Re: Guidance regarding Gender Pooling in Liberia 2013 DHS data [message #19993 is a reply to message #19992] |
Fri, 11 September 2020 18:28 ![Go to previous message Go to previous message](/theme/default/images/up.png) ![Go to next message Go to previous message](/theme/default/images/down.png) |
Bridgette-DHS
Messages: 3151 Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
You don't need to alter the weights at all. I would open the men's file, then enter "rename mv* v*" and "gen sex=1", then append the women's file, and enter "replace sex=2 if sex==.", "label define sex 1 "Man" 2 "Woman" " and "label values sex sex". I would define sex to be consistent with hv104 in the PR file. The IR labels will over-write the MR labels in the append because the last file in any set of appends will determine the variable and value labels. You should check your key variables; some may be different in the IR and MR files, which case you want to keep them as separate v and mv variables.
The weights in the two files have already been normalized to have a mean of 1 (1,000,000 with the usual factor). If you combine the files as described, the number of cases in the two files will differ, and that will accurately convey the different numbers of men and women in the household population, as identified in the PR files. The weights in the IR and MR files have also already been adjusted for any nonresponse (loss of cases when going from the PR file to the IR and MR files). For svy purposes you don't need to change the cluster or strata ID codes. So it's much simpler than pooling surveys from different countries or time periods.
|
|
|