Honest DID

[bcallaway11]

I added the wrapper to call code from the HonestDiD package. I'll write up more details before too long. But here is some quick example code:

##-----------------------------------------------------------------------------
## event study code
library(did)
library(HonestDiD)
library(dplyr)
data(mpdta)

cs_res <- att_gt(yname="lemp",
                 tname="year",
                 idname="countyreal",
                 gname="first.treat",
                 data=mpdta,
                 base_period="universal")

ggdid(cs_res, ylim=c(-.25,.15))

cs_es <- aggte(cs_res, type="dynamic")
ggdid(cs_es, ylim=c(-.25,.15))



## -----------------------------------------------------------------------------
# code for running honest_did
hd_cs_smooth <- honest_did(cs_es,
                           type="smoothness")
hd_cs_smooth


hd_cs_rm <- honest_did(cs_es, type="relative_magnitude")
hd_cs_rm


## -----------------------------------------------------------------------------
# make senstivity analysis plots
createSensitivityPlot(hd_cs_smooth$robust_ci,
                      hd_cs_smooth$orig_ci)

createSensitivityPlot_relativeMagnitudes(hd_cs_rm$robust_ci,
                                         hd_cs_rm$orig_ci)

[bcallaway11]

One issue that we'll need to eventually address is that HonestDiD is not on CRAN, so this will stop us from pushing to CRAN when this code is included.