| Re: Succeeding birth interval [message #9425 is a reply to message #9417] |
Fri, 25 March 2016 20:09  |
 |
user-rhs
Messages: 132
Registered: December 2013
|
Senior Member |
|
|
Gizachew,
If you take a look at the dataset in question, you will see that they are in wide format, i.e. one observation per woman, with information on each birth stored as separate variable, with the birth index (bidx) is the suffix to identify the birth. For example, b3_01 provides date of birth information on the first child in the birth index (the woman's youngest child, since bidx is the reverse birth order), and b3_02 provides the date of birth on the next to youngest child, and so on. Therefore, when the Recode Manual refers to a birth as "current," it's referring to that suffix. The "current" birth for b12_01 would be the youngest child, so b12_01 would *always* be missing, by definition. Say you're interested in the interval between the youngest child and the one before. You would look at b12_02, as the "current" birth is the next to last child (bidx==2), and the value in b12_02 tells you the interval between child with bidx==2 and the "following" birth, i.e. the child with bidx==1 (the youngest).
HTH,
rhs
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
