I am trying to estimate a policy effect on farmers' land sizes. The policy (treatment) is at the village level. I could simply use a diff-in-diff approach by comparing farmers' lands across treatment and control villages. However, the problem is that some farmers have two lands in two different villages, one of them is in the treatment, and the other is in the control village. I am struggling to estimate the model (or organization of the dataset maybe). Here is the dataset example
Farmer_ID | Village | LandSize | Treatment Status |
---|---|---|---|
1 | A | 10 | Treated |
1 | B | 20 | Control |
2 | A | 15 | Treated |