Science question: does correlation without causation ever mean anything?

Put another way: are things that correlate, if there is no causation, anything more than curious?

Here’s the pic that has me thinking about this, it’s the (now widely distributed) NC country gay marriage amendment passage/refutation:

which I pulled from the Atlantic.

I think people are all over this graphic because there is a clear correlation between where the NC universities are and where the amendment did not pass.

So?

This is purely curiosity unless you are trying to make the claim that education (or something else, such as per capita income, unemployment levels, per capita # of books, etc., all of which might be very different in the counties which have universities than those which do not) has something to do with willingness to pass gay marriage laws.

I think it foolhardy and insulting to imply that there is any causation and I think those that show data like this are cowardly if they don’t also claim causation. Correlation without causation (as is VERY OFTEN the case with correlations) is meaningless. Showing graphs of correlations without explaining the causation (if there is any) explicitly can serve no purpose other than to incite and inflame, and the cowardice comes from the incorrect belief that all one is doing is presenting “the data”. No, one is not presenting “the data” because in these cases the data is meaningless.

Am I wrong? Please tell me.

-JD Cross

 

Advertisements