[This blog has been dormant for VERY long. Thought I would stir up something:)]
Q1: How do you kill a blue elephant?
A: with a blue elephant gun
If only life, data and analysis were to be as simple as that. As I see it, my data is the blue elephant. And the blue elephant gun is that miraculous hypothesis that comes closest to explaining the pattern in the complicated chaos called the dry forest. Lesson no. 1: it’s not exactly that straightforward to kill a blue elephant with a blue elephant gun. Reasons? You are not really sure whether your blue elephant is actually a blue elephant or not. What if it is a red elephant. Most likely, it is!
Q2: How do you kill a red elephant?
A: Strangle it till it turns blue. Then kill it with a blue elephant gun
Really? But hey, that makes things better now doesn’t it? So I have a red elephant. A lump of data that does not behave in any civilized fashion, graphically it’s just a mass of spots that are all over the place. Plus, all the great variables that you measured thinking that they were explanatory are suddenly just a bunch of variables. Now what? That thin rope of logic called ‘statistics’ helps you successfully strangle the red elephant. And voila! It is now blue enough to kill with a blue elephant gun. (maybe they should rename this science as ‘sadistics’ instead of ‘statistics’). But what if it’s a green elephant?
Q3: How do you kill a green elephant.
A. You make the green elephant really angry, till it turns red, then strangle it till it turns blue. Then kill it with a blue elephant gun
Phew! And I really thought I was in trouble here. You don’t see your red elephant? Log-transform your green elephant into it! Or do a square root. Or half a dozen other things that can convince the average reader that the green elephant is actually red. It’s fairly simple then on. Repeat steps mentioned in Q2.
Done all that?
Q4: Where is the blue elephant gun?
A: I wish I knew
So in the midst of converting various elephants into various other elephants, one loses track of the blue elephant gun. One must not blame oneself for that should one? After all, turning red elephants to blue elephants takes a lot of work. And that’s when u realize that that’s just the tiniest tip of the iceberg. One has to constantly reinvent the blue gun to kill the newly generated blue elephant. And that loop tends to infinity.
Q5: when do you stop?