| Re: Total Fertility Rates for India [message #18384 is a reply to message #18345] |
Thu, 21 November 2019 08:08   |
schoumaker
Messages: 66
Registered: May 2013
Location: Belgium
|
Senior Member |
|
|
Using tfr2, if the survey is conducted among all women, you can compute general age-specific rates in this way:
by v130 v024, sort : tfr2
You could extend the time window to 5 years (or more) for the computation of rates to have a larger number of events
by v130 v024, sort : tfr2, len(5)
You can also take into account of clustering with the cluster option.
However, you should be careful because:
- sample sizes will be too small in some cases (and you will get very large confidence intervals). Even with the latest DHS (with a huge sample size), in some states you will run into problems of sample size for some religions. You should consider grouping them, or grouping some small states.
- if you work with surveys conducted among ever married women, you will need to use an all-women factor specific to religion and state. To my knowledge, this is not available in the datasets, so you would have to compute them (but then you would need information on religion in the household file, I do not know if this is available).
Best,
Bruno
Bruno Schoumaker
Centre for Demographic Research
Université catholique de Louvain
[Updated on: Thu, 21 November 2019 08:11] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
