8 - Learning from Economic Data

Summary

The use of statistics and econometrics in economic applications is a broad field. Rather than try to cover a representative sample or even an unrepresentative sample of the different statistical applications in economics, I will pick one area where basic statistical and regression results are very useful and then work through a number of examples of applications within that area. These applications and examples I think you will find interesting and timely; in addition, they will nicely illustrate some of the things we will cover later on. They also have some special interest for me.

CONTROLLED EXPERIMENTS ARE RARE

Let's talk about some issues that come up in many applications of statistical and econometric tools when they are applied to economics. These are issues that provide a challenge for doing statistical or empirical work in economics. The main issue that we have to confront in economics is that we rarely have controlled experiments in which we can analyze the effect of the change in one variable on another variable. I say “rarely” because there are occasional exceptions. A basic paradigm of classic statistical analysis that is very prevalent in experimental psychology, as well as in medical and clinical applications, is the idea of testing a hypothesis about the effect of a certain drug or treatment by experimenting on one group of patients while using another group as a control. The control group should be as similar as possible to the treated group.