Statistical Ignorance

To start off, I am not a statistical expert.  I did poorly in my statistics classes, mostly because the first half was so easy I only started going after it suddenly became much harder.  I don't know what a chi squared regression is, or even a correlation coefficient.  This is very common with me, I am not an expert in any field, but I have a better grasp of the basics of many different things than most people.  I am a generalist.

So what do I know that most people don't know about statistics?  I know the difference between an anecdote and a double blind study testing the correlation between two characteristics.  Most people actually value their own anecdotal experience over a statistical study.  So, if all the climate scientists in the world agree that the world is warming, a person experiencing a cooler winter than normal in one part of the world is likely to say that the world isn't warming because they go outside and it is cooler.

Here is an absolutely classic case of this, from an intelligent person with an advanced degree, "The research may say that X has nothing to do with Y, but I have noted with A that an episode of Y, sometimes happens after X."

So what we have here are some data points that a person feels apply (it seems like Y follows X), and some other data points which don't apply (Y does not follow X) and (Y happens without X).  We don't know any of the numbers involved.  This means that we have stuff that wouldn't even count as data for a study and, even if it did, would not have anything like enough data to draw a conclusion. 

On the other hand we have a statistical analysis of multiple studies, with multiple people, with clear parameters, from people unable to have bias towards the people involved by the nature of the studies.


Imagine you have dangerous cancer.  You know a friend who says that she thinks she survived cancer by eating lots of bananas.  The doctor says bananas don't do anything for cancer, and gives you the results of multiple studies that say so.  Would you then eat bananas and not go with the medical advice?  Of course not, because choosing anecdotal evidence over sophisticated statistical analysis is stupid.

Why do people do this?  Because the brain is wired up to do this.  Before there were statistical studies, how people learned things was by noticing what happened to them, and the stories of what happened to people they knew.  People actively looked for patterns in this information in order to be better prepared for the future.  This is a very sensible and efficient means of finding out about the world if you don't have better methods.  It's like having a hoe for planting crops is really good if you don't have a plough.

This method of thinking is innate in people.  But as I have said repeatedly on this blog, lots of innate things in people have been surpassed by better things.  looking for patterns in the world around you is smart, and what we do naturally.  Deliberately choosing this method over statistical analysis is simply dumb.  the world would be a better place if we stopped doing it.