The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset use in Stata » DHS Sample Weights & Fixed Effects Controls (DHS Sample Weights & Fixed Effects Controls)
DHS Sample Weights & Fixed Effects Controls [message #22058] Tue, 26 January 2021 02:51 Go to next message
Bianca is currently offline  Bianca
Messages: 5
Registered: December 2020
Member
Dear DHS - User,

in my repeated cross-sectional analysis, I am using IR files for surveys between 2000 - 2018 which were conducted in countries of SSA.

Thereby, I followed the DHS sample weights procedure:

gen wgt = v005/1000000
egen stratum = group (v024 v025)
svyset [pw = wgt], psu(v021) strata (stratum)

Now, I am wondering whether - due to the svy command - I am already controlling for region-fixed effects or whether I have to include a control variable for the region (v000 v024)? It is a rather trivial question but since it is my first time working with DHS data, I wanted to kindly ask you for your advice.

Thank you very much in advance!

Warm regards,
Bianca
Re: DHS Sample Weights & Fixed Effects Controls [message #22062 is a reply to message #22058] Tue, 26 January 2021 09:43 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2270
Registered: February 2013
Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

No, you do not need to adjust or control for region. The strata are usually defined by combinations of v024 and v025, but not always. The attached file: "Survey_strata.do" gives the strata for each survey. I suggest that you add "singleunit(centered)" to the svyset command to prevent a type of problem that may cause a crash.

[Updated on: Mon, 08 February 2021 09:52]

Report message to a moderator

Re: DHS Sample Weights & Fixed Effects Controls [message #22063 is a reply to message #22062] Tue, 26 January 2021 10:35 Go to previous messageGo to next message
Bianca is currently offline  Bianca
Messages: 5
Registered: December 2020
Member
Dear Mr Pullum,

Thank you very much for the quick & very helpful answer as well as the do-file!

May I ask you one more question:

After running your dofile, the strata variable for the two surveys "CD5" and "UG6" is missing. In this case, it is valid to replace them by your strata_v024 (group(v024 v025)) variable, right?

Thank you very much for your support. Warm greetings from Germany!
Re: DHS Sample Weights & Fixed Effects Controls [message #22064 is a reply to message #22063] Tue, 26 January 2021 10:56 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2270
Registered: February 2013
Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Yes, you can do that. The strata (and clusters) only affect the standard errors of the estimates. You will always be very close to the correct adjustment for strata if you use the combinations of v024 and v025.

Re: DHS Sample Weights & Fixed Effects Controls [message #22065 is a reply to message #22064] Tue, 26 January 2021 10:59 Go to previous message
Bianca is currently offline  Bianca
Messages: 5
Registered: December 2020
Member
Great, thank you so much!
Previous Topic: How to use cluster and strata variables in ever married surveys?
Next Topic: Intimate partner Violence
Goto Forum:
  


Current Time: Wed May 12 10:03:25 Coordinated Universal Time 2021