need help replicating Childhood mortality statistics from Nigeria [message #2676] |
Wed, 06 August 2014 08:24 |
purnima1921@gmail.com
Messages: 3 Registered: August 2014
|
Member |
|
|
Hi,
I am trying to replicate country wide under 5 and infant mortality rates in the Nigeria DHS report 2013 and then create state wide statistics for the same using the same code(so I know I am doing it correctly) ( http://dhsprogram.com/publications/publication-FR293-DHS-Fin al-Reports.cfm ). I would really appreciate your help with the following questions.
I used SAS to extract the numerators and denominators for each cohort (A,B,C) and age group and then used excel to calculate the final numerators and denominators as well as statistics. The codes are attached below in a word file with details of what I did. I come close and was able to replicate the neonatal mortality rate but have not been able to replicate the exact statistics for the others and was wondering if I am missing something important.
A related question is if I am then calculating the same values for states and use a 10 year time period instead of five, should I do it as a single rate for a 10 year period or calculate it separately for 2 five year periods and then combine it? Is there something I should keep in mind while trying to break these down by states (besides the small sample sizes creating large standard errors?)
Finally-- how would you code for standard error's or confidence intervals around these estimates
Thanks in advance,
Purnima
|
|
|