People say this so much that is now used as a way to dismiss any study. It’s almost becoming a meme.
I think is clear that the baseline of any study is that correlation does not imply causation. You start there. That’s a basic intelligent start. You never take for granted that one thing causes the other. The ideas is to try to identify if there are any potential paths for causation.
Otherwise there’s not even a point in sharing or discussing these studies.
I’m not saying that bringing this up is wrong, but does it need to be said so much? I think it’s more interesting to analyze beyond that and try to see if there’s causation. We already know there’s an abysmal amount of correlated things but there must be a percentage of those things which are in fact caused by another at a certain level. I would love these discussions to be more focused on this and not just doing the plain classic dismissal of “correlation doesn’t mean causation”.
>t's fair to dismiss it with the "correlation does not imply causation" argument.
In almost every thread for an article of this sort, someone pops in to say "correlation isn't causation", as if people doing research don't understand this concept.
The whole "correlation does not imply causation" thing is completely misunderstood.
The issue is with the meaning of the word "imply"; when used in the formal sense as it appears is Classical Logic, correlation does indeed not imply causation.
In common parlance however, "imply" is often used to mean "provides evidence for", and correlation can indeed provide (potentially strong) evidence for a hypothesised causal link; the problem lies in people reading "correlation does not imply causation", assuming the informal meaning of "imply", and then going on to reject any notion of causation which uses observed correlation as evidence.
Pretty much every empirical science uses notions of correlation (in its various formal statistical guises) to provide support for causation, indeed to reject such reasoning would be to invalidate huge swathes of mainstream accepted science; half the battle in these instances is making the leap from correlation to causation in a manner which is considered scientifically sound.
> There's no way a single study can prove causality
Why? If you're talking about epidemiological studies looking for associations between variables in big datasets, okay. But that's just one type of study, and it doesn't get better by throwing many of them together in a meta study.
You absolutely can show causality in a single study by doing experiments.
I'm feeling rather the opposite really, "correlation is not causation" is a beaten horse and will come up in every discussion where it may or may not fit on this website.
When you have a correlation plus a plausible mechanism for causation, it's not appropriate to invoke "correlation is not causation". It's literally true, but distracting. That's where you do additional studies.
This is a vacuous statement in this context and adds nothing to the discussion, because you can never prove causation via a study/experiment, only reject the null hypothesis.
So, what's your point? We have all had this ingrained in us since our first statistics class. Saying that correlation does not equal causation really says nothing. It does not agree. It does not disagree. The only thing this does is easily dismiss something of importance, which by definition has a connection to the discussion.
Correlation != causation is laziness at it finest.
Mainly because people so commonly dismiss a study by stating that 'correlation doesn't imply causation,' while going about their lives with their own ideas and biases based on anecdotal evidence just the same. It's the tool of a common course of visible hypocrisy. Maybe it's just a personal thing.
This certainly is a true and tested adage, but I've got a feeling that the correlation-backlash has gone too far in the other direction. Don't just repeat it without getting it. If substantiated through proposal of a suitable mechanism, correlation is solid corroboration of a causality-claim.
More precisely, correlation most definitely never implies absence of causation.
I don't get this attitude, if things are correlated, it should at least make a scientist wonder why. It certainly could be random chance, but correlation can also lead to establishing a causal model or discovering a third variable. If two things keep happening in conjunction, it at least merits further investigation.
It seems like there's this extreme reaction against people behaving like correlation equals causation, but instead of over-emphasizing correlation, it gets dismissed entirely.
Agreed. The "correlation != causation" mantra does not match the context here.
It's true that correlation studies are not as conclusive as laboratory experiments in general, but when we're talking about negative long-term health effects on humans, you can't establish causation without doing something unethical. But does that mean science becomes useless? I don't think so.
There are times when it's appropriate to infer causation from a correlation and act on that conclusion. For example, the famous lead toxicity studies are correlation studies. One such study, cited over 1000 times, [0] meets only 6 out of 9 criteria for inferring causation [1], and people are comfortable making that inference.
A good next step would be to go through the PowerWatch list of studies [2] and evaluate these studies based on these criteria (or a similar list).
I think this is taking "correlation doesn't always prove causation" a few steps too far, right off a cliff. There are difficulties with reasoning from correlation, but saying it "usually doesn't" indicate correlation is too bold, especially in experiments where confounders have been controlled for. [1]
[1] (And don't get me started with the semi-clever types who ask if this or that variable has been controlled for when the study or the article about the study explicitly says it has. Save us all from halfwits trying to teach experienced researchers their business.)
The host of scientists who conducted their studies know this. Which is why they did many such studies (some of them very large scale), controlled for confounding variables, modified the parameters across time in both direction (watching more and watching less this and that kind of show), and of course reproduced most of the results.
Go tell them "correlation does not prove causation", and they will laugh in your face. We're long past correlation by now. Causation itself has been thoroughly demonstrated.
Example: People are citing that birth rates decline in wealthier countries, yet fail to realize that wealth in those countries is a direct function of time. People exchange their time, for wealth.
It is not enough to simply have wealth. You also need time.
Most of these "studies" people keep linking, are looking at a single variable's correlation and not controlling for anything else. Amazing to me the "science" people will accept.
The editorialized title of the submission may have its problems, but "Correlation != Causation" and equivalent statements need to be consistently downvoted because they, as on every other site where they have for a while now been equally if not even more overused, have become the lazy man's catch-all content-free retort.
Much of what we take as certain once started with observed correlation. This trend of immediately and automatically discounting observed correlations will likely do more harm than good over time, as one would be hard-pressed to hypothesize and later prove (or at least fail to disprove, in the scientific sense) causation if one's habit were to always blurt "correlation != causation" in the face of new information.
I think is clear that the baseline of any study is that correlation does not imply causation. You start there. That’s a basic intelligent start. You never take for granted that one thing causes the other. The ideas is to try to identify if there are any potential paths for causation.
Otherwise there’s not even a point in sharing or discussing these studies.
I’m not saying that bringing this up is wrong, but does it need to be said so much? I think it’s more interesting to analyze beyond that and try to see if there’s causation. We already know there’s an abysmal amount of correlated things but there must be a percentage of those things which are in fact caused by another at a certain level. I would love these discussions to be more focused on this and not just doing the plain classic dismissal of “correlation doesn’t mean causation”.
reply