Install csdid2 package

[Luis Miguel García]

Dear all,

I´m trying to install the csdid2 package from github: https://github.com/friosavila/csdid2.

However my code is not working. This is:

net install github, from("https://haghish.github.io/github/")
github install friosavila/csdid2

Using the checking command, the install option is not available. How can I install the package?

github search csdid2, in(all) language(Stata) all

Thanks.

[FernandoRios]

Hi Luis,
the latest is installed as follows

net install csdid2, from("https://friosavila.github.io/stpackages")

This is the place where i have all my latest updates, both released, or just experimental

[Luis Miguel García]

Thanks Fernando!

[Laurence Jones]

Hi, FernandoRios this install method no longer seems to be working (at least not form me). However, I found the approach listed here (https://economistwritingeveryday.com/tag/honestdid/) works.

program fra
    syntax anything, [all replace force]
    local from "https://friosavila.github.io/stpackages"
    tokenize `anything'
    if "`1'`2'"==""  net from `from'
    else if !inlist("`1'","describe", "install", "get") {
        display as error "`1' invalid subcommand"
    }
    else {
        net `1' `2', `all' `replace' from(`from')
    }
    qui:net from http://www.stata.com/
end
fra install fra, replace
fra install csdid2
ssc install coefplot

[Julie Brun Bjorkheim]

Hi Fernando

I am trying to install csdid2 using net install csdid2, from("https://friosavila.github.io/stpackages") as you suggested above. It dosent complain when doing this and says all files are up to date. But when using the command firstly I cannot open a helpfile at all. It dosent recognize any helpfile for csdid2.

When running a simple csdid2 lemp, ivar(countyreal) tvar(year) gvar(first) it starts producing the regression, but haults without any error message, see screen shot.

Have you run into this problem? Any thoughts?

[FernandoRios]

Hi Julie
That is not an error
I just havent written a helpfile for csdid2. Most of it works just like csdid, and the repository has extra notes on how to add other restrictions when doing aggregations.
One of my "soft goals" this year is to write its helpfile, add an extra extension (more efficient use of memory). But its still on the to-do list.
F

[Chen Samulsion]

Fernando haven't written a helpfile for cidid2, maybe you can use csdid.sthlp temporarily.

When running a simple csdid2 lemp, ivar(countyreal) tvar(year) gvar(first) it starts producing the regression, but haults without any error message, see screen shot.
Have you run into this problem? Any thoughts?

After running csdid2, you can just use estat to produce outcomes you want

estat event

[Paola Rossi]

Hello Fernando.
I am using csdid2 that I find very efficient. The only problem is that I cannot manage to save the plot as I used to do with csdid. Can you help me with that? Thank you Paola

[FernandoRios]

Need more detailed question

[Jiancong Liu]

Hi Fernando,

Thank you very much for creating csdid2 command, which saves me tons of time. I have a question regarding the difference with csdid command, though. To illustrate my question, I'm using the example data from the help file of csdid.

use https://friosavila.github.io/playingwithstata/drdid/mpdta.dta, clear

I tried to run the same specification using csdid and csdid2:

csdid  lemp lpop , ivar(countyreal) time(year) gvar(first_treat) method(dripw) agg(simple) notyet
estat event

csdid2  lemp lpop , ivar(countyreal) time(year) gvar(first_treat) method(dripw) agg(simple) notyet
estat event

Although the simple ATT estimates and post average estimates are the same across two commands, the pre average estimates and the estimates for pre-shock periods are different. I'm attaching the results below.

. use https://friosavila.github.io/playingwithstata/drdid/mpdta.dta, clear
(Written by R.              )

. csdid  lemp lpop , ivar(countyreal) time(year) gvar(first_treat) method(dripw) agg(simple) notyet
............
Difference-in-difference with Multiple Time Periods

                                                         Number of obs = 2,500
Outcome model  : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         ATT |  -.0413516   .0114278    -3.62   0.000    -.0637498   -.0189535
------------------------------------------------------------------------------
Control: Not yet Treated

See Callaway and Sant'Anna (2021) for details

. estat event
ATT by Periods Before and After treatment
Event Study:Dynamic effects
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
     Pre_avg |  -.0004609   .0072816    -0.06   0.950    -.0147326    .0138108
    Post_avg |  -.0799926    .018494    -4.33   0.000    -.1162402    -.043745
         Tm3 |   .0267278   .0140657     1.90   0.057    -.0008404     .054296
         Tm2 |  -.0052499   .0130015    -0.40   0.686    -.0307325    .0202326
         Tm1 |  -.0228606   .0146118    -1.56   0.118    -.0514991     .005778
         Tp0 |  -.0201445   .0116323    -1.73   0.083    -.0429433    .0026544
         Tp1 |  -.0547303      .0164    -3.34   0.001    -.0868738   -.0225869
         Tp2 |  -.1381918    .034228    -4.04   0.000    -.2052775   -.0711061
         Tp3 |  -.1069039   .0328865    -3.25   0.001    -.1713602   -.0424476
------------------------------------------------------------------------------

. csdid2  lemp lpop , ivar(countyreal) time(year) gvar(first_treat) method(dripw) agg(simple) notyet
Producing Long Gaps by default
Using method dripw
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
............
Difference-in-difference with Multiple Time Periods
Outcome model  : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
             |               Robust
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   SimpleATT |  -.0413516   .0114278    -3.62   0.000    -.0637498   -.0189535
------------------------------------------------------------------------------

. estat event
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
     Pre_avg |   .0190891   .0169212     1.13   0.259    -.0140758    .0522541
    Post_avg |  -.0799926    .018494    -4.33   0.000    -.1162402    -.043745
         tm4 |   .0062963   .0245367     0.26   0.797    -.0417948    .0543873
         tm3 |   .0281105   .0176754     1.59   0.112    -.0065326    .0627536
         tm2 |   .0228606   .0146118     1.56   0.118     -.005778    .0514991
         tp0 |  -.0201445   .0116323    -1.73   0.083    -.0429433    .0026544
         tp1 |  -.0547303      .0164    -3.34   0.001    -.0868738   -.0225869
         tp2 |  -.1381918    .034228    -4.04   0.000    -.2052775   -.0711061
         tp3 |  -.1069039   .0328865    -3.25   0.001    -.1713602   -.0424476
------------------------------------------------------------------------------

Could you kindly share some insights on how I should interpret the difference between them? Thanks a lot in advance!

[FernandoRios]

For thy to match you need to use the long2 option in csdid
that is the default with csdid2