The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Nepal » NDHS 2022 analysed data matching with the country report (about NDHS 2022)
NDHS 2022 analysed data matching with the country report [message #30238] Mon, 21 October 2024 10:40 Go to next message
Suna is currently offline  Suna
Messages: 5
Registered: October 2024
Member
I hope someone can find this post.
I am analysing the Nepal DHS 2022 data.
I thought I had finished cleaning the data, but the numbers in the cleaned data differ from those in the NDHS2022 report.

So, I have two questions.
1. In general, when researchers analyse DHS data, do they adjust it to match the numbers in the country reports?
2. If I limit the women to those aged 15-49, I think the numbers in the data analysed by STATA will match those in Tables 9.1 and 9.2 in the country reports (2022 NDHS). However, both numbers did not match. I would appreciate it if you could tell me the correct STATA code.

I apologise for the difficult work.
Re: NDHS 2022 analysed data matching with the country report [message #30277 is a reply to message #30238] Fri, 25 October 2024 14:39 Go to previous messageGo to next message
Janet-DHS is currently offline  Janet-DHS
Messages: 899
Registered: April 2022
Senior Member
Following is a response from DHS staff member, Tom Pullum:

I think you are trying to get the total numbers of births and stillbirths for tables 9.1 and 9.2.  In Stata you can do the following:

use "...NPNR82FL.DTA", clear
sort v001 v002 v003 pidx
egen sequence=seq() if p32<=2 & p19<24, by(v001 v002 v003)
tab sequence

* Totals for table 9.1 and 9.2
tab p32 if sequence==1 [iweight=v005/1000000]

Note that you would use the new NR file, and select on p32 and p19. The "sequence" line selects the most recent birth or stillbirth in the past two years. You need to weight by v005.

The subtotals I get are 1932 and 17, which add to 1949. The final report gives 1933 and 18, which are different from 1932 and 17. It gives a total of 1949, which is the sum of 1932 and 17 but is not the sum of 1933 and 18. A puzzling discrepancy but negligible.

Let us know if you have trouble matching the rest of the table.

You ask "In general, when researchers analyse DHS data, do they adjust it to match the numbers in the country reports?" The answer is no. You want to match the report, if possible, but if you can't get a match, we would not recommend that you make adjustments.
Re: NDHS 2022 analysed data matching with the country report [message #30286 is a reply to message #30277] Sat, 26 October 2024 09:14 Go to previous messageGo to next message
Suna is currently offline  Suna
Messages: 5
Registered: October 2024
Member
Thank you for answering my question.

I understand how to make the correct number.

Regarding my last question, although the numbers still have discrepancies, I will continue the analysis according to the research plan.

It was very helpful.
Re: NDHS 2022 analysed data matching with the country report [message #30405 is a reply to message #30238] Sun, 24 November 2024 21:16 Go to previous messageGo to next message
pastoralworth is currently offline  pastoralworth
Messages: 1
Registered: November 2024
Member
Discrepancies between raw DHS data and report figures are common due to adjustments made in the reports, such as applying weights and accounting for sampling design. Researchers typically adjust the data to match the numbers in country reports by using the appropriate survey weights and design features. To match the figures from Tables 9.1 and 9.2 retro bowl in the 2022 NDHS, ensure you apply the correct weights in STATA using the svyset command, like this: svyset [pweight=weight_variable], strata(strata_variable) fpc(fpc_variable). This will align your analysis with the weighted data used in the report.






Re: NDHS 2022 analysed data matching with the country report [message #30423 is a reply to message #30405] Wed, 27 November 2024 00:01 Go to previous message
Suna is currently offline  Suna
Messages: 5
Registered: October 2024
Member
Dear Pastoralworth,


Thank you for your support.

"svyset" command helped me to make correct results.


Best regards,

Suna
Previous Topic: number of women with a live birth in the 2 years preceding the survey
Next Topic: Linking SPA datasets & SPA geolocation information
Goto Forum:
  


Current Time: Sun Dec 8 05:33:15 Coordinated Universal Time 2024