xuyiqing / fect

fixed effect counterfactual estimators
Other
57 stars 20 forks source link

Why do IFe and MC methods give me NA in F test? #36

Open JingYDENG opened 11 months ago

JingYDENG commented 11 months ago

Hi Prof XU,

Thanks for developing this tool. I am a new user and I try to do the counterfactual estimation for my staggered DD. I met a problem, I can't generate a numeric F-test p-value in the pre-trend test. I used an unbalanced panel data in this analysis and the number of observations is about 70000. What should I do?

截屏2023-12-07 00 51 24
JingYDENG commented 11 months ago

My code is : out.ife <- fect(y~ treat, data = sampled_vine, index = c("id","time"), force = "two-way", method = "ife", CV = TRUE, r = c(0, 5), se = TRUE, nboots = 200, parallel = TRUE, min.T0=8)

JingYDENG commented 11 months ago

At the same time, Tr units equally weighted in ATT is also NA.

lzy318 commented 10 months ago

Could you attach a treatment status plot generated by PanelView?