Correlation-Causation Fallacy

Author: Ali Samy Rezk

From: Egypt

In November 2016, a wave of fires began in Israel from south of the Beersheva area to Nahariya. Some of the fires occurred naturally; others were arson attacks. The largest fire occurred in Haifa, where 527 apartments were destroyed among 77 buildings, leaving 1,600 people homeless. 75,000 residents, about a quarter of the city's population, were evacuated from 11 neighborhoods. (see Fig1)


The matter was spread tremendously on social media, and some in the Arab world asked, why might this happen? Is there a specific message from this fire? All these questions were on the minds of some, especially since a few weeks before this fire, the call to prayer had been prohibited. Therefore, the minds of some were wondering, could this be divine anger or what? Finally, most people have concluded

that there is a causal relationship between preventing ears and the occurrence of these fires. Are these ideas correct or not? The truth is, I do not know. But what I know very well is that there is a very famous gruesome mistake that happened in economics.


In Russian Folklore, there was a disease killing the people of Russia, and the King of Russia wanted to eliminate this disease. The king gathered a group of state scholars and asked them what the cause of this disease could be. He could have reached the puzzle! The king did not believe this, but after he looked at the maps again, he noticed it! The place where there are many doctors, there are also many patients. So, do doctors are the cause of illness? - The king concluded that and ordered the killing of all the doctors.


I am happy to inform you that if you are convinced of this logic, then this is a problem or almost a disaster! It is normal that there are many patients in a place where there are many doctors. You are mixing correlation with causation!!


Correlation means that two things are happening together. That is, when one thing happened, the other happened. This is not necessary to be a causal relation between them. Saying that a specific thing is the cause of another is a very difficult phenomenon. This is not a trivial matter, many calamities and disasters occurred because of confusion between those things!!


For instance, before the nineties in New York, the percentage of crimes was extremely high. The percentage of crimes had risen to 80%. This means that there is a disaster happening! At that time, Analysts published some research and papers saying that the crime rate would increase more and more. And suddenly something weird happened! At the beginning of the 1990s, the crime rate decreased abnormally until it reached the same rate it was 40 years ago!! When the press and the media heard about that, they began to analyze these things. Most of the analysts believed that the reason for the decrease in the crime rate was that the increased number of police officers and soldiers. They again fell into the same mistake as the king!! Ok, it is normal that because the crimes were increasing, the officers were also increasing. But does that mean that the police were the reason for the decreasing crimes?! They then discovered that the reason has no relation with the police! Of course, it had some way of relation, but it was not the "reason"!


I know you will be surprised but the most important reason was the abortion! The effect of legalized abortion on crime (also the Donohue–Levitt hypothesis) is a controversial hypothesis about the reduction in crime in the decades following the legalization of abortion. Proponents argue that the availability of abortion resulted in fewer births of children at the highest risk of committing the crime. The earliest research suggesting such an effect was a 1966 study in Sweden. In 2001, Steven Levitt of the University of Chicago and John Donohue of Yale University argued, citing their research and earlier studies, that children who are unwanted or whose parents cannot support them are likelier to become criminals. This idea was further popularized by its inclusion in the book Freakonomics, which Levitt co-wrote.


What I want to reach out to is the idea of causation is a very difficult one. Emphasizing that there is a certain thing is the reason for another thing is an idea that needs to be studied, analyzed, and a black hole! Most of the things we say that have a causal relationship are dependent upon our observations. Our observations are usually not correct! Thus, when scientists need to prove that relation, they use the “control experiment”. It is a scientific test that is directly manipulated by a scientist, in order to test a single variable at a time.


This experiment simply means that if you say that Mohammed Salah is the reason for Liverpool’s win upon Man United, you need to replay the match again with the same of everything and just remove Mo Salah from Liverpool. Then you have to analyze the two matches' results and if Liverpool wins again, Mo salah does not really affect. But if Man united wins, now you can conclude that, someway, Mo salah was a huge reason for the win. This is a relative paradox because of course, no one can return the match with all its entirety and details!


Unfortunately, there are many things that sometimes happen in our daily life, but we have no reason for them. We can’t replay it again to make sure that a specific reason is a true reason! Therefore, most of us resort to find a correlation: two things happened together and then say that this is the reason. But as I said before, these observational things can have nothing with the real reasons!


One statistician published a very nice article. He came up with things that had to do with each other. For example, there is a relationship between US spending on science, space, and technology with suicide (see fig2). There is a relationship between the number of people drowning in the pool and the number of films in which the actor Nicolas Cage appears. There is a relationship between the consumption of cheese and the number of people who died by becoming tangled in their bedsheets. There is also a relationship between mozzarella cheese and the number of doctorate degrees in civil engineering. I have not finished yet!


There is also a relationship between the number of people drowning and the divorce rates in Kentucky. If you want to read more, you will find that there is some relationship between the import of oil from Norway and the number of people who die in train accidents. Personally, I opened the site and discovered my own studies, and found that there is some relationship between milk consumption and divorce rates in Alabama. These accidents are almost entirely related to each other. There is a strange relationship between eating ice cream and the number of people drowning. Once you imagine it with your observational mind, you will begin thinking about going and forbidding the ice cream to save the people from drowning!!!! The matter is just that the two things are happening in Summer whether eating ice cream or drowning and this is why their rates are increasing together. This is just a correlation.

In a nutshell, when two things are happening together or that something happens before the other, this does not mean that there is necessary to be causation between them. Before you say that there is causation, you have whether to make your experiment and control the other variables or to find so many same cases but different in the variable you are studying. Of course, all of these things will not make you 100% right that there is causation, you just try to increase the chances of being.

 

Author: Ali Samy Rezk


Ali Samy Rezk is 17 years old. He is currently a G12 student at Dakahlia STEM School - Egypt. He is interested in engineering, writing, football, history, politics and pop-science. In the future, he hopes to be an engineer besides being a pop-science presenter.

 

Resources:

629 views1 comment