Regression to the Mean
There was an experiment to praise and criticize pilots. When the student performed a good landing, the flight instructor praised. Then, the next landing got worse. But when the student performed pretty badly, the instructor criticized. Then, the next flight got better. So the researcher concluded that criticism works better than praise.
But, in fact, criticism and praise have very little to do with the landing. The main factor is the “regression to the mean”. Suppose you perform a set of experiments. If one sample fall very far from the mean, the next one would be closer to the mean to make up for the large deviation from the mean. So, when the student performed particularly bad, the next landing tended to be better, irrespective of what the instructor might say.
Sample Bias
When doing a poll, we need to take samples from a group with great diversity so that the sample set can represent the entire population. But that’s not how the media work. A story of patients sick because of bad medication sells better than a story of a patient cured expectedly by the same medication. So, the media tends to show only the bad story. And, again, our feeling is more sensitive to media than numbers. We believe the media, not the fact.
The same principle applies to basketball player who taking a good shot for few consecutive time, performing much better than his average. Should we be more worried that he’s not gonna score the next shot because of ‘regression to the mean’.
Gambler Fallacy/ Hot Hand Fallacy
A fair coin would turn heads and tails with equal chance. If you see heads 5 times in a row, the next flip is still has equal chance of turning head and turning tail. But, our feeling doesn’t believe in this. Our feeling tells us that the chance of tail is greater as the coin turns heads more and more.The same principle applies to basketball player who taking a good shot for few consecutive time, performing much better than his average. Should we be more worried that he’s not gonna score the next shot because of ‘regression to the mean’.
---------------------------------------------------------------------------------------------------------
Book or Audiobooks?
Personally, I prefer audiobooks. It's fun, and I can listen when I'm doing something else. It also makes other activities (e.g., jogging) a lot more fun. For more detail about audiobooks, please read [this post].

However, when the material is an audiobook, it is extremely hard to locate a specific part of content. Most likely you will have to listen to the entire audiobook once again.
This book summary will help solve the pain of having to go through the book all over again.
I am leaving out the details of the books. Most books have interesting examples and case studies, not included here. Reading the original book would be much more entertaining and enlightening. If you like the summary, you may want to get the original from the source below.
No comments:
Post a Comment