| Re: how to get the right case numbers in child recode, & then restrict further [message #958 is a reply to message #878] |
Tue, 03 December 2013 15:25  |
Sarah B
Messages: 23
Registered: June 2013
|
Member |
|
|
Can you tell us which table you are looking at to get your denominator? Are you looking at children of interviewed mothers (that's who is in the KR file), or children living in households (which would be in the PR file)?
I just went back and realized you do specify the table in your earlier message. Table 2.3 is de jure children living in the household -- i.e., it is not restricted to children of interviewed mothers. To match the sample sizes here, you need to use the PR file and restrict to children under 5 who are usual residents. Given that you want children of interviewed mothers, though, you do want the KR file. For matching denominators (which is a great starting point), I suggest you use tables in chapters 10 and 11. Note that sample sizes are weighted.
By the way, the reason your command
keep if B9==0
doesn't work is that Stata is case-sensitive. It reads b9 and B9 as two different variables -- you need to use b9. This is different than SPSS, which considers b9 and B9 to be the same.
Happy Stata-ing!
[Updated on: Sun, 02 February 2014 20:19] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
