This book opened my eyes to the differences between bias and noise, and how they are equally proportional to disrupting our desired outcome. That is to say, imagine missing a target by some random “noisy” amount, which widens the bell curve of outcomes. The distance, on average, away from the target result would be just as bad as an equal error magnitude in bias. The only difference being that the bias is easier to see, because it is clustering away from the target. Thus, bias is often the focus in reducing error for our desired outcomes, when really, it should be considered equally with noise, if our goal is error reduction.
Simple techniques for reducing error through noise include:
- a noise audit,
- collecting multiple opinions (even yourself multiple times), without priming the answer of individuals, and
- regularly measuring noise
Three Principles of Decision Hygiene
- “Resist premature intuitions.”
- “Think like a statistician.”
- “Accuracy, not individual expression.”
See a quick summary of these ideas here, or review the book for greater detail!
5/5.