The Add Health survey has a complex design, and they provide two sets of weights for a 2 level model: one for the school level selection process and one for selection of individuals within schools: w3_2_wc is the individual level weight, and schlwt1 is the school level weight. psuscid is the school level sampling unit variable, and aid is the individual level ID variable.
Using stata 14, I used this code, and received the error message shown:
. svyset psuscid [pweight=schwt1], strata(region) vce(linearized) singleunit(missing) || aid [pweight=w3_2_wc]
Note: Stage 1 is sampled with replacement; further stages will be ignored for variance estimation.
sampling weights may only be specified once
Also tried this, and got error:
. svyset psuscid [pweight=schwt1], strata (region)///aid [pweight=w3_2_wc]
option / not allowed
When I try to use the drop-down box for svyset, there is a spot for a weight for each stage, but when I enter a weight for each stage, I get the error message: sampling weights and replication weights are not allowed with stage level weights.
I appreciate any suggestions. Thank you.
Barbara Wise
Using stata 14, I used this code, and received the error message shown:
. svyset psuscid [pweight=schwt1], strata(region) vce(linearized) singleunit(missing) || aid [pweight=w3_2_wc]
Note: Stage 1 is sampled with replacement; further stages will be ignored for variance estimation.
sampling weights may only be specified once
Also tried this, and got error:
. svyset psuscid [pweight=schwt1], strata (region)///aid [pweight=w3_2_wc]
option / not allowed
When I try to use the drop-down box for svyset, there is a spot for a weight for each stage, but when I enter a weight for each stage, I get the error message: sampling weights and replication weights are not allowed with stage level weights.
I appreciate any suggestions. Thank you.
Barbara Wise
Comment