[心得分享] Calling Bullshit-It is not about making you look smarter, it is about making "others" smarter

 

Calling Bullshit: The Art of Skepticism in a Data-Driven World

作者： Carl Bergstrom and Jevin West

推薦分數：⭐⭐⭐⭐☆

英文難度：⭐⭐⭐⭐☆

（GJ的英文程度-英文指考75分，多益875分，單字量沒有高中多，口說可以表達自己但絕對不到native）

People love to talk bullshits for lots of different reasons. Maybe it’s because they can receive benefits, or maybe they accidentally said something wrong, or even just want to make fun of it. But things get worse after the boom of social media, fake news fly fast and truth never. We get eyes on dramatic headlines and share those information without confirming its authenticity, and even after there are following information that corrects this news, it doesn’t matter. Because the ordinary story is what you think is more interesting. This causes the fake news like NFL athletes burning American flags, manipulation of USA election and lots of different impacts. That’s why noticing and calling bullshit is so important and I would like to share some of the tips in the book.

First let’s talk about selection bias. This is a common bias you know if you read about statistics, but the selection I want to talk here is a bit tricky. There was a news talking about New Taipei City had the most numbers of crime among whole Taiwan in the past month, and it concludes that it is the worst city to live( I may over exaggerate it but it is something like it), the mayor of the city did a bad job. But if you live in Taiwan, you would know that this might be bullshit information because, the number of people living in New Taipei City is the most among Taiwan as well, having the greatest number of crimes is a bit rational. This is something we call it selection bias, people choose to select the information that benefits them and ignore the data that seems more important to others. In order to avoid this kind of mistakes, think more and share less when you receive any information. You will find it useful before you press the share key because you had really think through it.

Second is the Goodhart’s Law. When we set a target, and try to measure the result, we usually get better result then it should be. For example, if the Ministry of Education announced that it will reward schools which have the highest score of their students test result. What might happen if you are the president of the school? If you want to get the rewards, you have the motivation to increase the test score result, but just pushing student isn’t enough, you may accidentally tell teachers to share the quiz paper or cheat during the test. The result may seem good to the Ministry of Education, but it is not what it truly is. The way to rethink about this is remember a rule of thumb, when things or too good or too bad to be true, there must be more information for you to dig in.

At last, I want to share Fermi Estimation to you for you to spot bullshit numbers. Fermi Estimation is an estimation to calculate a nearly good enough result to check the numbers. We usually face this kind of questions during interview. For example, can you estimate how many gas stations there are in Taiwan. You have to start to estimate, there are 23 millions of people in Taiwan. Let’s say 1/3 people have cars, a car goes to gas station once in 2 weeks, a gas station handles a car averaging 10 mins. Now we can estimate the number, and get the result 4000 gas stations in Taiwan. If the number is way too high or low against this number, you need to be alert and check deeper from the data source. I would like to end with a sentence wrote by the author, calling bullshit is not about making you look smarter, it is about making “others” smarter. Don’t think you are the one and only one, influencing others to be better is way more important than what you think you are.

“Calling Bullshit” 2022/08/14