Regression to the mean
Psychology of effective training
One of the questions I was asking myself was the pyshology of effective training, how to train skills efficiently ?
One of the important principle of skill training is rewards for improved performance work better than punishment for mistakes. This proposition is supported by much evidence from research on birds,animals and humans.
Well some of you may argue that this is not correct. Reward is not better than punishment, well this may be true for animals and birds but not for humans. Praises are usually followed by a bad performance while punishment are followed by significant improvement.
The observation of people of second opinion is is astute and correct, but the inference they had drawn is wrong .
What they had observerd is known as randomness to the mean, which is randomness in the fluctuation of quality of performance.
Understanding Regression
Whether undected or wrongly explained, the phenomenon of regression is strange to human mind. So strange that it was identified and understood only two hundred years after invention of gravitational theory and calculus.
It was identified and discovered by sir Francis Galton. A half cousin of Charles Darwin and renowned mathematician of 19th century.
The correlation coeffiecient between two measures, which varies between 0 and 1 , is a measure of the relative weight of factors they share. For example we all share half of our genes from our parents, so the correlation from a parent to child is not far from .50.
It took Galton several years to identify that the correlation and regression are not two concepts they are different perspectives of the same concept.
- When the correlation between two scores is imperfect there will be regression to mean
Praises are usually followed by a bad performance while punishment are folllowed by significant improvement.
Now we can explain why the may be correct and inference is wrong the correlation between punishment and performance is is not perfect but in some cases we can observe this scenarion that punishment leading to better performace.
He said that experiece taught him that criticism is more effective than praise, what he doesn’t understand is that it’s all due to regression to mean
The illusion of understanding
The falwed stories of the past shape our views of the world and our expectations of the future, these are called narrative fallacies .
narrative fallacies continously arise from our attempt to make sense of the world. The narrative fallacies can arise due to halo effect. If you like the president’s politics, you probably like his voice and appearance as well. The tendency to like (or dislike) a person including things you have not observered is known as halo effect.
The examples of halo effect can be seen everywhere if a sports person is handome and attractive you may think that he is likely to perform well in the match and vise versa.
How to tame the Halo effect
One of the best stratigies used to tame halo effect is to use a general principal called decorelate error.
This principal is easy to explain with an example, if you are assigned with a task to count the number of coins in a jar, for an individual the judgement for this kind of task can be poor but if you average out the opinion of a group of people your judgement can be remarkably well.
Some people may underestimate and some may overestimate but if we average out all of their opinion it will give better judgement.