| Harmonizing state identifiers in pooled NFHS-4 and NFHS-5 data [message #31923] |
Fri, 19 June 2026 10:19  |
farihakabir148
Messages: 6
Registered: February 2024
|
Member |
|
|
I am pooling the NFHS-4 (201516) and NFHS-5 (201921) women's IR datasets and using state fixed effects, using IV for my analysis. However, I noticed that the state variables (v024/v101) are not directly comparable because of administrative changes between the two rounds.
For example, NFHS-4 has Jammu and Kashmir as a single unit, while NFHS-5 separates Jammu & Kashmir and Ladakh. Similarly, NFHS-4 treats Dadra and Nagar Haveli and Daman and Diu as separate union territories, whereas NFHS-5 combines them.
For those who have pooled NFHS-4 and NFHS-5 data, what is the recommended way to harmonize the state variable for fixed-effects analyses? Is there an official DHS/NFHS recommendation or an accepted best practice?
If anyone has Stata code they have used for this harmonization and would be willing to share it, I would greatly appreciate it.
Thank you.
|
|
|
|
| Re: Harmonizing state identifiers in pooled NFHS-4 and NFHS-5 data [message #31924 is a reply to message #31923] |
Fri, 19 June 2026 14:05  |
farihakabir148
Messages: 6
Registered: February 2024
|
Member |
|
|
gen state = v024
label var state "State (harmonised NFHS-4/5)"
** gen state = v024
** label var state "State (for fixed effects)"
* Harmonisation adjustments (NFHS-4/5 comparability)
* Ladakh → Jammu & Kashmir
replace state = 14 if Year==2021 & v024==37
* Dadra & Nagar Haveli + Daman & Diu merged in NFHS-5
replace state = 8 if Year==2021 & v024==25
Does this make sense?
|
|
|
|
Members
Search
Help
Register
Login
Home
