The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Mortality » Cox Model and Recoded Variables From BR File (Negative Coding for Dotted (.) Observation)
Cox Model and Recoded Variables From BR File [message #20206] Thu, 08 October 2020 01:41 Go to next message
shujaat.smc@gmail.com is currently offline  shujaat.smc@gmail.com
Messages: 75
Registered: July 2020
Senior Member
Dear DHS Representative,

I am working on Sub-population of neonates with Bw2500gms and above, looking for determinants of mortality among them in four provinces of Pakistan.

I recode the following variables using "Negative Coding" for dotted observations and Low birth weight babies using following commands;

****Size of child at birth****
recode m18 (4/.=-1 "NotApp&Lower") (3=1 "Average") (2=2 "LargerThanAvg") (1=3 "VeryLarge"), gen (percievedsizeofbaby)

*******When the Child First put to breats *****

recode m34 (0/201=1 "DayofBirth") (202/207=2 "Between2nd-7thDayofBirth") (208/230=3 "From8th-30thDayofBirth") (.=-1 "NotApp"), gen (breastfeedingpractice)

**** Variable Birthorder bord recoding *********

recode bord (1=1 "First") (2=2 "Second") (3=3 "Third") (4=4 "Fourth") (5/15=5 "FifthorAbove"), gen (BirthorderofNewborn)

************************************************************ *****************
I use these new variables for cox-model using command

svy linearized, subpop(subpopulationbw2) : stcox ib(last).breastfeedingpractice ib(last).BirthorderofNewborn ib(first).percievedsizeofbaby

******* My Question is on the following error message from stata 13.1 ****

🚩Note: subpop() takes on values other than 0 and 1
subpop() != 0 indicates subpopulation
(running stcox on estimation sample)

breastfeedingpractice: factor variables may not contain negative values
an error occurred when svy executed stcox 🚩r(7);


Kindly explain what mistake I am making in recoding & generating new variables and why got this error in stata.

Truely obliged and appreciate your quick responses for my questions asked earlier.

Best Regards

Dr. Hussain
Re: Cox Model and Recoded Variables From BR File [message #20210 is a reply to message #20206] Thu, 08 October 2020 07:45 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3016
Registered: February 2013
Senior Member

Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

Your recodes look fine. The problem is just with your specification of the subpop option for svy. What Stata expects for the argument (you have entered "subpopulationbw2") is a variable that has the value 1 for the subpopulation you want to restrict to. For example, you could have this line, "gen Punjab=1 if v024==1" and then specify the option with "subpop(Punjab)". Then to run the model for Sindh you would have "gen Sindh=1 if v024==2" and then specify the option with "subpop(Sindh)", etc. (If this doesn't work, try recoding as 0 or 1, but I think . or 1 is ok.) It's good that you are running the model with the subpop option rather than with "...if v024==1", etc. It's just that the specification of the argument is off.

Previous Topic: maternal mortality
Next Topic: Previous pregnancy birth interval in months
Goto Forum:
  


Current Time: Thu Mar 28 06:04:33 Coordinated Universal Time 2024