"“Evidence-based policy” is becoming increasingly important (at least as a goal), yet few policy makers (or even policy analysts) know how to evaluate evidence impartially. Many policy makers only believe evidence that reinforces their beliefs and supports decisions they have already made. Or they only trust experts they know personally. Or they only look at the prestige of the researcher and publication, and not at the evidence itself. None of these shortcuts leads reliably to truth. None leads reliably to good policy decisions.
In this class, you’ll learn to evaluate quality of evidence. You’ll learn that no evidence is perfect, but some forms of evidence are more prone to reaching false conclusions than others. You’ll learn what kinds of evidence can demonstrate cause and effect more or less convincingly . And you’ll learn what kinds of evidence lead to conclusions that generalize beyond the time and place where the evidence was first collected.
You’ll learn by doing. Just as writing your own songs helps you better appreciate and understand songs written by others, analyzing your own data helps you to appreciate the strengths and weaknesses of analyses presented by others--and to learn what real data looks like (not the cleaned up toy datasets we use in textbooks). We use Stata, because it’s an easier language than SAS or R. Some graduates have gone on to get jobs as data analysts and researchers."