Eli5: In science, can you establish causation without knowing the causal mechanism?

896 views

Pretty much the title

In: 210

31 Answers

Anonymous 0 Comments

Causation can be established by running an experiment with a test group and a control group, where the *only* difference is the presumed cause.

Suppose I grow two sets of plants in the same low-quality soil, and I treat both groups the same, but I add ammonium sulfate to one set. I call them the test group and the control group. To be certain that both groups are treated the same, I make sure that the person tending the plants and taking the growth measurements doesn’t know which plants had the ammonium sulfate added, so they don’t unconsciously treat the two groups differently.

I notice that the test group (the one with the ammonium sulfate) grows better. I can be pretty confident that something in the ammonium sulfate *caused* the improvement in growth, since that was the only difference.

I can perform additional studies to narrow down what specifically, and in this case I’d find many other nitrogen-rich compounds that help my plants grow. I might still not have the faintest idea *why* nitrogen makes plants grow better, but I can be pretty sure that’s the underlying cause.