Friday, February 3, 2012

Crime rates very high when the moon is full? It is a matter for the werewolves statistics among you

I was a kid

particularly loose. I'm not that brave or an adult, but the themes of my cowardice have changed a bit. As an adult, I am more afraid of losing my job in a recession or my Internet identity theft. As a child, was terrified of werewolves. Every full moon, I do not worry about being on the wrong side of crackling, sharp teeth.

course you should not have been more worried when the moon was full from any other night. Or what? A classic article in the British Medical Journal has tried to answer a similar question: crime rates highest on the full moon

Science is all about formulating and testing hypotheses. In this case, the assumption would be: "Crime rates are higher when the full moon." Often, scientists test the "null hypothesis": the default declaration which, if true, suggests that the experiment did not detect a real effect. In this case, the null hypothesis can be expressed as follows: "There is no difference in crime rates when the moon is full, compared to other nights."

The problem with data such as crime rates is that it contains random noise - the patterns can appear and disappear as if by accident. Therefore, we must first ask ourselves how we want that there is no real difference between crime rates and the full moon than at any other night. Very safe? Pretty sure? Almost sure?

In other words, how high we set the bar? Statisticians to quantify the degree of certainty they want to reach the significance level of a threshold between 0 and 1, where the closer to 0, the more confident you can be no real difference between two sets of numbers . This is what scientists mean when they talk about how "significant" results are.

The significance level can also be expressed as a percentage between 0 and 100. For example, if we want to be at least 95% as no real difference between the sets of numbers, this is called a significance level of 5% (0.05). In other words, if one compares the number of crimes in a lot of the full moon and full moon nights, the numbers would be different at least 95% of the time.

And that's all. Now you have the tools to answer the original question, because when the p value is below a level of significance, we can say with some statistical confidence that crime rates rise when the full moon.

Applying these methods to their data, researchers found that there was a significant difference between the number of crimes committed when the moon was full and the number of crimes in the evening other : P value was less than 0001 (usually written as p
In other words, the researchers showed that if it were to repeat this analysis several times, eg every year, we hope to obtain a difference in the number of crimes between the full moon and no the full moon most of the time.


Blog Archive