## The Wrong Question

At some point in your average statistics class, you learn about hypothesis testing. Creating and testing hypothesis is one of the most basic concepts in science, so we’d better do it right.

The way hypothesis testing is usually done is by first collecting data. Then we use a statistical test to ask for the probability of our data, given the hypothesis. Wait … what?!

We already know the probability of our data: it is 100% — that’s why we call it data. And we want to test our hypothesis, so we can’t assume it is true. What we really want to know is how likely our hypothesis is, given the data. Clearly, we were asking the wrong question and started with invalid assumptions.

How will we ever be able to teach people to challenge their assumptions when we also teach them to make invalid ones? How will we teach people to question everything when we also teach them to ask the wrong questions?

Published
May
2nd,
2012 by Sebastian — All Posts — Impressum