My N=1 vinegar study pretty definitively answers one question: “Will my body odour reduce at this specific date, time, and location after I scrub concentrated vinegar under my armpits with a paper towel?” However, it doesn’t tell me whether it’s causal
I don’t understand why you say that it doesn’t tell you whether it’s causal, when you previously concluded that it was causal:
Even at this point, I was extremely confident that, even if I got the mechanism wrong, this remedy worked. Why the confidence off of so little evidence? Well, what are the chances that the smell just happened to go away on its own, right at that exact moment? Basically zero.
Sorry, I wrote this out in an hour, so it could have been a bit clearer. The data alone does not imply causality.
Data that supports hypothesis & causal hypothesis & no competing hypothesis ⇒ causality.
The causal hypothesis doesn’t need to be a detailed mechanistic hypothesis (it might just be “vinegar will remove the smell”). As long as nothing else could have caused it, then you know what the cause is, even if you’re unsure of the underlying mechanics.
So for an example where I have the same(ish) data but wouldn’t be highly confident of causality (because I have more than one hypothesis), let’s say I have a light headache, and someone gives me an unspecified pill for it. I take it, and five minutes later, my headache is gone. That is some evidence that the pill worked, but I’m not highly confident because of competing hypotheses: For example, the placebo effect, or maybe the headache would’ve gone away on its own.
Let me know if this makes sense, and I’ll update the post.
I agree that you don’t need a mechanism, and I agree that if there is nothing else that could have caused it, then that permits you to infer causation. More specifically, I think it’s the determinism here that allows you to infer it. (Which you also say in the post.)
I don’t understand why you say that it doesn’t tell you whether it’s causal, when you previously concluded that it was causal:
Sorry, I wrote this out in an hour, so it could have been a bit clearer. The data alone does not imply causality.
Data that supports hypothesis & causal hypothesis & no competing hypothesis ⇒ causality.
The causal hypothesis doesn’t need to be a detailed mechanistic hypothesis (it might just be “vinegar will remove the smell”). As long as nothing else could have caused it, then you know what the cause is, even if you’re unsure of the underlying mechanics.
So for an example where I have the same(ish) data but wouldn’t be highly confident of causality (because I have more than one hypothesis), let’s say I have a light headache, and someone gives me an unspecified pill for it. I take it, and five minutes later, my headache is gone. That is some evidence that the pill worked, but I’m not highly confident because of competing hypotheses: For example, the placebo effect, or maybe the headache would’ve gone away on its own.
Let me know if this makes sense, and I’ll update the post.
I agree that you don’t need a mechanism, and I agree that if there is nothing else that could have caused it, then that permits you to infer causation. More specifically, I think it’s the determinism here that allows you to infer it. (Which you also say in the post.)