8. Glossary

Degrees of freedom

This is the difference between the number of equations used to interpret a phenomenon and the number of parameters involved in the final equation.

For example, if we estimate that only 3 parameters can explain the variations of a phenomenon, to within ε, for which we have 10 observations, we'll say that we have (10-3) degrees of freedom... On the other hand, if you want to introduce 10 parameters (or interactions, if applicable) into the final equation, you'll be able to explain the whole thing without any degrees of freedom; in other words, it's good enough if you're sure you've chosen the right parameters, but it's no good at all if you've made a mistake in your choice. This validity can be easily verified by observing an eleventh experiment: if it confirms the hypothesis, validity is almost certain; if it refutes the...