facerizon wrote:A classic example from the world of Sociological statistics: In the mid-20th century a correlation was drawn between Ice Cream sales and murder rates in the inner-city. Statistics showed that every time ice cream increased in quantities sold, more murders occurred. On the surface, this lead some to ponder if ice cream consumption was actually the root cause of homicide. Logically, this was not the case, and later became known as "cum hoc ergo propter hoc" or a logical fallacy. Ice cream sales increased simply because it was hot outside. Murders rate also increased because it was hot outside, as people's tension levels were unable to adapt to their suddenly uncomfortable environments. Ice cream however, regardless of the correlational statistic, did not contribute to higher frequencies in homicidal tendencies.
The prevailing argument here and against stats in general on this thread is that statistics can be used badly. That doesn't discredit the stats themselves. It's the same as the prevailing religious argument against Science, "oh well, Science used to think the world was flat, what does it know?". Numbers are evidence. It's up to you to use those numbers to form an argument and a basis for your "observations". If I use the numbers as evidence for my claim, and I've engaged in some kind of logical fallacy, it's up to you to call me out on it.
"Observation" is not a legitimate evidence. If your claim is "Curry is a weak defender because he lets defenders blow by him", I want you to watch a dozen Curry games and record and jot down every time he lets a defender blow by him, then do the same thing for every other point guard and find out whether he's above or below the average. Of course you're not going to do that, so you go and look at statistics that others have compiled that do it for you.
An observation is
not evidence for anything.
)
