Why are further post-hoc analysis required? I guess the question would be is what is correlation?
In: Mathematics
It is more likely that more people carry umbrellas when the weather forecast says that it will be rainy. Since weather forecasts are more accurate than not, a rainy weather forecasts is correlated to rain. However neither carrying umbrellas nor rainy weather forecasts cause rain to fall.
There are a lot of things that look like they are related, but actually aren’t. I’m sure someone will give a better explanation. I just want to share spurious correlations. There’s a [website](https://www.tylervigen.com/spurious-correlations) with plenty of statistics that are correlated but don’t have causal relations, like number of people who drowned while falling in a pool seems correlated to number of films Nic Cage appeared in.
Correlation is a mutual relationship or connection between two or more things. Funny examples of correlation can be found on [https://www.tylervigen.com/spurious-correlations](https://www.tylervigen.com/spurious-correlations) .
However, you may notice that, even if there is a correlation (Both graphics are pretty similar), there isn’t a causation. This means, the decrease in the divorce rate in Maine didn’t cause the decrease in the consumption of margarine, even if both graphics are pretty similar (They can be found on the link above)
When we say that ” **Correlation isn’t equal to causation** “, we mean that the fact that two things are correlated, as seen before, doesn’t mean that one of them caused the other.
Correlation just means that two data points tend to move together
Lets look at Ice Cream sales and drownings. If you were to run the numbers you’d see that these two things are very well correlated, ice cream sales and drownings are both low through winter, start climbing in late spring, peak in the summer, and decrease through the fall so does that mean that selling ice cream causes kids to drown?!?!
No
What it does mean is that there’s a common factor. In this case it would be outside temperature which is also correlated to both. As it gets hotter outside people start buying more ice cream, they also start swimming in pools more, and the more time spent in pools the more likely there is to be a drowning; but banning the sale of ice cream won’t have an impact on drownings even though they’re correlated because its not a *causal* relationship
Correlation is a simple link between two things. For example : 100% of people who drank water in their life died or will die. There is a correlation between drinking water and dying. But there is of course no causality. Drinking water doesn’t kill you. Hope this simple example helped.
Latest Answers