Home » Data » Weighting data » Clustered Standard Errors
Re: Clustered Standard Errors [message #22167 is a reply to message #22150] |
Tue, 09 February 2021 05:04 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to previous message](/theme/default/images/down.png) |
analyst_till
Messages: 10 Registered: January 2021
|
Member |
|
|
Dear Mr. Pullum,
thanks a lot again, the study you mentioned will probably be interesting for me to look at.
I am sorry, my formulation was not well put. In fact, I just wanted to say that I will compare observations in the same districts across time, with districts being my panel variable. However, I do not aggregate my individual observations at district level in the dataset, such that the final outcome Y is the outcome of individual i in district j at time t. Also, I proceeded exactly as you mentioned, such that my regression equation looks as follows (with "post" instead of "survey"):
svy: reg y i.post##i.Treatment i.district
However, I have have two small questions left:
1. If you recommend sticking to "svyset": how can I make sure that I cluster standard errors at district-level then? It does not seem to be possible within the "svyset"/"svy:" framework, or am I wrong?
2. Including robustness checks, I will probably use 4 different DHS rounds for Ethiopia. I noticed that the strata variable v023 was not always coded in the same way (e.g. in one round 22, in the other round 21 strata). Is it best in this case to rely on "egen strata=group(v024 v025)" in each of the rounds, to make sure that the strata variable is indexed in the same way across surveys and afterwards proceed as noted above in the pooled data?
Best,
Till
|
|
|
Goto Forum:
Current Time: Fri Aug 30 23:27:56 Coordinated Universal Time 2024
|