Appropriate 'weight' [message #27931] |
Mon, 23 October 2023 13:40 ![Go to next message Go to next message](/theme/default/images/down.png) |
Rupon
Messages: 16 Registered: October 2023
|
Member |
|
|
Dear all!
We plan to conduct the following analysis for a study on maternal care utilisation in India (using NFHS 5)
1. How a set of SELECT states differ from the rest of India in maternal care utilisation in India.
2. What determines maternal care utilisation in the SELECT states in India.
Which weight should we go with? Is it within-state weight for (2)?
Where can we use weight (rural/urban)?
Thank you
Regards
Rupon
|
|
|
Re: Appropriate 'weight' [message #27939 is a reply to message #27931] |
Tue, 24 October 2023 09:32 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to next message](/theme/default/images/down.png) |
Janet-DHS
Messages: 816 Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
State weights are used for the analysis of a single state. If you are analyzing groups of states, as in your proposed research, then you will want to use the national weights, because they weight each state in proportion to its actual population. If you used the state weights, then you would be weighting each state in proportion to its sample size. Smaller states tended to be over-sampled, and larger states tended to be under-sampled (this is typical in stratified samples), and as a result the state weights would over-represent the smaller states and under-represent the larger states, within each group of states. So I recommend that you use the national weight, v005 or hv005, in your analysis.
|
|
|
|