| Natural/Quasi | In non-experimental settings, sometimes implicit randomization occurs, and the treatment occurs “as if” it is random | Uni admission cutoff provides a natural experiment on uni education. Students just above/below are likely to be very similar. For these students, uni education is “as if” random. Comparing these students (ones that went to uni/not) produces an estimate of the causal effect of college education. | |
| Regression Discontinuity Design | Discrete treatment status determined by an underlying continuous variable, which is used for quasi experiments | |  |
| Differences-in-Differences | 2 time-series process \(y_1\) and \(y_2\) have the factors affecting them | |  |
RCT
(Randomized Control Trials) | | | |