Tuesday, August 7, 2012

The power of average

I think one of the most abused statistical metrix since the beginning of time is this thing called the mean, also known as the average. Everybody wants to know what the average performance is, and whether they are above average. If they are above average, everything is surely fine. We are above average afterall. If they are under average, they can be driven to feeling absolutely guilty for being under average. It is interpreted as a sure sign that they are not putting in enough effort. How can you not perform on average?

I don't know what gives average so much power. I get that there is an inherent need in human being to constantly compare itself to the majority and if they are slightly better than the majority, then that's great because it means they are not that much different,thus they fit in; but they are not so ordinary, because they are above average.

I do admit to have succumbed to the opportunity of making a lot of people feel really guilty when they perform below average. So in this respect, as much as I personally do not feel good (and think that it is silly to derive satisfaction) from being above average, I do feel bad (and can make others feel bad) for being below average.

Of course all of these are silly, and irrational at best. Just because everyone else is performing at 75% does not mean that you have to perform that 75% minimum. Afterall, the way the average score is calculated, you would understand that it is just an indication of middle score and nothing else.

But it is so common that it has become acceptable with very little, if any, questions. The average tells us basically almost nothing. Take for example, the performance of a student across his subjects in a given semester. He performs at 90% in one, 80% in another, 70% and 60% in the rest. This gives him an average of 75%, which is interpreted to indicate that he is likely to score 75% in his subsequent subjects, but in no way explain the spread in his performance in the semester just passed. Change the student performance to company performance or personal savings goal and you get the drift.

The point is that deviation from the average is an acceptable, given fact that almost always happens in every context imaginable. It makes predictions more challenging, but also more interesting all the same.

I am currently thinking of a business issue that has got nothing to do with averages computations, but for some reason, I started thinking about averages and how it is often used in modelling and the plethora of people who caution against relying too heavily on it. I just need to get this off my head.

On a totally different note, does anyone know that the multiply platform is going to be closed down permanently on 1 Dec 2012? I used to write in that platform, I guess I must start moving my pieces. I find this exercise rather annoying, but it is also one of the reasons why I use a platform like google, which is highly unlikely to close down permanently.

No comments:

Post a Comment