Correlation & Causation Pt. 1: Exploitation by Media
This is the first in a two-part series on correlation and causation, and how conflating the two can dramatically distort our perceptions and thinking.
The first part will be about the exploitation by media, and the second part will be about the related philosophical and theological “bugs” that we must work to root-out.
Let’s take a look at event patterns in the world.
When we notice, through some study, that pattern A is correlated with pattern B, what does it tell us?
- Does it mean that A makes B more likely?
- Does it mean that B makes A more likely?
- Does it mean that a third pattern, C, makes both A and B more likely?
- Or, was the coincidence of A and B simply random noise?
The sad fact is that we don’t know which of these causal possibilities (and a non-causal possibility) is actually true. This is why we say, “Correlation does not equal causation.”
It turns out, though, that the brain is very eager — dopamine stimulated — to jump to causal conclusions. Furthermore, we’re most excited at the possibility — of the above four — that is most startling and weird.
It “feels” like new revelation.
And Commercial Media Knows This
Let’s say there’s a study that correlates non-married cohabitation with a higher incidence of physical abuse. What is the most sensible, boring explanatory possibility?
The most sensible, boring explanatory possibility is that there is a third factor C — likely something to do with socio-economic status, and population density, and the cultural byproducts therefrom — that makes non-married cohabitation and physical abuse rise in tandem, without being causally related to one another with any statistical significance.
But that’s “boring.” It doesn’t sell tabloid newspapers and doesn’t serve as social clickbait.
The more “exciting” possibility is that being non-married causes abuse.
This would have us conclude, “To lower partner abuse, those partners should get married.”
Which is, of course, precisely the opposite strategy one should employ.
The problem is, again, that we love counterintuitive revelations. There’s a measured “second opinion bias” that has us feel excited about having the “privilege” of being an honest devil’s advocate. As soon as we’re tricked into thinking that some bizarre claim is merely misunderstood, or deceived into thinking it has statistical backing, its “bizarreness” becomes extra fuel to fight on its side with conviction.
Whenever you read, or hear from a friend, about a study showing some A-to-B causation, do the following in your brain:
- Un-cause the causation. Separate the two parts into “A” and “B.”
- Run through the four explanatory possibilities mentioned before.
- Evaluate which possibility makes the most sense. Extra points if it’s also boring, which is a memetic weakness.
But Remember: Common Sense Isn’t Infallible
At the same time, remember that certain systems in the world really are very complex, and can yield causal relationships that are both counterintuitive and true. Common sense gets you past the exploitive headlines, but it’s no replacement for an actual, knee-deep understanding of complicated systems.
In other words, common sense isn’t “common,” but it also isn’t always “sense.”
And Remember: Studies Can Be Awesome or Crap
When I did product management for social games, one of my jobs was user experimentation and data analysis in order to make design decisions that optimized the interests of the customer and company.
The reality is this: Data is absolutely vital for getting a correlative, and ultimately causal, understanding of how the world works.
At the same time: There’s no shortage of ways to screw it up, and even maliciously fudge the data, and perhaps even get away with it. We’ve seen this happen with the “vaccines cause autism” fiasco, where an atrocious study catalyzed a tragic memetic bloom that, today, continues to threaten the health of our children.
I think, if we all get a healthy scrutiny against urban myths about causal claims, and fight hard for the “boring,” we’ll go a long way toward killing memetically strong falsehoods, which is necessary to optimize peace, wisdom, virtue, and charity in the world.
For more about memetics and how to avoid the value pitfalls therein, rewind to the four-part series we ran earlier this year.