We're continuing our series on practices that promote wise thinking. (See
introduction and
Part 1.)
Practice #2: Distinguish Correlation and Causation
A sleight of hand used to deceive people is to present correlation of two facts as causation.
Let's look at some definitions:
Correlation: A causal, complementary, parallel, or reciprocal relationship, especially a structural, functional, or qualitative correspondence between two comparable entities: a correlation between drug abuse and crime.
Let's examine a clear example first. Here are two facts, indisputable:
a) In the last three months, I have more gray hairs.
b) Gasoline prices have increased $0.62 in the last three months.
It would be ridiculous to conclude that increasing gas prices lead to more gray hairs, or more gray hairs on my pitiful head cause the price of gas to rise. Why?
Because acts (a) and (b) are correlations, but there is no causation here (in either direction). There is no reason to connect the natural process of my graying hair with the swings in gas prices.
This one is easy to see, right? Can you think of other equally ridiculous correlations?
Let's look at another example.
Two more facts:
a) In the past three months I have been doing pushups and situps almost daily.
b) I can do more pushups and situps now than I could four months ago.
Again, simple right. (b) is a result of (a) -- (a) caused (b); this is not just a correlation, there is a strong relationship in one direction between the two facts.
Review the possible connections below. Are these correlations or causations? What would you conclude?
We spend more on educating children in Washington, D.C. than anywhere else in the US. Children in Washington, D.C. have lower standardized test scores than most other states.
Atmospheric CO2 levels are higher now than in the recent past. Average global temperature levels are higher now than in the recent past.
There is three times more safety equipment installed in cars now than in cars 20 years ago. Deaths in car accidents have declined slightly.
In all three cases, you need more information in order to discern whether there is correlation or causation between these things.
You will often be presented with facts as examples of causation. Men, be instantly suspicious! Causation is less frequent and more difficult to prove than most people initially expect. Correlations are rampant, causation infrequent.
Why are causations difficult to discern? Because the real world is full of complex, interlocking systems with multiple feedback loops of differing strengths. (I hope to cover some of this in a future post.)
It's a huge step forward if you can recognize when you're being asked to believe a causation (a causes b) without sufficient evidence. Help your children learn this as well, through simple examples.
Think wisely!
Please comment, and let's sharpen one another.