| Creating Time Variable For Survival Analysis & Missing(Dots) Cells [message #20177] |
Tue, 06 October 2020 01:19  |
shujaat.smc@gmail.com
Messages: 75
Registered: July 2020
|
Senior Member |
|
|
Dear DHS Representative,
I have created a time variable for neonatal death using the b6 variable, please have a look at the tabular chart of "dayofneonataldeath", missing 31,683 what should I do with this big number, putting this new time variable in Stata stset shows that the 31,683 observations are missing.
Should I use negative coding for these missing numbers, .=-1 or use code Not applicable .=NA ??? or recode time variable as (202/.=Month2&above) so it includes all the deaths in months & missing values.
Please see the tab of my new variable (dayofneonataldeath) below.
Waiting for your reply.
Best Regards
Dr. Hussain
RECODE of b6 |
(age at |
death) | Freq. Percent Cum.
-------------+-----------------------------------
Day0 | 360 1.05 1.05
Day1 | 275 0.80 1.84
Day2 | 120 0.35 2.19
Day3 | 149 0.43 2.63
Day4 | 74 0.21 2.84
Day5 | 67 0.19 3.04
Day6 | 60 0.17 3.21
Day7 | 46 0.13 3.34
Day8 | 32 0.09 3.44
Day9 | 15 0.04 3.48
Day10 | 34 0.10 3.58
Day11 | 16 0.05 3.63
Day12 | 15 0.04 3.67
Day13 | 15 0.04 3.71
Day14 | 8 0.02 3.74
Day15 | 32 0.09 3.83
Day16 | 6 0.02 3.85
Day17 | 5 0.01 3.86
Day18 | 4 0.01 3.87
Day19 | 2 0.01 3.88
Day20 | 25 0.07 3.95
Day21 | 7 0.02 3.97
Day22 | 8 0.02 3.99
Day23 | 2 0.01 4.00
Day24 | 4 0.01 4.01
Day25 | 10 0.03 4.04
Day26 | 1 0.00 4.04
Day27 | 2 0.01 4.05
Day28 | 6 0.02 4.07
Day30 | 149 0.43 4.50
Month2&Above | 1,193 3.47 7.97
. | 31,683 92.03 100.00
-------------+-----------------------------------
Total | 34,425 100.00
|
|
|
|
| Re: Creating Time Variable For Survival Analysis & Missing(Dots) Cells [message #20199 is a reply to message #20177] |
Wed, 07 October 2020 08:10  |
Bridgette-DHS
Messages: 3214
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
The variables b6 and b7 are Not Applicable if b5=1, that is, if the child is alive at the time of the survey. They are not "missing". If you want a binary variable for a neonatal death, the easiest thing is to do something like this:
gen nndeath=0
replace nndeath=1 if b7==0
Sometimes you do need to use b6, rather than b7, for example to identify early or late neonatal deaths. The important point, however, is that when b6 and b7 are NA, the child is alive and your outcome would be coded 0.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
