Someone asked me whether it was okay to define a correlation coefficient as a descriptive statistic. It depends on whether the correlation is being used to describe a population or if it is intended to test a particular hypothesis. I also warned that these types of labels (inferential, descriptive) aren’t really needed most of the time.

After I responded to her, I realized that I didn’t have a decent definition of inferential or descriptive statistics, so I add them and found a few examples on the BioMed Central website.

You can find an earlier version of this page on my original website.