I am using Stata 15 on Windows 8. I would like to conduct MCMC multiple imputation. I am able to reproduce the examples given in the .pdf documentation after issuing "help mi impute mvn". I encounter a problem when I apply identical code to my own data, and require the output of the "worst linear function" (WLF). The code I type (from the examples on the .pdf manual) is:
. mi impute mvn y1 y2 = x1 x2 x3 x4, mcmconly burnin(500) rseed(121) savewlf(wlf.dta, replace) initmcmc(em)
The file wlf.dta gets saved; however, the wlf variable is equal to 0 throughout. When, as an alternative, I save the trace with the saveptrace() option, I do get regular output of the individual coefficient estimates. However, given the problem with the WLF, I wonder if I am doing something wrong.
. mi impute mvn y1 y2 = x1 x2 x3 x4, mcmconly burnin(500) rseed(121) savewlf(wlf.dta, replace) initmcmc(em)
The file wlf.dta gets saved; however, the wlf variable is equal to 0 throughout. When, as an alternative, I save the trace with the saveptrace() option, I do get regular output of the individual coefficient estimates. However, given the problem with the WLF, I wonder if I am doing something wrong.
