No, ‘causation’ doesn’t imply correlation either

It’s proverbially true that an observed correlation between two variables does not imply there is some causal relation between the two…but what about the other way round? What if there is a known relation between two variables (for example from physics, laboratory experiments etc) does this imply there will be a correlation between the variables?

The answer is no, sort of. The point about correlation is that it is an observation of apparent relations between data sets. Two phenomenon can have an actual connection but the outcome might not be observable. Why not? Well, just because something exists doesn’t mean that your data collection methods can observe it: they might not be sensitive enough (e.g. your telescope might not be big enough) and closely related to that lots of other things will also have causal connections.

For example, the physics connecting CO2 and atmospheric temperature is very old, much older than the consensus among climate scientists that CO2 emissions are causing global warming. Why? Well in part because for much of the Twentieth-Century the connection was not easily observable — greenhouse gases are not the only drivers of surface temperature. Postwar data collection improved and the impact of rising CO2 (and other greenhouse gases) also became more prominent in surface temperatures and the impact of other factors (such as particles causing cooling) became better understood.

Put another way causation implies correlation but only if all other factors can be controlled (or accounted for) and your data collection is good enough.


11 thoughts on “No, ‘causation’ doesn’t imply correlation either

  1. Offtopic Puppy news: Jon Bromfield, noted for the thread on Scalzi’s blog in which Scalzi and MZW took turns dunking on him (and led to Bromfield’s banning back in 2007 popped up on File770 to let us all know that Scalzi’s still living rent-free in Bromfield’s head.

    Liked by 1 person

    1. Your post got got by the spam filter, apparently twice because it went in there after I clicked ‘approve’.

      Anyway – thanks for the link back to that 2007 Whatever post. These things fascinate me.


      1. I saw that meaningless drive-by comment. It always mystifies me why people try to convince us how much they don’t care by endlessly repeating how much they don’t care.

        Liked by 2 people

    2. That’s interesting. I thought he was just a random minor drive-by puppy. Didn’t know he had a “history” of sorts (if getting banned from Whatever 12 years ago counts as history) with Scalzi.

      On the other hand, there is an occasionally genre critic who still has me blocked on Twitter because of an argument we had on a long defunct site back in 2006, so some people do have long memories.

      Liked by 1 person

  2. Lo this wintry Friday night, I am enjoying thinking about how various constituencies across the spectrum might talk about any causation/correlation between Reason and Hell (ie: post’s tag)

    Liked by 1 person

    1. I checked the Lady Biz spreadsheet and it’s there. Good.

      No listing of the Doctor Who episodes, though! For shame! I’m not going to stand for more of that “Good Place” crap.

      Liked by 1 person

Comments are closed.