Pearson Coefficient

Not to be confused with the production possibilities curve, the Pearson correlation coefficient (the other “PPC”) is a measurement in statistics of the linear correlation between X and Y.

Imagine a bunch of dots on a graph, spread out everywhere pretty evenly. That’s a pretty weak correlation of the two variables, which means the Pearson coefficient is close to zero. Now imagine those dots practically creating an upward-sloping line...that’s a positive correlation, which means it’s close to one.

If the dots make up the line pretty well, but it’s downward-sloping instead of upward-sloping, we’ve got just as good of a correlation, but it’s negative instead (close to negative one).

Since the Pearson coefficient is only measuring linear correlation, it’s pretty straightforward...no crazy curves to deal with, as would be the case with a logistic function. Yet it means that, for some X and Y relationships, a linear correlation might be the wrong tool to use. For many relationships though, it fits the statistical bill. There are some tests you can do, like T-tests, to see if you should be using another statistical tool or not.

The Pearson coefficient has a lot of (good) baggage attached to it: it’s used in regression analysis...yep, that’s machine learning...and has cool mathematical applications for populations and samples (a part of a population). It helps economists, statisticians, and researchers estimate inferences based on their data. Basically, it helps them see if a relationship is likely there, or not. X and Y are either peas in a pod, or X friendzones Y...because it was a love triangle with the null hypothesis.