Thinking fast and slow 17. Regression to the mean

Following our reading calendar

It is almost as if Kahneman had read Carlos comments on the former chapter and what do we have here? A story!

After my first reading of this book, one of the things I do remember more deeply is the story that opens this chapter and brought Kahneman to propose his “regression to the mean” bias.

So the story goes, while presenting some of his ideas to flight instructors one of them challenged him saying that rewarding good behavior was a bad strategy.

“On many ocasions I have praised flight cadets for clean execution of some acrobatic maneuver. The next time they try the same maneuver they usually do worse. On the other hand, I have often screamed into a cadet’s earphone for bad execution, and in general he does better […] So please don’t tell us that reward works and punishment does not, because the opposite is the case”

There is another explanation of course. The flight instructor was just observing a regression to the mean. Performances are not steady, they fluctuate, so after some good maneuver is not strange to expect a less than perfect one, and viceversa.

They way Kahneman teaches their mistake to the instructors is also cool. He made them throw two coins without looking and see if they can hit a target. Those that did it better the first time tend to perform less well in the second try, and viceversa

After the not so convincing biases of the former two chapters, here we return to something that makes sense. Luck, randomness is a very relevant force in our daily lives, but we are not designed to accept randomness, we want causal explanations for everything. So if an cadet performs poorly with his plane last time and now he is doing better, it is because the instructor yelled to him, not because there was a regression to the mean. Very Talebian again.

With a little bit technical analysis, as he explains Francis Galton’s discovery of regression to the mean, and the main teaching of this chapter”Whenever the correlation between two scores is imperfect, there will be regression to the mean”

One of the examples he uses to explain is not ok, though. He says:

“Highly intelligent women tend to marry men who are less intelligent than they are” and invites readers to find an explanation.

If you use the word “tend” most English speakers would understand it as “there is a connection” no matter whether in Statisticianish -thank you Carlos for the word- it means “there is a correlation between two magnitudes but not necessarily a causal link”. And if you ask for an “explanation” then people will look for a “causal explanation”.  And no, for most people in Planet Earth “The correlation between the intelligence scores of spouses is less than perfect” is not an explanation.

However, in this case, this example doesn’t jeopardize the main conclusions. When judging regression to the mean we find it very difficult, and we should keep this possibility in mind all times instead of trying to find a causal explanation every time there is a flux in a distribution.

I think the key is that this time Kahneman is not revising how we use terms like “probable” but he is talking about how things, animals and people behave in real life. The world is a lot more random than we think it is.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s