Difference-in-Differences (DiD) design combines both before-after and treatment-control comparison when doing causal inference.
The problems that DiD solves are:
- A treatment-control comparison is not necessarily a causal comparison because of the potential systematic differences between two groups
- A unit is arguably the “best match” for itself
- A before-after comparison (of the same units) is not necessarily a causal comparison because of the potential change in time