| Creating a variable for disease [message #15034] |
Sat, 26 May 2018 12:08  |
anikhpg42@gmail.com
Messages: 38
Registered: December 2017
Location: Bangladesh
|
Member |
|
|
Hi,
I am working with the malnutrition status of children under age 5.
Recently I have generated an idea but don't know, is it possible to work out with BDHS data.
You know that, regarding child disease, there are some variables such as, had diarrhea/fever/ARI recently? (Yes, No..)
Now, I wanted to combine these 3 variables into one named "Had diseases recently"= "Yes, No" [combination of diarrhea,fever and ARI by recoding]
(just like the variables of ''media exposure'', 'complementary foods' etc)
What I want to know that, will it be granted to all and everywhere (journals/experts) if I combine these three into one variable?
By combining them into one variable, I want to investigate it as one of the possible determinants
(among with other socio-demographic determinants) of child malnutrition in my logistic regression.
If I can not do it such this way, is there any possible way I can combine them in one variable?
Best regards,
Anik
ASIBUL ISLAM ANIK
|
|
|
|
| Re: Creating a variable for disease [message #15073 is a reply to message #15034] |
Wed, 30 May 2018 07:54  |
Bridgette-DHS
Messages: 3218
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
Yes, you definitely can construct such a variable. It would be 1 if any of the three relevant variables is "yes" and 0 otherwise, except that it would be NA (dot) if all of the three variables are NA. But it would seem much more appropriate to me to think of illness as a consequence, rather than as a determinant, of malnutrition.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
