Correlation is not causation!
“Just because I have strong evidence does not mean I am wrong!”
Lately I have found myself saying this with disturbing frequency in debates involving correlation. It sounds absurd, but it reflects the fact that many people dismiss evidence out of hand with the catch all phrase “correlation is not causation.” The statement is true and valid, but they use it as an excuse to dismiss evidence and to reject the most obvious hypothesis that explains it.
Here is a common example. It’s a bundle of correlations between economic freedom and really nice social outcomes. Economic freedom (read: capitalism) is correlated with wealth, literacy, creativity, happiness, competitiveness, tolerance and human development and it is anti-correlated with corruption, crime and inequality. That’s good, right? Wrong! Correlation is not causation! No need to look more at this! No need to challenge the belief in socialism and the welfare state. Moving on, nothing to see here. Next question, please?
Anyone who has tried presenting these correlations to a left-leaning crowd has experienced these types of responses. At first, I did what all humble, truth seeking, rational people with integrity do: I looked to myself and took their criticism seriously. I tried to be as self-critical as possible to make sure that the evidence I presented was strong and the conclusion well-founded. Then I returned to that crowd eager to show them how strong my position was, only to hear the same mantra all over again: “correlation is not causation.”
After a while it dawned on me that no matter how much evidence I presented to them, they were not willing to engage in it. They were not really engaged in a scientific debate, but in an internal rationalization scheme. They were not looking for the truth, but for ways to dismiss conclusions they didn’t like. No matter how much evidence I presented, they would never change their minds.
Realizing this, I changed my focus. That’s when I started saying “just because I have strong evidence doesn’t mean I am wrong.” I realized that I could not let them get away with dismissing empirical evidence. I can talk about correlation and statistics, but I think it is equally important to talk about the human story underlying the evidence.
Take the strong correlation between economic freedom and wealth, for instance. Don’t think of it as cold, inanimate data. Think about what human tragedies had to occur to create that correlation. Each datapoint often represents millions of human deaths and sufferings. Millions of people living in poverty. Millions of malnourished children whose teeth fall out because they cannot afford to go to a dentist.
Think about that for a moment. When a leftist dismisses evidence because “correlation is not causation,” he is dismissing people. He is dismissing human lives and saying their misery doesn’t matter. They are irrelevant to him. When he is demanding more evidence in order to even bother to look at it, he is calling for more blood, more deaths, more suffering.
For someone who claims to be world champions of empathy, that’s pretty harsh. Now, I don’t think that they actually go through a mental process of dismissing human lives. I think that they just don’t understand that the data represents real people and real successes and tragedies. They really do see themselves as morally superior and the most compassionate of all. So in order to get them to connect with the evidence and take it seriously, all that is needed is to get them to connect with the human stories underlying the data. They need to understand that lives are at stake and that if they truly care about alleviating suffering and preventing poverty and death, they should take the data seriously. “Correlation is not causation” is simply not a good enough excuse, because it is a matter of life and death.
In the area of climate change, leftists have embraced a mode of thinking that follows this line of reasoning. Climate change, they say, is about life and death and therefore we should act, even if we lack sufficient evidence. This is known as the precautionary principle. That’s an even stronger position than the one I present here. I am not calling for actions. I reject the precautionary principle. No-one should act on lacking evidence. I am only saying that once you have proved a significant correlation, you actually have real empirical evidence and that evidence demands an honest inquiry.
So while a correlation in and of itself is not enough to conclude with causation, once it has reached the level of significance any honest truth seeking person needs to take it seriously. They need to provide arguments for why the obvious interpretation of it is not correct, and if they cannot dismiss the obvious interpretation, then they should either accept that theory, or they should start actively contribute to funding and promoting science projects that can settle the question. That’s what honest, decent, empathic people who care about lives do.
So the next time you meet someone who dismisses the highly significant correlation between economic freedom and other social parameters with the phrase “correlation is not causation,” tell them that they are pissing on the graves of the victims of socialism. Lives matter, and in order for the deaths of these victims not to be meaningless, we have an obligation to take the data — and the victims they represent — seriously.
philosophical, technological, political and economical ponderings