Bayesian Thinking and How It Can Help You Avoid 4 Important Cognitive Biases
Using probability and statistics for better decision making
“You have changed. You were a different person back then.”
My reaction to such comments used to be defensive where I would eventually succumb to my inability of defending my inconsistency.
“Consistency is valued and adaptive and inconsistency is commonly thought to be an undesirable personality trait” — Allgeier, Byrne, Brooks, & Revnes
My new line of defense — Bayesian thinking! To those new to the concept, and to put it as succinctly and practically as possible:
Bayesian thinking extends the popular Bayes’ formula to critical thinking and urges us to incrementally update our probabilities as we encounter new information.
A commonly used example to understand this is the ‘Good Driver Conundrum.’ Let’s say that you believe you are a good driver. But one day you get into a car accident. Your instinctive reaction might be either of the two: Dismiss the evidence and defend your prior belief (“It was that crazy person’s fault”) or give new evidence importance like nothing else matters (“I was wrong, I am a bad driver”)
But if you analyse it from the lens of Bayesian thinking, you will look at the prior belief (“I am a good driver”) in the light of new evidence (getting into an accident) — one instance of accident should not change your belief but should update your confidence in it. It is not black and white or 0% and 100% that you are a good driver, it is more gray scale and you assign it a more practical probability.
The 4 Cognitive Biases And How Bayesian Thinking Can Be Helpful
1. Commitment & Consistency
One of the six principles established by Cialdini (1984) in his book Influence, it describes the human tendency to believe and act in a way that is consistent with their values and self image.
When we take a stand or “commit” to something, we will be inclined to behave consistently with that commitment. We will keep fooling ourselves to keep our thoughts, and actions consistent with…