“Are crime rates really higher when there’s a full moon?“ 

This vintage ’84 study said that it “could be something to do with biochemical changes in the body under the influence of the moon’s gravity.”

Other than there’s no data (that I know off hand) of to back this up:

The problem with data like crime rates is that it contains random noise – patterns can appear and disappear by chance alone. So we first need to ask ourselves how sure we want to be that there is a real difference between crime rates on a full moon and those on any other night. Quite sure? Fairly sure? Almost certain?

Applying these methods to their data, the researchers found there was a significant difference between the number of crimes committed when there was a full moon and the number of crimes on other nights: their p-value was much smaller than 0.001 (usually written as p<0.001). This means there was less than a 0.1% chance of the full moon and non-full moon crime rates actually being the same, despite the numbers that were observed in this particular analysis. (via by @n8thangreen)

Sometimes, I just grab a p-value, slap it around with the significance level and either accept or reject the null hypothesis. You can throw in a critical difference value for good times and don’t forget the F test, those feel nice. Ya know, sometimes it’s all about your confidence intervals. Then, the happy ending is you can proudly proclaim an interpretation of the data that may, depending on certain conditions, be unlikely to have occurred by chance alone. Ahhh.

img