Correlation CAN Mean Causation

OK.  For whatever reason I've heard this phrase a lot: "Correlation does not mean causation."

And as a psychology major who had to learn about research methods in EVERY SINGLE CLASS, I must speak up about this. 

First of all, yes, correlation does not automatically mean that one item correlated with another caused the other.  But it doesn't automatically mean there is no causation at hand.  It means we have to look critically at the data and use other studies and repeat studies to really find out what is going on here.  ONE study alone is never enough to prove anything. 

You can't always set up a study to prove causation because there are often ethical guidelines that get in the way.  Or you might be looking at historical data and cannot set it up to prove causation. 

For instance people who are taller tend to also be heavier.  And people who are heavier tend to also be taller.  But gaining weight doesn't make you taller.  But getting taller tends to make you gain weight.  But then again we all know someone who is much taller than us, but who weighs less than us, and sometimes we are surprised that someone so short can weigh so much.  Height and weight are correlated, but there's also some causation going on and some other factors too. 

This is how it is for almost everything.  Nothing is a one-to-one kind of a thing.  We need to be critical of correlational studies.

BUT it does not mean that we ignore correlations simply because they can't prove causation all by themselves.  They are important. And if they show a correlation between two things, we need to investigate further, especially when they don't make sense.  But simply saying "correlation doesn't prove causation" seems just dismissive to me.   It could be a fluke, there could be a third variable we need to look at, and yes, it could be that one is causing the other.

It's important to not jump to conclusions over ONE study.  It's also important to not ignore the overwhelming evidence for something shown in many studies.  Both are bad ideas.

Magazines are probably the worst at jumping to conclusions about one study.  Some newspaper articles are also bad.  Some study comes out and journalists jump at it declaring something silly.  It's important to go to the actual study.  I read so many research papers as an undergrad.  A good research paper will talk about the limitations of the study and about the possibilities of other things going on contradictory to what the data shows.  Nothing is ever cut and dry. 

So don't read too much into one study and also don't automatically dismiss a study either.  The more studies, the better.  There are always going to be studies that end up showing some kind of a fluke that goes against the norm, this is why they need to be repeated and why we all need to keep an open and a critical mind.