Hi,
I’m working on a two-dimensional fuzzy RD with two treatments that affect the outcome differently. Both running variables are discrete. In 1D RD, we use the triangular weight w(x)=1−∣x∣/h. For 2D, is it appropriate to use a product kernel w(x,y)=[1−∣x∣/h_x]*[1−∣y∣/h_y]? Any suggestions are welcome.
Rio
I’m working on a two-dimensional fuzzy RD with two treatments that affect the outcome differently. Both running variables are discrete. In 1D RD, we use the triangular weight w(x)=1−∣x∣/h. For 2D, is it appropriate to use a product kernel w(x,y)=[1−∣x∣/h_x]*[1−∣y∣/h_y]? Any suggestions are welcome.
Rio
