I’ve seen this referenced a lot, especially with psychology, but can someone explain what exactly it means? How does correlation not imply causation? Sometimes, does correlation ever imply causation?
In: 160
Just because two things happen at the same time does not *necessarily* mean that one *caused* the other.
Often with such fallacies, it can *look* like those two things are linked, since they are contextually related. However, with a bit of critical thinking, you can see that they happened completely independent of each other and one did not *cause* the other.
two things can happen in some manner that brings them “close”, like happen closely in time, but that doesn’t mean that one causes the other. as an example, the number of films Nicolas Cage has participated in correlates with the number of people drowning in swimming pools… does this mean that one of those causes the other? also cheese consumption correlates with the number of people who died tangled in their bedsheets… again, it’s not very likely that one of those causes the other.
you can see a whole bunch of such things at https://tylervigen.com/spurious-correlations.
John and Mike are walking down the street. Mike falls into a hole. This happens every day for a year. The correlation is that John was with Mike every time Mike falls in a hole.
Saying “Mike shouldn’t walk with John anymore, every time he does he falls in a hole” requires more evidence than is provided. The cause of Mike falling in the hole is not demonstrated. It could be that John pushes Mike, or it could be that Mike is blind and can’t see the warning sign in front of the hole.
Correlation – B happened after A
Causation – B happened because of A
“Correlation does not imply causation” tells you that one thing happening before another doesn’t mean the first thing caused the second thing.
https://www.psychologytoday.com/us/blog/how-we-do-it/201307/the-stork-and-baby-trap
“Pioneering statistician George Udny Yule, author of the seminal 1911 textbook Introduction to the Theory of Statistics, explained confounding factors with a pleasing reference to reproduction. He noted that in Alsatian villages numbers of human newborns are correlated with numbers of storks nesting locally. It is tempting to conclude that storks do actually deliver babies, but the real explanation is far more mundane. Larger villages have more houses with chimneys for storks to build nests, and more babies are of course delivered in larger villages.”
A simple explanation:
A shaman gets up every day and chants to the horizon. As he chants, the sun come sup. He says the sun rose because of his chants.
Correlation: Every time the guy chants, the sun comes up. It’s been doing so every day for years.
Causation: There is none. He did not cause the sun to rise. He’s full of shit.
Important to remember that the phrase should really be “correlation does not *necessarily* imply causation.” Correlation *can* be evidence of a causative link. This happens when the correlation is seen consistently, potential confounding factors are controlled for, there’s a mechanistic reason to believe the link is causal, and there’s other experimental evidence that implies a link.
Smoking causing cancer is a good example. All we really have for this in humans in correlational data, because it’s unethical to perform direct experiments on humans. But the correlational data is very high quality, there’s a biological reason to think the causal link exists, and animal experiments show a direct relationship.
https://www.tylervigen.com/spurious-correlations
The spurious correlation website has some very funny examples of things that are correlated but clearly can’t be causative.
Something people like to throw out there to dismiss something that doesn’t align with their beliefs
It’s also known as ‘post hoc, ergo propter hoc’: ‘from this, therefore because of this’.
*Post hoc* arguments are fallacy, because such is almost never the case: that two events occur simultaneously does not necessarily mean that one of those events causes the other.
For example, the Church of the Flying Spaghetti Monster facetiously claims (as a mockery of organized religion) that global warming is caused by the decline of piracy following the end of the age of sail. They make the connection by charting the decline of piracy and corelating it to an increase in the average global temperature as piracy declines.
In reality, the timing of the two events is entirely coincidental; the decline in piracy has nothing to do with any change in the Earth’s average temperature.
That a correlation between the two events *exists* does not imply causation.
Latest Answers