Both eyes blinded by data

01

On October 23, 2015, the "Science" magazine published a positive article on some Indian children, which tells the story of these children undergoing cataract surgery and seeing light.

On the surface, there is nothing special about this news, because for us, receiving cataract surgery is a rare thing.

However, the facts are not so simple. These Indian children suffer from cataracts in their lives, and their vision has not been clear. When the disease was diagnosed, local doctors told their parents (they all came from remote and poor areas and families who had not received education). These children had passed the key formation period of vision, and now the treatment is late.

Fortunately, Indian ophthalmological experts successfully performed cataract surgery for a ten -year -old Indian child. Hundreds of children with cataracts are now recovering their vision. A 22 -year -old young man could ride a bicycle through a lively market after receiving cataract surgery a few years ago.

02

The concept of "key formation of vision" comes from the research made by David Hubel and Torstine Bozeel on cats and monkeys. The study shows that in a period of a key growth period of animals, if their visual signals lack visual signals, vision will be defective for life.

For humans, this critical period was originally considered to be 8 years old (for moral reasons, no researchers compared through human research).

Hubel and Vestel won the Nobel Prize in 1981 for this study. Since then, doctors around the world have no longer performed cataract surgery for children over 8 years old.

However, although this data is clear, it is wrong. Indian children's cataract surgery shows that the data of critical periods is wrong. From this perspective, the "positive" news on the surface is actually "negative" news. Because this means that many children over 8 years old are refused to perform cataract surgery, and these children are blind for life because of the doctors' over superstitions.

Another news in 2015 also reflects people's over superstition about data. Brene Norteck and a research team tried to repeat the 100 psychological experiments that caused attention in 2008, and published the final experimental results on August 28, 2015 to the magazine of Science. In all repeated experiments, about 1/3 of the original experiments can reproduce the original results.

Even so, the results of the experiment that can be reproduced are much less than the original experiments. Similar problems have occurred in other fields.

A few years ago, an article in Nature Magazine revealed that most of the research on cancer could not be reproduced. In October 2015, "Nature" magazine specially published a special issue to discuss how to reduce the number of unreproducible research. Many people have also begun to consider how to reduce the probability of inconceivable data.

I think this move is wrong. This is an example of basic prejudice, which expresses the desire for reliable evidence that can be regarded as inference. Scientists are willing to weigh between the first type of errors (the result of existence -fake negatives) and the second type of error (discovery is not a real result -false positive).

In fact, when you are committed to reducing the first type of errors, the second type of errors may increase, thereby missing some discoveries. Therefore, we can reduce the required requirements from 0.05 to 0.01, or even 0.001 to reduce the probability of false positive. However, in this way, the probability of false negatives will rise sharply.

Basic prejudice prompts us to do our best to eliminate false positives, but this progress will be slow.

I think there is a better way to give up the pursuit of certainty and realize that any data may have errors. After all, suspicionism is the pillar of science.

This reminds me of a conversation with a researcher. He insisted that we cannot believe in intuition, but should believe in data. Although I agree, I never believe in intuition (should appreciate intuition and make judgments), but I do n’t agree to believe in data.

03

As mentioned above, there are too many examples to prove that data will blind people's eyes. The ability we need to have is to use them without the effectiveness of related data. We should be able to draw conclusions and speculation in the situation of ambiguous and uncertainty.

To do this, we must overcome basic prejudice and liberate ourselves from the expectations of the data we trust. What I mean is not to say that errors in research are reasonable -think of the encounters of Indian children who have caused the blindness because of the data over the data.

What I want to express is that we should not ignore the possibility of errors in data. Indian ophthalmological experts reflected on vision recovery cases and discussed the benefits of cataract surgery after missing the critical period.

The study of the inspiration method and prejudice gave us a deep understanding of the limitations of inspiration and intuition. Scientists also need to work hard to make people realize the limitations of data, such as conducting research to show people that if they trust the data too much, they will have serious consequences. This study also analyzes the potential causes of basic prejudice and the way to prevent prejudice. Some cognitive scientists are already studying the difficulty of processing fuzzy data.

However, I think scientists need to do more, that is, larger cooperation research. Such research may have a profound impact outside the scientific community. We are living in the era of big data, quantitative investors are taking over Wall Street, and all decision -making formulation is based on data. In a world with more and more data -centered, learning how to deal with imperfect data is of great significance.

What do you think is the most important scientific discovery at present? In this fascinating book, Edge founder John Brockman joined hands with philosophers and psychologists Stephen Pin Keke, celestial physicist Lee Smolin; cosmologist Paul Davis; sociology Family Nicholas Kristekis; educatist Howard Gardner; future scholar Kevin Kelly; gene editor pioneer George Church, and 186 thinkers in many fields such as artists, inventors, entrepreneurs , Bring you a big thinking to expand scientific thinking.

186 thinkers from different fields have proposed 186 new concepts of science, involving philosophy, psychology, physics, astronomy, biology, neuroscience, linguistics, etc. Essence

- END -