This article is from WeChat public account:Immature Research (Public ID: PSR-26) , author: Luo Yuxiang

When the Nobel Prize is presented each year, the circle of friends will be screened. However, since most people, like me, the level of physical sciences such as physical chemistry is in a state of decreasing year after year in high school, it is generally only when the Nobel Prize is announced that it can only blaze the literary awards, and follow the trend(number) (落)After Haruki Murakami.

Economics is very strange.

For the sake of reason, it is a discipline about society. It should be very resonating, and anyone can talk about it. But given the depth of its theoretical (and mathematics), it makes it hard for laymen to think about it.

But this year’s Nobel Prize is a bit different (The winners are Abhijit Banerjee and Esther Duflo from MIT’s gods and Michael Kremer next door. ). The official awards are:

“Recognize its experimental approach to global poverty alleviation (for their experimental approach to alleviating global poverty).

, the first group accepts policy A, the second group does not accept, then you can have a very good counterfactual.

This is because, although each student has their own differences, the students in the first group and the second group will be very similar.

Randomized, the grouping of students is not affected by any real-world factors, and the probability of everyone accepting Policy A is 50%. So random allocation can guarantee:

Half of the boys in the school will accept Policy A and the other half will not accept it;

Girls will accept Policy A in half and the other half will not accept it;

Schoolmaster will accept Policy A in half and the other half will not accept it;

Fujiazi will accept 50% of Policy A and the other half will not accept it;

The cat family will receive 50% of the policy A and the other half will not accept it;

The dog breed will receive 50% of the policy A and the other half will not accept it;

O type blood will receive 50% of policy A and the other half will not accept it;

Golden Taurus will receive 50% of Policy A and the other half will not accept…

In other words, the group of students who accepted Policy A had an equal composition in terms of gender, learning ability, family income, blood type, preferences, and constellation compared with the control group. At the same time, because the grouping is completely random, in addition to these factors, other externally unobserved student traits (such as the length of night dreaming) will also be evenly distributed between the two groups.

Therefore, the two groups of students randomly assigned will have the same average value in all indicators, which is equivalent to Xiaoming and Xiaoming 2 of the parallel universe. It is a good counterfactual reasoning – between the two groups of students If there is any difference in the passing rate of the test, it can be considered to be caused by Policy A, because there is no difference between the two groups of students in addition to Policy A, which can explain the difference in test scores.

Because “random” can achieve the effect of simulating parallel universes and obtaining high-quality counterfactual reasoning, in economics, randomized controlled experiments are considered to be the golden criterion for establishing causality and testing policy effects.

In fact, randomized controlled trials were originally born in the medical world and were originally used to study whether drugs can actually cure diseases. Using the stochastic principle in social sciences such as economics, let scholars perform experiments like doctors in real society in addition to deducting social phenomena at the theoretical level, and observe the results and understand the causal relationship between things. And judge whether the policy is useful or not.

3. The real-life operation of the Nobel Prize winner

Abdul Latif Jameel Poverty Action Lab(J-PAL), established by the 2019 Nobel Prize in Economics, Banerjee and Duflo This uses randomized controlled trials to study the principles of global poverty and to test the effectiveness of various poverty alleviation policies.

The research content of J-PAL Lab covers a wide range of topics. Here are some of their recent topics:

  • Agriculture: Can rainwater harvest help small farmers get rid of poverty? (Study on 2,880 farmers in Niger, Africa)

  • Criminal: Can summer jobs reduce teenage crime? (for 5444 high school students in Chicago, USA)

  • Education: Can high-tech educational tools improve student achievement?