How I Know What Is True

Causes Versus Things That Happen Together

Also called: Causation versus correlation

Established idea
Formal theory
Working interpretation

Noticing that two things tend to happen together is different from knowing what would happen if I changed one of them. That difference matters in factories, in psychology and in how I make sense of my own experience.

What could I change on purpose to tell a real cause from two things that merely go together?

Why it attracts me

I am wary of explanations that rest only on two things happening together. They feel convincing precisely because the pattern is real. What is not yet known is the cause.

The idea

A correlation says two things move together. A cause says that changing one would change the other. A hidden shared cause can produce the first without the second. The cleanest test is an intervention: change one factor on purpose, keep track of everything else, and see what follows.

An example

The Factorial Experiment Lab is built around this. A heat-sealing machine is making weak seals, and three settings are suspects: temperature, how long the jaws stay closed, and pressure. Instead of combing through old production records, you deliberately run all eight combinations of low and high settings, in a random order. Running them in a random order guards against slow changes during the day being mistaken for the effect of a setting. Because you chose the settings, the differences you see can be traced back to them.

Where it connects

This clarifies Rehearsing the Future in My Head: practicing something in imagination may influence behavior, but that cause has to be shown, not assumed. It explains why Digital Copies for Testing What-Ifs and Simulating a System Event by Event are valued for "what if" questions: they let you intervene in a model when you cannot intervene in the world. And it keeps Adjusting My Confidence as Evidence Arrives honest, because evidence that two things go together should move my confidence about a cause only so far.

Questions I'm still exploring

Which of my beliefs about myself rest on things that just happened together?
How can I test causes in my own life when I cannot run a clean experiment?
When is a strong pattern good enough to act on even if the cause is unclear?

Sources and further reading

Christopher Hitchcock, "Causal Models", Stanford Encyclopedia of Philosophy
Judea Pearl and Dana Mackenzie, The Book of Why (Basic Books, 2018)

Working interpretation: drafted from my notes and interests for review. It is not a direct quotation, and I may still change it.

In short

Two things can rise and fall together without one causing the other. Often a third factor drives both, the way hot weather raises both ice cream sales and swimming accidents. Correlation means things move together. Causation means changing one would change the other. The surest way to tell them apart is to intervene: change one thing on purpose, ideally choosing at random which cases get the change, and watch what follows.

Why this matters to me

In production work, acting on a false cause wastes time and can make things worse. The same mistake is easy to make with my own habits and moods: I notice two things together and build a story about why. I want to keep asking what would actually happen if I changed something, and to treat my personal explanations as guesses until they are tested.

Connections

Rehearsing the Future in My HeadClarifies

Practicing an action in imagination may influence behavior, but that cause has to be shown by testing, not assumed because the two seem linked.

Reasoning From the Ground UpPart of

Getting down to what is really true means separating real causes from things that merely happen together.