Difference-in-Difference
The difference-in-difference empirical strategy leverages u
neven treatment in natural experiments and randomized-control trials. The specification is quite simple and requires a weak but untestable assumption, that there is no differential trend between the control and treatment groups.
One must be careful in this area however, correlated shocks can lead to statistical significance if the econometrician doesn't cluster the standard errors at the state/classroom/firm level. This is a helpful and important paper which developed this central point.
The difference-in-difference is a very useful tool for economists working on any sort of policy evaluation. The difference-in-difference-in-difference involves another layer of differential treatment, as in, within states, across states, and over time. This is intuitively a stronger set-up.
Vita, Curriculum vitae, cv, Andrew Johnston, Andrew, Johnston, economics, applied economics, economist, microeconomics, empirics, empirical economics, Wharton, Andrew Johnston, Andrew, Johnston, economics, applied economics, economist, microeconomics, empirics, empirical economics, Wharton, Andrew Johnston, Economics, Andrew Johnston economics
