@jaj@mastodon.social @riverpunk@defcon.social Okay so. What we're looking at in this image is what I'd like to call "coincidence". Two things that look like they're connected, but are actually unrelated. A lot of people think that "correlation is not causation" means "correlation can be coincidence". That we have to account for the possibility that the data just happened to line up by chance. And to be clear, yes, we do have to account for the possibility that the data happened to line up a bit and it means nothing. But that's not what "correlation is not causation" means! When the correlation is strong enough, you can rule out coincidence. With proper statistical analysis, you can say "this correlation cannot have happened by chance". And even then, correlation is not causation. If you have "A correlates to B (statistically meaningful)", it can't be a coincidence, but it doesn't imply "A causes B". It could be "B causes A", or "A and B have a shared cause", or a number of other things. I've seen a lot of people say "I know correlation is not causation, but this correlation is very strong, it's got to be causation". That's the kind of misunderstanding you get when you misunderstand "correlation is not causation".