Dear all,
I have a question regarding the use of "nlcom" with the user-written Stata command "mixlogitwtp" (Hole, 2016).
In the help file for "mixlogitwtp", it is explained how we can use "nlcom" to calculate the mean and SD of the log-normally distributed price coefficient. The help file provides the following two formulae:
For the mean: nlcom (Mean_price: -exp([Mean]mprice + 0.5*[SD]mprice^2))
For the SD: nlcom (SD_price: exp([Mean]mprice + 0.5*[SD]mprice^2) * sqrt(exp([SD]mprice^2)-1))
Now, my question is: Is there a similar formula for the case where we additionally use the "corr" option in the "mixlogitwtp" command?
Once we specify the random parameters to be correlated (i.e. by using the "corr" option), Stata's output no longer includes the estimated standard deviations of the estimated mean parameters. I understand that we need to use the post-estimation command "mixlcov, sd" in order to get them. However, the formulae mentioned above with "nlcom" no longer work for models with correlations.
I would appreciate any help. Thank you in advance and apologies if I was not specific enough. If this is the case, please let me know what additional information you may need.
Best,
Kumai
I have a question regarding the use of "nlcom" with the user-written Stata command "mixlogitwtp" (Hole, 2016).
In the help file for "mixlogitwtp", it is explained how we can use "nlcom" to calculate the mean and SD of the log-normally distributed price coefficient. The help file provides the following two formulae:
For the mean: nlcom (Mean_price: -exp([Mean]mprice + 0.5*[SD]mprice^2))
For the SD: nlcom (SD_price: exp([Mean]mprice + 0.5*[SD]mprice^2) * sqrt(exp([SD]mprice^2)-1))
Now, my question is: Is there a similar formula for the case where we additionally use the "corr" option in the "mixlogitwtp" command?
Once we specify the random parameters to be correlated (i.e. by using the "corr" option), Stata's output no longer includes the estimated standard deviations of the estimated mean parameters. I understand that we need to use the post-estimation command "mixlcov, sd" in order to get them. However, the formulae mentioned above with "nlcom" no longer work for models with correlations.
I would appreciate any help. Thank you in advance and apologies if I was not specific enough. If this is the case, please let me know what additional information you may need.
Best,
Kumai
