The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » India » Geographical units and sampling
Geographical units and sampling [message #1195] Mon, 20 January 2014 23:32 Go to next message
joyp is currently offline  joyp
Messages: 5
Registered: January 2014
Member
Dear Forum Users,

I am trying to understand sampling in DHS surveys. From the documentation, I understand that the surveys are nationally representative and also representative at state level. I was wondering if it is possible to use the data at finer (lower) geograhical level? From the forum discussions, I found that districts can be identified for NFHS1 and NFHS2 (but not NFHS 3 due to HIV module). If I calculate any descriptive statistics at district level, will it be representative of the district?

Similarly, since households are randomly selected within a PSU, can I use the descriptive statistics calculated for a particular PSU? I am worried that it may not be representative of the PSU.

Any clarification would be most useful. Please let me know if my query is unclear. I want to ensure that I understand this survey data collection well in order to use this data correctly.

Thank you.
Re: Geographical units and sampling [message #1223 is a reply to message #1195] Thu, 23 January 2014 23:49 Go to previous messageGo to next message
joyp is currently offline  joyp
Messages: 5
Registered: January 2014
Member
Dear Forum users,

I would be most grateful for any suggestions re my query below.

joyp wrote on Tue, 21 January 2014 15:32
Dear Forum Users,

I am trying to understand sampling in DHS surveys. From the documentation, I understand that the surveys are nationally representative and also representative at state level. I was wondering if it is possible to use the data at finer (lower) geograhical level? From the forum discussions, I found that districts can be identified for NFHS1 and NFHS2 (but not NFHS 3 due to HIV module). If I calculate any descriptive statistics at district level, will it be representative of the district?

Similarly, since households are randomly selected within a PSU, can I use the descriptive statistics calculated for a particular PSU? I am worried that it may not be representative of the PSU.

Any clarification would be most useful. Please let me know if my query is unclear. I want to ensure that I understand this survey data collection well in order to use this data correctly.

Thank you.

Re: Geographical units and sampling [message #1231 is a reply to message #1195] Sat, 25 January 2014 10:46 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User,
I am working on getting an answer for you.
Thanks!
Re: Geographical units and sampling [message #1234 is a reply to message #1231] Sun, 26 January 2014 10:08 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User,
Here is a response from one of our experts, Fred Arnold:

It's correct that the districts can be identified for NFHS-1 and NFHS-2, but the districts were not identified in NFHS-3. All three NFHS surveys are representative at the national and state level. The samples are not designed to be representative at the district level. It is not recommended that descriptive statistics be calculated at the district level. The PSUs are not identified in any of the NFHS surveys. The sample for each PSU is representative of the PSU, but there are too few households in any PSU to produce reliable statistics at that level. However, if you are doing a multi-level analysis, you could use PSU averages in the analysis.

Re: Geographical units and sampling [message #1277 is a reply to message #1234] Mon, 03 February 2014 22:33 Go to previous message
joyp is currently offline  joyp
Messages: 5
Registered: January 2014
Member
Thank you Fred and Liz. It answers my question perfectly. I am planning to conduct multilevel analysis using PSU averages. Your help is much appreciated.
Previous Topic: Health care in India
Next Topic: Sas and DHS
Goto Forum:
  


Current Time: Mon Nov 11 05:52:49 Coordinated Universal Time 2024