Hello stata experts,
I'm running a latent class analysis (gsem command) using NIH HINTS dataset (https://hints.cancer.gov/), whose manual says I have to use "svyset [pw=person_finwt0], jkrw(person_finwt1-person_finwt50,multiplier(0.98)) vce(jack) mse" for complex survey setting (person_finwt0 is the weight, and person_finwt1 - 50 are replication weights). So I did it and tried to run "svy: gsem " command and got a message that jackknife is not allowed.
------------------------------------------------------------------
prefix svy jackknife not allowed with latent classes
an error occurred when svy executed gsem
an error occurred when jackknife executed gsem
------------------------------------------------------------------
(1) In this case, what can I do in STATA? Would there be any other way to adjust for HINTS survey design? The HINTS survey was based on: stratified sample of addresses from file of residential addresses to oversample minority households, and surveying one adult in the households.
(2) Also, I converted the data file for MPlus, and tried to run LCA using TYPE = COMPLEX; REPSE = JACKKNIFE; but it requires a file of multipliers instead of designating multiplier(0.98) as in STATA. How could I generate a multiplier file in this case?
Thank you,
Hoon
I'm running a latent class analysis (gsem command) using NIH HINTS dataset (https://hints.cancer.gov/), whose manual says I have to use "svyset [pw=person_finwt0], jkrw(person_finwt1-person_finwt50,multiplier(0.98)) vce(jack) mse" for complex survey setting (person_finwt0 is the weight, and person_finwt1 - 50 are replication weights). So I did it and tried to run "svy: gsem " command and got a message that jackknife is not allowed.
------------------------------------------------------------------
prefix svy jackknife not allowed with latent classes
an error occurred when svy executed gsem
an error occurred when jackknife executed gsem
------------------------------------------------------------------
(1) In this case, what can I do in STATA? Would there be any other way to adjust for HINTS survey design? The HINTS survey was based on: stratified sample of addresses from file of residential addresses to oversample minority households, and surveying one adult in the households.
(2) Also, I converted the data file for MPlus, and tried to run LCA using TYPE = COMPLEX; REPSE = JACKKNIFE; but it requires a file of multipliers instead of designating multiplier(0.98) as in STATA. How could I generate a multiplier file in this case?
Thank you,
Hoon
Comment