| Regions as stratas [message #30203] |
Mon, 14 October 2024 14:48  |
Noorulain
Messages: 3
Registered: August 2024
|
Member |
|
|
Number of strata = 10 Number of obs = 4,516
Number of PSUs = 404 Population size = 4,461.1268
Design df = 394
F(11, 384) = 4.45
Prob > F = 0.0000
------------------------------------------------------------ ----------------------------
| Linearized
m34_new | Odds ratio std. err. t P>|t| [95% conf. interval]
-----------------------+------------------------------------ ----------------------------
place_of_delivery_cat |
Home | 1.056671 .1411085 0.41 0.680 .8126801 1.373916
Government | .996002 .1372553 -0.03 0.977 .7596225 1.305938
|
size_child_new |
Small | .4480244 .0928376 -3.87 0.000 .2981096 .6733289
Large | 2.385682 .6654784 3.12 0.002 1.378609 4.128423
|
highest_edu_cat |
No education | 1.468173 .2833936 1.99 0.047 1.004542 2.145786
Primary | 1.060413 .2033389 0.31 0.760 .7273613 1.545965
Secondary | .9829099 .1947128 -0.09 0.931 .6658441 1.450958
Initially variable named "region" having four categories (sindh, punjab, balochistan, KPK) was in my multivariable model, I removed it later on because I was told that the "number of stratas 10" signifies region variable, so I was told to remove region from my multivariable model.
is this the correct approach? what does stratas and PSU signifies here. what does 10 stratas and 404 PSU signifies? should I removed it from my multivariable model? Once I removed this, my primary exposure variables becomes insignificant which was earlier significant otherwise. Leads on this?
I am using PDHS IR File, 2017-2018.
2) "Total children ever born" this variable has only live births and no still births, right? Reference "PDHS IR FILE 2017-2018"
|
|
|
|
| Re: Regions as stratas [message #30223 is a reply to message #30203] |
Thu, 17 October 2024 09:53  |
Janet-DHS
Messages: 888
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
The number of children ever born refers only to live births. It does not include stillbirths. This question has come up several times on the forum. I am surprised that there is any uncertainty about it. "Birth" is equivalent to "live birth". It does not include stillbirths.
For your main question, please give us your svyset statement and the estimation command (something like "svy: logit...."). The strata in this survey are given by v022 or v023 (they are identical). The strata are combinations of region and urban/rural. It should be ok for a variable such as region to be on the right hand side of an estimation command AND to be part of the stratification. Who told you to remove region from the model?
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
