. In other words, our irrationality leads us to elect our new leaders without any objectivity in the process, and, I would even go as far as stating that we, never vote with the greater good in mind.
Overall, our system depends on the law of averages: the more votes there are, the closer we will get to a true average, and perhaps the correct choice for the majority of the population.
My system of Statistical Democracy proposes to school a random sample population to make the decision for the whole country. Thanks to statistics, we can almost be certain that the sample population we select will be representative of every Canadian within a certain percentage error (something we call Confidence Interval, and it has this funky looking formulas that allow us to calculate such things). Therefore, in this mind experiment, my sample population will make a better decision, not because of the complete number of voters (which never happens anyway), but because the sample population will be extremely well versed on each candidate's platforms and will make an objective selection, based on facts, that will benefit the majority of the population depending on the current needs of the country.
In this particular case, knowledge will be power and the sample population will chose a better leader because it will fully understand why Mr. X is the best choice to be the prime minister.
There are many issues that can come up with a system like this, however. For example, corruption could take hold of the process as lobbying of these voting citizens could lead to coerced results; the sense of empowerment of the actual voters could lead them to think of themselves a different class of citizens (perhaps better); resentment could arise, as not everyone would agree with the election results--but this already happens and yet we just deal with it; I can think of endless doom scenarios where anything like this could go terribly wrong--i.e., the anti-utopian utopia--so I'll stop. But I do think there could be ways to mitigate problems as they arise. In addition, the sample population will be anonymous, and the election should be carried out as a one side blind type of experiment: everyone (or whoever decides to vote) votes, but the sample population's vote are the only ones that count, and no one knows who the real voters are--except of course, the citizens who were fully and objectively educated with the different options.
I always argue that Statistics is the most applied of all Mathematics. I don't see why we couldn't use it to get a sample representation of our population and have them study the platforms thoroughly in order to exercise an objective, educated suffrage. I think it would yield a better leader, and everyone would probably benefit--even the ones who don't currently vote.
By the way, I am not saying that we should not have democratic elections or that I do not believe in having an individual voice of my own, on the contrary, we've come too far to just let go of the best system our societies have come up with in our written history, but I think it can always be improved for the betterment of the majority--of course, if everyone took the time to know everything about every option out there, I wouldn't even think of such dumb ideas of Statistical Democracy.
Comments: