Re: Query on weight variable [message #24713 is a reply to message #24703] |
Mon, 27 June 2022 15:22 |
Janet-DHS
Messages: 888 Registered: April 2022
|
Senior Member |
|
|
If you are using a combination of KR and PR variables, you should use the weights in the KR file, which (as you say) are the women's weights. The general rule, when combining data sets, is to use the weight for the data that is most (or more) at risk of nonresponse, and therefore is adjusted for such nonresponse. The weights in the PR file are adjusted for household-level nonresponse. The women's weights have an additional adjustment for women's nonresponse (within households). That's why I recommend that you use v005 rather than hv005.
|
|
|