The Apes—Points To Ponder

Cognitive Bias and Mathematical Manipulation

In formulating hypotheses and proposing scientific models of real-world events, scientists almost always encounter the fact that their carefully designed experiments do not produce the hoped-for result, but may provide something that is close to the hoped-for result. In this area we encounter the problem of “cognitive bias.”

The term “cognitive bias” describes errors in thinking processes caused by holding on to individual preferences in the face of contrary evidence. This could be described as “unintentional dishonesty,” in that the individual affected by it is completely unaware of their bias. Where it crops up in scientific experimentation (outside of psychology, where it is an area of study), it is described as “confirmation bias.” It is the tendency to interpret experimental results in a way that confirms one’s cherished hypotheses or even pre-existing beliefs. In science pre-existing beliefs are often just fashionable theories.

The scientific method is supposed to eliminate such bias by the process of peer review. Other experts in the field review the published results produced by a specific scientist or scientific team and offer critiques. However, peer review is only effective if the reviewers are not also suffering from the same confirmation bias.

As we review some of the theories of modern physics in the coming pages, we will encounter the existence of “adjustable parameters.” We can explain by example:

Consider the trajectory of a ball thrown at an upward angle through the air. It will follow a parabolic curve almost exactly, rising in the air at first and then falling. It’s position in the air at any point will depend on the initial upward angle of its trajectory and the time elapsed since it was thrown. If there were no other forces affecting the ball it would move in a perfect parabola. However, the resistance of the air to the movement of the ball inevitably distorts the parabola.

If we adjust the mathematical equation by adding an “adjustable parameter,” we can compensate for the air resistance. Adding a fixed parameter might do the trick, but air resistance can vary. It will be different at sea level than on a high mountain, and hence the parameter will need to be adjusted, for context.

This does not mean that the theory of the parabolic movement is incorrect, just that we need to adjust the model. The scientific problem is not that adjustable parameters are necessarily wrong—they may not be. But if you cannot assign a cause to the adjustable parameters in a model, then the model is clearly suspect. You can usually make inconvenient results look respectable by resorting to adjustable parameters.

16