Regression discontinuity design

In the cut-off trial there are then three regions for allocation. If the middle randomized region is eliminated so that the two outer regions meet, we then have the extreme example of a cut-off design, the regression discontinuity design. In analysing the cut-off... 

For a=0, which corresponds to the extreme case of a cut-off design in which no patients are randomized (the regression discontinuity design), E[A2 2.75. For a = oo we have a completely randomized design and A2 1. 

Credible evaluation uses robust statistical techniques to construct the counterfactual. These techniques include randomization or experimental design quasi-experiments such as matching, regression discontinuity design, or double differences and nonexperimental or instrumental variable methods (annex 6.3). All these techniques try to guarantee the comparability of the treatment and control groups by avoiding what statisticians refer to as bias (box 6.7). 

---