Correlation, Causality and What are you measuring?

Basic understanding of life, universe and the whole mess, involves understanding the difference between things that correlate and things that can demonstrate causality. This is a bit of a pet peeve of mine. In arguments, I tend to focus on the logic presented to me, more than the actual message. This can cause me to be somewhat of a bore in discussions, and I have a tendency to go over the top sometimes, but I think I have good reason. For it is precisely these fallacies of the mind that create a window of opportunity for hate mongers and others who have a need to shove their world view down our throats.

First: Say for instance that someone show that number of rapes in our society is up, and so is the number of immigrants we accept across our borders (this is a fictional case, I have no idea what the numbers really say on this, I choose this rather controversial case to prove a point).

Now, these statistics could possibly correlate (both numbers might have risen, perhaps even proportionally, and maybe during the same time period) but even if this were so, there is a huge leap to prove causality, i.e. demonstrating that rape statistics are up BECAUSE OF rising immigration numbers, a case which would allow for us to say that immigration somehow causes more rapes to be carried out.

Consider the possibility that during this same time period the price of peanuts has gone up, again proportionally to the number of rapes. This would mean that someone can say with the same confidence that we must keep prices low on peanuts to protect society from rapists. No one in their right mind would say this. That would be nuts.

Second: What’s being measured?

In this particular example, we can be quite sure that it is really the number of REPORTED RAPES that has gone up, since there is really no other way to measure this but to look at how many rapes are reported to the police. There is no way of telling if it is the inclination to report these crimes that has risen, or if the number of actual rapes has increased.

What we have here is a problem of understanding what’s been measured, what the numbers really show. This is self evident to many, but sadly this reasoning is lost on so many people that more often than not, further debate is futile.

To conclude: Demonstrating cause-and-effect relationships is very difficult since there’s an infinite number of possible (but not necessarily likely) explanations for any given phenomenon. And statistics are tricky.

And this is why everything you think you know is wrong.