This video is part of the Correlation Does Not Imply Causation, So What Does series, presented by Jim Colton, Lead Statistical Consultant at GraphPad.
Transcript:
Also on no correlation probably means no cause and effect. I'll say that with a lot of caveats though, I've got actually four. So, be car ...
This video is part of the Correlation Does Not Imply Causation, So What Does series, presented by Jim Colton, Lead Statistical Consultant at GraphPad.
Transcript:
Also on no correlation probably means no cause and effect. I'll say that with a lot of caveats though, I've got actually four. So, be careful if you see no correlation. There might be some confounding. What I mean, let me give you an example. You've drug dose on the X axis and survival in months on the Y. So, it looks like there's not really a correlation between drug dose and survival. This is the same data believe it or not, which shows you how graphs can really ... Naomi showed us a lot of examples of how graphs ... the way you display it really changed things.
So, here I've added another variable though, and I've got high severity in the reds, so people who are more severe, got the higher dose, makes sense. Maybe the doctor said, "Hey, we've got to give these patients a higher dose." And the low severity patients are in the black. So, drug dose does have a big impact on survival, it's just an un-balance in the data, or a confounding if you want to say, a third variable another way to say it, that made it look like there was no relationship.
So, no correlation doesn't necessarily mean no effect. Or maybe another way of thinking about it is you have to account for all the variables. Okay, another example where there's an exception to this no correlation means no causation, radiation exposure. At the bottom we have dental X-Ray which is 0.1 MSV's. I should know that, I don't know exactly the units there. Oh, this is a bad graph I think actually. It's not my graph, but 0.1 for chest X-Ray, 0.4 for mammogram. So, if you were going to study the effect of radiation on cancer risk and you just got a bunch of typical people and somehow measured how much exposure they had, chances are well, we know there's a correlation, but if you don't get some extreme cases you're probably not going to see it.
So, if you are only looking at people who are both low sum threshold of radiation exposure, there's probably going to be no way you're really going to see the relationship. So, you have to be a little careful about these threshold cases. Also, this is one that I added myself here, if you look at this top graph, what I have is a little experiment where I'm looking at plant growth and I've got plants that received almost no water and plants that received a lot of water, and there's really no difference. There's no correlation between water and plant growth. But in the middle you can see okay, well if we had collected data at moderate levels of watering, we'd see a decent growth on the plant.
So, there is a strong relationship, but you don't see it because you only looked at two data points and the relationship was quadratic. So, that's another way you can miss out when there's no correlation, miss a cause and effect. And the last one is a lot of times there's a lot of noise in the system, you can't see the cause and effect or correlation. An example I give her is pre-natal vitamins and birth defects, so there may be a cause and effect relationship between vitamins reducing birth defects, but there's so much variation if you look at birth defect data that it might be hard to see that unless some very specific trial was run.