Re: Data discrepancy-the Philippines 2022 [message #29480 is a reply to message #29461] |
Tue, 25 June 2024 10:16 |
Janet-DHS
Messages: 880 Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
In the IR file there are 27,821 women, as you say. The weights for the IR file are constructed such that the mean weight is exactly 1 (that's the mean of v005 when it is divided by 1,000,000). Because of that, the weighted and unweighted sample sizes, for all women, are equal.
However, for subpopulations, the weighted and unweighted totals will not be equal. In particular, that's the case for the women age 15-24. If you enter "tab v013 if v013<=2" you will get an unweighted total of 10,388, as you say. But if you enter "tab v013 [iweight=v005/1000000] if v013<=2" you will see that the weighted total is 10,207.501, which rounds (barely!) to 10,208. The frequencies given in the report, like this one, are weighted.
|
|
|