calculating SE and CIs for selected metrics (both proportions and rates) [message #27461] |
Wed, 16 August 2023 10:11 |
geoK
Messages: 56 Registered: May 2014
|
Senior Member |
|
|
Hello,
I am looking for ways to calculate SEs and CIs for rates (and in particular Stillbirth rate) and proportions (e.g. teenage pregnancy and unsafe abortion) at national, state, and district level (NFHS-4&5 India surveys). I came across the Report's appendix B where it is said:
"Taylor linearization method to estimate variances for survey estimates that are means, proportions, or ratios. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates."
Can i confirm that Taylor lin method was used to calculate SEs shows in Tables B for both rates and proportions, and is there any chance to access the SAS programs developed by ICF to calculate those SEs / CIs please?
Thank you!
PS: I already have programs to calculate mortality and fertility rates and their CIs, but do not have any program for stillbirth.
|
|
|
Re: calculating SE and CIs for selected metrics (both proportions and rates) [message #27463 is a reply to message #27461] |
Wed, 16 August 2023 12:34 |
Bridgette-DHS
Messages: 3199 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS staff member, Tom Pullum:
We cannot distribute the SAS programs you describe, but you don't need them. Stata has commands to calculate standard errors and confidence intervals. Here is an example, which gives the proportion of children born in the past 5 years who are boys or girls, in the NFHS-5, in each state, with standard errors and confidence intervals adjusted for the survey design with svyset and svy. In addition to "proportion", there are similar commands for the "mean" and "ratio".
use "...IAKR7EFL.DTA", clear
svyset v001 [pweight=v005], strata(v023) singleunit(centered)
svy: proportion b4, over(v024)
Your PS suggests that you want to analyze stillbirths. Please look at other forum posts on that topic, but if you still need help, please post another message with a different title.
|
|
|