Wednesday, December 17, 2014

Refine, Test, Refine More. Test More

A pattern is a much less rigorous form of a scientific theory. That is to say that the pattern is extracted from experience in an attempt to explain it economically. And the pattern, like a theory, must be hardened through testing. So, if you find a useful pattern you must test it in two ways. First, you must make sure that it does, indeed, explain the phenomenon. Second, you must use the pattern to predict future events and see if it holds up under that requirement. Supposing that it does, you must keep looking for ways to refute your pattern. Over time, if you cannot refute the pattern then it is probably a pretty good pattern. And if you do manage to refute, you might try refining it. However, as you attempt to refute your theory, you must be aware of some troublesome cognitive biases. That is we are inclined to accept evidence that supports our theory or pattern while dismissing data which refutes it. We want our pattern to be correct. But, in order to make it correct, we have to view it critically, rejecting aspects of it that don't fit and continuing to challenge the parts that seem to fit. Over time, if you objectively challenge the pattern, you will strengthen it and increase its utility.

No comments: