Page 522 -

C h a p t e r Experiments and

13 Quasi-Experiments

In many fields, such as psychology and medicine, causal effects are commonly
estimated using experiments. Before being approved for widespread medical
use, for example, a new drug must be subjected to experimental trials in which
some patients are randomly selected to receive the drug while others are given a
harmless ineffective substitute (a “placebo”); the drug is approved only if this
randomized controlled experiment provides convincing statistical evidence that
the drug is safe and effective.

There are three reasons to study randomized controlled experiments in an
econometrics course. First, an ideal randomized controlled experiment provides a
conceptual benchmark to judge estimates of causal effects made with observational
data. Second, the results of randomized controlled experiments, when conducted,
can be very influential, so it is important to understand the limitations and threats
to validity of actual experiments as well as their strengths. Third, external
circumstances sometimes produce what appears to be randomization; that is,
because of external events, the treatment of some individual occurs “as if” it is
random, possibly conditional on some control variables. This “as if” randomness
produces a “quasi-experiment” or “natural experiment,” and many of the methods
developed for analyzing randomized experiments can be applied (with some
modifications) to quasi-experiments.

This chapter examines experiments and quasi-experiments in economics. The
statistical tools used in this chapter are multiple regression analysis, regression
analysis of panel data, and instrumental variables (IV) regression. What distinguishes
the discussion in this chapter is not the tools used, but rather the type of data
analyzed and the special opportunities and challenges posed when analyzing
experiments and quasi-experiments.

The methods developed in this chapter are often used for evaluating social or
economic programs. Program evaluation is the field of study that concerns estimating
the effect of a program, policy, or some other intervention or “treatment.” What is
the effect on earnings of going through a job training program? What is the effect
on employment of low-skilled workers of an increase in the minimum wage? What
is the effect on college attendance of making low-cost student aid loans available

521