Hi

[Luzcristina]

Hi, apologies if this is obvious...I have a set of panel data and I am trying to evaluate changes in income caused by the staggered introduction of a policy change. When I try to add covariates to the csdid command I get the same ATT results that I obtained without any covariates. Is this normal?

I have replicated below what my code looks like when I add my covariates:

csdid income, ivar(pidp) time(interview_year)) gvar(g) indepvar (age_dv i.urban_dv i.employement_status i.benefit_income i.education_level) not yet method (dripw) agg(event) wboot replace rseed(1)

Many thanks in advance,

Cristina

[Fernando Rios-Avila]

Hi there, there is a small typo in how you are using the command. the correct syntax should be: csdid income age_dv i.urban_dv i.employement_status i.benefit_income i.education_level, ivar(pidp) time(interview_year)) gvar(g) notyet method(dripw) agg(event) wboot rseed(1) My other recommendation is to make sure your overlapping assumption holds for your dummies for instance, tabulate: tab benefit_income g If there are any zeroes there, you may have to think about dropping that G, or recoding benefit_income HTH F

…

On Thu, Mar 30, 2023 at 8:45 AM Luzcristina ***@***.***> wrote: Hi, apologies if this is obvious...I have a set of panel data and I am trying to evaluate changes in income caused by the staggered introduction of a policy change. When I try to add covariates to the csdid command I get the same ATT results that I obtained without any covariates. Is this normal? I have replicated below what my code looks like when I add my covariates: csdid income, ivar(pidp) time(interview_year)) gvar(g) indepvar (age_dv i.urban_dv i.employement_status i.benefit_income i.education_level) not yet method (dripw) agg(event) wboot replace rseed(1) Many thanks in advance, Cristina — Reply to this email directly, view it on GitHub <#35>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ASZKKFTDXED4FLAUNXVDJBDW6V56JANCNFSM6AAAAAAWNHZ2WE> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Luzcristina]

Excellent! Thank you so much Fernando, it seems to now be working and producing meaningful results too :-)
Kind regards,
Cristina